Subversion Repositories QNX 8.QNX8 LLVM/Clang compiler suite

Rev

Blame | Last modification | View Log | Download | RSS feed

  1. /*===---- arm_neon.h - ARM Neon intrinsics ---------------------------------===
  2.  *
  3.  * Permission is hereby granted, free of charge, to any person obtaining a copy
  4.  * of this software and associated documentation files (the "Software"), to deal
  5.  * in the Software without restriction, including without limitation the rights
  6.  * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
  7.  * copies of the Software, and to permit persons to whom the Software is
  8.  * furnished to do so, subject to the following conditions:
  9.  *
  10.  * The above copyright notice and this permission notice shall be included in
  11.  * all copies or substantial portions of the Software.
  12.  *
  13.  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  14.  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  15.  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  16.  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  17.  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
  18.  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  19.  * THE SOFTWARE.
  20.  *
  21.  *===-----------------------------------------------------------------------===
  22.  */
  23.  
  24. #ifndef __ARM_NEON_H
  25. #define __ARM_NEON_H
  26.  
  27. #ifndef __ARM_FP
  28. #error "NEON intrinsics not available with the soft-float ABI. Please use -mfloat-abi=softfp or -mfloat-abi=hard"
  29. #else
  30.  
  31. #if !defined(__ARM_NEON)
  32. #error "NEON support not enabled"
  33. #else
  34.  
  35. #include <stdint.h>
  36.  
  37. #include <arm_bf16.h>
  38. typedef __bf16 bfloat16_t;
  39. typedef float float32_t;
  40. typedef __fp16 float16_t;
  41. #ifdef __aarch64__
  42. typedef double float64_t;
  43. #endif
  44.  
  45. #ifdef __aarch64__
  46. typedef uint8_t poly8_t;
  47. typedef uint16_t poly16_t;
  48. typedef uint64_t poly64_t;
  49. typedef __uint128_t poly128_t;
  50. #else
  51. typedef int8_t poly8_t;
  52. typedef int16_t poly16_t;
  53. typedef int64_t poly64_t;
  54. #endif
  55. typedef __attribute__((neon_vector_type(8))) int8_t int8x8_t;
  56. typedef __attribute__((neon_vector_type(16))) int8_t int8x16_t;
  57. typedef __attribute__((neon_vector_type(4))) int16_t int16x4_t;
  58. typedef __attribute__((neon_vector_type(8))) int16_t int16x8_t;
  59. typedef __attribute__((neon_vector_type(2))) int32_t int32x2_t;
  60. typedef __attribute__((neon_vector_type(4))) int32_t int32x4_t;
  61. typedef __attribute__((neon_vector_type(1))) int64_t int64x1_t;
  62. typedef __attribute__((neon_vector_type(2))) int64_t int64x2_t;
  63. typedef __attribute__((neon_vector_type(8))) uint8_t uint8x8_t;
  64. typedef __attribute__((neon_vector_type(16))) uint8_t uint8x16_t;
  65. typedef __attribute__((neon_vector_type(4))) uint16_t uint16x4_t;
  66. typedef __attribute__((neon_vector_type(8))) uint16_t uint16x8_t;
  67. typedef __attribute__((neon_vector_type(2))) uint32_t uint32x2_t;
  68. typedef __attribute__((neon_vector_type(4))) uint32_t uint32x4_t;
  69. typedef __attribute__((neon_vector_type(1))) uint64_t uint64x1_t;
  70. typedef __attribute__((neon_vector_type(2))) uint64_t uint64x2_t;
  71. typedef __attribute__((neon_vector_type(4))) float16_t float16x4_t;
  72. typedef __attribute__((neon_vector_type(8))) float16_t float16x8_t;
  73. typedef __attribute__((neon_vector_type(2))) float32_t float32x2_t;
  74. typedef __attribute__((neon_vector_type(4))) float32_t float32x4_t;
  75. #ifdef __aarch64__
  76. typedef __attribute__((neon_vector_type(1))) float64_t float64x1_t;
  77. typedef __attribute__((neon_vector_type(2))) float64_t float64x2_t;
  78. #endif
  79. typedef __attribute__((neon_polyvector_type(8))) poly8_t poly8x8_t;
  80. typedef __attribute__((neon_polyvector_type(16))) poly8_t poly8x16_t;
  81. typedef __attribute__((neon_polyvector_type(4))) poly16_t poly16x4_t;
  82. typedef __attribute__((neon_polyvector_type(8))) poly16_t poly16x8_t;
  83. typedef __attribute__((neon_polyvector_type(1))) poly64_t poly64x1_t;
  84. typedef __attribute__((neon_polyvector_type(2))) poly64_t poly64x2_t;
  85.  
  86. typedef struct int8x8x2_t {
  87.   int8x8_t val[2];
  88. } int8x8x2_t;
  89.  
  90. typedef struct int8x16x2_t {
  91.   int8x16_t val[2];
  92. } int8x16x2_t;
  93.  
  94. typedef struct int16x4x2_t {
  95.   int16x4_t val[2];
  96. } int16x4x2_t;
  97.  
  98. typedef struct int16x8x2_t {
  99.   int16x8_t val[2];
  100. } int16x8x2_t;
  101.  
  102. typedef struct int32x2x2_t {
  103.   int32x2_t val[2];
  104. } int32x2x2_t;
  105.  
  106. typedef struct int32x4x2_t {
  107.   int32x4_t val[2];
  108. } int32x4x2_t;
  109.  
  110. typedef struct int64x1x2_t {
  111.   int64x1_t val[2];
  112. } int64x1x2_t;
  113.  
  114. typedef struct int64x2x2_t {
  115.   int64x2_t val[2];
  116. } int64x2x2_t;
  117.  
  118. typedef struct uint8x8x2_t {
  119.   uint8x8_t val[2];
  120. } uint8x8x2_t;
  121.  
  122. typedef struct uint8x16x2_t {
  123.   uint8x16_t val[2];
  124. } uint8x16x2_t;
  125.  
  126. typedef struct uint16x4x2_t {
  127.   uint16x4_t val[2];
  128. } uint16x4x2_t;
  129.  
  130. typedef struct uint16x8x2_t {
  131.   uint16x8_t val[2];
  132. } uint16x8x2_t;
  133.  
  134. typedef struct uint32x2x2_t {
  135.   uint32x2_t val[2];
  136. } uint32x2x2_t;
  137.  
  138. typedef struct uint32x4x2_t {
  139.   uint32x4_t val[2];
  140. } uint32x4x2_t;
  141.  
  142. typedef struct uint64x1x2_t {
  143.   uint64x1_t val[2];
  144. } uint64x1x2_t;
  145.  
  146. typedef struct uint64x2x2_t {
  147.   uint64x2_t val[2];
  148. } uint64x2x2_t;
  149.  
  150. typedef struct float16x4x2_t {
  151.   float16x4_t val[2];
  152. } float16x4x2_t;
  153.  
  154. typedef struct float16x8x2_t {
  155.   float16x8_t val[2];
  156. } float16x8x2_t;
  157.  
  158. typedef struct float32x2x2_t {
  159.   float32x2_t val[2];
  160. } float32x2x2_t;
  161.  
  162. typedef struct float32x4x2_t {
  163.   float32x4_t val[2];
  164. } float32x4x2_t;
  165.  
  166. #ifdef __aarch64__
  167. typedef struct float64x1x2_t {
  168.   float64x1_t val[2];
  169. } float64x1x2_t;
  170.  
  171. typedef struct float64x2x2_t {
  172.   float64x2_t val[2];
  173. } float64x2x2_t;
  174.  
  175. #endif
  176. typedef struct poly8x8x2_t {
  177.   poly8x8_t val[2];
  178. } poly8x8x2_t;
  179.  
  180. typedef struct poly8x16x2_t {
  181.   poly8x16_t val[2];
  182. } poly8x16x2_t;
  183.  
  184. typedef struct poly16x4x2_t {
  185.   poly16x4_t val[2];
  186. } poly16x4x2_t;
  187.  
  188. typedef struct poly16x8x2_t {
  189.   poly16x8_t val[2];
  190. } poly16x8x2_t;
  191.  
  192. typedef struct poly64x1x2_t {
  193.   poly64x1_t val[2];
  194. } poly64x1x2_t;
  195.  
  196. typedef struct poly64x2x2_t {
  197.   poly64x2_t val[2];
  198. } poly64x2x2_t;
  199.  
  200. typedef struct int8x8x3_t {
  201.   int8x8_t val[3];
  202. } int8x8x3_t;
  203.  
  204. typedef struct int8x16x3_t {
  205.   int8x16_t val[3];
  206. } int8x16x3_t;
  207.  
  208. typedef struct int16x4x3_t {
  209.   int16x4_t val[3];
  210. } int16x4x3_t;
  211.  
  212. typedef struct int16x8x3_t {
  213.   int16x8_t val[3];
  214. } int16x8x3_t;
  215.  
  216. typedef struct int32x2x3_t {
  217.   int32x2_t val[3];
  218. } int32x2x3_t;
  219.  
  220. typedef struct int32x4x3_t {
  221.   int32x4_t val[3];
  222. } int32x4x3_t;
  223.  
  224. typedef struct int64x1x3_t {
  225.   int64x1_t val[3];
  226. } int64x1x3_t;
  227.  
  228. typedef struct int64x2x3_t {
  229.   int64x2_t val[3];
  230. } int64x2x3_t;
  231.  
  232. typedef struct uint8x8x3_t {
  233.   uint8x8_t val[3];
  234. } uint8x8x3_t;
  235.  
  236. typedef struct uint8x16x3_t {
  237.   uint8x16_t val[3];
  238. } uint8x16x3_t;
  239.  
  240. typedef struct uint16x4x3_t {
  241.   uint16x4_t val[3];
  242. } uint16x4x3_t;
  243.  
  244. typedef struct uint16x8x3_t {
  245.   uint16x8_t val[3];
  246. } uint16x8x3_t;
  247.  
  248. typedef struct uint32x2x3_t {
  249.   uint32x2_t val[3];
  250. } uint32x2x3_t;
  251.  
  252. typedef struct uint32x4x3_t {
  253.   uint32x4_t val[3];
  254. } uint32x4x3_t;
  255.  
  256. typedef struct uint64x1x3_t {
  257.   uint64x1_t val[3];
  258. } uint64x1x3_t;
  259.  
  260. typedef struct uint64x2x3_t {
  261.   uint64x2_t val[3];
  262. } uint64x2x3_t;
  263.  
  264. typedef struct float16x4x3_t {
  265.   float16x4_t val[3];
  266. } float16x4x3_t;
  267.  
  268. typedef struct float16x8x3_t {
  269.   float16x8_t val[3];
  270. } float16x8x3_t;
  271.  
  272. typedef struct float32x2x3_t {
  273.   float32x2_t val[3];
  274. } float32x2x3_t;
  275.  
  276. typedef struct float32x4x3_t {
  277.   float32x4_t val[3];
  278. } float32x4x3_t;
  279.  
  280. #ifdef __aarch64__
  281. typedef struct float64x1x3_t {
  282.   float64x1_t val[3];
  283. } float64x1x3_t;
  284.  
  285. typedef struct float64x2x3_t {
  286.   float64x2_t val[3];
  287. } float64x2x3_t;
  288.  
  289. #endif
  290. typedef struct poly8x8x3_t {
  291.   poly8x8_t val[3];
  292. } poly8x8x3_t;
  293.  
  294. typedef struct poly8x16x3_t {
  295.   poly8x16_t val[3];
  296. } poly8x16x3_t;
  297.  
  298. typedef struct poly16x4x3_t {
  299.   poly16x4_t val[3];
  300. } poly16x4x3_t;
  301.  
  302. typedef struct poly16x8x3_t {
  303.   poly16x8_t val[3];
  304. } poly16x8x3_t;
  305.  
  306. typedef struct poly64x1x3_t {
  307.   poly64x1_t val[3];
  308. } poly64x1x3_t;
  309.  
  310. typedef struct poly64x2x3_t {
  311.   poly64x2_t val[3];
  312. } poly64x2x3_t;
  313.  
  314. typedef struct int8x8x4_t {
  315.   int8x8_t val[4];
  316. } int8x8x4_t;
  317.  
  318. typedef struct int8x16x4_t {
  319.   int8x16_t val[4];
  320. } int8x16x4_t;
  321.  
  322. typedef struct int16x4x4_t {
  323.   int16x4_t val[4];
  324. } int16x4x4_t;
  325.  
  326. typedef struct int16x8x4_t {
  327.   int16x8_t val[4];
  328. } int16x8x4_t;
  329.  
  330. typedef struct int32x2x4_t {
  331.   int32x2_t val[4];
  332. } int32x2x4_t;
  333.  
  334. typedef struct int32x4x4_t {
  335.   int32x4_t val[4];
  336. } int32x4x4_t;
  337.  
  338. typedef struct int64x1x4_t {
  339.   int64x1_t val[4];
  340. } int64x1x4_t;
  341.  
  342. typedef struct int64x2x4_t {
  343.   int64x2_t val[4];
  344. } int64x2x4_t;
  345.  
  346. typedef struct uint8x8x4_t {
  347.   uint8x8_t val[4];
  348. } uint8x8x4_t;
  349.  
  350. typedef struct uint8x16x4_t {
  351.   uint8x16_t val[4];
  352. } uint8x16x4_t;
  353.  
  354. typedef struct uint16x4x4_t {
  355.   uint16x4_t val[4];
  356. } uint16x4x4_t;
  357.  
  358. typedef struct uint16x8x4_t {
  359.   uint16x8_t val[4];
  360. } uint16x8x4_t;
  361.  
  362. typedef struct uint32x2x4_t {
  363.   uint32x2_t val[4];
  364. } uint32x2x4_t;
  365.  
  366. typedef struct uint32x4x4_t {
  367.   uint32x4_t val[4];
  368. } uint32x4x4_t;
  369.  
  370. typedef struct uint64x1x4_t {
  371.   uint64x1_t val[4];
  372. } uint64x1x4_t;
  373.  
  374. typedef struct uint64x2x4_t {
  375.   uint64x2_t val[4];
  376. } uint64x2x4_t;
  377.  
  378. typedef struct float16x4x4_t {
  379.   float16x4_t val[4];
  380. } float16x4x4_t;
  381.  
  382. typedef struct float16x8x4_t {
  383.   float16x8_t val[4];
  384. } float16x8x4_t;
  385.  
  386. typedef struct float32x2x4_t {
  387.   float32x2_t val[4];
  388. } float32x2x4_t;
  389.  
  390. typedef struct float32x4x4_t {
  391.   float32x4_t val[4];
  392. } float32x4x4_t;
  393.  
  394. #ifdef __aarch64__
  395. typedef struct float64x1x4_t {
  396.   float64x1_t val[4];
  397. } float64x1x4_t;
  398.  
  399. typedef struct float64x2x4_t {
  400.   float64x2_t val[4];
  401. } float64x2x4_t;
  402.  
  403. #endif
  404. typedef struct poly8x8x4_t {
  405.   poly8x8_t val[4];
  406. } poly8x8x4_t;
  407.  
  408. typedef struct poly8x16x4_t {
  409.   poly8x16_t val[4];
  410. } poly8x16x4_t;
  411.  
  412. typedef struct poly16x4x4_t {
  413.   poly16x4_t val[4];
  414. } poly16x4x4_t;
  415.  
  416. typedef struct poly16x8x4_t {
  417.   poly16x8_t val[4];
  418. } poly16x8x4_t;
  419.  
  420. typedef struct poly64x1x4_t {
  421.   poly64x1_t val[4];
  422. } poly64x1x4_t;
  423.  
  424. typedef struct poly64x2x4_t {
  425.   poly64x2_t val[4];
  426. } poly64x2x4_t;
  427.  
  428. typedef __attribute__((neon_vector_type(4))) bfloat16_t bfloat16x4_t;
  429. typedef __attribute__((neon_vector_type(8))) bfloat16_t bfloat16x8_t;
  430.  
  431. typedef struct bfloat16x4x2_t {
  432.   bfloat16x4_t val[2];
  433. } bfloat16x4x2_t;
  434.  
  435. typedef struct bfloat16x8x2_t {
  436.   bfloat16x8_t val[2];
  437. } bfloat16x8x2_t;
  438.  
  439. typedef struct bfloat16x4x3_t {
  440.   bfloat16x4_t val[3];
  441. } bfloat16x4x3_t;
  442.  
  443. typedef struct bfloat16x8x3_t {
  444.   bfloat16x8_t val[3];
  445. } bfloat16x8x3_t;
  446.  
  447. typedef struct bfloat16x4x4_t {
  448.   bfloat16x4_t val[4];
  449. } bfloat16x4x4_t;
  450.  
  451. typedef struct bfloat16x8x4_t {
  452.   bfloat16x8_t val[4];
  453. } bfloat16x8x4_t;
  454.  
  455. #define __ai static __inline__ __attribute__((__always_inline__, __nodebug__))
  456.  
  457. #ifdef __LITTLE_ENDIAN__
  458. #define splat_lane_p8(__p0, __p1) __extension__ ({ \
  459.   poly8x8_t __ret; \
  460.   poly8x8_t __s0 = __p0; \
  461.   __ret = (poly8x8_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 4); \
  462.   __ret; \
  463. })
  464. #else
  465. #define splat_lane_p8(__p0, __p1) __extension__ ({ \
  466.   poly8x8_t __ret; \
  467.   poly8x8_t __s0 = __p0; \
  468.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  469.   __ret = (poly8x8_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 4); \
  470.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  471.   __ret; \
  472. })
  473. #define __noswap_splat_lane_p8(__p0, __p1) __extension__ ({ \
  474.   poly8x8_t __ret; \
  475.   poly8x8_t __s0 = __p0; \
  476.   __ret = (poly8x8_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 4); \
  477.   __ret; \
  478. })
  479. #endif
  480.  
  481. #define splat_lane_p64(__p0, __p1) __extension__ ({ \
  482.   poly64x1_t __ret; \
  483.   poly64x1_t __s0 = __p0; \
  484.   __ret = (poly64x1_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 6); \
  485.   __ret; \
  486. })
  487. #ifdef __LITTLE_ENDIAN__
  488. #define splat_lane_p16(__p0, __p1) __extension__ ({ \
  489.   poly16x4_t __ret; \
  490.   poly16x4_t __s0 = __p0; \
  491.   __ret = (poly16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 5); \
  492.   __ret; \
  493. })
  494. #else
  495. #define splat_lane_p16(__p0, __p1) __extension__ ({ \
  496.   poly16x4_t __ret; \
  497.   poly16x4_t __s0 = __p0; \
  498.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  499.   __ret = (poly16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 5); \
  500.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  501.   __ret; \
  502. })
  503. #define __noswap_splat_lane_p16(__p0, __p1) __extension__ ({ \
  504.   poly16x4_t __ret; \
  505.   poly16x4_t __s0 = __p0; \
  506.   __ret = (poly16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 5); \
  507.   __ret; \
  508. })
  509. #endif
  510.  
  511. #ifdef __LITTLE_ENDIAN__
  512. #define splatq_lane_p8(__p0, __p1) __extension__ ({ \
  513.   poly8x16_t __ret; \
  514.   poly8x8_t __s0 = __p0; \
  515.   __ret = (poly8x16_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 4); \
  516.   __ret; \
  517. })
  518. #else
  519. #define splatq_lane_p8(__p0, __p1) __extension__ ({ \
  520.   poly8x16_t __ret; \
  521.   poly8x8_t __s0 = __p0; \
  522.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  523.   __ret = (poly8x16_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 4); \
  524.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  525.   __ret; \
  526. })
  527. #define __noswap_splatq_lane_p8(__p0, __p1) __extension__ ({ \
  528.   poly8x16_t __ret; \
  529.   poly8x8_t __s0 = __p0; \
  530.   __ret = (poly8x16_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 4); \
  531.   __ret; \
  532. })
  533. #endif
  534.  
  535. #ifdef __LITTLE_ENDIAN__
  536. #define splatq_lane_p64(__p0, __p1) __extension__ ({ \
  537.   poly64x2_t __ret; \
  538.   poly64x1_t __s0 = __p0; \
  539.   __ret = (poly64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 6); \
  540.   __ret; \
  541. })
  542. #else
  543. #define splatq_lane_p64(__p0, __p1) __extension__ ({ \
  544.   poly64x2_t __ret; \
  545.   poly64x1_t __s0 = __p0; \
  546.   __ret = (poly64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 6); \
  547.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  548.   __ret; \
  549. })
  550. #define __noswap_splatq_lane_p64(__p0, __p1) __extension__ ({ \
  551.   poly64x2_t __ret; \
  552.   poly64x1_t __s0 = __p0; \
  553.   __ret = (poly64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 6); \
  554.   __ret; \
  555. })
  556. #endif
  557.  
  558. #ifdef __LITTLE_ENDIAN__
  559. #define splatq_lane_p16(__p0, __p1) __extension__ ({ \
  560.   poly16x8_t __ret; \
  561.   poly16x4_t __s0 = __p0; \
  562.   __ret = (poly16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 5); \
  563.   __ret; \
  564. })
  565. #else
  566. #define splatq_lane_p16(__p0, __p1) __extension__ ({ \
  567.   poly16x8_t __ret; \
  568.   poly16x4_t __s0 = __p0; \
  569.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  570.   __ret = (poly16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 5); \
  571.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  572.   __ret; \
  573. })
  574. #define __noswap_splatq_lane_p16(__p0, __p1) __extension__ ({ \
  575.   poly16x8_t __ret; \
  576.   poly16x4_t __s0 = __p0; \
  577.   __ret = (poly16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 5); \
  578.   __ret; \
  579. })
  580. #endif
  581.  
  582. #ifdef __LITTLE_ENDIAN__
  583. #define splatq_lane_u8(__p0, __p1) __extension__ ({ \
  584.   uint8x16_t __ret; \
  585.   uint8x8_t __s0 = __p0; \
  586.   __ret = (uint8x16_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 16); \
  587.   __ret; \
  588. })
  589. #else
  590. #define splatq_lane_u8(__p0, __p1) __extension__ ({ \
  591.   uint8x16_t __ret; \
  592.   uint8x8_t __s0 = __p0; \
  593.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  594.   __ret = (uint8x16_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 16); \
  595.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  596.   __ret; \
  597. })
  598. #define __noswap_splatq_lane_u8(__p0, __p1) __extension__ ({ \
  599.   uint8x16_t __ret; \
  600.   uint8x8_t __s0 = __p0; \
  601.   __ret = (uint8x16_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 16); \
  602.   __ret; \
  603. })
  604. #endif
  605.  
  606. #ifdef __LITTLE_ENDIAN__
  607. #define splatq_lane_u32(__p0, __p1) __extension__ ({ \
  608.   uint32x4_t __ret; \
  609.   uint32x2_t __s0 = __p0; \
  610.   __ret = (uint32x4_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 18); \
  611.   __ret; \
  612. })
  613. #else
  614. #define splatq_lane_u32(__p0, __p1) __extension__ ({ \
  615.   uint32x4_t __ret; \
  616.   uint32x2_t __s0 = __p0; \
  617.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  618.   __ret = (uint32x4_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 18); \
  619.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  620.   __ret; \
  621. })
  622. #define __noswap_splatq_lane_u32(__p0, __p1) __extension__ ({ \
  623.   uint32x4_t __ret; \
  624.   uint32x2_t __s0 = __p0; \
  625.   __ret = (uint32x4_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 18); \
  626.   __ret; \
  627. })
  628. #endif
  629.  
  630. #ifdef __LITTLE_ENDIAN__
  631. #define splatq_lane_u64(__p0, __p1) __extension__ ({ \
  632.   uint64x2_t __ret; \
  633.   uint64x1_t __s0 = __p0; \
  634.   __ret = (uint64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 19); \
  635.   __ret; \
  636. })
  637. #else
  638. #define splatq_lane_u64(__p0, __p1) __extension__ ({ \
  639.   uint64x2_t __ret; \
  640.   uint64x1_t __s0 = __p0; \
  641.   __ret = (uint64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 19); \
  642.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  643.   __ret; \
  644. })
  645. #define __noswap_splatq_lane_u64(__p0, __p1) __extension__ ({ \
  646.   uint64x2_t __ret; \
  647.   uint64x1_t __s0 = __p0; \
  648.   __ret = (uint64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 19); \
  649.   __ret; \
  650. })
  651. #endif
  652.  
  653. #ifdef __LITTLE_ENDIAN__
  654. #define splatq_lane_u16(__p0, __p1) __extension__ ({ \
  655.   uint16x8_t __ret; \
  656.   uint16x4_t __s0 = __p0; \
  657.   __ret = (uint16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 17); \
  658.   __ret; \
  659. })
  660. #else
  661. #define splatq_lane_u16(__p0, __p1) __extension__ ({ \
  662.   uint16x8_t __ret; \
  663.   uint16x4_t __s0 = __p0; \
  664.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  665.   __ret = (uint16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 17); \
  666.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  667.   __ret; \
  668. })
  669. #define __noswap_splatq_lane_u16(__p0, __p1) __extension__ ({ \
  670.   uint16x8_t __ret; \
  671.   uint16x4_t __s0 = __p0; \
  672.   __ret = (uint16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 17); \
  673.   __ret; \
  674. })
  675. #endif
  676.  
  677. #ifdef __LITTLE_ENDIAN__
  678. #define splatq_lane_s8(__p0, __p1) __extension__ ({ \
  679.   int8x16_t __ret; \
  680.   int8x8_t __s0 = __p0; \
  681.   __ret = (int8x16_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 0); \
  682.   __ret; \
  683. })
  684. #else
  685. #define splatq_lane_s8(__p0, __p1) __extension__ ({ \
  686.   int8x16_t __ret; \
  687.   int8x8_t __s0 = __p0; \
  688.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  689.   __ret = (int8x16_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 0); \
  690.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  691.   __ret; \
  692. })
  693. #define __noswap_splatq_lane_s8(__p0, __p1) __extension__ ({ \
  694.   int8x16_t __ret; \
  695.   int8x8_t __s0 = __p0; \
  696.   __ret = (int8x16_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 0); \
  697.   __ret; \
  698. })
  699. #endif
  700.  
  701. #ifdef __LITTLE_ENDIAN__
  702. #define splatq_lane_f64(__p0, __p1) __extension__ ({ \
  703.   float64x2_t __ret; \
  704.   float64x1_t __s0 = __p0; \
  705.   __ret = (float64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 10); \
  706.   __ret; \
  707. })
  708. #else
  709. #define splatq_lane_f64(__p0, __p1) __extension__ ({ \
  710.   float64x2_t __ret; \
  711.   float64x1_t __s0 = __p0; \
  712.   __ret = (float64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 10); \
  713.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  714.   __ret; \
  715. })
  716. #define __noswap_splatq_lane_f64(__p0, __p1) __extension__ ({ \
  717.   float64x2_t __ret; \
  718.   float64x1_t __s0 = __p0; \
  719.   __ret = (float64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 10); \
  720.   __ret; \
  721. })
  722. #endif
  723.  
  724. #ifdef __LITTLE_ENDIAN__
  725. #define splatq_lane_f32(__p0, __p1) __extension__ ({ \
  726.   float32x4_t __ret; \
  727.   float32x2_t __s0 = __p0; \
  728.   __ret = (float32x4_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 9); \
  729.   __ret; \
  730. })
  731. #else
  732. #define splatq_lane_f32(__p0, __p1) __extension__ ({ \
  733.   float32x4_t __ret; \
  734.   float32x2_t __s0 = __p0; \
  735.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  736.   __ret = (float32x4_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 9); \
  737.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  738.   __ret; \
  739. })
  740. #define __noswap_splatq_lane_f32(__p0, __p1) __extension__ ({ \
  741.   float32x4_t __ret; \
  742.   float32x2_t __s0 = __p0; \
  743.   __ret = (float32x4_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 9); \
  744.   __ret; \
  745. })
  746. #endif
  747.  
  748. #ifdef __LITTLE_ENDIAN__
  749. #define splatq_lane_f16(__p0, __p1) __extension__ ({ \
  750.   float16x8_t __ret; \
  751.   float16x4_t __s0 = __p0; \
  752.   __ret = (float16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 8); \
  753.   __ret; \
  754. })
  755. #else
  756. #define splatq_lane_f16(__p0, __p1) __extension__ ({ \
  757.   float16x8_t __ret; \
  758.   float16x4_t __s0 = __p0; \
  759.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  760.   __ret = (float16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 8); \
  761.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  762.   __ret; \
  763. })
  764. #define __noswap_splatq_lane_f16(__p0, __p1) __extension__ ({ \
  765.   float16x8_t __ret; \
  766.   float16x4_t __s0 = __p0; \
  767.   __ret = (float16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 8); \
  768.   __ret; \
  769. })
  770. #endif
  771.  
  772. #ifdef __LITTLE_ENDIAN__
  773. #define splatq_lane_s32(__p0, __p1) __extension__ ({ \
  774.   int32x4_t __ret; \
  775.   int32x2_t __s0 = __p0; \
  776.   __ret = (int32x4_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 2); \
  777.   __ret; \
  778. })
  779. #else
  780. #define splatq_lane_s32(__p0, __p1) __extension__ ({ \
  781.   int32x4_t __ret; \
  782.   int32x2_t __s0 = __p0; \
  783.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  784.   __ret = (int32x4_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 2); \
  785.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  786.   __ret; \
  787. })
  788. #define __noswap_splatq_lane_s32(__p0, __p1) __extension__ ({ \
  789.   int32x4_t __ret; \
  790.   int32x2_t __s0 = __p0; \
  791.   __ret = (int32x4_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 2); \
  792.   __ret; \
  793. })
  794. #endif
  795.  
  796. #ifdef __LITTLE_ENDIAN__
  797. #define splatq_lane_s64(__p0, __p1) __extension__ ({ \
  798.   int64x2_t __ret; \
  799.   int64x1_t __s0 = __p0; \
  800.   __ret = (int64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 3); \
  801.   __ret; \
  802. })
  803. #else
  804. #define splatq_lane_s64(__p0, __p1) __extension__ ({ \
  805.   int64x2_t __ret; \
  806.   int64x1_t __s0 = __p0; \
  807.   __ret = (int64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 3); \
  808.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  809.   __ret; \
  810. })
  811. #define __noswap_splatq_lane_s64(__p0, __p1) __extension__ ({ \
  812.   int64x2_t __ret; \
  813.   int64x1_t __s0 = __p0; \
  814.   __ret = (int64x2_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 3); \
  815.   __ret; \
  816. })
  817. #endif
  818.  
  819. #ifdef __LITTLE_ENDIAN__
  820. #define splatq_lane_s16(__p0, __p1) __extension__ ({ \
  821.   int16x8_t __ret; \
  822.   int16x4_t __s0 = __p0; \
  823.   __ret = (int16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 1); \
  824.   __ret; \
  825. })
  826. #else
  827. #define splatq_lane_s16(__p0, __p1) __extension__ ({ \
  828.   int16x8_t __ret; \
  829.   int16x4_t __s0 = __p0; \
  830.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  831.   __ret = (int16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__rev0, __p1, 1); \
  832.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  833.   __ret; \
  834. })
  835. #define __noswap_splatq_lane_s16(__p0, __p1) __extension__ ({ \
  836.   int16x8_t __ret; \
  837.   int16x4_t __s0 = __p0; \
  838.   __ret = (int16x8_t) __builtin_neon_splatq_lane_v((int8x8_t)__s0, __p1, 1); \
  839.   __ret; \
  840. })
  841. #endif
  842.  
  843. #ifdef __LITTLE_ENDIAN__
  844. #define splat_lane_u8(__p0, __p1) __extension__ ({ \
  845.   uint8x8_t __ret; \
  846.   uint8x8_t __s0 = __p0; \
  847.   __ret = (uint8x8_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 16); \
  848.   __ret; \
  849. })
  850. #else
  851. #define splat_lane_u8(__p0, __p1) __extension__ ({ \
  852.   uint8x8_t __ret; \
  853.   uint8x8_t __s0 = __p0; \
  854.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  855.   __ret = (uint8x8_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 16); \
  856.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  857.   __ret; \
  858. })
  859. #define __noswap_splat_lane_u8(__p0, __p1) __extension__ ({ \
  860.   uint8x8_t __ret; \
  861.   uint8x8_t __s0 = __p0; \
  862.   __ret = (uint8x8_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 16); \
  863.   __ret; \
  864. })
  865. #endif
  866.  
  867. #ifdef __LITTLE_ENDIAN__
  868. #define splat_lane_u32(__p0, __p1) __extension__ ({ \
  869.   uint32x2_t __ret; \
  870.   uint32x2_t __s0 = __p0; \
  871.   __ret = (uint32x2_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 18); \
  872.   __ret; \
  873. })
  874. #else
  875. #define splat_lane_u32(__p0, __p1) __extension__ ({ \
  876.   uint32x2_t __ret; \
  877.   uint32x2_t __s0 = __p0; \
  878.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  879.   __ret = (uint32x2_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 18); \
  880.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  881.   __ret; \
  882. })
  883. #define __noswap_splat_lane_u32(__p0, __p1) __extension__ ({ \
  884.   uint32x2_t __ret; \
  885.   uint32x2_t __s0 = __p0; \
  886.   __ret = (uint32x2_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 18); \
  887.   __ret; \
  888. })
  889. #endif
  890.  
  891. #define splat_lane_u64(__p0, __p1) __extension__ ({ \
  892.   uint64x1_t __ret; \
  893.   uint64x1_t __s0 = __p0; \
  894.   __ret = (uint64x1_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 19); \
  895.   __ret; \
  896. })
  897. #ifdef __LITTLE_ENDIAN__
  898. #define splat_lane_u16(__p0, __p1) __extension__ ({ \
  899.   uint16x4_t __ret; \
  900.   uint16x4_t __s0 = __p0; \
  901.   __ret = (uint16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 17); \
  902.   __ret; \
  903. })
  904. #else
  905. #define splat_lane_u16(__p0, __p1) __extension__ ({ \
  906.   uint16x4_t __ret; \
  907.   uint16x4_t __s0 = __p0; \
  908.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  909.   __ret = (uint16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 17); \
  910.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  911.   __ret; \
  912. })
  913. #define __noswap_splat_lane_u16(__p0, __p1) __extension__ ({ \
  914.   uint16x4_t __ret; \
  915.   uint16x4_t __s0 = __p0; \
  916.   __ret = (uint16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 17); \
  917.   __ret; \
  918. })
  919. #endif
  920.  
  921. #ifdef __LITTLE_ENDIAN__
  922. #define splat_lane_s8(__p0, __p1) __extension__ ({ \
  923.   int8x8_t __ret; \
  924.   int8x8_t __s0 = __p0; \
  925.   __ret = (int8x8_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 0); \
  926.   __ret; \
  927. })
  928. #else
  929. #define splat_lane_s8(__p0, __p1) __extension__ ({ \
  930.   int8x8_t __ret; \
  931.   int8x8_t __s0 = __p0; \
  932.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  933.   __ret = (int8x8_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 0); \
  934.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  935.   __ret; \
  936. })
  937. #define __noswap_splat_lane_s8(__p0, __p1) __extension__ ({ \
  938.   int8x8_t __ret; \
  939.   int8x8_t __s0 = __p0; \
  940.   __ret = (int8x8_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 0); \
  941.   __ret; \
  942. })
  943. #endif
  944.  
  945. #define splat_lane_f64(__p0, __p1) __extension__ ({ \
  946.   float64x1_t __ret; \
  947.   float64x1_t __s0 = __p0; \
  948.   __ret = (float64x1_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 10); \
  949.   __ret; \
  950. })
  951. #ifdef __LITTLE_ENDIAN__
  952. #define splat_lane_f32(__p0, __p1) __extension__ ({ \
  953.   float32x2_t __ret; \
  954.   float32x2_t __s0 = __p0; \
  955.   __ret = (float32x2_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 9); \
  956.   __ret; \
  957. })
  958. #else
  959. #define splat_lane_f32(__p0, __p1) __extension__ ({ \
  960.   float32x2_t __ret; \
  961.   float32x2_t __s0 = __p0; \
  962.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  963.   __ret = (float32x2_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 9); \
  964.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  965.   __ret; \
  966. })
  967. #define __noswap_splat_lane_f32(__p0, __p1) __extension__ ({ \
  968.   float32x2_t __ret; \
  969.   float32x2_t __s0 = __p0; \
  970.   __ret = (float32x2_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 9); \
  971.   __ret; \
  972. })
  973. #endif
  974.  
  975. #ifdef __LITTLE_ENDIAN__
  976. #define splat_lane_f16(__p0, __p1) __extension__ ({ \
  977.   float16x4_t __ret; \
  978.   float16x4_t __s0 = __p0; \
  979.   __ret = (float16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 8); \
  980.   __ret; \
  981. })
  982. #else
  983. #define splat_lane_f16(__p0, __p1) __extension__ ({ \
  984.   float16x4_t __ret; \
  985.   float16x4_t __s0 = __p0; \
  986.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  987.   __ret = (float16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 8); \
  988.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  989.   __ret; \
  990. })
  991. #define __noswap_splat_lane_f16(__p0, __p1) __extension__ ({ \
  992.   float16x4_t __ret; \
  993.   float16x4_t __s0 = __p0; \
  994.   __ret = (float16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 8); \
  995.   __ret; \
  996. })
  997. #endif
  998.  
  999. #ifdef __LITTLE_ENDIAN__
  1000. #define splat_lane_s32(__p0, __p1) __extension__ ({ \
  1001.   int32x2_t __ret; \
  1002.   int32x2_t __s0 = __p0; \
  1003.   __ret = (int32x2_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 2); \
  1004.   __ret; \
  1005. })
  1006. #else
  1007. #define splat_lane_s32(__p0, __p1) __extension__ ({ \
  1008.   int32x2_t __ret; \
  1009.   int32x2_t __s0 = __p0; \
  1010.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  1011.   __ret = (int32x2_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 2); \
  1012.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  1013.   __ret; \
  1014. })
  1015. #define __noswap_splat_lane_s32(__p0, __p1) __extension__ ({ \
  1016.   int32x2_t __ret; \
  1017.   int32x2_t __s0 = __p0; \
  1018.   __ret = (int32x2_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 2); \
  1019.   __ret; \
  1020. })
  1021. #endif
  1022.  
  1023. #define splat_lane_s64(__p0, __p1) __extension__ ({ \
  1024.   int64x1_t __ret; \
  1025.   int64x1_t __s0 = __p0; \
  1026.   __ret = (int64x1_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 3); \
  1027.   __ret; \
  1028. })
  1029. #ifdef __LITTLE_ENDIAN__
  1030. #define splat_lane_s16(__p0, __p1) __extension__ ({ \
  1031.   int16x4_t __ret; \
  1032.   int16x4_t __s0 = __p0; \
  1033.   __ret = (int16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 1); \
  1034.   __ret; \
  1035. })
  1036. #else
  1037. #define splat_lane_s16(__p0, __p1) __extension__ ({ \
  1038.   int16x4_t __ret; \
  1039.   int16x4_t __s0 = __p0; \
  1040.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  1041.   __ret = (int16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__rev0, __p1, 1); \
  1042.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  1043.   __ret; \
  1044. })
  1045. #define __noswap_splat_lane_s16(__p0, __p1) __extension__ ({ \
  1046.   int16x4_t __ret; \
  1047.   int16x4_t __s0 = __p0; \
  1048.   __ret = (int16x4_t) __builtin_neon_splat_lane_v((int8x8_t)__s0, __p1, 1); \
  1049.   __ret; \
  1050. })
  1051. #endif
  1052.  
  1053. #ifdef __LITTLE_ENDIAN__
  1054. #define splat_laneq_p8(__p0, __p1) __extension__ ({ \
  1055.   poly8x8_t __ret; \
  1056.   poly8x16_t __s0 = __p0; \
  1057.   __ret = (poly8x8_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 36); \
  1058.   __ret; \
  1059. })
  1060. #else
  1061. #define splat_laneq_p8(__p0, __p1) __extension__ ({ \
  1062.   poly8x8_t __ret; \
  1063.   poly8x16_t __s0 = __p0; \
  1064.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  1065.   __ret = (poly8x8_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 36); \
  1066.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  1067.   __ret; \
  1068. })
  1069. #define __noswap_splat_laneq_p8(__p0, __p1) __extension__ ({ \
  1070.   poly8x8_t __ret; \
  1071.   poly8x16_t __s0 = __p0; \
  1072.   __ret = (poly8x8_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 36); \
  1073.   __ret; \
  1074. })
  1075. #endif
  1076.  
  1077. #ifdef __LITTLE_ENDIAN__
  1078. #define splat_laneq_p64(__p0, __p1) __extension__ ({ \
  1079.   poly64x1_t __ret; \
  1080.   poly64x2_t __s0 = __p0; \
  1081.   __ret = (poly64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 38); \
  1082.   __ret; \
  1083. })
  1084. #else
  1085. #define splat_laneq_p64(__p0, __p1) __extension__ ({ \
  1086.   poly64x1_t __ret; \
  1087.   poly64x2_t __s0 = __p0; \
  1088.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  1089.   __ret = (poly64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 38); \
  1090.   __ret; \
  1091. })
  1092. #define __noswap_splat_laneq_p64(__p0, __p1) __extension__ ({ \
  1093.   poly64x1_t __ret; \
  1094.   poly64x2_t __s0 = __p0; \
  1095.   __ret = (poly64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 38); \
  1096.   __ret; \
  1097. })
  1098. #endif
  1099.  
  1100. #ifdef __LITTLE_ENDIAN__
  1101. #define splat_laneq_p16(__p0, __p1) __extension__ ({ \
  1102.   poly16x4_t __ret; \
  1103.   poly16x8_t __s0 = __p0; \
  1104.   __ret = (poly16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 37); \
  1105.   __ret; \
  1106. })
  1107. #else
  1108. #define splat_laneq_p16(__p0, __p1) __extension__ ({ \
  1109.   poly16x4_t __ret; \
  1110.   poly16x8_t __s0 = __p0; \
  1111.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  1112.   __ret = (poly16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 37); \
  1113.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  1114.   __ret; \
  1115. })
  1116. #define __noswap_splat_laneq_p16(__p0, __p1) __extension__ ({ \
  1117.   poly16x4_t __ret; \
  1118.   poly16x8_t __s0 = __p0; \
  1119.   __ret = (poly16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 37); \
  1120.   __ret; \
  1121. })
  1122. #endif
  1123.  
  1124. #ifdef __LITTLE_ENDIAN__
  1125. #define splatq_laneq_p8(__p0, __p1) __extension__ ({ \
  1126.   poly8x16_t __ret; \
  1127.   poly8x16_t __s0 = __p0; \
  1128.   __ret = (poly8x16_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 36); \
  1129.   __ret; \
  1130. })
  1131. #else
  1132. #define splatq_laneq_p8(__p0, __p1) __extension__ ({ \
  1133.   poly8x16_t __ret; \
  1134.   poly8x16_t __s0 = __p0; \
  1135.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  1136.   __ret = (poly8x16_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 36); \
  1137.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  1138.   __ret; \
  1139. })
  1140. #define __noswap_splatq_laneq_p8(__p0, __p1) __extension__ ({ \
  1141.   poly8x16_t __ret; \
  1142.   poly8x16_t __s0 = __p0; \
  1143.   __ret = (poly8x16_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 36); \
  1144.   __ret; \
  1145. })
  1146. #endif
  1147.  
  1148. #ifdef __LITTLE_ENDIAN__
  1149. #define splatq_laneq_p64(__p0, __p1) __extension__ ({ \
  1150.   poly64x2_t __ret; \
  1151.   poly64x2_t __s0 = __p0; \
  1152.   __ret = (poly64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 38); \
  1153.   __ret; \
  1154. })
  1155. #else
  1156. #define splatq_laneq_p64(__p0, __p1) __extension__ ({ \
  1157.   poly64x2_t __ret; \
  1158.   poly64x2_t __s0 = __p0; \
  1159.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  1160.   __ret = (poly64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 38); \
  1161.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  1162.   __ret; \
  1163. })
  1164. #define __noswap_splatq_laneq_p64(__p0, __p1) __extension__ ({ \
  1165.   poly64x2_t __ret; \
  1166.   poly64x2_t __s0 = __p0; \
  1167.   __ret = (poly64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 38); \
  1168.   __ret; \
  1169. })
  1170. #endif
  1171.  
  1172. #ifdef __LITTLE_ENDIAN__
  1173. #define splatq_laneq_p16(__p0, __p1) __extension__ ({ \
  1174.   poly16x8_t __ret; \
  1175.   poly16x8_t __s0 = __p0; \
  1176.   __ret = (poly16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 37); \
  1177.   __ret; \
  1178. })
  1179. #else
  1180. #define splatq_laneq_p16(__p0, __p1) __extension__ ({ \
  1181.   poly16x8_t __ret; \
  1182.   poly16x8_t __s0 = __p0; \
  1183.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  1184.   __ret = (poly16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 37); \
  1185.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  1186.   __ret; \
  1187. })
  1188. #define __noswap_splatq_laneq_p16(__p0, __p1) __extension__ ({ \
  1189.   poly16x8_t __ret; \
  1190.   poly16x8_t __s0 = __p0; \
  1191.   __ret = (poly16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 37); \
  1192.   __ret; \
  1193. })
  1194. #endif
  1195.  
  1196. #ifdef __LITTLE_ENDIAN__
  1197. #define splatq_laneq_u8(__p0, __p1) __extension__ ({ \
  1198.   uint8x16_t __ret; \
  1199.   uint8x16_t __s0 = __p0; \
  1200.   __ret = (uint8x16_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 48); \
  1201.   __ret; \
  1202. })
  1203. #else
  1204. #define splatq_laneq_u8(__p0, __p1) __extension__ ({ \
  1205.   uint8x16_t __ret; \
  1206.   uint8x16_t __s0 = __p0; \
  1207.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  1208.   __ret = (uint8x16_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 48); \
  1209.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  1210.   __ret; \
  1211. })
  1212. #define __noswap_splatq_laneq_u8(__p0, __p1) __extension__ ({ \
  1213.   uint8x16_t __ret; \
  1214.   uint8x16_t __s0 = __p0; \
  1215.   __ret = (uint8x16_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 48); \
  1216.   __ret; \
  1217. })
  1218. #endif
  1219.  
  1220. #ifdef __LITTLE_ENDIAN__
  1221. #define splatq_laneq_u32(__p0, __p1) __extension__ ({ \
  1222.   uint32x4_t __ret; \
  1223.   uint32x4_t __s0 = __p0; \
  1224.   __ret = (uint32x4_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 50); \
  1225.   __ret; \
  1226. })
  1227. #else
  1228. #define splatq_laneq_u32(__p0, __p1) __extension__ ({ \
  1229.   uint32x4_t __ret; \
  1230.   uint32x4_t __s0 = __p0; \
  1231.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  1232.   __ret = (uint32x4_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 50); \
  1233.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  1234.   __ret; \
  1235. })
  1236. #define __noswap_splatq_laneq_u32(__p0, __p1) __extension__ ({ \
  1237.   uint32x4_t __ret; \
  1238.   uint32x4_t __s0 = __p0; \
  1239.   __ret = (uint32x4_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 50); \
  1240.   __ret; \
  1241. })
  1242. #endif
  1243.  
  1244. #ifdef __LITTLE_ENDIAN__
  1245. #define splatq_laneq_u64(__p0, __p1) __extension__ ({ \
  1246.   uint64x2_t __ret; \
  1247.   uint64x2_t __s0 = __p0; \
  1248.   __ret = (uint64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 51); \
  1249.   __ret; \
  1250. })
  1251. #else
  1252. #define splatq_laneq_u64(__p0, __p1) __extension__ ({ \
  1253.   uint64x2_t __ret; \
  1254.   uint64x2_t __s0 = __p0; \
  1255.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  1256.   __ret = (uint64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 51); \
  1257.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  1258.   __ret; \
  1259. })
  1260. #define __noswap_splatq_laneq_u64(__p0, __p1) __extension__ ({ \
  1261.   uint64x2_t __ret; \
  1262.   uint64x2_t __s0 = __p0; \
  1263.   __ret = (uint64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 51); \
  1264.   __ret; \
  1265. })
  1266. #endif
  1267.  
  1268. #ifdef __LITTLE_ENDIAN__
  1269. #define splatq_laneq_u16(__p0, __p1) __extension__ ({ \
  1270.   uint16x8_t __ret; \
  1271.   uint16x8_t __s0 = __p0; \
  1272.   __ret = (uint16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 49); \
  1273.   __ret; \
  1274. })
  1275. #else
  1276. #define splatq_laneq_u16(__p0, __p1) __extension__ ({ \
  1277.   uint16x8_t __ret; \
  1278.   uint16x8_t __s0 = __p0; \
  1279.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  1280.   __ret = (uint16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 49); \
  1281.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  1282.   __ret; \
  1283. })
  1284. #define __noswap_splatq_laneq_u16(__p0, __p1) __extension__ ({ \
  1285.   uint16x8_t __ret; \
  1286.   uint16x8_t __s0 = __p0; \
  1287.   __ret = (uint16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 49); \
  1288.   __ret; \
  1289. })
  1290. #endif
  1291.  
  1292. #ifdef __LITTLE_ENDIAN__
  1293. #define splatq_laneq_s8(__p0, __p1) __extension__ ({ \
  1294.   int8x16_t __ret; \
  1295.   int8x16_t __s0 = __p0; \
  1296.   __ret = (int8x16_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 32); \
  1297.   __ret; \
  1298. })
  1299. #else
  1300. #define splatq_laneq_s8(__p0, __p1) __extension__ ({ \
  1301.   int8x16_t __ret; \
  1302.   int8x16_t __s0 = __p0; \
  1303.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  1304.   __ret = (int8x16_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 32); \
  1305.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  1306.   __ret; \
  1307. })
  1308. #define __noswap_splatq_laneq_s8(__p0, __p1) __extension__ ({ \
  1309.   int8x16_t __ret; \
  1310.   int8x16_t __s0 = __p0; \
  1311.   __ret = (int8x16_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 32); \
  1312.   __ret; \
  1313. })
  1314. #endif
  1315.  
  1316. #ifdef __LITTLE_ENDIAN__
  1317. #define splatq_laneq_f64(__p0, __p1) __extension__ ({ \
  1318.   float64x2_t __ret; \
  1319.   float64x2_t __s0 = __p0; \
  1320.   __ret = (float64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 42); \
  1321.   __ret; \
  1322. })
  1323. #else
  1324. #define splatq_laneq_f64(__p0, __p1) __extension__ ({ \
  1325.   float64x2_t __ret; \
  1326.   float64x2_t __s0 = __p0; \
  1327.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  1328.   __ret = (float64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 42); \
  1329.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  1330.   __ret; \
  1331. })
  1332. #define __noswap_splatq_laneq_f64(__p0, __p1) __extension__ ({ \
  1333.   float64x2_t __ret; \
  1334.   float64x2_t __s0 = __p0; \
  1335.   __ret = (float64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 42); \
  1336.   __ret; \
  1337. })
  1338. #endif
  1339.  
  1340. #ifdef __LITTLE_ENDIAN__
  1341. #define splatq_laneq_f32(__p0, __p1) __extension__ ({ \
  1342.   float32x4_t __ret; \
  1343.   float32x4_t __s0 = __p0; \
  1344.   __ret = (float32x4_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 41); \
  1345.   __ret; \
  1346. })
  1347. #else
  1348. #define splatq_laneq_f32(__p0, __p1) __extension__ ({ \
  1349.   float32x4_t __ret; \
  1350.   float32x4_t __s0 = __p0; \
  1351.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  1352.   __ret = (float32x4_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 41); \
  1353.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  1354.   __ret; \
  1355. })
  1356. #define __noswap_splatq_laneq_f32(__p0, __p1) __extension__ ({ \
  1357.   float32x4_t __ret; \
  1358.   float32x4_t __s0 = __p0; \
  1359.   __ret = (float32x4_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 41); \
  1360.   __ret; \
  1361. })
  1362. #endif
  1363.  
  1364. #ifdef __LITTLE_ENDIAN__
  1365. #define splatq_laneq_f16(__p0, __p1) __extension__ ({ \
  1366.   float16x8_t __ret; \
  1367.   float16x8_t __s0 = __p0; \
  1368.   __ret = (float16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 40); \
  1369.   __ret; \
  1370. })
  1371. #else
  1372. #define splatq_laneq_f16(__p0, __p1) __extension__ ({ \
  1373.   float16x8_t __ret; \
  1374.   float16x8_t __s0 = __p0; \
  1375.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  1376.   __ret = (float16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 40); \
  1377.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  1378.   __ret; \
  1379. })
  1380. #define __noswap_splatq_laneq_f16(__p0, __p1) __extension__ ({ \
  1381.   float16x8_t __ret; \
  1382.   float16x8_t __s0 = __p0; \
  1383.   __ret = (float16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 40); \
  1384.   __ret; \
  1385. })
  1386. #endif
  1387.  
  1388. #ifdef __LITTLE_ENDIAN__
  1389. #define splatq_laneq_s32(__p0, __p1) __extension__ ({ \
  1390.   int32x4_t __ret; \
  1391.   int32x4_t __s0 = __p0; \
  1392.   __ret = (int32x4_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 34); \
  1393.   __ret; \
  1394. })
  1395. #else
  1396. #define splatq_laneq_s32(__p0, __p1) __extension__ ({ \
  1397.   int32x4_t __ret; \
  1398.   int32x4_t __s0 = __p0; \
  1399.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  1400.   __ret = (int32x4_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 34); \
  1401.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  1402.   __ret; \
  1403. })
  1404. #define __noswap_splatq_laneq_s32(__p0, __p1) __extension__ ({ \
  1405.   int32x4_t __ret; \
  1406.   int32x4_t __s0 = __p0; \
  1407.   __ret = (int32x4_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 34); \
  1408.   __ret; \
  1409. })
  1410. #endif
  1411.  
  1412. #ifdef __LITTLE_ENDIAN__
  1413. #define splatq_laneq_s64(__p0, __p1) __extension__ ({ \
  1414.   int64x2_t __ret; \
  1415.   int64x2_t __s0 = __p0; \
  1416.   __ret = (int64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 35); \
  1417.   __ret; \
  1418. })
  1419. #else
  1420. #define splatq_laneq_s64(__p0, __p1) __extension__ ({ \
  1421.   int64x2_t __ret; \
  1422.   int64x2_t __s0 = __p0; \
  1423.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  1424.   __ret = (int64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 35); \
  1425.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  1426.   __ret; \
  1427. })
  1428. #define __noswap_splatq_laneq_s64(__p0, __p1) __extension__ ({ \
  1429.   int64x2_t __ret; \
  1430.   int64x2_t __s0 = __p0; \
  1431.   __ret = (int64x2_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 35); \
  1432.   __ret; \
  1433. })
  1434. #endif
  1435.  
  1436. #ifdef __LITTLE_ENDIAN__
  1437. #define splatq_laneq_s16(__p0, __p1) __extension__ ({ \
  1438.   int16x8_t __ret; \
  1439.   int16x8_t __s0 = __p0; \
  1440.   __ret = (int16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 33); \
  1441.   __ret; \
  1442. })
  1443. #else
  1444. #define splatq_laneq_s16(__p0, __p1) __extension__ ({ \
  1445.   int16x8_t __ret; \
  1446.   int16x8_t __s0 = __p0; \
  1447.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  1448.   __ret = (int16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__rev0, __p1, 33); \
  1449.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  1450.   __ret; \
  1451. })
  1452. #define __noswap_splatq_laneq_s16(__p0, __p1) __extension__ ({ \
  1453.   int16x8_t __ret; \
  1454.   int16x8_t __s0 = __p0; \
  1455.   __ret = (int16x8_t) __builtin_neon_splatq_laneq_v((int8x16_t)__s0, __p1, 33); \
  1456.   __ret; \
  1457. })
  1458. #endif
  1459.  
  1460. #ifdef __LITTLE_ENDIAN__
  1461. #define splat_laneq_u8(__p0, __p1) __extension__ ({ \
  1462.   uint8x8_t __ret; \
  1463.   uint8x16_t __s0 = __p0; \
  1464.   __ret = (uint8x8_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 48); \
  1465.   __ret; \
  1466. })
  1467. #else
  1468. #define splat_laneq_u8(__p0, __p1) __extension__ ({ \
  1469.   uint8x8_t __ret; \
  1470.   uint8x16_t __s0 = __p0; \
  1471.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  1472.   __ret = (uint8x8_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 48); \
  1473.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  1474.   __ret; \
  1475. })
  1476. #define __noswap_splat_laneq_u8(__p0, __p1) __extension__ ({ \
  1477.   uint8x8_t __ret; \
  1478.   uint8x16_t __s0 = __p0; \
  1479.   __ret = (uint8x8_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 48); \
  1480.   __ret; \
  1481. })
  1482. #endif
  1483.  
  1484. #ifdef __LITTLE_ENDIAN__
  1485. #define splat_laneq_u32(__p0, __p1) __extension__ ({ \
  1486.   uint32x2_t __ret; \
  1487.   uint32x4_t __s0 = __p0; \
  1488.   __ret = (uint32x2_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 50); \
  1489.   __ret; \
  1490. })
  1491. #else
  1492. #define splat_laneq_u32(__p0, __p1) __extension__ ({ \
  1493.   uint32x2_t __ret; \
  1494.   uint32x4_t __s0 = __p0; \
  1495.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  1496.   __ret = (uint32x2_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 50); \
  1497.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  1498.   __ret; \
  1499. })
  1500. #define __noswap_splat_laneq_u32(__p0, __p1) __extension__ ({ \
  1501.   uint32x2_t __ret; \
  1502.   uint32x4_t __s0 = __p0; \
  1503.   __ret = (uint32x2_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 50); \
  1504.   __ret; \
  1505. })
  1506. #endif
  1507.  
  1508. #ifdef __LITTLE_ENDIAN__
  1509. #define splat_laneq_u64(__p0, __p1) __extension__ ({ \
  1510.   uint64x1_t __ret; \
  1511.   uint64x2_t __s0 = __p0; \
  1512.   __ret = (uint64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 51); \
  1513.   __ret; \
  1514. })
  1515. #else
  1516. #define splat_laneq_u64(__p0, __p1) __extension__ ({ \
  1517.   uint64x1_t __ret; \
  1518.   uint64x2_t __s0 = __p0; \
  1519.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  1520.   __ret = (uint64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 51); \
  1521.   __ret; \
  1522. })
  1523. #define __noswap_splat_laneq_u64(__p0, __p1) __extension__ ({ \
  1524.   uint64x1_t __ret; \
  1525.   uint64x2_t __s0 = __p0; \
  1526.   __ret = (uint64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 51); \
  1527.   __ret; \
  1528. })
  1529. #endif
  1530.  
  1531. #ifdef __LITTLE_ENDIAN__
  1532. #define splat_laneq_u16(__p0, __p1) __extension__ ({ \
  1533.   uint16x4_t __ret; \
  1534.   uint16x8_t __s0 = __p0; \
  1535.   __ret = (uint16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 49); \
  1536.   __ret; \
  1537. })
  1538. #else
  1539. #define splat_laneq_u16(__p0, __p1) __extension__ ({ \
  1540.   uint16x4_t __ret; \
  1541.   uint16x8_t __s0 = __p0; \
  1542.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  1543.   __ret = (uint16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 49); \
  1544.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  1545.   __ret; \
  1546. })
  1547. #define __noswap_splat_laneq_u16(__p0, __p1) __extension__ ({ \
  1548.   uint16x4_t __ret; \
  1549.   uint16x8_t __s0 = __p0; \
  1550.   __ret = (uint16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 49); \
  1551.   __ret; \
  1552. })
  1553. #endif
  1554.  
  1555. #ifdef __LITTLE_ENDIAN__
  1556. #define splat_laneq_s8(__p0, __p1) __extension__ ({ \
  1557.   int8x8_t __ret; \
  1558.   int8x16_t __s0 = __p0; \
  1559.   __ret = (int8x8_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 32); \
  1560.   __ret; \
  1561. })
  1562. #else
  1563. #define splat_laneq_s8(__p0, __p1) __extension__ ({ \
  1564.   int8x8_t __ret; \
  1565.   int8x16_t __s0 = __p0; \
  1566.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  1567.   __ret = (int8x8_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 32); \
  1568.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  1569.   __ret; \
  1570. })
  1571. #define __noswap_splat_laneq_s8(__p0, __p1) __extension__ ({ \
  1572.   int8x8_t __ret; \
  1573.   int8x16_t __s0 = __p0; \
  1574.   __ret = (int8x8_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 32); \
  1575.   __ret; \
  1576. })
  1577. #endif
  1578.  
  1579. #ifdef __LITTLE_ENDIAN__
  1580. #define splat_laneq_f64(__p0, __p1) __extension__ ({ \
  1581.   float64x1_t __ret; \
  1582.   float64x2_t __s0 = __p0; \
  1583.   __ret = (float64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 42); \
  1584.   __ret; \
  1585. })
  1586. #else
  1587. #define splat_laneq_f64(__p0, __p1) __extension__ ({ \
  1588.   float64x1_t __ret; \
  1589.   float64x2_t __s0 = __p0; \
  1590.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  1591.   __ret = (float64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 42); \
  1592.   __ret; \
  1593. })
  1594. #define __noswap_splat_laneq_f64(__p0, __p1) __extension__ ({ \
  1595.   float64x1_t __ret; \
  1596.   float64x2_t __s0 = __p0; \
  1597.   __ret = (float64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 42); \
  1598.   __ret; \
  1599. })
  1600. #endif
  1601.  
  1602. #ifdef __LITTLE_ENDIAN__
  1603. #define splat_laneq_f32(__p0, __p1) __extension__ ({ \
  1604.   float32x2_t __ret; \
  1605.   float32x4_t __s0 = __p0; \
  1606.   __ret = (float32x2_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 41); \
  1607.   __ret; \
  1608. })
  1609. #else
  1610. #define splat_laneq_f32(__p0, __p1) __extension__ ({ \
  1611.   float32x2_t __ret; \
  1612.   float32x4_t __s0 = __p0; \
  1613.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  1614.   __ret = (float32x2_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 41); \
  1615.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  1616.   __ret; \
  1617. })
  1618. #define __noswap_splat_laneq_f32(__p0, __p1) __extension__ ({ \
  1619.   float32x2_t __ret; \
  1620.   float32x4_t __s0 = __p0; \
  1621.   __ret = (float32x2_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 41); \
  1622.   __ret; \
  1623. })
  1624. #endif
  1625.  
  1626. #ifdef __LITTLE_ENDIAN__
  1627. #define splat_laneq_f16(__p0, __p1) __extension__ ({ \
  1628.   float16x4_t __ret; \
  1629.   float16x8_t __s0 = __p0; \
  1630.   __ret = (float16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 40); \
  1631.   __ret; \
  1632. })
  1633. #else
  1634. #define splat_laneq_f16(__p0, __p1) __extension__ ({ \
  1635.   float16x4_t __ret; \
  1636.   float16x8_t __s0 = __p0; \
  1637.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  1638.   __ret = (float16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 40); \
  1639.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  1640.   __ret; \
  1641. })
  1642. #define __noswap_splat_laneq_f16(__p0, __p1) __extension__ ({ \
  1643.   float16x4_t __ret; \
  1644.   float16x8_t __s0 = __p0; \
  1645.   __ret = (float16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 40); \
  1646.   __ret; \
  1647. })
  1648. #endif
  1649.  
  1650. #ifdef __LITTLE_ENDIAN__
  1651. #define splat_laneq_s32(__p0, __p1) __extension__ ({ \
  1652.   int32x2_t __ret; \
  1653.   int32x4_t __s0 = __p0; \
  1654.   __ret = (int32x2_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 34); \
  1655.   __ret; \
  1656. })
  1657. #else
  1658. #define splat_laneq_s32(__p0, __p1) __extension__ ({ \
  1659.   int32x2_t __ret; \
  1660.   int32x4_t __s0 = __p0; \
  1661.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  1662.   __ret = (int32x2_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 34); \
  1663.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  1664.   __ret; \
  1665. })
  1666. #define __noswap_splat_laneq_s32(__p0, __p1) __extension__ ({ \
  1667.   int32x2_t __ret; \
  1668.   int32x4_t __s0 = __p0; \
  1669.   __ret = (int32x2_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 34); \
  1670.   __ret; \
  1671. })
  1672. #endif
  1673.  
  1674. #ifdef __LITTLE_ENDIAN__
  1675. #define splat_laneq_s64(__p0, __p1) __extension__ ({ \
  1676.   int64x1_t __ret; \
  1677.   int64x2_t __s0 = __p0; \
  1678.   __ret = (int64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 35); \
  1679.   __ret; \
  1680. })
  1681. #else
  1682. #define splat_laneq_s64(__p0, __p1) __extension__ ({ \
  1683.   int64x1_t __ret; \
  1684.   int64x2_t __s0 = __p0; \
  1685.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  1686.   __ret = (int64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 35); \
  1687.   __ret; \
  1688. })
  1689. #define __noswap_splat_laneq_s64(__p0, __p1) __extension__ ({ \
  1690.   int64x1_t __ret; \
  1691.   int64x2_t __s0 = __p0; \
  1692.   __ret = (int64x1_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 35); \
  1693.   __ret; \
  1694. })
  1695. #endif
  1696.  
  1697. #ifdef __LITTLE_ENDIAN__
  1698. #define splat_laneq_s16(__p0, __p1) __extension__ ({ \
  1699.   int16x4_t __ret; \
  1700.   int16x8_t __s0 = __p0; \
  1701.   __ret = (int16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 33); \
  1702.   __ret; \
  1703. })
  1704. #else
  1705. #define splat_laneq_s16(__p0, __p1) __extension__ ({ \
  1706.   int16x4_t __ret; \
  1707.   int16x8_t __s0 = __p0; \
  1708.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  1709.   __ret = (int16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__rev0, __p1, 33); \
  1710.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  1711.   __ret; \
  1712. })
  1713. #define __noswap_splat_laneq_s16(__p0, __p1) __extension__ ({ \
  1714.   int16x4_t __ret; \
  1715.   int16x8_t __s0 = __p0; \
  1716.   __ret = (int16x4_t) __builtin_neon_splat_laneq_v((int8x16_t)__s0, __p1, 33); \
  1717.   __ret; \
  1718. })
  1719. #endif
  1720.  
  1721. #ifdef __LITTLE_ENDIAN__
  1722. __ai uint8x16_t vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  1723.   uint8x16_t __ret;
  1724.   __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  1725.   return __ret;
  1726. }
  1727. #else
  1728. __ai uint8x16_t vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  1729.   uint8x16_t __ret;
  1730.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  1731.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  1732.   __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  1733.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  1734.   return __ret;
  1735. }
  1736. __ai uint8x16_t __noswap_vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  1737.   uint8x16_t __ret;
  1738.   __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  1739.   return __ret;
  1740. }
  1741. #endif
  1742.  
  1743. #ifdef __LITTLE_ENDIAN__
  1744. __ai uint32x4_t vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  1745.   uint32x4_t __ret;
  1746.   __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  1747.   return __ret;
  1748. }
  1749. #else
  1750. __ai uint32x4_t vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  1751.   uint32x4_t __ret;
  1752.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  1753.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  1754.   __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  1755.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  1756.   return __ret;
  1757. }
  1758. __ai uint32x4_t __noswap_vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  1759.   uint32x4_t __ret;
  1760.   __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  1761.   return __ret;
  1762. }
  1763. #endif
  1764.  
  1765. #ifdef __LITTLE_ENDIAN__
  1766. __ai uint16x8_t vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  1767.   uint16x8_t __ret;
  1768.   __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  1769.   return __ret;
  1770. }
  1771. #else
  1772. __ai uint16x8_t vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  1773.   uint16x8_t __ret;
  1774.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  1775.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  1776.   __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  1777.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  1778.   return __ret;
  1779. }
  1780. __ai uint16x8_t __noswap_vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  1781.   uint16x8_t __ret;
  1782.   __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  1783.   return __ret;
  1784. }
  1785. #endif
  1786.  
  1787. #ifdef __LITTLE_ENDIAN__
  1788. __ai int8x16_t vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
  1789.   int8x16_t __ret;
  1790.   __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  1791.   return __ret;
  1792. }
  1793. #else
  1794. __ai int8x16_t vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
  1795.   int8x16_t __ret;
  1796.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  1797.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  1798.   __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  1799.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  1800.   return __ret;
  1801. }
  1802. __ai int8x16_t __noswap_vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
  1803.   int8x16_t __ret;
  1804.   __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  1805.   return __ret;
  1806. }
  1807. #endif
  1808.  
  1809. #ifdef __LITTLE_ENDIAN__
  1810. __ai float32x4_t vabdq_f32(float32x4_t __p0, float32x4_t __p1) {
  1811.   float32x4_t __ret;
  1812.   __ret = (float32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  1813.   return __ret;
  1814. }
  1815. #else
  1816. __ai float32x4_t vabdq_f32(float32x4_t __p0, float32x4_t __p1) {
  1817.   float32x4_t __ret;
  1818.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  1819.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  1820.   __ret = (float32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  1821.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  1822.   return __ret;
  1823. }
  1824. #endif
  1825.  
  1826. #ifdef __LITTLE_ENDIAN__
  1827. __ai int32x4_t vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
  1828.   int32x4_t __ret;
  1829.   __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  1830.   return __ret;
  1831. }
  1832. #else
  1833. __ai int32x4_t vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
  1834.   int32x4_t __ret;
  1835.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  1836.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  1837.   __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  1838.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  1839.   return __ret;
  1840. }
  1841. __ai int32x4_t __noswap_vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
  1842.   int32x4_t __ret;
  1843.   __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  1844.   return __ret;
  1845. }
  1846. #endif
  1847.  
  1848. #ifdef __LITTLE_ENDIAN__
  1849. __ai int16x8_t vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
  1850.   int16x8_t __ret;
  1851.   __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  1852.   return __ret;
  1853. }
  1854. #else
  1855. __ai int16x8_t vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
  1856.   int16x8_t __ret;
  1857.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  1858.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  1859.   __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  1860.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  1861.   return __ret;
  1862. }
  1863. __ai int16x8_t __noswap_vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
  1864.   int16x8_t __ret;
  1865.   __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  1866.   return __ret;
  1867. }
  1868. #endif
  1869.  
  1870. #ifdef __LITTLE_ENDIAN__
  1871. __ai uint8x8_t vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  1872.   uint8x8_t __ret;
  1873.   __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  1874.   return __ret;
  1875. }
  1876. #else
  1877. __ai uint8x8_t vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  1878.   uint8x8_t __ret;
  1879.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  1880.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  1881.   __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  1882.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  1883.   return __ret;
  1884. }
  1885. __ai uint8x8_t __noswap_vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  1886.   uint8x8_t __ret;
  1887.   __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  1888.   return __ret;
  1889. }
  1890. #endif
  1891.  
  1892. #ifdef __LITTLE_ENDIAN__
  1893. __ai uint32x2_t vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  1894.   uint32x2_t __ret;
  1895.   __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  1896.   return __ret;
  1897. }
  1898. #else
  1899. __ai uint32x2_t vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  1900.   uint32x2_t __ret;
  1901.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  1902.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  1903.   __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  1904.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  1905.   return __ret;
  1906. }
  1907. __ai uint32x2_t __noswap_vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  1908.   uint32x2_t __ret;
  1909.   __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  1910.   return __ret;
  1911. }
  1912. #endif
  1913.  
  1914. #ifdef __LITTLE_ENDIAN__
  1915. __ai uint16x4_t vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  1916.   uint16x4_t __ret;
  1917.   __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  1918.   return __ret;
  1919. }
  1920. #else
  1921. __ai uint16x4_t vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  1922.   uint16x4_t __ret;
  1923.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  1924.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  1925.   __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  1926.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  1927.   return __ret;
  1928. }
  1929. __ai uint16x4_t __noswap_vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  1930.   uint16x4_t __ret;
  1931.   __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  1932.   return __ret;
  1933. }
  1934. #endif
  1935.  
  1936. #ifdef __LITTLE_ENDIAN__
  1937. __ai int8x8_t vabd_s8(int8x8_t __p0, int8x8_t __p1) {
  1938.   int8x8_t __ret;
  1939.   __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  1940.   return __ret;
  1941. }
  1942. #else
  1943. __ai int8x8_t vabd_s8(int8x8_t __p0, int8x8_t __p1) {
  1944.   int8x8_t __ret;
  1945.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  1946.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  1947.   __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  1948.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  1949.   return __ret;
  1950. }
  1951. __ai int8x8_t __noswap_vabd_s8(int8x8_t __p0, int8x8_t __p1) {
  1952.   int8x8_t __ret;
  1953.   __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  1954.   return __ret;
  1955. }
  1956. #endif
  1957.  
  1958. #ifdef __LITTLE_ENDIAN__
  1959. __ai float32x2_t vabd_f32(float32x2_t __p0, float32x2_t __p1) {
  1960.   float32x2_t __ret;
  1961.   __ret = (float32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  1962.   return __ret;
  1963. }
  1964. #else
  1965. __ai float32x2_t vabd_f32(float32x2_t __p0, float32x2_t __p1) {
  1966.   float32x2_t __ret;
  1967.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  1968.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  1969.   __ret = (float32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  1970.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  1971.   return __ret;
  1972. }
  1973. #endif
  1974.  
  1975. #ifdef __LITTLE_ENDIAN__
  1976. __ai int32x2_t vabd_s32(int32x2_t __p0, int32x2_t __p1) {
  1977.   int32x2_t __ret;
  1978.   __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  1979.   return __ret;
  1980. }
  1981. #else
  1982. __ai int32x2_t vabd_s32(int32x2_t __p0, int32x2_t __p1) {
  1983.   int32x2_t __ret;
  1984.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  1985.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  1986.   __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  1987.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  1988.   return __ret;
  1989. }
  1990. __ai int32x2_t __noswap_vabd_s32(int32x2_t __p0, int32x2_t __p1) {
  1991.   int32x2_t __ret;
  1992.   __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  1993.   return __ret;
  1994. }
  1995. #endif
  1996.  
  1997. #ifdef __LITTLE_ENDIAN__
  1998. __ai int16x4_t vabd_s16(int16x4_t __p0, int16x4_t __p1) {
  1999.   int16x4_t __ret;
  2000.   __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  2001.   return __ret;
  2002. }
  2003. #else
  2004. __ai int16x4_t vabd_s16(int16x4_t __p0, int16x4_t __p1) {
  2005.   int16x4_t __ret;
  2006.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2007.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2008.   __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  2009.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2010.   return __ret;
  2011. }
  2012. __ai int16x4_t __noswap_vabd_s16(int16x4_t __p0, int16x4_t __p1) {
  2013.   int16x4_t __ret;
  2014.   __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  2015.   return __ret;
  2016. }
  2017. #endif
  2018.  
  2019. #ifdef __LITTLE_ENDIAN__
  2020. __ai int8x16_t vabsq_s8(int8x16_t __p0) {
  2021.   int8x16_t __ret;
  2022.   __ret = (int8x16_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 32);
  2023.   return __ret;
  2024. }
  2025. #else
  2026. __ai int8x16_t vabsq_s8(int8x16_t __p0) {
  2027.   int8x16_t __ret;
  2028.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2029.   __ret = (int8x16_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 32);
  2030.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2031.   return __ret;
  2032. }
  2033. #endif
  2034.  
  2035. #ifdef __LITTLE_ENDIAN__
  2036. __ai float32x4_t vabsq_f32(float32x4_t __p0) {
  2037.   float32x4_t __ret;
  2038.   __ret = (float32x4_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 41);
  2039.   return __ret;
  2040. }
  2041. #else
  2042. __ai float32x4_t vabsq_f32(float32x4_t __p0) {
  2043.   float32x4_t __ret;
  2044.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2045.   __ret = (float32x4_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 41);
  2046.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2047.   return __ret;
  2048. }
  2049. #endif
  2050.  
  2051. #ifdef __LITTLE_ENDIAN__
  2052. __ai int32x4_t vabsq_s32(int32x4_t __p0) {
  2053.   int32x4_t __ret;
  2054.   __ret = (int32x4_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 34);
  2055.   return __ret;
  2056. }
  2057. #else
  2058. __ai int32x4_t vabsq_s32(int32x4_t __p0) {
  2059.   int32x4_t __ret;
  2060.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2061.   __ret = (int32x4_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 34);
  2062.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2063.   return __ret;
  2064. }
  2065. #endif
  2066.  
  2067. #ifdef __LITTLE_ENDIAN__
  2068. __ai int16x8_t vabsq_s16(int16x8_t __p0) {
  2069.   int16x8_t __ret;
  2070.   __ret = (int16x8_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 33);
  2071.   return __ret;
  2072. }
  2073. #else
  2074. __ai int16x8_t vabsq_s16(int16x8_t __p0) {
  2075.   int16x8_t __ret;
  2076.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2077.   __ret = (int16x8_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 33);
  2078.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2079.   return __ret;
  2080. }
  2081. #endif
  2082.  
  2083. #ifdef __LITTLE_ENDIAN__
  2084. __ai int8x8_t vabs_s8(int8x8_t __p0) {
  2085.   int8x8_t __ret;
  2086.   __ret = (int8x8_t) __builtin_neon_vabs_v((int8x8_t)__p0, 0);
  2087.   return __ret;
  2088. }
  2089. #else
  2090. __ai int8x8_t vabs_s8(int8x8_t __p0) {
  2091.   int8x8_t __ret;
  2092.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2093.   __ret = (int8x8_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 0);
  2094.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2095.   return __ret;
  2096. }
  2097. #endif
  2098.  
  2099. #ifdef __LITTLE_ENDIAN__
  2100. __ai float32x2_t vabs_f32(float32x2_t __p0) {
  2101.   float32x2_t __ret;
  2102.   __ret = (float32x2_t) __builtin_neon_vabs_v((int8x8_t)__p0, 9);
  2103.   return __ret;
  2104. }
  2105. #else
  2106. __ai float32x2_t vabs_f32(float32x2_t __p0) {
  2107.   float32x2_t __ret;
  2108.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2109.   __ret = (float32x2_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 9);
  2110.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2111.   return __ret;
  2112. }
  2113. #endif
  2114.  
  2115. #ifdef __LITTLE_ENDIAN__
  2116. __ai int32x2_t vabs_s32(int32x2_t __p0) {
  2117.   int32x2_t __ret;
  2118.   __ret = (int32x2_t) __builtin_neon_vabs_v((int8x8_t)__p0, 2);
  2119.   return __ret;
  2120. }
  2121. #else
  2122. __ai int32x2_t vabs_s32(int32x2_t __p0) {
  2123.   int32x2_t __ret;
  2124.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2125.   __ret = (int32x2_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 2);
  2126.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2127.   return __ret;
  2128. }
  2129. #endif
  2130.  
  2131. #ifdef __LITTLE_ENDIAN__
  2132. __ai int16x4_t vabs_s16(int16x4_t __p0) {
  2133.   int16x4_t __ret;
  2134.   __ret = (int16x4_t) __builtin_neon_vabs_v((int8x8_t)__p0, 1);
  2135.   return __ret;
  2136. }
  2137. #else
  2138. __ai int16x4_t vabs_s16(int16x4_t __p0) {
  2139.   int16x4_t __ret;
  2140.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2141.   __ret = (int16x4_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 1);
  2142.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2143.   return __ret;
  2144. }
  2145. #endif
  2146.  
  2147. #ifdef __LITTLE_ENDIAN__
  2148. __ai uint8x16_t vaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  2149.   uint8x16_t __ret;
  2150.   __ret = __p0 + __p1;
  2151.   return __ret;
  2152. }
  2153. #else
  2154. __ai uint8x16_t vaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  2155.   uint8x16_t __ret;
  2156.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2157.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2158.   __ret = __rev0 + __rev1;
  2159.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2160.   return __ret;
  2161. }
  2162. #endif
  2163.  
  2164. #ifdef __LITTLE_ENDIAN__
  2165. __ai uint32x4_t vaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  2166.   uint32x4_t __ret;
  2167.   __ret = __p0 + __p1;
  2168.   return __ret;
  2169. }
  2170. #else
  2171. __ai uint32x4_t vaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  2172.   uint32x4_t __ret;
  2173.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2174.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2175.   __ret = __rev0 + __rev1;
  2176.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2177.   return __ret;
  2178. }
  2179. #endif
  2180.  
  2181. #ifdef __LITTLE_ENDIAN__
  2182. __ai uint64x2_t vaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  2183.   uint64x2_t __ret;
  2184.   __ret = __p0 + __p1;
  2185.   return __ret;
  2186. }
  2187. #else
  2188. __ai uint64x2_t vaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  2189.   uint64x2_t __ret;
  2190.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2191.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2192.   __ret = __rev0 + __rev1;
  2193.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2194.   return __ret;
  2195. }
  2196. #endif
  2197.  
  2198. #ifdef __LITTLE_ENDIAN__
  2199. __ai uint16x8_t vaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  2200.   uint16x8_t __ret;
  2201.   __ret = __p0 + __p1;
  2202.   return __ret;
  2203. }
  2204. #else
  2205. __ai uint16x8_t vaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  2206.   uint16x8_t __ret;
  2207.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2208.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2209.   __ret = __rev0 + __rev1;
  2210.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2211.   return __ret;
  2212. }
  2213. #endif
  2214.  
  2215. #ifdef __LITTLE_ENDIAN__
  2216. __ai int8x16_t vaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  2217.   int8x16_t __ret;
  2218.   __ret = __p0 + __p1;
  2219.   return __ret;
  2220. }
  2221. #else
  2222. __ai int8x16_t vaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  2223.   int8x16_t __ret;
  2224.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2225.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2226.   __ret = __rev0 + __rev1;
  2227.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2228.   return __ret;
  2229. }
  2230. #endif
  2231.  
  2232. #ifdef __LITTLE_ENDIAN__
  2233. __ai float32x4_t vaddq_f32(float32x4_t __p0, float32x4_t __p1) {
  2234.   float32x4_t __ret;
  2235.   __ret = __p0 + __p1;
  2236.   return __ret;
  2237. }
  2238. #else
  2239. __ai float32x4_t vaddq_f32(float32x4_t __p0, float32x4_t __p1) {
  2240.   float32x4_t __ret;
  2241.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2242.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2243.   __ret = __rev0 + __rev1;
  2244.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2245.   return __ret;
  2246. }
  2247. #endif
  2248.  
  2249. #ifdef __LITTLE_ENDIAN__
  2250. __ai int32x4_t vaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  2251.   int32x4_t __ret;
  2252.   __ret = __p0 + __p1;
  2253.   return __ret;
  2254. }
  2255. #else
  2256. __ai int32x4_t vaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  2257.   int32x4_t __ret;
  2258.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2259.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2260.   __ret = __rev0 + __rev1;
  2261.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2262.   return __ret;
  2263. }
  2264. #endif
  2265.  
  2266. #ifdef __LITTLE_ENDIAN__
  2267. __ai int64x2_t vaddq_s64(int64x2_t __p0, int64x2_t __p1) {
  2268.   int64x2_t __ret;
  2269.   __ret = __p0 + __p1;
  2270.   return __ret;
  2271. }
  2272. #else
  2273. __ai int64x2_t vaddq_s64(int64x2_t __p0, int64x2_t __p1) {
  2274.   int64x2_t __ret;
  2275.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2276.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2277.   __ret = __rev0 + __rev1;
  2278.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2279.   return __ret;
  2280. }
  2281. #endif
  2282.  
  2283. #ifdef __LITTLE_ENDIAN__
  2284. __ai int16x8_t vaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  2285.   int16x8_t __ret;
  2286.   __ret = __p0 + __p1;
  2287.   return __ret;
  2288. }
  2289. #else
  2290. __ai int16x8_t vaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  2291.   int16x8_t __ret;
  2292.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2293.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2294.   __ret = __rev0 + __rev1;
  2295.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2296.   return __ret;
  2297. }
  2298. #endif
  2299.  
  2300. #ifdef __LITTLE_ENDIAN__
  2301. __ai uint8x8_t vadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  2302.   uint8x8_t __ret;
  2303.   __ret = __p0 + __p1;
  2304.   return __ret;
  2305. }
  2306. #else
  2307. __ai uint8x8_t vadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  2308.   uint8x8_t __ret;
  2309.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2310.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2311.   __ret = __rev0 + __rev1;
  2312.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2313.   return __ret;
  2314. }
  2315. #endif
  2316.  
  2317. #ifdef __LITTLE_ENDIAN__
  2318. __ai uint32x2_t vadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  2319.   uint32x2_t __ret;
  2320.   __ret = __p0 + __p1;
  2321.   return __ret;
  2322. }
  2323. #else
  2324. __ai uint32x2_t vadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  2325.   uint32x2_t __ret;
  2326.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2327.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2328.   __ret = __rev0 + __rev1;
  2329.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2330.   return __ret;
  2331. }
  2332. #endif
  2333.  
  2334. __ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
  2335.   uint64x1_t __ret;
  2336.   __ret = __p0 + __p1;
  2337.   return __ret;
  2338. }
  2339. #ifdef __LITTLE_ENDIAN__
  2340. __ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  2341.   uint16x4_t __ret;
  2342.   __ret = __p0 + __p1;
  2343.   return __ret;
  2344. }
  2345. #else
  2346. __ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  2347.   uint16x4_t __ret;
  2348.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2349.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2350.   __ret = __rev0 + __rev1;
  2351.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2352.   return __ret;
  2353. }
  2354. #endif
  2355.  
  2356. #ifdef __LITTLE_ENDIAN__
  2357. __ai int8x8_t vadd_s8(int8x8_t __p0, int8x8_t __p1) {
  2358.   int8x8_t __ret;
  2359.   __ret = __p0 + __p1;
  2360.   return __ret;
  2361. }
  2362. #else
  2363. __ai int8x8_t vadd_s8(int8x8_t __p0, int8x8_t __p1) {
  2364.   int8x8_t __ret;
  2365.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2366.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2367.   __ret = __rev0 + __rev1;
  2368.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2369.   return __ret;
  2370. }
  2371. #endif
  2372.  
  2373. #ifdef __LITTLE_ENDIAN__
  2374. __ai float32x2_t vadd_f32(float32x2_t __p0, float32x2_t __p1) {
  2375.   float32x2_t __ret;
  2376.   __ret = __p0 + __p1;
  2377.   return __ret;
  2378. }
  2379. #else
  2380. __ai float32x2_t vadd_f32(float32x2_t __p0, float32x2_t __p1) {
  2381.   float32x2_t __ret;
  2382.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2383.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2384.   __ret = __rev0 + __rev1;
  2385.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2386.   return __ret;
  2387. }
  2388. #endif
  2389.  
  2390. #ifdef __LITTLE_ENDIAN__
  2391. __ai int32x2_t vadd_s32(int32x2_t __p0, int32x2_t __p1) {
  2392.   int32x2_t __ret;
  2393.   __ret = __p0 + __p1;
  2394.   return __ret;
  2395. }
  2396. #else
  2397. __ai int32x2_t vadd_s32(int32x2_t __p0, int32x2_t __p1) {
  2398.   int32x2_t __ret;
  2399.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2400.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2401.   __ret = __rev0 + __rev1;
  2402.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2403.   return __ret;
  2404. }
  2405. #endif
  2406.  
  2407. __ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
  2408.   int64x1_t __ret;
  2409.   __ret = __p0 + __p1;
  2410.   return __ret;
  2411. }
  2412. #ifdef __LITTLE_ENDIAN__
  2413. __ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
  2414.   int16x4_t __ret;
  2415.   __ret = __p0 + __p1;
  2416.   return __ret;
  2417. }
  2418. #else
  2419. __ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
  2420.   int16x4_t __ret;
  2421.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2422.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2423.   __ret = __rev0 + __rev1;
  2424.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2425.   return __ret;
  2426. }
  2427. #endif
  2428.  
  2429. #ifdef __LITTLE_ENDIAN__
  2430. __ai poly8x8_t vadd_p8(poly8x8_t __p0, poly8x8_t __p1) {
  2431.   poly8x8_t __ret;
  2432.   __ret = (poly8x8_t) __builtin_neon_vadd_v((int8x8_t)__p0, (int8x8_t)__p1, 4);
  2433.   return __ret;
  2434. }
  2435. #else
  2436. __ai poly8x8_t vadd_p8(poly8x8_t __p0, poly8x8_t __p1) {
  2437.   poly8x8_t __ret;
  2438.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2439.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2440.   __ret = (poly8x8_t) __builtin_neon_vadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 4);
  2441.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2442.   return __ret;
  2443. }
  2444. #endif
  2445.  
  2446. __ai poly64x1_t vadd_p64(poly64x1_t __p0, poly64x1_t __p1) {
  2447.   poly64x1_t __ret;
  2448.   __ret = (poly64x1_t) __builtin_neon_vadd_v((int8x8_t)__p0, (int8x8_t)__p1, 6);
  2449.   return __ret;
  2450. }
  2451. #ifdef __LITTLE_ENDIAN__
  2452. __ai poly16x4_t vadd_p16(poly16x4_t __p0, poly16x4_t __p1) {
  2453.   poly16x4_t __ret;
  2454.   __ret = (poly16x4_t) __builtin_neon_vadd_v((int8x8_t)__p0, (int8x8_t)__p1, 5);
  2455.   return __ret;
  2456. }
  2457. #else
  2458. __ai poly16x4_t vadd_p16(poly16x4_t __p0, poly16x4_t __p1) {
  2459.   poly16x4_t __ret;
  2460.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2461.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2462.   __ret = (poly16x4_t) __builtin_neon_vadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 5);
  2463.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2464.   return __ret;
  2465. }
  2466. #endif
  2467.  
  2468. #ifdef __LITTLE_ENDIAN__
  2469. __ai poly8x16_t vaddq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  2470.   poly8x16_t __ret;
  2471.   __ret = (poly8x16_t) __builtin_neon_vaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 36);
  2472.   return __ret;
  2473. }
  2474. #else
  2475. __ai poly8x16_t vaddq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  2476.   poly8x16_t __ret;
  2477.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2478.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2479.   __ret = (poly8x16_t) __builtin_neon_vaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 36);
  2480.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2481.   return __ret;
  2482. }
  2483. #endif
  2484.  
  2485. #ifdef __LITTLE_ENDIAN__
  2486. __ai poly64x2_t vaddq_p64(poly64x2_t __p0, poly64x2_t __p1) {
  2487.   poly64x2_t __ret;
  2488.   __ret = (poly64x2_t) __builtin_neon_vaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 38);
  2489.   return __ret;
  2490. }
  2491. #else
  2492. __ai poly64x2_t vaddq_p64(poly64x2_t __p0, poly64x2_t __p1) {
  2493.   poly64x2_t __ret;
  2494.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2495.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2496.   __ret = (poly64x2_t) __builtin_neon_vaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 38);
  2497.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2498.   return __ret;
  2499. }
  2500. #endif
  2501.  
  2502. #ifdef __LITTLE_ENDIAN__
  2503. __ai poly16x8_t vaddq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  2504.   poly16x8_t __ret;
  2505.   __ret = (poly16x8_t) __builtin_neon_vaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 37);
  2506.   return __ret;
  2507. }
  2508. #else
  2509. __ai poly16x8_t vaddq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  2510.   poly16x8_t __ret;
  2511.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2512.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2513.   __ret = (poly16x8_t) __builtin_neon_vaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 37);
  2514.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2515.   return __ret;
  2516. }
  2517. #endif
  2518.  
  2519. #ifdef __LITTLE_ENDIAN__
  2520. __ai uint16x4_t vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  2521.   uint16x4_t __ret;
  2522.   __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  2523.   return __ret;
  2524. }
  2525. #else
  2526. __ai uint16x4_t vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  2527.   uint16x4_t __ret;
  2528.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2529.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2530.   __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
  2531.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2532.   return __ret;
  2533. }
  2534. __ai uint16x4_t __noswap_vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  2535.   uint16x4_t __ret;
  2536.   __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  2537.   return __ret;
  2538. }
  2539. #endif
  2540.  
  2541. #ifdef __LITTLE_ENDIAN__
  2542. __ai uint32x2_t vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  2543.   uint32x2_t __ret;
  2544.   __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  2545.   return __ret;
  2546. }
  2547. #else
  2548. __ai uint32x2_t vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  2549.   uint32x2_t __ret;
  2550.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2551.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2552.   __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
  2553.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2554.   return __ret;
  2555. }
  2556. __ai uint32x2_t __noswap_vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  2557.   uint32x2_t __ret;
  2558.   __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  2559.   return __ret;
  2560. }
  2561. #endif
  2562.  
  2563. #ifdef __LITTLE_ENDIAN__
  2564. __ai uint8x8_t vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  2565.   uint8x8_t __ret;
  2566.   __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  2567.   return __ret;
  2568. }
  2569. #else
  2570. __ai uint8x8_t vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  2571.   uint8x8_t __ret;
  2572.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2573.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2574.   __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
  2575.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2576.   return __ret;
  2577. }
  2578. __ai uint8x8_t __noswap_vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  2579.   uint8x8_t __ret;
  2580.   __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  2581.   return __ret;
  2582. }
  2583. #endif
  2584.  
  2585. #ifdef __LITTLE_ENDIAN__
  2586. __ai int16x4_t vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
  2587.   int16x4_t __ret;
  2588.   __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  2589.   return __ret;
  2590. }
  2591. #else
  2592. __ai int16x4_t vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
  2593.   int16x4_t __ret;
  2594.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2595.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2596.   __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
  2597.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2598.   return __ret;
  2599. }
  2600. __ai int16x4_t __noswap_vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
  2601.   int16x4_t __ret;
  2602.   __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  2603.   return __ret;
  2604. }
  2605. #endif
  2606.  
  2607. #ifdef __LITTLE_ENDIAN__
  2608. __ai int32x2_t vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
  2609.   int32x2_t __ret;
  2610.   __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  2611.   return __ret;
  2612. }
  2613. #else
  2614. __ai int32x2_t vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
  2615.   int32x2_t __ret;
  2616.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2617.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2618.   __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
  2619.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2620.   return __ret;
  2621. }
  2622. __ai int32x2_t __noswap_vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
  2623.   int32x2_t __ret;
  2624.   __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  2625.   return __ret;
  2626. }
  2627. #endif
  2628.  
  2629. #ifdef __LITTLE_ENDIAN__
  2630. __ai int8x8_t vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
  2631.   int8x8_t __ret;
  2632.   __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  2633.   return __ret;
  2634. }
  2635. #else
  2636. __ai int8x8_t vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
  2637.   int8x8_t __ret;
  2638.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2639.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2640.   __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
  2641.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2642.   return __ret;
  2643. }
  2644. __ai int8x8_t __noswap_vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
  2645.   int8x8_t __ret;
  2646.   __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  2647.   return __ret;
  2648. }
  2649. #endif
  2650.  
  2651. #ifdef __LITTLE_ENDIAN__
  2652. __ai uint8x16_t vandq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  2653.   uint8x16_t __ret;
  2654.   __ret = __p0 & __p1;
  2655.   return __ret;
  2656. }
  2657. #else
  2658. __ai uint8x16_t vandq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  2659.   uint8x16_t __ret;
  2660.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2661.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2662.   __ret = __rev0 & __rev1;
  2663.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2664.   return __ret;
  2665. }
  2666. #endif
  2667.  
  2668. #ifdef __LITTLE_ENDIAN__
  2669. __ai uint32x4_t vandq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  2670.   uint32x4_t __ret;
  2671.   __ret = __p0 & __p1;
  2672.   return __ret;
  2673. }
  2674. #else
  2675. __ai uint32x4_t vandq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  2676.   uint32x4_t __ret;
  2677.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2678.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2679.   __ret = __rev0 & __rev1;
  2680.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2681.   return __ret;
  2682. }
  2683. #endif
  2684.  
  2685. #ifdef __LITTLE_ENDIAN__
  2686. __ai uint64x2_t vandq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  2687.   uint64x2_t __ret;
  2688.   __ret = __p0 & __p1;
  2689.   return __ret;
  2690. }
  2691. #else
  2692. __ai uint64x2_t vandq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  2693.   uint64x2_t __ret;
  2694.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2695.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2696.   __ret = __rev0 & __rev1;
  2697.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2698.   return __ret;
  2699. }
  2700. #endif
  2701.  
  2702. #ifdef __LITTLE_ENDIAN__
  2703. __ai uint16x8_t vandq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  2704.   uint16x8_t __ret;
  2705.   __ret = __p0 & __p1;
  2706.   return __ret;
  2707. }
  2708. #else
  2709. __ai uint16x8_t vandq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  2710.   uint16x8_t __ret;
  2711.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2712.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2713.   __ret = __rev0 & __rev1;
  2714.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2715.   return __ret;
  2716. }
  2717. #endif
  2718.  
  2719. #ifdef __LITTLE_ENDIAN__
  2720. __ai int8x16_t vandq_s8(int8x16_t __p0, int8x16_t __p1) {
  2721.   int8x16_t __ret;
  2722.   __ret = __p0 & __p1;
  2723.   return __ret;
  2724. }
  2725. #else
  2726. __ai int8x16_t vandq_s8(int8x16_t __p0, int8x16_t __p1) {
  2727.   int8x16_t __ret;
  2728.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2729.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2730.   __ret = __rev0 & __rev1;
  2731.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2732.   return __ret;
  2733. }
  2734. #endif
  2735.  
  2736. #ifdef __LITTLE_ENDIAN__
  2737. __ai int32x4_t vandq_s32(int32x4_t __p0, int32x4_t __p1) {
  2738.   int32x4_t __ret;
  2739.   __ret = __p0 & __p1;
  2740.   return __ret;
  2741. }
  2742. #else
  2743. __ai int32x4_t vandq_s32(int32x4_t __p0, int32x4_t __p1) {
  2744.   int32x4_t __ret;
  2745.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2746.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2747.   __ret = __rev0 & __rev1;
  2748.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2749.   return __ret;
  2750. }
  2751. #endif
  2752.  
  2753. #ifdef __LITTLE_ENDIAN__
  2754. __ai int64x2_t vandq_s64(int64x2_t __p0, int64x2_t __p1) {
  2755.   int64x2_t __ret;
  2756.   __ret = __p0 & __p1;
  2757.   return __ret;
  2758. }
  2759. #else
  2760. __ai int64x2_t vandq_s64(int64x2_t __p0, int64x2_t __p1) {
  2761.   int64x2_t __ret;
  2762.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2763.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2764.   __ret = __rev0 & __rev1;
  2765.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2766.   return __ret;
  2767. }
  2768. #endif
  2769.  
  2770. #ifdef __LITTLE_ENDIAN__
  2771. __ai int16x8_t vandq_s16(int16x8_t __p0, int16x8_t __p1) {
  2772.   int16x8_t __ret;
  2773.   __ret = __p0 & __p1;
  2774.   return __ret;
  2775. }
  2776. #else
  2777. __ai int16x8_t vandq_s16(int16x8_t __p0, int16x8_t __p1) {
  2778.   int16x8_t __ret;
  2779.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2780.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2781.   __ret = __rev0 & __rev1;
  2782.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2783.   return __ret;
  2784. }
  2785. #endif
  2786.  
  2787. #ifdef __LITTLE_ENDIAN__
  2788. __ai uint8x8_t vand_u8(uint8x8_t __p0, uint8x8_t __p1) {
  2789.   uint8x8_t __ret;
  2790.   __ret = __p0 & __p1;
  2791.   return __ret;
  2792. }
  2793. #else
  2794. __ai uint8x8_t vand_u8(uint8x8_t __p0, uint8x8_t __p1) {
  2795.   uint8x8_t __ret;
  2796.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2797.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2798.   __ret = __rev0 & __rev1;
  2799.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2800.   return __ret;
  2801. }
  2802. #endif
  2803.  
  2804. #ifdef __LITTLE_ENDIAN__
  2805. __ai uint32x2_t vand_u32(uint32x2_t __p0, uint32x2_t __p1) {
  2806.   uint32x2_t __ret;
  2807.   __ret = __p0 & __p1;
  2808.   return __ret;
  2809. }
  2810. #else
  2811. __ai uint32x2_t vand_u32(uint32x2_t __p0, uint32x2_t __p1) {
  2812.   uint32x2_t __ret;
  2813.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2814.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2815.   __ret = __rev0 & __rev1;
  2816.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2817.   return __ret;
  2818. }
  2819. #endif
  2820.  
  2821. __ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
  2822.   uint64x1_t __ret;
  2823.   __ret = __p0 & __p1;
  2824.   return __ret;
  2825. }
  2826. #ifdef __LITTLE_ENDIAN__
  2827. __ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
  2828.   uint16x4_t __ret;
  2829.   __ret = __p0 & __p1;
  2830.   return __ret;
  2831. }
  2832. #else
  2833. __ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
  2834.   uint16x4_t __ret;
  2835.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2836.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2837.   __ret = __rev0 & __rev1;
  2838.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2839.   return __ret;
  2840. }
  2841. #endif
  2842.  
  2843. #ifdef __LITTLE_ENDIAN__
  2844. __ai int8x8_t vand_s8(int8x8_t __p0, int8x8_t __p1) {
  2845.   int8x8_t __ret;
  2846.   __ret = __p0 & __p1;
  2847.   return __ret;
  2848. }
  2849. #else
  2850. __ai int8x8_t vand_s8(int8x8_t __p0, int8x8_t __p1) {
  2851.   int8x8_t __ret;
  2852.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2853.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2854.   __ret = __rev0 & __rev1;
  2855.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2856.   return __ret;
  2857. }
  2858. #endif
  2859.  
  2860. #ifdef __LITTLE_ENDIAN__
  2861. __ai int32x2_t vand_s32(int32x2_t __p0, int32x2_t __p1) {
  2862.   int32x2_t __ret;
  2863.   __ret = __p0 & __p1;
  2864.   return __ret;
  2865. }
  2866. #else
  2867. __ai int32x2_t vand_s32(int32x2_t __p0, int32x2_t __p1) {
  2868.   int32x2_t __ret;
  2869.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2870.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2871.   __ret = __rev0 & __rev1;
  2872.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2873.   return __ret;
  2874. }
  2875. #endif
  2876.  
  2877. __ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
  2878.   int64x1_t __ret;
  2879.   __ret = __p0 & __p1;
  2880.   return __ret;
  2881. }
  2882. #ifdef __LITTLE_ENDIAN__
  2883. __ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
  2884.   int16x4_t __ret;
  2885.   __ret = __p0 & __p1;
  2886.   return __ret;
  2887. }
  2888. #else
  2889. __ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
  2890.   int16x4_t __ret;
  2891.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2892.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2893.   __ret = __rev0 & __rev1;
  2894.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2895.   return __ret;
  2896. }
  2897. #endif
  2898.  
  2899. #ifdef __LITTLE_ENDIAN__
  2900. __ai uint8x16_t vbicq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  2901.   uint8x16_t __ret;
  2902.   __ret = __p0 & ~__p1;
  2903.   return __ret;
  2904. }
  2905. #else
  2906. __ai uint8x16_t vbicq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  2907.   uint8x16_t __ret;
  2908.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2909.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2910.   __ret = __rev0 & ~__rev1;
  2911.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2912.   return __ret;
  2913. }
  2914. #endif
  2915.  
  2916. #ifdef __LITTLE_ENDIAN__
  2917. __ai uint32x4_t vbicq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  2918.   uint32x4_t __ret;
  2919.   __ret = __p0 & ~__p1;
  2920.   return __ret;
  2921. }
  2922. #else
  2923. __ai uint32x4_t vbicq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  2924.   uint32x4_t __ret;
  2925.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2926.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2927.   __ret = __rev0 & ~__rev1;
  2928.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2929.   return __ret;
  2930. }
  2931. #endif
  2932.  
  2933. #ifdef __LITTLE_ENDIAN__
  2934. __ai uint64x2_t vbicq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  2935.   uint64x2_t __ret;
  2936.   __ret = __p0 & ~__p1;
  2937.   return __ret;
  2938. }
  2939. #else
  2940. __ai uint64x2_t vbicq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  2941.   uint64x2_t __ret;
  2942.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  2943.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  2944.   __ret = __rev0 & ~__rev1;
  2945.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  2946.   return __ret;
  2947. }
  2948. #endif
  2949.  
  2950. #ifdef __LITTLE_ENDIAN__
  2951. __ai uint16x8_t vbicq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  2952.   uint16x8_t __ret;
  2953.   __ret = __p0 & ~__p1;
  2954.   return __ret;
  2955. }
  2956. #else
  2957. __ai uint16x8_t vbicq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  2958.   uint16x8_t __ret;
  2959.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  2960.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  2961.   __ret = __rev0 & ~__rev1;
  2962.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  2963.   return __ret;
  2964. }
  2965. #endif
  2966.  
  2967. #ifdef __LITTLE_ENDIAN__
  2968. __ai int8x16_t vbicq_s8(int8x16_t __p0, int8x16_t __p1) {
  2969.   int8x16_t __ret;
  2970.   __ret = __p0 & ~__p1;
  2971.   return __ret;
  2972. }
  2973. #else
  2974. __ai int8x16_t vbicq_s8(int8x16_t __p0, int8x16_t __p1) {
  2975.   int8x16_t __ret;
  2976.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2977.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2978.   __ret = __rev0 & ~__rev1;
  2979.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  2980.   return __ret;
  2981. }
  2982. #endif
  2983.  
  2984. #ifdef __LITTLE_ENDIAN__
  2985. __ai int32x4_t vbicq_s32(int32x4_t __p0, int32x4_t __p1) {
  2986.   int32x4_t __ret;
  2987.   __ret = __p0 & ~__p1;
  2988.   return __ret;
  2989. }
  2990. #else
  2991. __ai int32x4_t vbicq_s32(int32x4_t __p0, int32x4_t __p1) {
  2992.   int32x4_t __ret;
  2993.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  2994.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  2995.   __ret = __rev0 & ~__rev1;
  2996.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  2997.   return __ret;
  2998. }
  2999. #endif
  3000.  
  3001. #ifdef __LITTLE_ENDIAN__
  3002. __ai int64x2_t vbicq_s64(int64x2_t __p0, int64x2_t __p1) {
  3003.   int64x2_t __ret;
  3004.   __ret = __p0 & ~__p1;
  3005.   return __ret;
  3006. }
  3007. #else
  3008. __ai int64x2_t vbicq_s64(int64x2_t __p0, int64x2_t __p1) {
  3009.   int64x2_t __ret;
  3010.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3011.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3012.   __ret = __rev0 & ~__rev1;
  3013.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3014.   return __ret;
  3015. }
  3016. #endif
  3017.  
  3018. #ifdef __LITTLE_ENDIAN__
  3019. __ai int16x8_t vbicq_s16(int16x8_t __p0, int16x8_t __p1) {
  3020.   int16x8_t __ret;
  3021.   __ret = __p0 & ~__p1;
  3022.   return __ret;
  3023. }
  3024. #else
  3025. __ai int16x8_t vbicq_s16(int16x8_t __p0, int16x8_t __p1) {
  3026.   int16x8_t __ret;
  3027.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3028.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3029.   __ret = __rev0 & ~__rev1;
  3030.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3031.   return __ret;
  3032. }
  3033. #endif
  3034.  
  3035. #ifdef __LITTLE_ENDIAN__
  3036. __ai uint8x8_t vbic_u8(uint8x8_t __p0, uint8x8_t __p1) {
  3037.   uint8x8_t __ret;
  3038.   __ret = __p0 & ~__p1;
  3039.   return __ret;
  3040. }
  3041. #else
  3042. __ai uint8x8_t vbic_u8(uint8x8_t __p0, uint8x8_t __p1) {
  3043.   uint8x8_t __ret;
  3044.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3045.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3046.   __ret = __rev0 & ~__rev1;
  3047.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3048.   return __ret;
  3049. }
  3050. #endif
  3051.  
  3052. #ifdef __LITTLE_ENDIAN__
  3053. __ai uint32x2_t vbic_u32(uint32x2_t __p0, uint32x2_t __p1) {
  3054.   uint32x2_t __ret;
  3055.   __ret = __p0 & ~__p1;
  3056.   return __ret;
  3057. }
  3058. #else
  3059. __ai uint32x2_t vbic_u32(uint32x2_t __p0, uint32x2_t __p1) {
  3060.   uint32x2_t __ret;
  3061.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3062.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3063.   __ret = __rev0 & ~__rev1;
  3064.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3065.   return __ret;
  3066. }
  3067. #endif
  3068.  
  3069. __ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
  3070.   uint64x1_t __ret;
  3071.   __ret = __p0 & ~__p1;
  3072.   return __ret;
  3073. }
  3074. #ifdef __LITTLE_ENDIAN__
  3075. __ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
  3076.   uint16x4_t __ret;
  3077.   __ret = __p0 & ~__p1;
  3078.   return __ret;
  3079. }
  3080. #else
  3081. __ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
  3082.   uint16x4_t __ret;
  3083.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3084.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3085.   __ret = __rev0 & ~__rev1;
  3086.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3087.   return __ret;
  3088. }
  3089. #endif
  3090.  
  3091. #ifdef __LITTLE_ENDIAN__
  3092. __ai int8x8_t vbic_s8(int8x8_t __p0, int8x8_t __p1) {
  3093.   int8x8_t __ret;
  3094.   __ret = __p0 & ~__p1;
  3095.   return __ret;
  3096. }
  3097. #else
  3098. __ai int8x8_t vbic_s8(int8x8_t __p0, int8x8_t __p1) {
  3099.   int8x8_t __ret;
  3100.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3101.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3102.   __ret = __rev0 & ~__rev1;
  3103.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3104.   return __ret;
  3105. }
  3106. #endif
  3107.  
  3108. #ifdef __LITTLE_ENDIAN__
  3109. __ai int32x2_t vbic_s32(int32x2_t __p0, int32x2_t __p1) {
  3110.   int32x2_t __ret;
  3111.   __ret = __p0 & ~__p1;
  3112.   return __ret;
  3113. }
  3114. #else
  3115. __ai int32x2_t vbic_s32(int32x2_t __p0, int32x2_t __p1) {
  3116.   int32x2_t __ret;
  3117.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3118.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3119.   __ret = __rev0 & ~__rev1;
  3120.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3121.   return __ret;
  3122. }
  3123. #endif
  3124.  
  3125. __ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
  3126.   int64x1_t __ret;
  3127.   __ret = __p0 & ~__p1;
  3128.   return __ret;
  3129. }
  3130. #ifdef __LITTLE_ENDIAN__
  3131. __ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
  3132.   int16x4_t __ret;
  3133.   __ret = __p0 & ~__p1;
  3134.   return __ret;
  3135. }
  3136. #else
  3137. __ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
  3138.   int16x4_t __ret;
  3139.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3140.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3141.   __ret = __rev0 & ~__rev1;
  3142.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3143.   return __ret;
  3144. }
  3145. #endif
  3146.  
  3147. #ifdef __LITTLE_ENDIAN__
  3148. __ai poly8x8_t vbsl_p8(uint8x8_t __p0, poly8x8_t __p1, poly8x8_t __p2) {
  3149.   poly8x8_t __ret;
  3150.   __ret = (poly8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 4);
  3151.   return __ret;
  3152. }
  3153. #else
  3154. __ai poly8x8_t vbsl_p8(uint8x8_t __p0, poly8x8_t __p1, poly8x8_t __p2) {
  3155.   poly8x8_t __ret;
  3156.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3157.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3158.   poly8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  3159.   __ret = (poly8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 4);
  3160.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3161.   return __ret;
  3162. }
  3163. #endif
  3164.  
  3165. #ifdef __LITTLE_ENDIAN__
  3166. __ai poly16x4_t vbsl_p16(uint16x4_t __p0, poly16x4_t __p1, poly16x4_t __p2) {
  3167.   poly16x4_t __ret;
  3168.   __ret = (poly16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 5);
  3169.   return __ret;
  3170. }
  3171. #else
  3172. __ai poly16x4_t vbsl_p16(uint16x4_t __p0, poly16x4_t __p1, poly16x4_t __p2) {
  3173.   poly16x4_t __ret;
  3174.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3175.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3176.   poly16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  3177.   __ret = (poly16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 5);
  3178.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3179.   return __ret;
  3180. }
  3181. #endif
  3182.  
  3183. #ifdef __LITTLE_ENDIAN__
  3184. __ai poly8x16_t vbslq_p8(uint8x16_t __p0, poly8x16_t __p1, poly8x16_t __p2) {
  3185.   poly8x16_t __ret;
  3186.   __ret = (poly8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 36);
  3187.   return __ret;
  3188. }
  3189. #else
  3190. __ai poly8x16_t vbslq_p8(uint8x16_t __p0, poly8x16_t __p1, poly8x16_t __p2) {
  3191.   poly8x16_t __ret;
  3192.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3193.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3194.   poly8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3195.   __ret = (poly8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 36);
  3196.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3197.   return __ret;
  3198. }
  3199. #endif
  3200.  
  3201. #ifdef __LITTLE_ENDIAN__
  3202. __ai poly16x8_t vbslq_p16(uint16x8_t __p0, poly16x8_t __p1, poly16x8_t __p2) {
  3203.   poly16x8_t __ret;
  3204.   __ret = (poly16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 37);
  3205.   return __ret;
  3206. }
  3207. #else
  3208. __ai poly16x8_t vbslq_p16(uint16x8_t __p0, poly16x8_t __p1, poly16x8_t __p2) {
  3209.   poly16x8_t __ret;
  3210.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3211.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3212.   poly16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  3213.   __ret = (poly16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 37);
  3214.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3215.   return __ret;
  3216. }
  3217. #endif
  3218.  
  3219. #ifdef __LITTLE_ENDIAN__
  3220. __ai uint8x16_t vbslq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  3221.   uint8x16_t __ret;
  3222.   __ret = (uint8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 48);
  3223.   return __ret;
  3224. }
  3225. #else
  3226. __ai uint8x16_t vbslq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  3227.   uint8x16_t __ret;
  3228.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3229.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3230.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3231.   __ret = (uint8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 48);
  3232.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3233.   return __ret;
  3234. }
  3235. #endif
  3236.  
  3237. #ifdef __LITTLE_ENDIAN__
  3238. __ai uint32x4_t vbslq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  3239.   uint32x4_t __ret;
  3240.   __ret = (uint32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  3241.   return __ret;
  3242. }
  3243. #else
  3244. __ai uint32x4_t vbslq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  3245.   uint32x4_t __ret;
  3246.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3247.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3248.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  3249.   __ret = (uint32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  3250.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3251.   return __ret;
  3252. }
  3253. #endif
  3254.  
  3255. #ifdef __LITTLE_ENDIAN__
  3256. __ai uint64x2_t vbslq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  3257.   uint64x2_t __ret;
  3258.   __ret = (uint64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 51);
  3259.   return __ret;
  3260. }
  3261. #else
  3262. __ai uint64x2_t vbslq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  3263.   uint64x2_t __ret;
  3264.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3265.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3266.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  3267.   __ret = (uint64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 51);
  3268.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3269.   return __ret;
  3270. }
  3271. #endif
  3272.  
  3273. #ifdef __LITTLE_ENDIAN__
  3274. __ai uint16x8_t vbslq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  3275.   uint16x8_t __ret;
  3276.   __ret = (uint16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 49);
  3277.   return __ret;
  3278. }
  3279. #else
  3280. __ai uint16x8_t vbslq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  3281.   uint16x8_t __ret;
  3282.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3283.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3284.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  3285.   __ret = (uint16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 49);
  3286.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3287.   return __ret;
  3288. }
  3289. #endif
  3290.  
  3291. #ifdef __LITTLE_ENDIAN__
  3292. __ai int8x16_t vbslq_s8(uint8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  3293.   int8x16_t __ret;
  3294.   __ret = (int8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 32);
  3295.   return __ret;
  3296. }
  3297. #else
  3298. __ai int8x16_t vbslq_s8(uint8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  3299.   int8x16_t __ret;
  3300.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3301.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3302.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3303.   __ret = (int8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 32);
  3304.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3305.   return __ret;
  3306. }
  3307. #endif
  3308.  
  3309. #ifdef __LITTLE_ENDIAN__
  3310. __ai float32x4_t vbslq_f32(uint32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  3311.   float32x4_t __ret;
  3312.   __ret = (float32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  3313.   return __ret;
  3314. }
  3315. #else
  3316. __ai float32x4_t vbslq_f32(uint32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  3317.   float32x4_t __ret;
  3318.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3319.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3320.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  3321.   __ret = (float32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  3322.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3323.   return __ret;
  3324. }
  3325. #endif
  3326.  
  3327. #ifdef __LITTLE_ENDIAN__
  3328. __ai int32x4_t vbslq_s32(uint32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  3329.   int32x4_t __ret;
  3330.   __ret = (int32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  3331.   return __ret;
  3332. }
  3333. #else
  3334. __ai int32x4_t vbslq_s32(uint32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  3335.   int32x4_t __ret;
  3336.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3337.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3338.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  3339.   __ret = (int32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  3340.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3341.   return __ret;
  3342. }
  3343. #endif
  3344.  
  3345. #ifdef __LITTLE_ENDIAN__
  3346. __ai int64x2_t vbslq_s64(uint64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  3347.   int64x2_t __ret;
  3348.   __ret = (int64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 35);
  3349.   return __ret;
  3350. }
  3351. #else
  3352. __ai int64x2_t vbslq_s64(uint64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  3353.   int64x2_t __ret;
  3354.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3355.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3356.   int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  3357.   __ret = (int64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 35);
  3358.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3359.   return __ret;
  3360. }
  3361. #endif
  3362.  
  3363. #ifdef __LITTLE_ENDIAN__
  3364. __ai int16x8_t vbslq_s16(uint16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  3365.   int16x8_t __ret;
  3366.   __ret = (int16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
  3367.   return __ret;
  3368. }
  3369. #else
  3370. __ai int16x8_t vbslq_s16(uint16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  3371.   int16x8_t __ret;
  3372.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3373.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3374.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  3375.   __ret = (int16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 33);
  3376.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3377.   return __ret;
  3378. }
  3379. #endif
  3380.  
  3381. #ifdef __LITTLE_ENDIAN__
  3382. __ai uint8x8_t vbsl_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  3383.   uint8x8_t __ret;
  3384.   __ret = (uint8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 16);
  3385.   return __ret;
  3386. }
  3387. #else
  3388. __ai uint8x8_t vbsl_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  3389.   uint8x8_t __ret;
  3390.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3391.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3392.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  3393.   __ret = (uint8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 16);
  3394.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3395.   return __ret;
  3396. }
  3397. #endif
  3398.  
  3399. #ifdef __LITTLE_ENDIAN__
  3400. __ai uint32x2_t vbsl_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  3401.   uint32x2_t __ret;
  3402.   __ret = (uint32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 18);
  3403.   return __ret;
  3404. }
  3405. #else
  3406. __ai uint32x2_t vbsl_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  3407.   uint32x2_t __ret;
  3408.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3409.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3410.   uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  3411.   __ret = (uint32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 18);
  3412.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3413.   return __ret;
  3414. }
  3415. #endif
  3416.  
  3417. __ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
  3418.   uint64x1_t __ret;
  3419.   __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
  3420.   return __ret;
  3421. }
  3422. #ifdef __LITTLE_ENDIAN__
  3423. __ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  3424.   uint16x4_t __ret;
  3425.   __ret = (uint16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 17);
  3426.   return __ret;
  3427. }
  3428. #else
  3429. __ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  3430.   uint16x4_t __ret;
  3431.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3432.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3433.   uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  3434.   __ret = (uint16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 17);
  3435.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3436.   return __ret;
  3437. }
  3438. #endif
  3439.  
  3440. #ifdef __LITTLE_ENDIAN__
  3441. __ai int8x8_t vbsl_s8(uint8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  3442.   int8x8_t __ret;
  3443.   __ret = (int8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 0);
  3444.   return __ret;
  3445. }
  3446. #else
  3447. __ai int8x8_t vbsl_s8(uint8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  3448.   int8x8_t __ret;
  3449.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3450.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3451.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  3452.   __ret = (int8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 0);
  3453.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3454.   return __ret;
  3455. }
  3456. #endif
  3457.  
  3458. #ifdef __LITTLE_ENDIAN__
  3459. __ai float32x2_t vbsl_f32(uint32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  3460.   float32x2_t __ret;
  3461.   __ret = (float32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  3462.   return __ret;
  3463. }
  3464. #else
  3465. __ai float32x2_t vbsl_f32(uint32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  3466.   float32x2_t __ret;
  3467.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3468.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3469.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  3470.   __ret = (float32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  3471.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3472.   return __ret;
  3473. }
  3474. #endif
  3475.  
  3476. #ifdef __LITTLE_ENDIAN__
  3477. __ai int32x2_t vbsl_s32(uint32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  3478.   int32x2_t __ret;
  3479.   __ret = (int32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  3480.   return __ret;
  3481. }
  3482. #else
  3483. __ai int32x2_t vbsl_s32(uint32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  3484.   int32x2_t __ret;
  3485.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3486.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3487.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  3488.   __ret = (int32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 2);
  3489.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3490.   return __ret;
  3491. }
  3492. #endif
  3493.  
  3494. __ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
  3495.   int64x1_t __ret;
  3496.   __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
  3497.   return __ret;
  3498. }
  3499. #ifdef __LITTLE_ENDIAN__
  3500. __ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  3501.   int16x4_t __ret;
  3502.   __ret = (int16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 1);
  3503.   return __ret;
  3504. }
  3505. #else
  3506. __ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  3507.   int16x4_t __ret;
  3508.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3509.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3510.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  3511.   __ret = (int16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 1);
  3512.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3513.   return __ret;
  3514. }
  3515. #endif
  3516.  
  3517. #ifdef __LITTLE_ENDIAN__
  3518. __ai uint32x4_t vcageq_f32(float32x4_t __p0, float32x4_t __p1) {
  3519.   uint32x4_t __ret;
  3520.   __ret = (uint32x4_t) __builtin_neon_vcageq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  3521.   return __ret;
  3522. }
  3523. #else
  3524. __ai uint32x4_t vcageq_f32(float32x4_t __p0, float32x4_t __p1) {
  3525.   uint32x4_t __ret;
  3526.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3527.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3528.   __ret = (uint32x4_t) __builtin_neon_vcageq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  3529.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3530.   return __ret;
  3531. }
  3532. #endif
  3533.  
  3534. #ifdef __LITTLE_ENDIAN__
  3535. __ai uint32x2_t vcage_f32(float32x2_t __p0, float32x2_t __p1) {
  3536.   uint32x2_t __ret;
  3537.   __ret = (uint32x2_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  3538.   return __ret;
  3539. }
  3540. #else
  3541. __ai uint32x2_t vcage_f32(float32x2_t __p0, float32x2_t __p1) {
  3542.   uint32x2_t __ret;
  3543.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3544.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3545.   __ret = (uint32x2_t) __builtin_neon_vcage_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  3546.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3547.   return __ret;
  3548. }
  3549. #endif
  3550.  
  3551. #ifdef __LITTLE_ENDIAN__
  3552. __ai uint32x4_t vcagtq_f32(float32x4_t __p0, float32x4_t __p1) {
  3553.   uint32x4_t __ret;
  3554.   __ret = (uint32x4_t) __builtin_neon_vcagtq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  3555.   return __ret;
  3556. }
  3557. #else
  3558. __ai uint32x4_t vcagtq_f32(float32x4_t __p0, float32x4_t __p1) {
  3559.   uint32x4_t __ret;
  3560.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3561.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3562.   __ret = (uint32x4_t) __builtin_neon_vcagtq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  3563.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3564.   return __ret;
  3565. }
  3566. #endif
  3567.  
  3568. #ifdef __LITTLE_ENDIAN__
  3569. __ai uint32x2_t vcagt_f32(float32x2_t __p0, float32x2_t __p1) {
  3570.   uint32x2_t __ret;
  3571.   __ret = (uint32x2_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  3572.   return __ret;
  3573. }
  3574. #else
  3575. __ai uint32x2_t vcagt_f32(float32x2_t __p0, float32x2_t __p1) {
  3576.   uint32x2_t __ret;
  3577.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3578.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3579.   __ret = (uint32x2_t) __builtin_neon_vcagt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  3580.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3581.   return __ret;
  3582. }
  3583. #endif
  3584.  
  3585. #ifdef __LITTLE_ENDIAN__
  3586. __ai uint32x4_t vcaleq_f32(float32x4_t __p0, float32x4_t __p1) {
  3587.   uint32x4_t __ret;
  3588.   __ret = (uint32x4_t) __builtin_neon_vcaleq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  3589.   return __ret;
  3590. }
  3591. #else
  3592. __ai uint32x4_t vcaleq_f32(float32x4_t __p0, float32x4_t __p1) {
  3593.   uint32x4_t __ret;
  3594.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3595.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3596.   __ret = (uint32x4_t) __builtin_neon_vcaleq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  3597.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3598.   return __ret;
  3599. }
  3600. #endif
  3601.  
  3602. #ifdef __LITTLE_ENDIAN__
  3603. __ai uint32x2_t vcale_f32(float32x2_t __p0, float32x2_t __p1) {
  3604.   uint32x2_t __ret;
  3605.   __ret = (uint32x2_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  3606.   return __ret;
  3607. }
  3608. #else
  3609. __ai uint32x2_t vcale_f32(float32x2_t __p0, float32x2_t __p1) {
  3610.   uint32x2_t __ret;
  3611.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3612.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3613.   __ret = (uint32x2_t) __builtin_neon_vcale_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  3614.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3615.   return __ret;
  3616. }
  3617. #endif
  3618.  
  3619. #ifdef __LITTLE_ENDIAN__
  3620. __ai uint32x4_t vcaltq_f32(float32x4_t __p0, float32x4_t __p1) {
  3621.   uint32x4_t __ret;
  3622.   __ret = (uint32x4_t) __builtin_neon_vcaltq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  3623.   return __ret;
  3624. }
  3625. #else
  3626. __ai uint32x4_t vcaltq_f32(float32x4_t __p0, float32x4_t __p1) {
  3627.   uint32x4_t __ret;
  3628.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3629.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3630.   __ret = (uint32x4_t) __builtin_neon_vcaltq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  3631.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3632.   return __ret;
  3633. }
  3634. #endif
  3635.  
  3636. #ifdef __LITTLE_ENDIAN__
  3637. __ai uint32x2_t vcalt_f32(float32x2_t __p0, float32x2_t __p1) {
  3638.   uint32x2_t __ret;
  3639.   __ret = (uint32x2_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  3640.   return __ret;
  3641. }
  3642. #else
  3643. __ai uint32x2_t vcalt_f32(float32x2_t __p0, float32x2_t __p1) {
  3644.   uint32x2_t __ret;
  3645.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3646.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3647.   __ret = (uint32x2_t) __builtin_neon_vcalt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  3648.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3649.   return __ret;
  3650. }
  3651. #endif
  3652.  
  3653. #ifdef __LITTLE_ENDIAN__
  3654. __ai uint8x8_t vceq_p8(poly8x8_t __p0, poly8x8_t __p1) {
  3655.   uint8x8_t __ret;
  3656.   __ret = (uint8x8_t)(__p0 == __p1);
  3657.   return __ret;
  3658. }
  3659. #else
  3660. __ai uint8x8_t vceq_p8(poly8x8_t __p0, poly8x8_t __p1) {
  3661.   uint8x8_t __ret;
  3662.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3663.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3664.   __ret = (uint8x8_t)(__rev0 == __rev1);
  3665.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3666.   return __ret;
  3667. }
  3668. #endif
  3669.  
  3670. #ifdef __LITTLE_ENDIAN__
  3671. __ai uint8x16_t vceqq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  3672.   uint8x16_t __ret;
  3673.   __ret = (uint8x16_t)(__p0 == __p1);
  3674.   return __ret;
  3675. }
  3676. #else
  3677. __ai uint8x16_t vceqq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  3678.   uint8x16_t __ret;
  3679.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3680.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3681.   __ret = (uint8x16_t)(__rev0 == __rev1);
  3682.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3683.   return __ret;
  3684. }
  3685. #endif
  3686.  
  3687. #ifdef __LITTLE_ENDIAN__
  3688. __ai uint8x16_t vceqq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  3689.   uint8x16_t __ret;
  3690.   __ret = (uint8x16_t)(__p0 == __p1);
  3691.   return __ret;
  3692. }
  3693. #else
  3694. __ai uint8x16_t vceqq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  3695.   uint8x16_t __ret;
  3696.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3697.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3698.   __ret = (uint8x16_t)(__rev0 == __rev1);
  3699.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3700.   return __ret;
  3701. }
  3702. #endif
  3703.  
  3704. #ifdef __LITTLE_ENDIAN__
  3705. __ai uint32x4_t vceqq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  3706.   uint32x4_t __ret;
  3707.   __ret = (uint32x4_t)(__p0 == __p1);
  3708.   return __ret;
  3709. }
  3710. #else
  3711. __ai uint32x4_t vceqq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  3712.   uint32x4_t __ret;
  3713.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3714.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3715.   __ret = (uint32x4_t)(__rev0 == __rev1);
  3716.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3717.   return __ret;
  3718. }
  3719. #endif
  3720.  
  3721. #ifdef __LITTLE_ENDIAN__
  3722. __ai uint16x8_t vceqq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  3723.   uint16x8_t __ret;
  3724.   __ret = (uint16x8_t)(__p0 == __p1);
  3725.   return __ret;
  3726. }
  3727. #else
  3728. __ai uint16x8_t vceqq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  3729.   uint16x8_t __ret;
  3730.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3731.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3732.   __ret = (uint16x8_t)(__rev0 == __rev1);
  3733.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3734.   return __ret;
  3735. }
  3736. #endif
  3737.  
  3738. #ifdef __LITTLE_ENDIAN__
  3739. __ai uint8x16_t vceqq_s8(int8x16_t __p0, int8x16_t __p1) {
  3740.   uint8x16_t __ret;
  3741.   __ret = (uint8x16_t)(__p0 == __p1);
  3742.   return __ret;
  3743. }
  3744. #else
  3745. __ai uint8x16_t vceqq_s8(int8x16_t __p0, int8x16_t __p1) {
  3746.   uint8x16_t __ret;
  3747.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3748.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3749.   __ret = (uint8x16_t)(__rev0 == __rev1);
  3750.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3751.   return __ret;
  3752. }
  3753. #endif
  3754.  
  3755. #ifdef __LITTLE_ENDIAN__
  3756. __ai uint32x4_t vceqq_f32(float32x4_t __p0, float32x4_t __p1) {
  3757.   uint32x4_t __ret;
  3758.   __ret = (uint32x4_t)(__p0 == __p1);
  3759.   return __ret;
  3760. }
  3761. #else
  3762. __ai uint32x4_t vceqq_f32(float32x4_t __p0, float32x4_t __p1) {
  3763.   uint32x4_t __ret;
  3764.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3765.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3766.   __ret = (uint32x4_t)(__rev0 == __rev1);
  3767.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3768.   return __ret;
  3769. }
  3770. #endif
  3771.  
  3772. #ifdef __LITTLE_ENDIAN__
  3773. __ai uint32x4_t vceqq_s32(int32x4_t __p0, int32x4_t __p1) {
  3774.   uint32x4_t __ret;
  3775.   __ret = (uint32x4_t)(__p0 == __p1);
  3776.   return __ret;
  3777. }
  3778. #else
  3779. __ai uint32x4_t vceqq_s32(int32x4_t __p0, int32x4_t __p1) {
  3780.   uint32x4_t __ret;
  3781.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3782.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3783.   __ret = (uint32x4_t)(__rev0 == __rev1);
  3784.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3785.   return __ret;
  3786. }
  3787. #endif
  3788.  
  3789. #ifdef __LITTLE_ENDIAN__
  3790. __ai uint16x8_t vceqq_s16(int16x8_t __p0, int16x8_t __p1) {
  3791.   uint16x8_t __ret;
  3792.   __ret = (uint16x8_t)(__p0 == __p1);
  3793.   return __ret;
  3794. }
  3795. #else
  3796. __ai uint16x8_t vceqq_s16(int16x8_t __p0, int16x8_t __p1) {
  3797.   uint16x8_t __ret;
  3798.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3799.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3800.   __ret = (uint16x8_t)(__rev0 == __rev1);
  3801.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3802.   return __ret;
  3803. }
  3804. #endif
  3805.  
  3806. #ifdef __LITTLE_ENDIAN__
  3807. __ai uint8x8_t vceq_u8(uint8x8_t __p0, uint8x8_t __p1) {
  3808.   uint8x8_t __ret;
  3809.   __ret = (uint8x8_t)(__p0 == __p1);
  3810.   return __ret;
  3811. }
  3812. #else
  3813. __ai uint8x8_t vceq_u8(uint8x8_t __p0, uint8x8_t __p1) {
  3814.   uint8x8_t __ret;
  3815.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3816.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3817.   __ret = (uint8x8_t)(__rev0 == __rev1);
  3818.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3819.   return __ret;
  3820. }
  3821. #endif
  3822.  
  3823. #ifdef __LITTLE_ENDIAN__
  3824. __ai uint32x2_t vceq_u32(uint32x2_t __p0, uint32x2_t __p1) {
  3825.   uint32x2_t __ret;
  3826.   __ret = (uint32x2_t)(__p0 == __p1);
  3827.   return __ret;
  3828. }
  3829. #else
  3830. __ai uint32x2_t vceq_u32(uint32x2_t __p0, uint32x2_t __p1) {
  3831.   uint32x2_t __ret;
  3832.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3833.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3834.   __ret = (uint32x2_t)(__rev0 == __rev1);
  3835.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3836.   return __ret;
  3837. }
  3838. #endif
  3839.  
  3840. #ifdef __LITTLE_ENDIAN__
  3841. __ai uint16x4_t vceq_u16(uint16x4_t __p0, uint16x4_t __p1) {
  3842.   uint16x4_t __ret;
  3843.   __ret = (uint16x4_t)(__p0 == __p1);
  3844.   return __ret;
  3845. }
  3846. #else
  3847. __ai uint16x4_t vceq_u16(uint16x4_t __p0, uint16x4_t __p1) {
  3848.   uint16x4_t __ret;
  3849.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3850.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3851.   __ret = (uint16x4_t)(__rev0 == __rev1);
  3852.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3853.   return __ret;
  3854. }
  3855. #endif
  3856.  
  3857. #ifdef __LITTLE_ENDIAN__
  3858. __ai uint8x8_t vceq_s8(int8x8_t __p0, int8x8_t __p1) {
  3859.   uint8x8_t __ret;
  3860.   __ret = (uint8x8_t)(__p0 == __p1);
  3861.   return __ret;
  3862. }
  3863. #else
  3864. __ai uint8x8_t vceq_s8(int8x8_t __p0, int8x8_t __p1) {
  3865.   uint8x8_t __ret;
  3866.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3867.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3868.   __ret = (uint8x8_t)(__rev0 == __rev1);
  3869.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3870.   return __ret;
  3871. }
  3872. #endif
  3873.  
  3874. #ifdef __LITTLE_ENDIAN__
  3875. __ai uint32x2_t vceq_f32(float32x2_t __p0, float32x2_t __p1) {
  3876.   uint32x2_t __ret;
  3877.   __ret = (uint32x2_t)(__p0 == __p1);
  3878.   return __ret;
  3879. }
  3880. #else
  3881. __ai uint32x2_t vceq_f32(float32x2_t __p0, float32x2_t __p1) {
  3882.   uint32x2_t __ret;
  3883.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3884.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3885.   __ret = (uint32x2_t)(__rev0 == __rev1);
  3886.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3887.   return __ret;
  3888. }
  3889. #endif
  3890.  
  3891. #ifdef __LITTLE_ENDIAN__
  3892. __ai uint32x2_t vceq_s32(int32x2_t __p0, int32x2_t __p1) {
  3893.   uint32x2_t __ret;
  3894.   __ret = (uint32x2_t)(__p0 == __p1);
  3895.   return __ret;
  3896. }
  3897. #else
  3898. __ai uint32x2_t vceq_s32(int32x2_t __p0, int32x2_t __p1) {
  3899.   uint32x2_t __ret;
  3900.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  3901.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  3902.   __ret = (uint32x2_t)(__rev0 == __rev1);
  3903.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  3904.   return __ret;
  3905. }
  3906. #endif
  3907.  
  3908. #ifdef __LITTLE_ENDIAN__
  3909. __ai uint16x4_t vceq_s16(int16x4_t __p0, int16x4_t __p1) {
  3910.   uint16x4_t __ret;
  3911.   __ret = (uint16x4_t)(__p0 == __p1);
  3912.   return __ret;
  3913. }
  3914. #else
  3915. __ai uint16x4_t vceq_s16(int16x4_t __p0, int16x4_t __p1) {
  3916.   uint16x4_t __ret;
  3917.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3918.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3919.   __ret = (uint16x4_t)(__rev0 == __rev1);
  3920.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3921.   return __ret;
  3922. }
  3923. #endif
  3924.  
  3925. #ifdef __LITTLE_ENDIAN__
  3926. __ai uint8x16_t vcgeq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  3927.   uint8x16_t __ret;
  3928.   __ret = (uint8x16_t)(__p0 >= __p1);
  3929.   return __ret;
  3930. }
  3931. #else
  3932. __ai uint8x16_t vcgeq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  3933.   uint8x16_t __ret;
  3934.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3935.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3936.   __ret = (uint8x16_t)(__rev0 >= __rev1);
  3937.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3938.   return __ret;
  3939. }
  3940. #endif
  3941.  
  3942. #ifdef __LITTLE_ENDIAN__
  3943. __ai uint32x4_t vcgeq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  3944.   uint32x4_t __ret;
  3945.   __ret = (uint32x4_t)(__p0 >= __p1);
  3946.   return __ret;
  3947. }
  3948. #else
  3949. __ai uint32x4_t vcgeq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  3950.   uint32x4_t __ret;
  3951.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  3952.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  3953.   __ret = (uint32x4_t)(__rev0 >= __rev1);
  3954.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  3955.   return __ret;
  3956. }
  3957. #endif
  3958.  
  3959. #ifdef __LITTLE_ENDIAN__
  3960. __ai uint16x8_t vcgeq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  3961.   uint16x8_t __ret;
  3962.   __ret = (uint16x8_t)(__p0 >= __p1);
  3963.   return __ret;
  3964. }
  3965. #else
  3966. __ai uint16x8_t vcgeq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  3967.   uint16x8_t __ret;
  3968.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  3969.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  3970.   __ret = (uint16x8_t)(__rev0 >= __rev1);
  3971.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  3972.   return __ret;
  3973. }
  3974. #endif
  3975.  
  3976. #ifdef __LITTLE_ENDIAN__
  3977. __ai uint8x16_t vcgeq_s8(int8x16_t __p0, int8x16_t __p1) {
  3978.   uint8x16_t __ret;
  3979.   __ret = (uint8x16_t)(__p0 >= __p1);
  3980.   return __ret;
  3981. }
  3982. #else
  3983. __ai uint8x16_t vcgeq_s8(int8x16_t __p0, int8x16_t __p1) {
  3984.   uint8x16_t __ret;
  3985.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3986.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3987.   __ret = (uint8x16_t)(__rev0 >= __rev1);
  3988.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  3989.   return __ret;
  3990. }
  3991. #endif
  3992.  
  3993. #ifdef __LITTLE_ENDIAN__
  3994. __ai uint32x4_t vcgeq_f32(float32x4_t __p0, float32x4_t __p1) {
  3995.   uint32x4_t __ret;
  3996.   __ret = (uint32x4_t)(__p0 >= __p1);
  3997.   return __ret;
  3998. }
  3999. #else
  4000. __ai uint32x4_t vcgeq_f32(float32x4_t __p0, float32x4_t __p1) {
  4001.   uint32x4_t __ret;
  4002.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4003.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4004.   __ret = (uint32x4_t)(__rev0 >= __rev1);
  4005.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4006.   return __ret;
  4007. }
  4008. #endif
  4009.  
  4010. #ifdef __LITTLE_ENDIAN__
  4011. __ai uint32x4_t vcgeq_s32(int32x4_t __p0, int32x4_t __p1) {
  4012.   uint32x4_t __ret;
  4013.   __ret = (uint32x4_t)(__p0 >= __p1);
  4014.   return __ret;
  4015. }
  4016. #else
  4017. __ai uint32x4_t vcgeq_s32(int32x4_t __p0, int32x4_t __p1) {
  4018.   uint32x4_t __ret;
  4019.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4020.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4021.   __ret = (uint32x4_t)(__rev0 >= __rev1);
  4022.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4023.   return __ret;
  4024. }
  4025. #endif
  4026.  
  4027. #ifdef __LITTLE_ENDIAN__
  4028. __ai uint16x8_t vcgeq_s16(int16x8_t __p0, int16x8_t __p1) {
  4029.   uint16x8_t __ret;
  4030.   __ret = (uint16x8_t)(__p0 >= __p1);
  4031.   return __ret;
  4032. }
  4033. #else
  4034. __ai uint16x8_t vcgeq_s16(int16x8_t __p0, int16x8_t __p1) {
  4035.   uint16x8_t __ret;
  4036.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4037.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4038.   __ret = (uint16x8_t)(__rev0 >= __rev1);
  4039.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4040.   return __ret;
  4041. }
  4042. #endif
  4043.  
  4044. #ifdef __LITTLE_ENDIAN__
  4045. __ai uint8x8_t vcge_u8(uint8x8_t __p0, uint8x8_t __p1) {
  4046.   uint8x8_t __ret;
  4047.   __ret = (uint8x8_t)(__p0 >= __p1);
  4048.   return __ret;
  4049. }
  4050. #else
  4051. __ai uint8x8_t vcge_u8(uint8x8_t __p0, uint8x8_t __p1) {
  4052.   uint8x8_t __ret;
  4053.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4054.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4055.   __ret = (uint8x8_t)(__rev0 >= __rev1);
  4056.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4057.   return __ret;
  4058. }
  4059. #endif
  4060.  
  4061. #ifdef __LITTLE_ENDIAN__
  4062. __ai uint32x2_t vcge_u32(uint32x2_t __p0, uint32x2_t __p1) {
  4063.   uint32x2_t __ret;
  4064.   __ret = (uint32x2_t)(__p0 >= __p1);
  4065.   return __ret;
  4066. }
  4067. #else
  4068. __ai uint32x2_t vcge_u32(uint32x2_t __p0, uint32x2_t __p1) {
  4069.   uint32x2_t __ret;
  4070.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4071.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4072.   __ret = (uint32x2_t)(__rev0 >= __rev1);
  4073.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4074.   return __ret;
  4075. }
  4076. #endif
  4077.  
  4078. #ifdef __LITTLE_ENDIAN__
  4079. __ai uint16x4_t vcge_u16(uint16x4_t __p0, uint16x4_t __p1) {
  4080.   uint16x4_t __ret;
  4081.   __ret = (uint16x4_t)(__p0 >= __p1);
  4082.   return __ret;
  4083. }
  4084. #else
  4085. __ai uint16x4_t vcge_u16(uint16x4_t __p0, uint16x4_t __p1) {
  4086.   uint16x4_t __ret;
  4087.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4088.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4089.   __ret = (uint16x4_t)(__rev0 >= __rev1);
  4090.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4091.   return __ret;
  4092. }
  4093. #endif
  4094.  
  4095. #ifdef __LITTLE_ENDIAN__
  4096. __ai uint8x8_t vcge_s8(int8x8_t __p0, int8x8_t __p1) {
  4097.   uint8x8_t __ret;
  4098.   __ret = (uint8x8_t)(__p0 >= __p1);
  4099.   return __ret;
  4100. }
  4101. #else
  4102. __ai uint8x8_t vcge_s8(int8x8_t __p0, int8x8_t __p1) {
  4103.   uint8x8_t __ret;
  4104.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4105.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4106.   __ret = (uint8x8_t)(__rev0 >= __rev1);
  4107.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4108.   return __ret;
  4109. }
  4110. #endif
  4111.  
  4112. #ifdef __LITTLE_ENDIAN__
  4113. __ai uint32x2_t vcge_f32(float32x2_t __p0, float32x2_t __p1) {
  4114.   uint32x2_t __ret;
  4115.   __ret = (uint32x2_t)(__p0 >= __p1);
  4116.   return __ret;
  4117. }
  4118. #else
  4119. __ai uint32x2_t vcge_f32(float32x2_t __p0, float32x2_t __p1) {
  4120.   uint32x2_t __ret;
  4121.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4122.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4123.   __ret = (uint32x2_t)(__rev0 >= __rev1);
  4124.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4125.   return __ret;
  4126. }
  4127. #endif
  4128.  
  4129. #ifdef __LITTLE_ENDIAN__
  4130. __ai uint32x2_t vcge_s32(int32x2_t __p0, int32x2_t __p1) {
  4131.   uint32x2_t __ret;
  4132.   __ret = (uint32x2_t)(__p0 >= __p1);
  4133.   return __ret;
  4134. }
  4135. #else
  4136. __ai uint32x2_t vcge_s32(int32x2_t __p0, int32x2_t __p1) {
  4137.   uint32x2_t __ret;
  4138.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4139.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4140.   __ret = (uint32x2_t)(__rev0 >= __rev1);
  4141.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4142.   return __ret;
  4143. }
  4144. #endif
  4145.  
  4146. #ifdef __LITTLE_ENDIAN__
  4147. __ai uint16x4_t vcge_s16(int16x4_t __p0, int16x4_t __p1) {
  4148.   uint16x4_t __ret;
  4149.   __ret = (uint16x4_t)(__p0 >= __p1);
  4150.   return __ret;
  4151. }
  4152. #else
  4153. __ai uint16x4_t vcge_s16(int16x4_t __p0, int16x4_t __p1) {
  4154.   uint16x4_t __ret;
  4155.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4156.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4157.   __ret = (uint16x4_t)(__rev0 >= __rev1);
  4158.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4159.   return __ret;
  4160. }
  4161. #endif
  4162.  
  4163. #ifdef __LITTLE_ENDIAN__
  4164. __ai uint8x16_t vcgtq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  4165.   uint8x16_t __ret;
  4166.   __ret = (uint8x16_t)(__p0 > __p1);
  4167.   return __ret;
  4168. }
  4169. #else
  4170. __ai uint8x16_t vcgtq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  4171.   uint8x16_t __ret;
  4172.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4173.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4174.   __ret = (uint8x16_t)(__rev0 > __rev1);
  4175.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4176.   return __ret;
  4177. }
  4178. #endif
  4179.  
  4180. #ifdef __LITTLE_ENDIAN__
  4181. __ai uint32x4_t vcgtq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  4182.   uint32x4_t __ret;
  4183.   __ret = (uint32x4_t)(__p0 > __p1);
  4184.   return __ret;
  4185. }
  4186. #else
  4187. __ai uint32x4_t vcgtq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  4188.   uint32x4_t __ret;
  4189.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4190.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4191.   __ret = (uint32x4_t)(__rev0 > __rev1);
  4192.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4193.   return __ret;
  4194. }
  4195. #endif
  4196.  
  4197. #ifdef __LITTLE_ENDIAN__
  4198. __ai uint16x8_t vcgtq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  4199.   uint16x8_t __ret;
  4200.   __ret = (uint16x8_t)(__p0 > __p1);
  4201.   return __ret;
  4202. }
  4203. #else
  4204. __ai uint16x8_t vcgtq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  4205.   uint16x8_t __ret;
  4206.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4207.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4208.   __ret = (uint16x8_t)(__rev0 > __rev1);
  4209.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4210.   return __ret;
  4211. }
  4212. #endif
  4213.  
  4214. #ifdef __LITTLE_ENDIAN__
  4215. __ai uint8x16_t vcgtq_s8(int8x16_t __p0, int8x16_t __p1) {
  4216.   uint8x16_t __ret;
  4217.   __ret = (uint8x16_t)(__p0 > __p1);
  4218.   return __ret;
  4219. }
  4220. #else
  4221. __ai uint8x16_t vcgtq_s8(int8x16_t __p0, int8x16_t __p1) {
  4222.   uint8x16_t __ret;
  4223.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4224.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4225.   __ret = (uint8x16_t)(__rev0 > __rev1);
  4226.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4227.   return __ret;
  4228. }
  4229. #endif
  4230.  
  4231. #ifdef __LITTLE_ENDIAN__
  4232. __ai uint32x4_t vcgtq_f32(float32x4_t __p0, float32x4_t __p1) {
  4233.   uint32x4_t __ret;
  4234.   __ret = (uint32x4_t)(__p0 > __p1);
  4235.   return __ret;
  4236. }
  4237. #else
  4238. __ai uint32x4_t vcgtq_f32(float32x4_t __p0, float32x4_t __p1) {
  4239.   uint32x4_t __ret;
  4240.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4241.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4242.   __ret = (uint32x4_t)(__rev0 > __rev1);
  4243.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4244.   return __ret;
  4245. }
  4246. #endif
  4247.  
  4248. #ifdef __LITTLE_ENDIAN__
  4249. __ai uint32x4_t vcgtq_s32(int32x4_t __p0, int32x4_t __p1) {
  4250.   uint32x4_t __ret;
  4251.   __ret = (uint32x4_t)(__p0 > __p1);
  4252.   return __ret;
  4253. }
  4254. #else
  4255. __ai uint32x4_t vcgtq_s32(int32x4_t __p0, int32x4_t __p1) {
  4256.   uint32x4_t __ret;
  4257.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4258.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4259.   __ret = (uint32x4_t)(__rev0 > __rev1);
  4260.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4261.   return __ret;
  4262. }
  4263. #endif
  4264.  
  4265. #ifdef __LITTLE_ENDIAN__
  4266. __ai uint16x8_t vcgtq_s16(int16x8_t __p0, int16x8_t __p1) {
  4267.   uint16x8_t __ret;
  4268.   __ret = (uint16x8_t)(__p0 > __p1);
  4269.   return __ret;
  4270. }
  4271. #else
  4272. __ai uint16x8_t vcgtq_s16(int16x8_t __p0, int16x8_t __p1) {
  4273.   uint16x8_t __ret;
  4274.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4275.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4276.   __ret = (uint16x8_t)(__rev0 > __rev1);
  4277.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4278.   return __ret;
  4279. }
  4280. #endif
  4281.  
  4282. #ifdef __LITTLE_ENDIAN__
  4283. __ai uint8x8_t vcgt_u8(uint8x8_t __p0, uint8x8_t __p1) {
  4284.   uint8x8_t __ret;
  4285.   __ret = (uint8x8_t)(__p0 > __p1);
  4286.   return __ret;
  4287. }
  4288. #else
  4289. __ai uint8x8_t vcgt_u8(uint8x8_t __p0, uint8x8_t __p1) {
  4290.   uint8x8_t __ret;
  4291.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4292.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4293.   __ret = (uint8x8_t)(__rev0 > __rev1);
  4294.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4295.   return __ret;
  4296. }
  4297. #endif
  4298.  
  4299. #ifdef __LITTLE_ENDIAN__
  4300. __ai uint32x2_t vcgt_u32(uint32x2_t __p0, uint32x2_t __p1) {
  4301.   uint32x2_t __ret;
  4302.   __ret = (uint32x2_t)(__p0 > __p1);
  4303.   return __ret;
  4304. }
  4305. #else
  4306. __ai uint32x2_t vcgt_u32(uint32x2_t __p0, uint32x2_t __p1) {
  4307.   uint32x2_t __ret;
  4308.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4309.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4310.   __ret = (uint32x2_t)(__rev0 > __rev1);
  4311.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4312.   return __ret;
  4313. }
  4314. #endif
  4315.  
  4316. #ifdef __LITTLE_ENDIAN__
  4317. __ai uint16x4_t vcgt_u16(uint16x4_t __p0, uint16x4_t __p1) {
  4318.   uint16x4_t __ret;
  4319.   __ret = (uint16x4_t)(__p0 > __p1);
  4320.   return __ret;
  4321. }
  4322. #else
  4323. __ai uint16x4_t vcgt_u16(uint16x4_t __p0, uint16x4_t __p1) {
  4324.   uint16x4_t __ret;
  4325.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4326.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4327.   __ret = (uint16x4_t)(__rev0 > __rev1);
  4328.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4329.   return __ret;
  4330. }
  4331. #endif
  4332.  
  4333. #ifdef __LITTLE_ENDIAN__
  4334. __ai uint8x8_t vcgt_s8(int8x8_t __p0, int8x8_t __p1) {
  4335.   uint8x8_t __ret;
  4336.   __ret = (uint8x8_t)(__p0 > __p1);
  4337.   return __ret;
  4338. }
  4339. #else
  4340. __ai uint8x8_t vcgt_s8(int8x8_t __p0, int8x8_t __p1) {
  4341.   uint8x8_t __ret;
  4342.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4343.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4344.   __ret = (uint8x8_t)(__rev0 > __rev1);
  4345.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4346.   return __ret;
  4347. }
  4348. #endif
  4349.  
  4350. #ifdef __LITTLE_ENDIAN__
  4351. __ai uint32x2_t vcgt_f32(float32x2_t __p0, float32x2_t __p1) {
  4352.   uint32x2_t __ret;
  4353.   __ret = (uint32x2_t)(__p0 > __p1);
  4354.   return __ret;
  4355. }
  4356. #else
  4357. __ai uint32x2_t vcgt_f32(float32x2_t __p0, float32x2_t __p1) {
  4358.   uint32x2_t __ret;
  4359.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4360.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4361.   __ret = (uint32x2_t)(__rev0 > __rev1);
  4362.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4363.   return __ret;
  4364. }
  4365. #endif
  4366.  
  4367. #ifdef __LITTLE_ENDIAN__
  4368. __ai uint32x2_t vcgt_s32(int32x2_t __p0, int32x2_t __p1) {
  4369.   uint32x2_t __ret;
  4370.   __ret = (uint32x2_t)(__p0 > __p1);
  4371.   return __ret;
  4372. }
  4373. #else
  4374. __ai uint32x2_t vcgt_s32(int32x2_t __p0, int32x2_t __p1) {
  4375.   uint32x2_t __ret;
  4376.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4377.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4378.   __ret = (uint32x2_t)(__rev0 > __rev1);
  4379.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4380.   return __ret;
  4381. }
  4382. #endif
  4383.  
  4384. #ifdef __LITTLE_ENDIAN__
  4385. __ai uint16x4_t vcgt_s16(int16x4_t __p0, int16x4_t __p1) {
  4386.   uint16x4_t __ret;
  4387.   __ret = (uint16x4_t)(__p0 > __p1);
  4388.   return __ret;
  4389. }
  4390. #else
  4391. __ai uint16x4_t vcgt_s16(int16x4_t __p0, int16x4_t __p1) {
  4392.   uint16x4_t __ret;
  4393.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4394.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4395.   __ret = (uint16x4_t)(__rev0 > __rev1);
  4396.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4397.   return __ret;
  4398. }
  4399. #endif
  4400.  
  4401. #ifdef __LITTLE_ENDIAN__
  4402. __ai uint8x16_t vcleq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  4403.   uint8x16_t __ret;
  4404.   __ret = (uint8x16_t)(__p0 <= __p1);
  4405.   return __ret;
  4406. }
  4407. #else
  4408. __ai uint8x16_t vcleq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  4409.   uint8x16_t __ret;
  4410.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4411.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4412.   __ret = (uint8x16_t)(__rev0 <= __rev1);
  4413.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4414.   return __ret;
  4415. }
  4416. #endif
  4417.  
  4418. #ifdef __LITTLE_ENDIAN__
  4419. __ai uint32x4_t vcleq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  4420.   uint32x4_t __ret;
  4421.   __ret = (uint32x4_t)(__p0 <= __p1);
  4422.   return __ret;
  4423. }
  4424. #else
  4425. __ai uint32x4_t vcleq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  4426.   uint32x4_t __ret;
  4427.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4428.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4429.   __ret = (uint32x4_t)(__rev0 <= __rev1);
  4430.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4431.   return __ret;
  4432. }
  4433. #endif
  4434.  
  4435. #ifdef __LITTLE_ENDIAN__
  4436. __ai uint16x8_t vcleq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  4437.   uint16x8_t __ret;
  4438.   __ret = (uint16x8_t)(__p0 <= __p1);
  4439.   return __ret;
  4440. }
  4441. #else
  4442. __ai uint16x8_t vcleq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  4443.   uint16x8_t __ret;
  4444.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4445.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4446.   __ret = (uint16x8_t)(__rev0 <= __rev1);
  4447.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4448.   return __ret;
  4449. }
  4450. #endif
  4451.  
  4452. #ifdef __LITTLE_ENDIAN__
  4453. __ai uint8x16_t vcleq_s8(int8x16_t __p0, int8x16_t __p1) {
  4454.   uint8x16_t __ret;
  4455.   __ret = (uint8x16_t)(__p0 <= __p1);
  4456.   return __ret;
  4457. }
  4458. #else
  4459. __ai uint8x16_t vcleq_s8(int8x16_t __p0, int8x16_t __p1) {
  4460.   uint8x16_t __ret;
  4461.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4462.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4463.   __ret = (uint8x16_t)(__rev0 <= __rev1);
  4464.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4465.   return __ret;
  4466. }
  4467. #endif
  4468.  
  4469. #ifdef __LITTLE_ENDIAN__
  4470. __ai uint32x4_t vcleq_f32(float32x4_t __p0, float32x4_t __p1) {
  4471.   uint32x4_t __ret;
  4472.   __ret = (uint32x4_t)(__p0 <= __p1);
  4473.   return __ret;
  4474. }
  4475. #else
  4476. __ai uint32x4_t vcleq_f32(float32x4_t __p0, float32x4_t __p1) {
  4477.   uint32x4_t __ret;
  4478.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4479.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4480.   __ret = (uint32x4_t)(__rev0 <= __rev1);
  4481.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4482.   return __ret;
  4483. }
  4484. #endif
  4485.  
  4486. #ifdef __LITTLE_ENDIAN__
  4487. __ai uint32x4_t vcleq_s32(int32x4_t __p0, int32x4_t __p1) {
  4488.   uint32x4_t __ret;
  4489.   __ret = (uint32x4_t)(__p0 <= __p1);
  4490.   return __ret;
  4491. }
  4492. #else
  4493. __ai uint32x4_t vcleq_s32(int32x4_t __p0, int32x4_t __p1) {
  4494.   uint32x4_t __ret;
  4495.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4496.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4497.   __ret = (uint32x4_t)(__rev0 <= __rev1);
  4498.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4499.   return __ret;
  4500. }
  4501. #endif
  4502.  
  4503. #ifdef __LITTLE_ENDIAN__
  4504. __ai uint16x8_t vcleq_s16(int16x8_t __p0, int16x8_t __p1) {
  4505.   uint16x8_t __ret;
  4506.   __ret = (uint16x8_t)(__p0 <= __p1);
  4507.   return __ret;
  4508. }
  4509. #else
  4510. __ai uint16x8_t vcleq_s16(int16x8_t __p0, int16x8_t __p1) {
  4511.   uint16x8_t __ret;
  4512.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4513.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4514.   __ret = (uint16x8_t)(__rev0 <= __rev1);
  4515.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4516.   return __ret;
  4517. }
  4518. #endif
  4519.  
  4520. #ifdef __LITTLE_ENDIAN__
  4521. __ai uint8x8_t vcle_u8(uint8x8_t __p0, uint8x8_t __p1) {
  4522.   uint8x8_t __ret;
  4523.   __ret = (uint8x8_t)(__p0 <= __p1);
  4524.   return __ret;
  4525. }
  4526. #else
  4527. __ai uint8x8_t vcle_u8(uint8x8_t __p0, uint8x8_t __p1) {
  4528.   uint8x8_t __ret;
  4529.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4530.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4531.   __ret = (uint8x8_t)(__rev0 <= __rev1);
  4532.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4533.   return __ret;
  4534. }
  4535. #endif
  4536.  
  4537. #ifdef __LITTLE_ENDIAN__
  4538. __ai uint32x2_t vcle_u32(uint32x2_t __p0, uint32x2_t __p1) {
  4539.   uint32x2_t __ret;
  4540.   __ret = (uint32x2_t)(__p0 <= __p1);
  4541.   return __ret;
  4542. }
  4543. #else
  4544. __ai uint32x2_t vcle_u32(uint32x2_t __p0, uint32x2_t __p1) {
  4545.   uint32x2_t __ret;
  4546.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4547.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4548.   __ret = (uint32x2_t)(__rev0 <= __rev1);
  4549.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4550.   return __ret;
  4551. }
  4552. #endif
  4553.  
  4554. #ifdef __LITTLE_ENDIAN__
  4555. __ai uint16x4_t vcle_u16(uint16x4_t __p0, uint16x4_t __p1) {
  4556.   uint16x4_t __ret;
  4557.   __ret = (uint16x4_t)(__p0 <= __p1);
  4558.   return __ret;
  4559. }
  4560. #else
  4561. __ai uint16x4_t vcle_u16(uint16x4_t __p0, uint16x4_t __p1) {
  4562.   uint16x4_t __ret;
  4563.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4564.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4565.   __ret = (uint16x4_t)(__rev0 <= __rev1);
  4566.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4567.   return __ret;
  4568. }
  4569. #endif
  4570.  
  4571. #ifdef __LITTLE_ENDIAN__
  4572. __ai uint8x8_t vcle_s8(int8x8_t __p0, int8x8_t __p1) {
  4573.   uint8x8_t __ret;
  4574.   __ret = (uint8x8_t)(__p0 <= __p1);
  4575.   return __ret;
  4576. }
  4577. #else
  4578. __ai uint8x8_t vcle_s8(int8x8_t __p0, int8x8_t __p1) {
  4579.   uint8x8_t __ret;
  4580.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4581.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4582.   __ret = (uint8x8_t)(__rev0 <= __rev1);
  4583.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4584.   return __ret;
  4585. }
  4586. #endif
  4587.  
  4588. #ifdef __LITTLE_ENDIAN__
  4589. __ai uint32x2_t vcle_f32(float32x2_t __p0, float32x2_t __p1) {
  4590.   uint32x2_t __ret;
  4591.   __ret = (uint32x2_t)(__p0 <= __p1);
  4592.   return __ret;
  4593. }
  4594. #else
  4595. __ai uint32x2_t vcle_f32(float32x2_t __p0, float32x2_t __p1) {
  4596.   uint32x2_t __ret;
  4597.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4598.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4599.   __ret = (uint32x2_t)(__rev0 <= __rev1);
  4600.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4601.   return __ret;
  4602. }
  4603. #endif
  4604.  
  4605. #ifdef __LITTLE_ENDIAN__
  4606. __ai uint32x2_t vcle_s32(int32x2_t __p0, int32x2_t __p1) {
  4607.   uint32x2_t __ret;
  4608.   __ret = (uint32x2_t)(__p0 <= __p1);
  4609.   return __ret;
  4610. }
  4611. #else
  4612. __ai uint32x2_t vcle_s32(int32x2_t __p0, int32x2_t __p1) {
  4613.   uint32x2_t __ret;
  4614.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4615.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4616.   __ret = (uint32x2_t)(__rev0 <= __rev1);
  4617.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4618.   return __ret;
  4619. }
  4620. #endif
  4621.  
  4622. #ifdef __LITTLE_ENDIAN__
  4623. __ai uint16x4_t vcle_s16(int16x4_t __p0, int16x4_t __p1) {
  4624.   uint16x4_t __ret;
  4625.   __ret = (uint16x4_t)(__p0 <= __p1);
  4626.   return __ret;
  4627. }
  4628. #else
  4629. __ai uint16x4_t vcle_s16(int16x4_t __p0, int16x4_t __p1) {
  4630.   uint16x4_t __ret;
  4631.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4632.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4633.   __ret = (uint16x4_t)(__rev0 <= __rev1);
  4634.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4635.   return __ret;
  4636. }
  4637. #endif
  4638.  
  4639. #ifdef __LITTLE_ENDIAN__
  4640. __ai int8x16_t vclsq_u8(uint8x16_t __p0) {
  4641.   int8x16_t __ret;
  4642.   __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 32);
  4643.   return __ret;
  4644. }
  4645. #else
  4646. __ai int8x16_t vclsq_u8(uint8x16_t __p0) {
  4647.   int8x16_t __ret;
  4648.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4649.   __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 32);
  4650.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4651.   return __ret;
  4652. }
  4653. #endif
  4654.  
  4655. #ifdef __LITTLE_ENDIAN__
  4656. __ai int32x4_t vclsq_u32(uint32x4_t __p0) {
  4657.   int32x4_t __ret;
  4658.   __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 34);
  4659.   return __ret;
  4660. }
  4661. #else
  4662. __ai int32x4_t vclsq_u32(uint32x4_t __p0) {
  4663.   int32x4_t __ret;
  4664.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4665.   __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 34);
  4666.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4667.   return __ret;
  4668. }
  4669. #endif
  4670.  
  4671. #ifdef __LITTLE_ENDIAN__
  4672. __ai int16x8_t vclsq_u16(uint16x8_t __p0) {
  4673.   int16x8_t __ret;
  4674.   __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 33);
  4675.   return __ret;
  4676. }
  4677. #else
  4678. __ai int16x8_t vclsq_u16(uint16x8_t __p0) {
  4679.   int16x8_t __ret;
  4680.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4681.   __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 33);
  4682.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4683.   return __ret;
  4684. }
  4685. #endif
  4686.  
  4687. #ifdef __LITTLE_ENDIAN__
  4688. __ai int8x16_t vclsq_s8(int8x16_t __p0) {
  4689.   int8x16_t __ret;
  4690.   __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 32);
  4691.   return __ret;
  4692. }
  4693. #else
  4694. __ai int8x16_t vclsq_s8(int8x16_t __p0) {
  4695.   int8x16_t __ret;
  4696.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4697.   __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 32);
  4698.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4699.   return __ret;
  4700. }
  4701. #endif
  4702.  
  4703. #ifdef __LITTLE_ENDIAN__
  4704. __ai int32x4_t vclsq_s32(int32x4_t __p0) {
  4705.   int32x4_t __ret;
  4706.   __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 34);
  4707.   return __ret;
  4708. }
  4709. #else
  4710. __ai int32x4_t vclsq_s32(int32x4_t __p0) {
  4711.   int32x4_t __ret;
  4712.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4713.   __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 34);
  4714.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4715.   return __ret;
  4716. }
  4717. #endif
  4718.  
  4719. #ifdef __LITTLE_ENDIAN__
  4720. __ai int16x8_t vclsq_s16(int16x8_t __p0) {
  4721.   int16x8_t __ret;
  4722.   __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 33);
  4723.   return __ret;
  4724. }
  4725. #else
  4726. __ai int16x8_t vclsq_s16(int16x8_t __p0) {
  4727.   int16x8_t __ret;
  4728.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4729.   __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 33);
  4730.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4731.   return __ret;
  4732. }
  4733. #endif
  4734.  
  4735. #ifdef __LITTLE_ENDIAN__
  4736. __ai int8x8_t vcls_u8(uint8x8_t __p0) {
  4737.   int8x8_t __ret;
  4738.   __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__p0, 0);
  4739.   return __ret;
  4740. }
  4741. #else
  4742. __ai int8x8_t vcls_u8(uint8x8_t __p0) {
  4743.   int8x8_t __ret;
  4744.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4745.   __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 0);
  4746.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4747.   return __ret;
  4748. }
  4749. #endif
  4750.  
  4751. #ifdef __LITTLE_ENDIAN__
  4752. __ai int32x2_t vcls_u32(uint32x2_t __p0) {
  4753.   int32x2_t __ret;
  4754.   __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__p0, 2);
  4755.   return __ret;
  4756. }
  4757. #else
  4758. __ai int32x2_t vcls_u32(uint32x2_t __p0) {
  4759.   int32x2_t __ret;
  4760.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4761.   __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 2);
  4762.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4763.   return __ret;
  4764. }
  4765. #endif
  4766.  
  4767. #ifdef __LITTLE_ENDIAN__
  4768. __ai int16x4_t vcls_u16(uint16x4_t __p0) {
  4769.   int16x4_t __ret;
  4770.   __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__p0, 1);
  4771.   return __ret;
  4772. }
  4773. #else
  4774. __ai int16x4_t vcls_u16(uint16x4_t __p0) {
  4775.   int16x4_t __ret;
  4776.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4777.   __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 1);
  4778.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4779.   return __ret;
  4780. }
  4781. #endif
  4782.  
  4783. #ifdef __LITTLE_ENDIAN__
  4784. __ai int8x8_t vcls_s8(int8x8_t __p0) {
  4785.   int8x8_t __ret;
  4786.   __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__p0, 0);
  4787.   return __ret;
  4788. }
  4789. #else
  4790. __ai int8x8_t vcls_s8(int8x8_t __p0) {
  4791.   int8x8_t __ret;
  4792.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4793.   __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 0);
  4794.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4795.   return __ret;
  4796. }
  4797. #endif
  4798.  
  4799. #ifdef __LITTLE_ENDIAN__
  4800. __ai int32x2_t vcls_s32(int32x2_t __p0) {
  4801.   int32x2_t __ret;
  4802.   __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__p0, 2);
  4803.   return __ret;
  4804. }
  4805. #else
  4806. __ai int32x2_t vcls_s32(int32x2_t __p0) {
  4807.   int32x2_t __ret;
  4808.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4809.   __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 2);
  4810.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4811.   return __ret;
  4812. }
  4813. #endif
  4814.  
  4815. #ifdef __LITTLE_ENDIAN__
  4816. __ai int16x4_t vcls_s16(int16x4_t __p0) {
  4817.   int16x4_t __ret;
  4818.   __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__p0, 1);
  4819.   return __ret;
  4820. }
  4821. #else
  4822. __ai int16x4_t vcls_s16(int16x4_t __p0) {
  4823.   int16x4_t __ret;
  4824.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4825.   __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 1);
  4826.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4827.   return __ret;
  4828. }
  4829. #endif
  4830.  
  4831. #ifdef __LITTLE_ENDIAN__
  4832. __ai uint8x16_t vcltq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  4833.   uint8x16_t __ret;
  4834.   __ret = (uint8x16_t)(__p0 < __p1);
  4835.   return __ret;
  4836. }
  4837. #else
  4838. __ai uint8x16_t vcltq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  4839.   uint8x16_t __ret;
  4840.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4841.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4842.   __ret = (uint8x16_t)(__rev0 < __rev1);
  4843.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4844.   return __ret;
  4845. }
  4846. #endif
  4847.  
  4848. #ifdef __LITTLE_ENDIAN__
  4849. __ai uint32x4_t vcltq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  4850.   uint32x4_t __ret;
  4851.   __ret = (uint32x4_t)(__p0 < __p1);
  4852.   return __ret;
  4853. }
  4854. #else
  4855. __ai uint32x4_t vcltq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  4856.   uint32x4_t __ret;
  4857.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4858.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4859.   __ret = (uint32x4_t)(__rev0 < __rev1);
  4860.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4861.   return __ret;
  4862. }
  4863. #endif
  4864.  
  4865. #ifdef __LITTLE_ENDIAN__
  4866. __ai uint16x8_t vcltq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  4867.   uint16x8_t __ret;
  4868.   __ret = (uint16x8_t)(__p0 < __p1);
  4869.   return __ret;
  4870. }
  4871. #else
  4872. __ai uint16x8_t vcltq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  4873.   uint16x8_t __ret;
  4874.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4875.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4876.   __ret = (uint16x8_t)(__rev0 < __rev1);
  4877.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4878.   return __ret;
  4879. }
  4880. #endif
  4881.  
  4882. #ifdef __LITTLE_ENDIAN__
  4883. __ai uint8x16_t vcltq_s8(int8x16_t __p0, int8x16_t __p1) {
  4884.   uint8x16_t __ret;
  4885.   __ret = (uint8x16_t)(__p0 < __p1);
  4886.   return __ret;
  4887. }
  4888. #else
  4889. __ai uint8x16_t vcltq_s8(int8x16_t __p0, int8x16_t __p1) {
  4890.   uint8x16_t __ret;
  4891.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4892.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4893.   __ret = (uint8x16_t)(__rev0 < __rev1);
  4894.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  4895.   return __ret;
  4896. }
  4897. #endif
  4898.  
  4899. #ifdef __LITTLE_ENDIAN__
  4900. __ai uint32x4_t vcltq_f32(float32x4_t __p0, float32x4_t __p1) {
  4901.   uint32x4_t __ret;
  4902.   __ret = (uint32x4_t)(__p0 < __p1);
  4903.   return __ret;
  4904. }
  4905. #else
  4906. __ai uint32x4_t vcltq_f32(float32x4_t __p0, float32x4_t __p1) {
  4907.   uint32x4_t __ret;
  4908.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4909.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4910.   __ret = (uint32x4_t)(__rev0 < __rev1);
  4911.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4912.   return __ret;
  4913. }
  4914. #endif
  4915.  
  4916. #ifdef __LITTLE_ENDIAN__
  4917. __ai uint32x4_t vcltq_s32(int32x4_t __p0, int32x4_t __p1) {
  4918.   uint32x4_t __ret;
  4919.   __ret = (uint32x4_t)(__p0 < __p1);
  4920.   return __ret;
  4921. }
  4922. #else
  4923. __ai uint32x4_t vcltq_s32(int32x4_t __p0, int32x4_t __p1) {
  4924.   uint32x4_t __ret;
  4925.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4926.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4927.   __ret = (uint32x4_t)(__rev0 < __rev1);
  4928.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4929.   return __ret;
  4930. }
  4931. #endif
  4932.  
  4933. #ifdef __LITTLE_ENDIAN__
  4934. __ai uint16x8_t vcltq_s16(int16x8_t __p0, int16x8_t __p1) {
  4935.   uint16x8_t __ret;
  4936.   __ret = (uint16x8_t)(__p0 < __p1);
  4937.   return __ret;
  4938. }
  4939. #else
  4940. __ai uint16x8_t vcltq_s16(int16x8_t __p0, int16x8_t __p1) {
  4941.   uint16x8_t __ret;
  4942.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4943.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4944.   __ret = (uint16x8_t)(__rev0 < __rev1);
  4945.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4946.   return __ret;
  4947. }
  4948. #endif
  4949.  
  4950. #ifdef __LITTLE_ENDIAN__
  4951. __ai uint8x8_t vclt_u8(uint8x8_t __p0, uint8x8_t __p1) {
  4952.   uint8x8_t __ret;
  4953.   __ret = (uint8x8_t)(__p0 < __p1);
  4954.   return __ret;
  4955. }
  4956. #else
  4957. __ai uint8x8_t vclt_u8(uint8x8_t __p0, uint8x8_t __p1) {
  4958.   uint8x8_t __ret;
  4959.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  4960.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  4961.   __ret = (uint8x8_t)(__rev0 < __rev1);
  4962.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  4963.   return __ret;
  4964. }
  4965. #endif
  4966.  
  4967. #ifdef __LITTLE_ENDIAN__
  4968. __ai uint32x2_t vclt_u32(uint32x2_t __p0, uint32x2_t __p1) {
  4969.   uint32x2_t __ret;
  4970.   __ret = (uint32x2_t)(__p0 < __p1);
  4971.   return __ret;
  4972. }
  4973. #else
  4974. __ai uint32x2_t vclt_u32(uint32x2_t __p0, uint32x2_t __p1) {
  4975.   uint32x2_t __ret;
  4976.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  4977.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  4978.   __ret = (uint32x2_t)(__rev0 < __rev1);
  4979.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  4980.   return __ret;
  4981. }
  4982. #endif
  4983.  
  4984. #ifdef __LITTLE_ENDIAN__
  4985. __ai uint16x4_t vclt_u16(uint16x4_t __p0, uint16x4_t __p1) {
  4986.   uint16x4_t __ret;
  4987.   __ret = (uint16x4_t)(__p0 < __p1);
  4988.   return __ret;
  4989. }
  4990. #else
  4991. __ai uint16x4_t vclt_u16(uint16x4_t __p0, uint16x4_t __p1) {
  4992.   uint16x4_t __ret;
  4993.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  4994.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  4995.   __ret = (uint16x4_t)(__rev0 < __rev1);
  4996.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  4997.   return __ret;
  4998. }
  4999. #endif
  5000.  
  5001. #ifdef __LITTLE_ENDIAN__
  5002. __ai uint8x8_t vclt_s8(int8x8_t __p0, int8x8_t __p1) {
  5003.   uint8x8_t __ret;
  5004.   __ret = (uint8x8_t)(__p0 < __p1);
  5005.   return __ret;
  5006. }
  5007. #else
  5008. __ai uint8x8_t vclt_s8(int8x8_t __p0, int8x8_t __p1) {
  5009.   uint8x8_t __ret;
  5010.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5011.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  5012.   __ret = (uint8x8_t)(__rev0 < __rev1);
  5013.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5014.   return __ret;
  5015. }
  5016. #endif
  5017.  
  5018. #ifdef __LITTLE_ENDIAN__
  5019. __ai uint32x2_t vclt_f32(float32x2_t __p0, float32x2_t __p1) {
  5020.   uint32x2_t __ret;
  5021.   __ret = (uint32x2_t)(__p0 < __p1);
  5022.   return __ret;
  5023. }
  5024. #else
  5025. __ai uint32x2_t vclt_f32(float32x2_t __p0, float32x2_t __p1) {
  5026.   uint32x2_t __ret;
  5027.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5028.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  5029.   __ret = (uint32x2_t)(__rev0 < __rev1);
  5030.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5031.   return __ret;
  5032. }
  5033. #endif
  5034.  
  5035. #ifdef __LITTLE_ENDIAN__
  5036. __ai uint32x2_t vclt_s32(int32x2_t __p0, int32x2_t __p1) {
  5037.   uint32x2_t __ret;
  5038.   __ret = (uint32x2_t)(__p0 < __p1);
  5039.   return __ret;
  5040. }
  5041. #else
  5042. __ai uint32x2_t vclt_s32(int32x2_t __p0, int32x2_t __p1) {
  5043.   uint32x2_t __ret;
  5044.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5045.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  5046.   __ret = (uint32x2_t)(__rev0 < __rev1);
  5047.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5048.   return __ret;
  5049. }
  5050. #endif
  5051.  
  5052. #ifdef __LITTLE_ENDIAN__
  5053. __ai uint16x4_t vclt_s16(int16x4_t __p0, int16x4_t __p1) {
  5054.   uint16x4_t __ret;
  5055.   __ret = (uint16x4_t)(__p0 < __p1);
  5056.   return __ret;
  5057. }
  5058. #else
  5059. __ai uint16x4_t vclt_s16(int16x4_t __p0, int16x4_t __p1) {
  5060.   uint16x4_t __ret;
  5061.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5062.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  5063.   __ret = (uint16x4_t)(__rev0 < __rev1);
  5064.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5065.   return __ret;
  5066. }
  5067. #endif
  5068.  
  5069. #ifdef __LITTLE_ENDIAN__
  5070. __ai uint8x16_t vclzq_u8(uint8x16_t __p0) {
  5071.   uint8x16_t __ret;
  5072.   __ret = (uint8x16_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 48);
  5073.   return __ret;
  5074. }
  5075. #else
  5076. __ai uint8x16_t vclzq_u8(uint8x16_t __p0) {
  5077.   uint8x16_t __ret;
  5078.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5079.   __ret = (uint8x16_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 48);
  5080.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5081.   return __ret;
  5082. }
  5083. #endif
  5084.  
  5085. #ifdef __LITTLE_ENDIAN__
  5086. __ai uint32x4_t vclzq_u32(uint32x4_t __p0) {
  5087.   uint32x4_t __ret;
  5088.   __ret = (uint32x4_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 50);
  5089.   return __ret;
  5090. }
  5091. #else
  5092. __ai uint32x4_t vclzq_u32(uint32x4_t __p0) {
  5093.   uint32x4_t __ret;
  5094.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5095.   __ret = (uint32x4_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 50);
  5096.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5097.   return __ret;
  5098. }
  5099. #endif
  5100.  
  5101. #ifdef __LITTLE_ENDIAN__
  5102. __ai uint16x8_t vclzq_u16(uint16x8_t __p0) {
  5103.   uint16x8_t __ret;
  5104.   __ret = (uint16x8_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 49);
  5105.   return __ret;
  5106. }
  5107. #else
  5108. __ai uint16x8_t vclzq_u16(uint16x8_t __p0) {
  5109.   uint16x8_t __ret;
  5110.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5111.   __ret = (uint16x8_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 49);
  5112.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5113.   return __ret;
  5114. }
  5115. #endif
  5116.  
  5117. #ifdef __LITTLE_ENDIAN__
  5118. __ai int8x16_t vclzq_s8(int8x16_t __p0) {
  5119.   int8x16_t __ret;
  5120.   __ret = (int8x16_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 32);
  5121.   return __ret;
  5122. }
  5123. #else
  5124. __ai int8x16_t vclzq_s8(int8x16_t __p0) {
  5125.   int8x16_t __ret;
  5126.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5127.   __ret = (int8x16_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 32);
  5128.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5129.   return __ret;
  5130. }
  5131. #endif
  5132.  
  5133. #ifdef __LITTLE_ENDIAN__
  5134. __ai int32x4_t vclzq_s32(int32x4_t __p0) {
  5135.   int32x4_t __ret;
  5136.   __ret = (int32x4_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 34);
  5137.   return __ret;
  5138. }
  5139. #else
  5140. __ai int32x4_t vclzq_s32(int32x4_t __p0) {
  5141.   int32x4_t __ret;
  5142.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5143.   __ret = (int32x4_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 34);
  5144.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5145.   return __ret;
  5146. }
  5147. #endif
  5148.  
  5149. #ifdef __LITTLE_ENDIAN__
  5150. __ai int16x8_t vclzq_s16(int16x8_t __p0) {
  5151.   int16x8_t __ret;
  5152.   __ret = (int16x8_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 33);
  5153.   return __ret;
  5154. }
  5155. #else
  5156. __ai int16x8_t vclzq_s16(int16x8_t __p0) {
  5157.   int16x8_t __ret;
  5158.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5159.   __ret = (int16x8_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 33);
  5160.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5161.   return __ret;
  5162. }
  5163. #endif
  5164.  
  5165. #ifdef __LITTLE_ENDIAN__
  5166. __ai uint8x8_t vclz_u8(uint8x8_t __p0) {
  5167.   uint8x8_t __ret;
  5168.   __ret = (uint8x8_t) __builtin_neon_vclz_v((int8x8_t)__p0, 16);
  5169.   return __ret;
  5170. }
  5171. #else
  5172. __ai uint8x8_t vclz_u8(uint8x8_t __p0) {
  5173.   uint8x8_t __ret;
  5174.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5175.   __ret = (uint8x8_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 16);
  5176.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5177.   return __ret;
  5178. }
  5179. #endif
  5180.  
  5181. #ifdef __LITTLE_ENDIAN__
  5182. __ai uint32x2_t vclz_u32(uint32x2_t __p0) {
  5183.   uint32x2_t __ret;
  5184.   __ret = (uint32x2_t) __builtin_neon_vclz_v((int8x8_t)__p0, 18);
  5185.   return __ret;
  5186. }
  5187. #else
  5188. __ai uint32x2_t vclz_u32(uint32x2_t __p0) {
  5189.   uint32x2_t __ret;
  5190.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5191.   __ret = (uint32x2_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 18);
  5192.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5193.   return __ret;
  5194. }
  5195. #endif
  5196.  
  5197. #ifdef __LITTLE_ENDIAN__
  5198. __ai uint16x4_t vclz_u16(uint16x4_t __p0) {
  5199.   uint16x4_t __ret;
  5200.   __ret = (uint16x4_t) __builtin_neon_vclz_v((int8x8_t)__p0, 17);
  5201.   return __ret;
  5202. }
  5203. #else
  5204. __ai uint16x4_t vclz_u16(uint16x4_t __p0) {
  5205.   uint16x4_t __ret;
  5206.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5207.   __ret = (uint16x4_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 17);
  5208.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5209.   return __ret;
  5210. }
  5211. #endif
  5212.  
  5213. #ifdef __LITTLE_ENDIAN__
  5214. __ai int8x8_t vclz_s8(int8x8_t __p0) {
  5215.   int8x8_t __ret;
  5216.   __ret = (int8x8_t) __builtin_neon_vclz_v((int8x8_t)__p0, 0);
  5217.   return __ret;
  5218. }
  5219. #else
  5220. __ai int8x8_t vclz_s8(int8x8_t __p0) {
  5221.   int8x8_t __ret;
  5222.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5223.   __ret = (int8x8_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 0);
  5224.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5225.   return __ret;
  5226. }
  5227. #endif
  5228.  
  5229. #ifdef __LITTLE_ENDIAN__
  5230. __ai int32x2_t vclz_s32(int32x2_t __p0) {
  5231.   int32x2_t __ret;
  5232.   __ret = (int32x2_t) __builtin_neon_vclz_v((int8x8_t)__p0, 2);
  5233.   return __ret;
  5234. }
  5235. #else
  5236. __ai int32x2_t vclz_s32(int32x2_t __p0) {
  5237.   int32x2_t __ret;
  5238.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5239.   __ret = (int32x2_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 2);
  5240.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5241.   return __ret;
  5242. }
  5243. #endif
  5244.  
  5245. #ifdef __LITTLE_ENDIAN__
  5246. __ai int16x4_t vclz_s16(int16x4_t __p0) {
  5247.   int16x4_t __ret;
  5248.   __ret = (int16x4_t) __builtin_neon_vclz_v((int8x8_t)__p0, 1);
  5249.   return __ret;
  5250. }
  5251. #else
  5252. __ai int16x4_t vclz_s16(int16x4_t __p0) {
  5253.   int16x4_t __ret;
  5254.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5255.   __ret = (int16x4_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 1);
  5256.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5257.   return __ret;
  5258. }
  5259. #endif
  5260.  
  5261. #ifdef __LITTLE_ENDIAN__
  5262. __ai poly8x8_t vcnt_p8(poly8x8_t __p0) {
  5263.   poly8x8_t __ret;
  5264.   __ret = (poly8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 4);
  5265.   return __ret;
  5266. }
  5267. #else
  5268. __ai poly8x8_t vcnt_p8(poly8x8_t __p0) {
  5269.   poly8x8_t __ret;
  5270.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5271.   __ret = (poly8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 4);
  5272.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5273.   return __ret;
  5274. }
  5275. #endif
  5276.  
  5277. #ifdef __LITTLE_ENDIAN__
  5278. __ai poly8x16_t vcntq_p8(poly8x16_t __p0) {
  5279.   poly8x16_t __ret;
  5280.   __ret = (poly8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 36);
  5281.   return __ret;
  5282. }
  5283. #else
  5284. __ai poly8x16_t vcntq_p8(poly8x16_t __p0) {
  5285.   poly8x16_t __ret;
  5286.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5287.   __ret = (poly8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 36);
  5288.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5289.   return __ret;
  5290. }
  5291. #endif
  5292.  
  5293. #ifdef __LITTLE_ENDIAN__
  5294. __ai uint8x16_t vcntq_u8(uint8x16_t __p0) {
  5295.   uint8x16_t __ret;
  5296.   __ret = (uint8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 48);
  5297.   return __ret;
  5298. }
  5299. #else
  5300. __ai uint8x16_t vcntq_u8(uint8x16_t __p0) {
  5301.   uint8x16_t __ret;
  5302.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5303.   __ret = (uint8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 48);
  5304.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5305.   return __ret;
  5306. }
  5307. #endif
  5308.  
  5309. #ifdef __LITTLE_ENDIAN__
  5310. __ai int8x16_t vcntq_s8(int8x16_t __p0) {
  5311.   int8x16_t __ret;
  5312.   __ret = (int8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 32);
  5313.   return __ret;
  5314. }
  5315. #else
  5316. __ai int8x16_t vcntq_s8(int8x16_t __p0) {
  5317.   int8x16_t __ret;
  5318.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5319.   __ret = (int8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 32);
  5320.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5321.   return __ret;
  5322. }
  5323. #endif
  5324.  
  5325. #ifdef __LITTLE_ENDIAN__
  5326. __ai uint8x8_t vcnt_u8(uint8x8_t __p0) {
  5327.   uint8x8_t __ret;
  5328.   __ret = (uint8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 16);
  5329.   return __ret;
  5330. }
  5331. #else
  5332. __ai uint8x8_t vcnt_u8(uint8x8_t __p0) {
  5333.   uint8x8_t __ret;
  5334.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5335.   __ret = (uint8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 16);
  5336.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5337.   return __ret;
  5338. }
  5339. #endif
  5340.  
  5341. #ifdef __LITTLE_ENDIAN__
  5342. __ai int8x8_t vcnt_s8(int8x8_t __p0) {
  5343.   int8x8_t __ret;
  5344.   __ret = (int8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 0);
  5345.   return __ret;
  5346. }
  5347. #else
  5348. __ai int8x8_t vcnt_s8(int8x8_t __p0) {
  5349.   int8x8_t __ret;
  5350.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5351.   __ret = (int8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 0);
  5352.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5353.   return __ret;
  5354. }
  5355. #endif
  5356.  
  5357. #ifdef __LITTLE_ENDIAN__
  5358. __ai poly8x16_t vcombine_p8(poly8x8_t __p0, poly8x8_t __p1) {
  5359.   poly8x16_t __ret;
  5360.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  5361.   return __ret;
  5362. }
  5363. #else
  5364. __ai poly8x16_t vcombine_p8(poly8x8_t __p0, poly8x8_t __p1) {
  5365.   poly8x16_t __ret;
  5366.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5367.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  5368.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  5369.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5370.   return __ret;
  5371. }
  5372. #endif
  5373.  
  5374. #ifdef __LITTLE_ENDIAN__
  5375. __ai poly16x8_t vcombine_p16(poly16x4_t __p0, poly16x4_t __p1) {
  5376.   poly16x8_t __ret;
  5377.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  5378.   return __ret;
  5379. }
  5380. #else
  5381. __ai poly16x8_t vcombine_p16(poly16x4_t __p0, poly16x4_t __p1) {
  5382.   poly16x8_t __ret;
  5383.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5384.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  5385.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
  5386.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5387.   return __ret;
  5388. }
  5389. #endif
  5390.  
  5391. #ifdef __LITTLE_ENDIAN__
  5392. __ai uint8x16_t vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
  5393.   uint8x16_t __ret;
  5394.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  5395.   return __ret;
  5396. }
  5397. #else
  5398. __ai uint8x16_t vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
  5399.   uint8x16_t __ret;
  5400.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5401.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  5402.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  5403.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5404.   return __ret;
  5405. }
  5406. __ai uint8x16_t __noswap_vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
  5407.   uint8x16_t __ret;
  5408.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  5409.   return __ret;
  5410. }
  5411. #endif
  5412.  
  5413. #ifdef __LITTLE_ENDIAN__
  5414. __ai uint32x4_t vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
  5415.   uint32x4_t __ret;
  5416.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  5417.   return __ret;
  5418. }
  5419. #else
  5420. __ai uint32x4_t vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
  5421.   uint32x4_t __ret;
  5422.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5423.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  5424.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
  5425.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5426.   return __ret;
  5427. }
  5428. __ai uint32x4_t __noswap_vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
  5429.   uint32x4_t __ret;
  5430.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  5431.   return __ret;
  5432. }
  5433. #endif
  5434.  
  5435. #ifdef __LITTLE_ENDIAN__
  5436. __ai uint64x2_t vcombine_u64(uint64x1_t __p0, uint64x1_t __p1) {
  5437.   uint64x2_t __ret;
  5438.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  5439.   return __ret;
  5440. }
  5441. #else
  5442. __ai uint64x2_t vcombine_u64(uint64x1_t __p0, uint64x1_t __p1) {
  5443.   uint64x2_t __ret;
  5444.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  5445.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5446.   return __ret;
  5447. }
  5448. #endif
  5449.  
  5450. #ifdef __LITTLE_ENDIAN__
  5451. __ai uint16x8_t vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
  5452.   uint16x8_t __ret;
  5453.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  5454.   return __ret;
  5455. }
  5456. #else
  5457. __ai uint16x8_t vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
  5458.   uint16x8_t __ret;
  5459.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5460.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  5461.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
  5462.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5463.   return __ret;
  5464. }
  5465. __ai uint16x8_t __noswap_vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
  5466.   uint16x8_t __ret;
  5467.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  5468.   return __ret;
  5469. }
  5470. #endif
  5471.  
  5472. #ifdef __LITTLE_ENDIAN__
  5473. __ai int8x16_t vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
  5474.   int8x16_t __ret;
  5475.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  5476.   return __ret;
  5477. }
  5478. #else
  5479. __ai int8x16_t vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
  5480.   int8x16_t __ret;
  5481.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  5482.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  5483.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  5484.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  5485.   return __ret;
  5486. }
  5487. __ai int8x16_t __noswap_vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
  5488.   int8x16_t __ret;
  5489.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
  5490.   return __ret;
  5491. }
  5492. #endif
  5493.  
  5494. #ifdef __LITTLE_ENDIAN__
  5495. __ai float32x4_t vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
  5496.   float32x4_t __ret;
  5497.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  5498.   return __ret;
  5499. }
  5500. #else
  5501. __ai float32x4_t vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
  5502.   float32x4_t __ret;
  5503.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5504.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  5505.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
  5506.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5507.   return __ret;
  5508. }
  5509. __ai float32x4_t __noswap_vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
  5510.   float32x4_t __ret;
  5511.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  5512.   return __ret;
  5513. }
  5514. #endif
  5515.  
  5516. #ifdef __LITTLE_ENDIAN__
  5517. __ai float16x8_t vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
  5518.   float16x8_t __ret;
  5519.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  5520.   return __ret;
  5521. }
  5522. #else
  5523. __ai float16x8_t vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
  5524.   float16x8_t __ret;
  5525.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5526.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  5527.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
  5528.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5529.   return __ret;
  5530. }
  5531. __ai float16x8_t __noswap_vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
  5532.   float16x8_t __ret;
  5533.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  5534.   return __ret;
  5535. }
  5536. #endif
  5537.  
  5538. #ifdef __LITTLE_ENDIAN__
  5539. __ai int32x4_t vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
  5540.   int32x4_t __ret;
  5541.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  5542.   return __ret;
  5543. }
  5544. #else
  5545. __ai int32x4_t vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
  5546.   int32x4_t __ret;
  5547.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5548.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  5549.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
  5550.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5551.   return __ret;
  5552. }
  5553. __ai int32x4_t __noswap_vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
  5554.   int32x4_t __ret;
  5555.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
  5556.   return __ret;
  5557. }
  5558. #endif
  5559.  
  5560. #ifdef __LITTLE_ENDIAN__
  5561. __ai int64x2_t vcombine_s64(int64x1_t __p0, int64x1_t __p1) {
  5562.   int64x2_t __ret;
  5563.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  5564.   return __ret;
  5565. }
  5566. #else
  5567. __ai int64x2_t vcombine_s64(int64x1_t __p0, int64x1_t __p1) {
  5568.   int64x2_t __ret;
  5569.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  5570.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5571.   return __ret;
  5572. }
  5573. #endif
  5574.  
  5575. #ifdef __LITTLE_ENDIAN__
  5576. __ai int16x8_t vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
  5577.   int16x8_t __ret;
  5578.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  5579.   return __ret;
  5580. }
  5581. #else
  5582. __ai int16x8_t vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
  5583.   int16x8_t __ret;
  5584.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5585.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  5586.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
  5587.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  5588.   return __ret;
  5589. }
  5590. __ai int16x8_t __noswap_vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
  5591.   int16x8_t __ret;
  5592.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  5593.   return __ret;
  5594. }
  5595. #endif
  5596.  
  5597. #define vcreate_p8(__p0) __extension__ ({ \
  5598.   poly8x8_t __ret; \
  5599.   uint64_t __promote = __p0; \
  5600.   __ret = (poly8x8_t)(__promote); \
  5601.   __ret; \
  5602. })
  5603. #define vcreate_p16(__p0) __extension__ ({ \
  5604.   poly16x4_t __ret; \
  5605.   uint64_t __promote = __p0; \
  5606.   __ret = (poly16x4_t)(__promote); \
  5607.   __ret; \
  5608. })
  5609. #define vcreate_u8(__p0) __extension__ ({ \
  5610.   uint8x8_t __ret; \
  5611.   uint64_t __promote = __p0; \
  5612.   __ret = (uint8x8_t)(__promote); \
  5613.   __ret; \
  5614. })
  5615. #define vcreate_u32(__p0) __extension__ ({ \
  5616.   uint32x2_t __ret; \
  5617.   uint64_t __promote = __p0; \
  5618.   __ret = (uint32x2_t)(__promote); \
  5619.   __ret; \
  5620. })
  5621. #define vcreate_u64(__p0) __extension__ ({ \
  5622.   uint64x1_t __ret; \
  5623.   uint64_t __promote = __p0; \
  5624.   __ret = (uint64x1_t)(__promote); \
  5625.   __ret; \
  5626. })
  5627. #define vcreate_u16(__p0) __extension__ ({ \
  5628.   uint16x4_t __ret; \
  5629.   uint64_t __promote = __p0; \
  5630.   __ret = (uint16x4_t)(__promote); \
  5631.   __ret; \
  5632. })
  5633. #define vcreate_s8(__p0) __extension__ ({ \
  5634.   int8x8_t __ret; \
  5635.   uint64_t __promote = __p0; \
  5636.   __ret = (int8x8_t)(__promote); \
  5637.   __ret; \
  5638. })
  5639. #define vcreate_f32(__p0) __extension__ ({ \
  5640.   float32x2_t __ret; \
  5641.   uint64_t __promote = __p0; \
  5642.   __ret = (float32x2_t)(__promote); \
  5643.   __ret; \
  5644. })
  5645. #define vcreate_f16(__p0) __extension__ ({ \
  5646.   float16x4_t __ret; \
  5647.   uint64_t __promote = __p0; \
  5648.   __ret = (float16x4_t)(__promote); \
  5649.   __ret; \
  5650. })
  5651. #define vcreate_s32(__p0) __extension__ ({ \
  5652.   int32x2_t __ret; \
  5653.   uint64_t __promote = __p0; \
  5654.   __ret = (int32x2_t)(__promote); \
  5655.   __ret; \
  5656. })
  5657. #define vcreate_s64(__p0) __extension__ ({ \
  5658.   int64x1_t __ret; \
  5659.   uint64_t __promote = __p0; \
  5660.   __ret = (int64x1_t)(__promote); \
  5661.   __ret; \
  5662. })
  5663. #define vcreate_s16(__p0) __extension__ ({ \
  5664.   int16x4_t __ret; \
  5665.   uint64_t __promote = __p0; \
  5666.   __ret = (int16x4_t)(__promote); \
  5667.   __ret; \
  5668. })
  5669. #ifdef __LITTLE_ENDIAN__
  5670. __ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
  5671.   float32x4_t __ret;
  5672.   __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__p0, 50);
  5673.   return __ret;
  5674. }
  5675. #else
  5676. __ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
  5677.   float32x4_t __ret;
  5678.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5679.   __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__rev0, 50);
  5680.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5681.   return __ret;
  5682. }
  5683. #endif
  5684.  
  5685. #ifdef __LITTLE_ENDIAN__
  5686. __ai float32x4_t vcvtq_f32_s32(int32x4_t __p0) {
  5687.   float32x4_t __ret;
  5688.   __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__p0, 34);
  5689.   return __ret;
  5690. }
  5691. #else
  5692. __ai float32x4_t vcvtq_f32_s32(int32x4_t __p0) {
  5693.   float32x4_t __ret;
  5694.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5695.   __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__rev0, 34);
  5696.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5697.   return __ret;
  5698. }
  5699. #endif
  5700.  
  5701. #ifdef __LITTLE_ENDIAN__
  5702. __ai float32x2_t vcvt_f32_u32(uint32x2_t __p0) {
  5703.   float32x2_t __ret;
  5704.   __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__p0, 18);
  5705.   return __ret;
  5706. }
  5707. #else
  5708. __ai float32x2_t vcvt_f32_u32(uint32x2_t __p0) {
  5709.   float32x2_t __ret;
  5710.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5711.   __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__rev0, 18);
  5712.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5713.   return __ret;
  5714. }
  5715. #endif
  5716.  
  5717. #ifdef __LITTLE_ENDIAN__
  5718. __ai float32x2_t vcvt_f32_s32(int32x2_t __p0) {
  5719.   float32x2_t __ret;
  5720.   __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__p0, 2);
  5721.   return __ret;
  5722. }
  5723. #else
  5724. __ai float32x2_t vcvt_f32_s32(int32x2_t __p0) {
  5725.   float32x2_t __ret;
  5726.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5727.   __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__rev0, 2);
  5728.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5729.   return __ret;
  5730. }
  5731. #endif
  5732.  
  5733. #ifdef __LITTLE_ENDIAN__
  5734. #define vcvtq_n_f32_u32(__p0, __p1) __extension__ ({ \
  5735.   float32x4_t __ret; \
  5736.   uint32x4_t __s0 = __p0; \
  5737.   __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__s0, __p1, 50); \
  5738.   __ret; \
  5739. })
  5740. #else
  5741. #define vcvtq_n_f32_u32(__p0, __p1) __extension__ ({ \
  5742.   float32x4_t __ret; \
  5743.   uint32x4_t __s0 = __p0; \
  5744.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  5745.   __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__rev0, __p1, 50); \
  5746.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  5747.   __ret; \
  5748. })
  5749. #endif
  5750.  
  5751. #ifdef __LITTLE_ENDIAN__
  5752. #define vcvtq_n_f32_s32(__p0, __p1) __extension__ ({ \
  5753.   float32x4_t __ret; \
  5754.   int32x4_t __s0 = __p0; \
  5755.   __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__s0, __p1, 34); \
  5756.   __ret; \
  5757. })
  5758. #else
  5759. #define vcvtq_n_f32_s32(__p0, __p1) __extension__ ({ \
  5760.   float32x4_t __ret; \
  5761.   int32x4_t __s0 = __p0; \
  5762.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  5763.   __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__rev0, __p1, 34); \
  5764.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  5765.   __ret; \
  5766. })
  5767. #endif
  5768.  
  5769. #ifdef __LITTLE_ENDIAN__
  5770. #define vcvt_n_f32_u32(__p0, __p1) __extension__ ({ \
  5771.   float32x2_t __ret; \
  5772.   uint32x2_t __s0 = __p0; \
  5773.   __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__s0, __p1, 18); \
  5774.   __ret; \
  5775. })
  5776. #else
  5777. #define vcvt_n_f32_u32(__p0, __p1) __extension__ ({ \
  5778.   float32x2_t __ret; \
  5779.   uint32x2_t __s0 = __p0; \
  5780.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  5781.   __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__rev0, __p1, 18); \
  5782.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  5783.   __ret; \
  5784. })
  5785. #endif
  5786.  
  5787. #ifdef __LITTLE_ENDIAN__
  5788. #define vcvt_n_f32_s32(__p0, __p1) __extension__ ({ \
  5789.   float32x2_t __ret; \
  5790.   int32x2_t __s0 = __p0; \
  5791.   __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__s0, __p1, 2); \
  5792.   __ret; \
  5793. })
  5794. #else
  5795. #define vcvt_n_f32_s32(__p0, __p1) __extension__ ({ \
  5796.   float32x2_t __ret; \
  5797.   int32x2_t __s0 = __p0; \
  5798.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  5799.   __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__rev0, __p1, 2); \
  5800.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  5801.   __ret; \
  5802. })
  5803. #endif
  5804.  
  5805. #ifdef __LITTLE_ENDIAN__
  5806. #define vcvtq_n_s32_f32(__p0, __p1) __extension__ ({ \
  5807.   int32x4_t __ret; \
  5808.   float32x4_t __s0 = __p0; \
  5809.   __ret = (int32x4_t) __builtin_neon_vcvtq_n_s32_v((int8x16_t)__s0, __p1, 34); \
  5810.   __ret; \
  5811. })
  5812. #else
  5813. #define vcvtq_n_s32_f32(__p0, __p1) __extension__ ({ \
  5814.   int32x4_t __ret; \
  5815.   float32x4_t __s0 = __p0; \
  5816.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  5817.   __ret = (int32x4_t) __builtin_neon_vcvtq_n_s32_v((int8x16_t)__rev0, __p1, 34); \
  5818.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  5819.   __ret; \
  5820. })
  5821. #endif
  5822.  
  5823. #ifdef __LITTLE_ENDIAN__
  5824. #define vcvt_n_s32_f32(__p0, __p1) __extension__ ({ \
  5825.   int32x2_t __ret; \
  5826.   float32x2_t __s0 = __p0; \
  5827.   __ret = (int32x2_t) __builtin_neon_vcvt_n_s32_v((int8x8_t)__s0, __p1, 2); \
  5828.   __ret; \
  5829. })
  5830. #else
  5831. #define vcvt_n_s32_f32(__p0, __p1) __extension__ ({ \
  5832.   int32x2_t __ret; \
  5833.   float32x2_t __s0 = __p0; \
  5834.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  5835.   __ret = (int32x2_t) __builtin_neon_vcvt_n_s32_v((int8x8_t)__rev0, __p1, 2); \
  5836.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  5837.   __ret; \
  5838. })
  5839. #endif
  5840.  
  5841. #ifdef __LITTLE_ENDIAN__
  5842. #define vcvtq_n_u32_f32(__p0, __p1) __extension__ ({ \
  5843.   uint32x4_t __ret; \
  5844.   float32x4_t __s0 = __p0; \
  5845.   __ret = (uint32x4_t) __builtin_neon_vcvtq_n_u32_v((int8x16_t)__s0, __p1, 50); \
  5846.   __ret; \
  5847. })
  5848. #else
  5849. #define vcvtq_n_u32_f32(__p0, __p1) __extension__ ({ \
  5850.   uint32x4_t __ret; \
  5851.   float32x4_t __s0 = __p0; \
  5852.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  5853.   __ret = (uint32x4_t) __builtin_neon_vcvtq_n_u32_v((int8x16_t)__rev0, __p1, 50); \
  5854.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  5855.   __ret; \
  5856. })
  5857. #endif
  5858.  
  5859. #ifdef __LITTLE_ENDIAN__
  5860. #define vcvt_n_u32_f32(__p0, __p1) __extension__ ({ \
  5861.   uint32x2_t __ret; \
  5862.   float32x2_t __s0 = __p0; \
  5863.   __ret = (uint32x2_t) __builtin_neon_vcvt_n_u32_v((int8x8_t)__s0, __p1, 18); \
  5864.   __ret; \
  5865. })
  5866. #else
  5867. #define vcvt_n_u32_f32(__p0, __p1) __extension__ ({ \
  5868.   uint32x2_t __ret; \
  5869.   float32x2_t __s0 = __p0; \
  5870.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  5871.   __ret = (uint32x2_t) __builtin_neon_vcvt_n_u32_v((int8x8_t)__rev0, __p1, 18); \
  5872.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  5873.   __ret; \
  5874. })
  5875. #endif
  5876.  
  5877. #ifdef __LITTLE_ENDIAN__
  5878. __ai int32x4_t vcvtq_s32_f32(float32x4_t __p0) {
  5879.   int32x4_t __ret;
  5880.   __ret = (int32x4_t) __builtin_neon_vcvtq_s32_v((int8x16_t)__p0, 34);
  5881.   return __ret;
  5882. }
  5883. #else
  5884. __ai int32x4_t vcvtq_s32_f32(float32x4_t __p0) {
  5885.   int32x4_t __ret;
  5886.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5887.   __ret = (int32x4_t) __builtin_neon_vcvtq_s32_v((int8x16_t)__rev0, 34);
  5888.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5889.   return __ret;
  5890. }
  5891. #endif
  5892.  
  5893. #ifdef __LITTLE_ENDIAN__
  5894. __ai int32x2_t vcvt_s32_f32(float32x2_t __p0) {
  5895.   int32x2_t __ret;
  5896.   __ret = (int32x2_t) __builtin_neon_vcvt_s32_v((int8x8_t)__p0, 2);
  5897.   return __ret;
  5898. }
  5899. #else
  5900. __ai int32x2_t vcvt_s32_f32(float32x2_t __p0) {
  5901.   int32x2_t __ret;
  5902.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5903.   __ret = (int32x2_t) __builtin_neon_vcvt_s32_v((int8x8_t)__rev0, 2);
  5904.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5905.   return __ret;
  5906. }
  5907. #endif
  5908.  
  5909. #ifdef __LITTLE_ENDIAN__
  5910. __ai uint32x4_t vcvtq_u32_f32(float32x4_t __p0) {
  5911.   uint32x4_t __ret;
  5912.   __ret = (uint32x4_t) __builtin_neon_vcvtq_u32_v((int8x16_t)__p0, 50);
  5913.   return __ret;
  5914. }
  5915. #else
  5916. __ai uint32x4_t vcvtq_u32_f32(float32x4_t __p0) {
  5917.   uint32x4_t __ret;
  5918.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  5919.   __ret = (uint32x4_t) __builtin_neon_vcvtq_u32_v((int8x16_t)__rev0, 50);
  5920.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  5921.   return __ret;
  5922. }
  5923. #endif
  5924.  
  5925. #ifdef __LITTLE_ENDIAN__
  5926. __ai uint32x2_t vcvt_u32_f32(float32x2_t __p0) {
  5927.   uint32x2_t __ret;
  5928.   __ret = (uint32x2_t) __builtin_neon_vcvt_u32_v((int8x8_t)__p0, 18);
  5929.   return __ret;
  5930. }
  5931. #else
  5932. __ai uint32x2_t vcvt_u32_f32(float32x2_t __p0) {
  5933.   uint32x2_t __ret;
  5934.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  5935.   __ret = (uint32x2_t) __builtin_neon_vcvt_u32_v((int8x8_t)__rev0, 18);
  5936.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  5937.   return __ret;
  5938. }
  5939. #endif
  5940.  
  5941. #ifdef __LITTLE_ENDIAN__
  5942. #define vdup_lane_p8(__p0_0, __p1_0) __extension__ ({ \
  5943.   poly8x8_t __ret_0; \
  5944.   poly8x8_t __s0_0 = __p0_0; \
  5945.   __ret_0 = splat_lane_p8(__s0_0, __p1_0); \
  5946.   __ret_0; \
  5947. })
  5948. #else
  5949. #define vdup_lane_p8(__p0_1, __p1_1) __extension__ ({ \
  5950.   poly8x8_t __ret_1; \
  5951.   poly8x8_t __s0_1 = __p0_1; \
  5952.   poly8x8_t __rev0_1;  __rev0_1 = __builtin_shufflevector(__s0_1, __s0_1, 7, 6, 5, 4, 3, 2, 1, 0); \
  5953.   __ret_1 = __noswap_splat_lane_p8(__rev0_1, __p1_1); \
  5954.   __ret_1 = __builtin_shufflevector(__ret_1, __ret_1, 7, 6, 5, 4, 3, 2, 1, 0); \
  5955.   __ret_1; \
  5956. })
  5957. #endif
  5958.  
  5959. #ifdef __LITTLE_ENDIAN__
  5960. #define vdup_lane_p16(__p0_2, __p1_2) __extension__ ({ \
  5961.   poly16x4_t __ret_2; \
  5962.   poly16x4_t __s0_2 = __p0_2; \
  5963.   __ret_2 = splat_lane_p16(__s0_2, __p1_2); \
  5964.   __ret_2; \
  5965. })
  5966. #else
  5967. #define vdup_lane_p16(__p0_3, __p1_3) __extension__ ({ \
  5968.   poly16x4_t __ret_3; \
  5969.   poly16x4_t __s0_3 = __p0_3; \
  5970.   poly16x4_t __rev0_3;  __rev0_3 = __builtin_shufflevector(__s0_3, __s0_3, 3, 2, 1, 0); \
  5971.   __ret_3 = __noswap_splat_lane_p16(__rev0_3, __p1_3); \
  5972.   __ret_3 = __builtin_shufflevector(__ret_3, __ret_3, 3, 2, 1, 0); \
  5973.   __ret_3; \
  5974. })
  5975. #endif
  5976.  
  5977. #ifdef __LITTLE_ENDIAN__
  5978. #define vdupq_lane_p8(__p0_4, __p1_4) __extension__ ({ \
  5979.   poly8x16_t __ret_4; \
  5980.   poly8x8_t __s0_4 = __p0_4; \
  5981.   __ret_4 = splatq_lane_p8(__s0_4, __p1_4); \
  5982.   __ret_4; \
  5983. })
  5984. #else
  5985. #define vdupq_lane_p8(__p0_5, __p1_5) __extension__ ({ \
  5986.   poly8x16_t __ret_5; \
  5987.   poly8x8_t __s0_5 = __p0_5; \
  5988.   poly8x8_t __rev0_5;  __rev0_5 = __builtin_shufflevector(__s0_5, __s0_5, 7, 6, 5, 4, 3, 2, 1, 0); \
  5989.   __ret_5 = __noswap_splatq_lane_p8(__rev0_5, __p1_5); \
  5990.   __ret_5 = __builtin_shufflevector(__ret_5, __ret_5, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  5991.   __ret_5; \
  5992. })
  5993. #endif
  5994.  
  5995. #ifdef __LITTLE_ENDIAN__
  5996. #define vdupq_lane_p16(__p0_6, __p1_6) __extension__ ({ \
  5997.   poly16x8_t __ret_6; \
  5998.   poly16x4_t __s0_6 = __p0_6; \
  5999.   __ret_6 = splatq_lane_p16(__s0_6, __p1_6); \
  6000.   __ret_6; \
  6001. })
  6002. #else
  6003. #define vdupq_lane_p16(__p0_7, __p1_7) __extension__ ({ \
  6004.   poly16x8_t __ret_7; \
  6005.   poly16x4_t __s0_7 = __p0_7; \
  6006.   poly16x4_t __rev0_7;  __rev0_7 = __builtin_shufflevector(__s0_7, __s0_7, 3, 2, 1, 0); \
  6007.   __ret_7 = __noswap_splatq_lane_p16(__rev0_7, __p1_7); \
  6008.   __ret_7 = __builtin_shufflevector(__ret_7, __ret_7, 7, 6, 5, 4, 3, 2, 1, 0); \
  6009.   __ret_7; \
  6010. })
  6011. #endif
  6012.  
  6013. #ifdef __LITTLE_ENDIAN__
  6014. #define vdupq_lane_u8(__p0_8, __p1_8) __extension__ ({ \
  6015.   uint8x16_t __ret_8; \
  6016.   uint8x8_t __s0_8 = __p0_8; \
  6017.   __ret_8 = splatq_lane_u8(__s0_8, __p1_8); \
  6018.   __ret_8; \
  6019. })
  6020. #else
  6021. #define vdupq_lane_u8(__p0_9, __p1_9) __extension__ ({ \
  6022.   uint8x16_t __ret_9; \
  6023.   uint8x8_t __s0_9 = __p0_9; \
  6024.   uint8x8_t __rev0_9;  __rev0_9 = __builtin_shufflevector(__s0_9, __s0_9, 7, 6, 5, 4, 3, 2, 1, 0); \
  6025.   __ret_9 = __noswap_splatq_lane_u8(__rev0_9, __p1_9); \
  6026.   __ret_9 = __builtin_shufflevector(__ret_9, __ret_9, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  6027.   __ret_9; \
  6028. })
  6029. #endif
  6030.  
  6031. #ifdef __LITTLE_ENDIAN__
  6032. #define vdupq_lane_u32(__p0_10, __p1_10) __extension__ ({ \
  6033.   uint32x4_t __ret_10; \
  6034.   uint32x2_t __s0_10 = __p0_10; \
  6035.   __ret_10 = splatq_lane_u32(__s0_10, __p1_10); \
  6036.   __ret_10; \
  6037. })
  6038. #else
  6039. #define vdupq_lane_u32(__p0_11, __p1_11) __extension__ ({ \
  6040.   uint32x4_t __ret_11; \
  6041.   uint32x2_t __s0_11 = __p0_11; \
  6042.   uint32x2_t __rev0_11;  __rev0_11 = __builtin_shufflevector(__s0_11, __s0_11, 1, 0); \
  6043.   __ret_11 = __noswap_splatq_lane_u32(__rev0_11, __p1_11); \
  6044.   __ret_11 = __builtin_shufflevector(__ret_11, __ret_11, 3, 2, 1, 0); \
  6045.   __ret_11; \
  6046. })
  6047. #endif
  6048.  
  6049. #ifdef __LITTLE_ENDIAN__
  6050. #define vdupq_lane_u64(__p0_12, __p1_12) __extension__ ({ \
  6051.   uint64x2_t __ret_12; \
  6052.   uint64x1_t __s0_12 = __p0_12; \
  6053.   __ret_12 = splatq_lane_u64(__s0_12, __p1_12); \
  6054.   __ret_12; \
  6055. })
  6056. #else
  6057. #define vdupq_lane_u64(__p0_13, __p1_13) __extension__ ({ \
  6058.   uint64x2_t __ret_13; \
  6059.   uint64x1_t __s0_13 = __p0_13; \
  6060.   __ret_13 = __noswap_splatq_lane_u64(__s0_13, __p1_13); \
  6061.   __ret_13 = __builtin_shufflevector(__ret_13, __ret_13, 1, 0); \
  6062.   __ret_13; \
  6063. })
  6064. #endif
  6065.  
  6066. #ifdef __LITTLE_ENDIAN__
  6067. #define vdupq_lane_u16(__p0_14, __p1_14) __extension__ ({ \
  6068.   uint16x8_t __ret_14; \
  6069.   uint16x4_t __s0_14 = __p0_14; \
  6070.   __ret_14 = splatq_lane_u16(__s0_14, __p1_14); \
  6071.   __ret_14; \
  6072. })
  6073. #else
  6074. #define vdupq_lane_u16(__p0_15, __p1_15) __extension__ ({ \
  6075.   uint16x8_t __ret_15; \
  6076.   uint16x4_t __s0_15 = __p0_15; \
  6077.   uint16x4_t __rev0_15;  __rev0_15 = __builtin_shufflevector(__s0_15, __s0_15, 3, 2, 1, 0); \
  6078.   __ret_15 = __noswap_splatq_lane_u16(__rev0_15, __p1_15); \
  6079.   __ret_15 = __builtin_shufflevector(__ret_15, __ret_15, 7, 6, 5, 4, 3, 2, 1, 0); \
  6080.   __ret_15; \
  6081. })
  6082. #endif
  6083.  
  6084. #ifdef __LITTLE_ENDIAN__
  6085. #define vdupq_lane_s8(__p0_16, __p1_16) __extension__ ({ \
  6086.   int8x16_t __ret_16; \
  6087.   int8x8_t __s0_16 = __p0_16; \
  6088.   __ret_16 = splatq_lane_s8(__s0_16, __p1_16); \
  6089.   __ret_16; \
  6090. })
  6091. #else
  6092. #define vdupq_lane_s8(__p0_17, __p1_17) __extension__ ({ \
  6093.   int8x16_t __ret_17; \
  6094.   int8x8_t __s0_17 = __p0_17; \
  6095.   int8x8_t __rev0_17;  __rev0_17 = __builtin_shufflevector(__s0_17, __s0_17, 7, 6, 5, 4, 3, 2, 1, 0); \
  6096.   __ret_17 = __noswap_splatq_lane_s8(__rev0_17, __p1_17); \
  6097.   __ret_17 = __builtin_shufflevector(__ret_17, __ret_17, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  6098.   __ret_17; \
  6099. })
  6100. #endif
  6101.  
  6102. #ifdef __LITTLE_ENDIAN__
  6103. #define vdupq_lane_f32(__p0_18, __p1_18) __extension__ ({ \
  6104.   float32x4_t __ret_18; \
  6105.   float32x2_t __s0_18 = __p0_18; \
  6106.   __ret_18 = splatq_lane_f32(__s0_18, __p1_18); \
  6107.   __ret_18; \
  6108. })
  6109. #else
  6110. #define vdupq_lane_f32(__p0_19, __p1_19) __extension__ ({ \
  6111.   float32x4_t __ret_19; \
  6112.   float32x2_t __s0_19 = __p0_19; \
  6113.   float32x2_t __rev0_19;  __rev0_19 = __builtin_shufflevector(__s0_19, __s0_19, 1, 0); \
  6114.   __ret_19 = __noswap_splatq_lane_f32(__rev0_19, __p1_19); \
  6115.   __ret_19 = __builtin_shufflevector(__ret_19, __ret_19, 3, 2, 1, 0); \
  6116.   __ret_19; \
  6117. })
  6118. #endif
  6119.  
  6120. #ifdef __LITTLE_ENDIAN__
  6121. #define vdupq_lane_f16(__p0_20, __p1_20) __extension__ ({ \
  6122.   float16x8_t __ret_20; \
  6123.   float16x4_t __s0_20 = __p0_20; \
  6124.   __ret_20 = splatq_lane_f16(__s0_20, __p1_20); \
  6125.   __ret_20; \
  6126. })
  6127. #else
  6128. #define vdupq_lane_f16(__p0_21, __p1_21) __extension__ ({ \
  6129.   float16x8_t __ret_21; \
  6130.   float16x4_t __s0_21 = __p0_21; \
  6131.   float16x4_t __rev0_21;  __rev0_21 = __builtin_shufflevector(__s0_21, __s0_21, 3, 2, 1, 0); \
  6132.   __ret_21 = __noswap_splatq_lane_f16(__rev0_21, __p1_21); \
  6133.   __ret_21 = __builtin_shufflevector(__ret_21, __ret_21, 7, 6, 5, 4, 3, 2, 1, 0); \
  6134.   __ret_21; \
  6135. })
  6136. #endif
  6137.  
  6138. #ifdef __LITTLE_ENDIAN__
  6139. #define vdupq_lane_s32(__p0_22, __p1_22) __extension__ ({ \
  6140.   int32x4_t __ret_22; \
  6141.   int32x2_t __s0_22 = __p0_22; \
  6142.   __ret_22 = splatq_lane_s32(__s0_22, __p1_22); \
  6143.   __ret_22; \
  6144. })
  6145. #else
  6146. #define vdupq_lane_s32(__p0_23, __p1_23) __extension__ ({ \
  6147.   int32x4_t __ret_23; \
  6148.   int32x2_t __s0_23 = __p0_23; \
  6149.   int32x2_t __rev0_23;  __rev0_23 = __builtin_shufflevector(__s0_23, __s0_23, 1, 0); \
  6150.   __ret_23 = __noswap_splatq_lane_s32(__rev0_23, __p1_23); \
  6151.   __ret_23 = __builtin_shufflevector(__ret_23, __ret_23, 3, 2, 1, 0); \
  6152.   __ret_23; \
  6153. })
  6154. #endif
  6155.  
  6156. #ifdef __LITTLE_ENDIAN__
  6157. #define vdupq_lane_s64(__p0_24, __p1_24) __extension__ ({ \
  6158.   int64x2_t __ret_24; \
  6159.   int64x1_t __s0_24 = __p0_24; \
  6160.   __ret_24 = splatq_lane_s64(__s0_24, __p1_24); \
  6161.   __ret_24; \
  6162. })
  6163. #else
  6164. #define vdupq_lane_s64(__p0_25, __p1_25) __extension__ ({ \
  6165.   int64x2_t __ret_25; \
  6166.   int64x1_t __s0_25 = __p0_25; \
  6167.   __ret_25 = __noswap_splatq_lane_s64(__s0_25, __p1_25); \
  6168.   __ret_25 = __builtin_shufflevector(__ret_25, __ret_25, 1, 0); \
  6169.   __ret_25; \
  6170. })
  6171. #endif
  6172.  
  6173. #ifdef __LITTLE_ENDIAN__
  6174. #define vdupq_lane_s16(__p0_26, __p1_26) __extension__ ({ \
  6175.   int16x8_t __ret_26; \
  6176.   int16x4_t __s0_26 = __p0_26; \
  6177.   __ret_26 = splatq_lane_s16(__s0_26, __p1_26); \
  6178.   __ret_26; \
  6179. })
  6180. #else
  6181. #define vdupq_lane_s16(__p0_27, __p1_27) __extension__ ({ \
  6182.   int16x8_t __ret_27; \
  6183.   int16x4_t __s0_27 = __p0_27; \
  6184.   int16x4_t __rev0_27;  __rev0_27 = __builtin_shufflevector(__s0_27, __s0_27, 3, 2, 1, 0); \
  6185.   __ret_27 = __noswap_splatq_lane_s16(__rev0_27, __p1_27); \
  6186.   __ret_27 = __builtin_shufflevector(__ret_27, __ret_27, 7, 6, 5, 4, 3, 2, 1, 0); \
  6187.   __ret_27; \
  6188. })
  6189. #endif
  6190.  
  6191. #ifdef __LITTLE_ENDIAN__
  6192. #define vdup_lane_u8(__p0_28, __p1_28) __extension__ ({ \
  6193.   uint8x8_t __ret_28; \
  6194.   uint8x8_t __s0_28 = __p0_28; \
  6195.   __ret_28 = splat_lane_u8(__s0_28, __p1_28); \
  6196.   __ret_28; \
  6197. })
  6198. #else
  6199. #define vdup_lane_u8(__p0_29, __p1_29) __extension__ ({ \
  6200.   uint8x8_t __ret_29; \
  6201.   uint8x8_t __s0_29 = __p0_29; \
  6202.   uint8x8_t __rev0_29;  __rev0_29 = __builtin_shufflevector(__s0_29, __s0_29, 7, 6, 5, 4, 3, 2, 1, 0); \
  6203.   __ret_29 = __noswap_splat_lane_u8(__rev0_29, __p1_29); \
  6204.   __ret_29 = __builtin_shufflevector(__ret_29, __ret_29, 7, 6, 5, 4, 3, 2, 1, 0); \
  6205.   __ret_29; \
  6206. })
  6207. #endif
  6208.  
  6209. #ifdef __LITTLE_ENDIAN__
  6210. #define vdup_lane_u32(__p0_30, __p1_30) __extension__ ({ \
  6211.   uint32x2_t __ret_30; \
  6212.   uint32x2_t __s0_30 = __p0_30; \
  6213.   __ret_30 = splat_lane_u32(__s0_30, __p1_30); \
  6214.   __ret_30; \
  6215. })
  6216. #else
  6217. #define vdup_lane_u32(__p0_31, __p1_31) __extension__ ({ \
  6218.   uint32x2_t __ret_31; \
  6219.   uint32x2_t __s0_31 = __p0_31; \
  6220.   uint32x2_t __rev0_31;  __rev0_31 = __builtin_shufflevector(__s0_31, __s0_31, 1, 0); \
  6221.   __ret_31 = __noswap_splat_lane_u32(__rev0_31, __p1_31); \
  6222.   __ret_31 = __builtin_shufflevector(__ret_31, __ret_31, 1, 0); \
  6223.   __ret_31; \
  6224. })
  6225. #endif
  6226.  
  6227. #define vdup_lane_u64(__p0_32, __p1_32) __extension__ ({ \
  6228.   uint64x1_t __ret_32; \
  6229.   uint64x1_t __s0_32 = __p0_32; \
  6230.   __ret_32 = splat_lane_u64(__s0_32, __p1_32); \
  6231.   __ret_32; \
  6232. })
  6233. #ifdef __LITTLE_ENDIAN__
  6234. #define vdup_lane_u16(__p0_33, __p1_33) __extension__ ({ \
  6235.   uint16x4_t __ret_33; \
  6236.   uint16x4_t __s0_33 = __p0_33; \
  6237.   __ret_33 = splat_lane_u16(__s0_33, __p1_33); \
  6238.   __ret_33; \
  6239. })
  6240. #else
  6241. #define vdup_lane_u16(__p0_34, __p1_34) __extension__ ({ \
  6242.   uint16x4_t __ret_34; \
  6243.   uint16x4_t __s0_34 = __p0_34; \
  6244.   uint16x4_t __rev0_34;  __rev0_34 = __builtin_shufflevector(__s0_34, __s0_34, 3, 2, 1, 0); \
  6245.   __ret_34 = __noswap_splat_lane_u16(__rev0_34, __p1_34); \
  6246.   __ret_34 = __builtin_shufflevector(__ret_34, __ret_34, 3, 2, 1, 0); \
  6247.   __ret_34; \
  6248. })
  6249. #endif
  6250.  
  6251. #ifdef __LITTLE_ENDIAN__
  6252. #define vdup_lane_s8(__p0_35, __p1_35) __extension__ ({ \
  6253.   int8x8_t __ret_35; \
  6254.   int8x8_t __s0_35 = __p0_35; \
  6255.   __ret_35 = splat_lane_s8(__s0_35, __p1_35); \
  6256.   __ret_35; \
  6257. })
  6258. #else
  6259. #define vdup_lane_s8(__p0_36, __p1_36) __extension__ ({ \
  6260.   int8x8_t __ret_36; \
  6261.   int8x8_t __s0_36 = __p0_36; \
  6262.   int8x8_t __rev0_36;  __rev0_36 = __builtin_shufflevector(__s0_36, __s0_36, 7, 6, 5, 4, 3, 2, 1, 0); \
  6263.   __ret_36 = __noswap_splat_lane_s8(__rev0_36, __p1_36); \
  6264.   __ret_36 = __builtin_shufflevector(__ret_36, __ret_36, 7, 6, 5, 4, 3, 2, 1, 0); \
  6265.   __ret_36; \
  6266. })
  6267. #endif
  6268.  
  6269. #ifdef __LITTLE_ENDIAN__
  6270. #define vdup_lane_f32(__p0_37, __p1_37) __extension__ ({ \
  6271.   float32x2_t __ret_37; \
  6272.   float32x2_t __s0_37 = __p0_37; \
  6273.   __ret_37 = splat_lane_f32(__s0_37, __p1_37); \
  6274.   __ret_37; \
  6275. })
  6276. #else
  6277. #define vdup_lane_f32(__p0_38, __p1_38) __extension__ ({ \
  6278.   float32x2_t __ret_38; \
  6279.   float32x2_t __s0_38 = __p0_38; \
  6280.   float32x2_t __rev0_38;  __rev0_38 = __builtin_shufflevector(__s0_38, __s0_38, 1, 0); \
  6281.   __ret_38 = __noswap_splat_lane_f32(__rev0_38, __p1_38); \
  6282.   __ret_38 = __builtin_shufflevector(__ret_38, __ret_38, 1, 0); \
  6283.   __ret_38; \
  6284. })
  6285. #endif
  6286.  
  6287. #ifdef __LITTLE_ENDIAN__
  6288. #define vdup_lane_f16(__p0_39, __p1_39) __extension__ ({ \
  6289.   float16x4_t __ret_39; \
  6290.   float16x4_t __s0_39 = __p0_39; \
  6291.   __ret_39 = splat_lane_f16(__s0_39, __p1_39); \
  6292.   __ret_39; \
  6293. })
  6294. #else
  6295. #define vdup_lane_f16(__p0_40, __p1_40) __extension__ ({ \
  6296.   float16x4_t __ret_40; \
  6297.   float16x4_t __s0_40 = __p0_40; \
  6298.   float16x4_t __rev0_40;  __rev0_40 = __builtin_shufflevector(__s0_40, __s0_40, 3, 2, 1, 0); \
  6299.   __ret_40 = __noswap_splat_lane_f16(__rev0_40, __p1_40); \
  6300.   __ret_40 = __builtin_shufflevector(__ret_40, __ret_40, 3, 2, 1, 0); \
  6301.   __ret_40; \
  6302. })
  6303. #endif
  6304.  
  6305. #ifdef __LITTLE_ENDIAN__
  6306. #define vdup_lane_s32(__p0_41, __p1_41) __extension__ ({ \
  6307.   int32x2_t __ret_41; \
  6308.   int32x2_t __s0_41 = __p0_41; \
  6309.   __ret_41 = splat_lane_s32(__s0_41, __p1_41); \
  6310.   __ret_41; \
  6311. })
  6312. #else
  6313. #define vdup_lane_s32(__p0_42, __p1_42) __extension__ ({ \
  6314.   int32x2_t __ret_42; \
  6315.   int32x2_t __s0_42 = __p0_42; \
  6316.   int32x2_t __rev0_42;  __rev0_42 = __builtin_shufflevector(__s0_42, __s0_42, 1, 0); \
  6317.   __ret_42 = __noswap_splat_lane_s32(__rev0_42, __p1_42); \
  6318.   __ret_42 = __builtin_shufflevector(__ret_42, __ret_42, 1, 0); \
  6319.   __ret_42; \
  6320. })
  6321. #endif
  6322.  
  6323. #define vdup_lane_s64(__p0_43, __p1_43) __extension__ ({ \
  6324.   int64x1_t __ret_43; \
  6325.   int64x1_t __s0_43 = __p0_43; \
  6326.   __ret_43 = splat_lane_s64(__s0_43, __p1_43); \
  6327.   __ret_43; \
  6328. })
  6329. #ifdef __LITTLE_ENDIAN__
  6330. #define vdup_lane_s16(__p0_44, __p1_44) __extension__ ({ \
  6331.   int16x4_t __ret_44; \
  6332.   int16x4_t __s0_44 = __p0_44; \
  6333.   __ret_44 = splat_lane_s16(__s0_44, __p1_44); \
  6334.   __ret_44; \
  6335. })
  6336. #else
  6337. #define vdup_lane_s16(__p0_45, __p1_45) __extension__ ({ \
  6338.   int16x4_t __ret_45; \
  6339.   int16x4_t __s0_45 = __p0_45; \
  6340.   int16x4_t __rev0_45;  __rev0_45 = __builtin_shufflevector(__s0_45, __s0_45, 3, 2, 1, 0); \
  6341.   __ret_45 = __noswap_splat_lane_s16(__rev0_45, __p1_45); \
  6342.   __ret_45 = __builtin_shufflevector(__ret_45, __ret_45, 3, 2, 1, 0); \
  6343.   __ret_45; \
  6344. })
  6345. #endif
  6346.  
  6347. #ifdef __LITTLE_ENDIAN__
  6348. __ai poly8x8_t vdup_n_p8(poly8_t __p0) {
  6349.   poly8x8_t __ret;
  6350.   __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6351.   return __ret;
  6352. }
  6353. #else
  6354. __ai poly8x8_t vdup_n_p8(poly8_t __p0) {
  6355.   poly8x8_t __ret;
  6356.   __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6357.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6358.   return __ret;
  6359. }
  6360. #endif
  6361.  
  6362. #ifdef __LITTLE_ENDIAN__
  6363. __ai poly16x4_t vdup_n_p16(poly16_t __p0) {
  6364.   poly16x4_t __ret;
  6365.   __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
  6366.   return __ret;
  6367. }
  6368. #else
  6369. __ai poly16x4_t vdup_n_p16(poly16_t __p0) {
  6370.   poly16x4_t __ret;
  6371.   __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
  6372.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6373.   return __ret;
  6374. }
  6375. #endif
  6376.  
  6377. #ifdef __LITTLE_ENDIAN__
  6378. __ai poly8x16_t vdupq_n_p8(poly8_t __p0) {
  6379.   poly8x16_t __ret;
  6380.   __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6381.   return __ret;
  6382. }
  6383. #else
  6384. __ai poly8x16_t vdupq_n_p8(poly8_t __p0) {
  6385.   poly8x16_t __ret;
  6386.   __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6387.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  6388.   return __ret;
  6389. }
  6390. #endif
  6391.  
  6392. #ifdef __LITTLE_ENDIAN__
  6393. __ai poly16x8_t vdupq_n_p16(poly16_t __p0) {
  6394.   poly16x8_t __ret;
  6395.   __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6396.   return __ret;
  6397. }
  6398. #else
  6399. __ai poly16x8_t vdupq_n_p16(poly16_t __p0) {
  6400.   poly16x8_t __ret;
  6401.   __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6402.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6403.   return __ret;
  6404. }
  6405. #endif
  6406.  
  6407. #ifdef __LITTLE_ENDIAN__
  6408. __ai uint8x16_t vdupq_n_u8(uint8_t __p0) {
  6409.   uint8x16_t __ret;
  6410.   __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6411.   return __ret;
  6412. }
  6413. #else
  6414. __ai uint8x16_t vdupq_n_u8(uint8_t __p0) {
  6415.   uint8x16_t __ret;
  6416.   __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6417.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  6418.   return __ret;
  6419. }
  6420. #endif
  6421.  
  6422. #ifdef __LITTLE_ENDIAN__
  6423. __ai uint32x4_t vdupq_n_u32(uint32_t __p0) {
  6424.   uint32x4_t __ret;
  6425.   __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
  6426.   return __ret;
  6427. }
  6428. #else
  6429. __ai uint32x4_t vdupq_n_u32(uint32_t __p0) {
  6430.   uint32x4_t __ret;
  6431.   __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
  6432.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6433.   return __ret;
  6434. }
  6435. #endif
  6436.  
  6437. #ifdef __LITTLE_ENDIAN__
  6438. __ai uint64x2_t vdupq_n_u64(uint64_t __p0) {
  6439.   uint64x2_t __ret;
  6440.   __ret = (uint64x2_t) {__p0, __p0};
  6441.   return __ret;
  6442. }
  6443. #else
  6444. __ai uint64x2_t vdupq_n_u64(uint64_t __p0) {
  6445.   uint64x2_t __ret;
  6446.   __ret = (uint64x2_t) {__p0, __p0};
  6447.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  6448.   return __ret;
  6449. }
  6450. #endif
  6451.  
  6452. #ifdef __LITTLE_ENDIAN__
  6453. __ai uint16x8_t vdupq_n_u16(uint16_t __p0) {
  6454.   uint16x8_t __ret;
  6455.   __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6456.   return __ret;
  6457. }
  6458. #else
  6459. __ai uint16x8_t vdupq_n_u16(uint16_t __p0) {
  6460.   uint16x8_t __ret;
  6461.   __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6462.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6463.   return __ret;
  6464. }
  6465. #endif
  6466.  
  6467. #ifdef __LITTLE_ENDIAN__
  6468. __ai int8x16_t vdupq_n_s8(int8_t __p0) {
  6469.   int8x16_t __ret;
  6470.   __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6471.   return __ret;
  6472. }
  6473. #else
  6474. __ai int8x16_t vdupq_n_s8(int8_t __p0) {
  6475.   int8x16_t __ret;
  6476.   __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6477.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  6478.   return __ret;
  6479. }
  6480. #endif
  6481.  
  6482. #ifdef __LITTLE_ENDIAN__
  6483. __ai float32x4_t vdupq_n_f32(float32_t __p0) {
  6484.   float32x4_t __ret;
  6485.   __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
  6486.   return __ret;
  6487. }
  6488. #else
  6489. __ai float32x4_t vdupq_n_f32(float32_t __p0) {
  6490.   float32x4_t __ret;
  6491.   __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
  6492.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6493.   return __ret;
  6494. }
  6495. #endif
  6496.  
  6497. #ifdef __LITTLE_ENDIAN__
  6498. #define vdupq_n_f16(__p0) __extension__ ({ \
  6499.   float16x8_t __ret; \
  6500.   float16_t __s0 = __p0; \
  6501.   __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
  6502.   __ret; \
  6503. })
  6504. #else
  6505. #define vdupq_n_f16(__p0) __extension__ ({ \
  6506.   float16x8_t __ret; \
  6507.   float16_t __s0 = __p0; \
  6508.   __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
  6509.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  6510.   __ret; \
  6511. })
  6512. #endif
  6513.  
  6514. #ifdef __LITTLE_ENDIAN__
  6515. __ai int32x4_t vdupq_n_s32(int32_t __p0) {
  6516.   int32x4_t __ret;
  6517.   __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
  6518.   return __ret;
  6519. }
  6520. #else
  6521. __ai int32x4_t vdupq_n_s32(int32_t __p0) {
  6522.   int32x4_t __ret;
  6523.   __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
  6524.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6525.   return __ret;
  6526. }
  6527. #endif
  6528.  
  6529. #ifdef __LITTLE_ENDIAN__
  6530. __ai int64x2_t vdupq_n_s64(int64_t __p0) {
  6531.   int64x2_t __ret;
  6532.   __ret = (int64x2_t) {__p0, __p0};
  6533.   return __ret;
  6534. }
  6535. #else
  6536. __ai int64x2_t vdupq_n_s64(int64_t __p0) {
  6537.   int64x2_t __ret;
  6538.   __ret = (int64x2_t) {__p0, __p0};
  6539.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  6540.   return __ret;
  6541. }
  6542. #endif
  6543.  
  6544. #ifdef __LITTLE_ENDIAN__
  6545. __ai int16x8_t vdupq_n_s16(int16_t __p0) {
  6546.   int16x8_t __ret;
  6547.   __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6548.   return __ret;
  6549. }
  6550. #else
  6551. __ai int16x8_t vdupq_n_s16(int16_t __p0) {
  6552.   int16x8_t __ret;
  6553.   __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6554.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6555.   return __ret;
  6556. }
  6557. #endif
  6558.  
  6559. #ifdef __LITTLE_ENDIAN__
  6560. __ai uint8x8_t vdup_n_u8(uint8_t __p0) {
  6561.   uint8x8_t __ret;
  6562.   __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6563.   return __ret;
  6564. }
  6565. #else
  6566. __ai uint8x8_t vdup_n_u8(uint8_t __p0) {
  6567.   uint8x8_t __ret;
  6568.   __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6569.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6570.   return __ret;
  6571. }
  6572. #endif
  6573.  
  6574. #ifdef __LITTLE_ENDIAN__
  6575. __ai uint32x2_t vdup_n_u32(uint32_t __p0) {
  6576.   uint32x2_t __ret;
  6577.   __ret = (uint32x2_t) {__p0, __p0};
  6578.   return __ret;
  6579. }
  6580. #else
  6581. __ai uint32x2_t vdup_n_u32(uint32_t __p0) {
  6582.   uint32x2_t __ret;
  6583.   __ret = (uint32x2_t) {__p0, __p0};
  6584.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  6585.   return __ret;
  6586. }
  6587. #endif
  6588.  
  6589. __ai uint64x1_t vdup_n_u64(uint64_t __p0) {
  6590.   uint64x1_t __ret;
  6591.   __ret = (uint64x1_t) {__p0};
  6592.   return __ret;
  6593. }
  6594. #ifdef __LITTLE_ENDIAN__
  6595. __ai uint16x4_t vdup_n_u16(uint16_t __p0) {
  6596.   uint16x4_t __ret;
  6597.   __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
  6598.   return __ret;
  6599. }
  6600. #else
  6601. __ai uint16x4_t vdup_n_u16(uint16_t __p0) {
  6602.   uint16x4_t __ret;
  6603.   __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
  6604.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6605.   return __ret;
  6606. }
  6607. #endif
  6608.  
  6609. #ifdef __LITTLE_ENDIAN__
  6610. __ai int8x8_t vdup_n_s8(int8_t __p0) {
  6611.   int8x8_t __ret;
  6612.   __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6613.   return __ret;
  6614. }
  6615. #else
  6616. __ai int8x8_t vdup_n_s8(int8_t __p0) {
  6617.   int8x8_t __ret;
  6618.   __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  6619.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6620.   return __ret;
  6621. }
  6622. #endif
  6623.  
  6624. #ifdef __LITTLE_ENDIAN__
  6625. __ai float32x2_t vdup_n_f32(float32_t __p0) {
  6626.   float32x2_t __ret;
  6627.   __ret = (float32x2_t) {__p0, __p0};
  6628.   return __ret;
  6629. }
  6630. #else
  6631. __ai float32x2_t vdup_n_f32(float32_t __p0) {
  6632.   float32x2_t __ret;
  6633.   __ret = (float32x2_t) {__p0, __p0};
  6634.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  6635.   return __ret;
  6636. }
  6637. #endif
  6638.  
  6639. #ifdef __LITTLE_ENDIAN__
  6640. #define vdup_n_f16(__p0) __extension__ ({ \
  6641.   float16x4_t __ret; \
  6642.   float16_t __s0 = __p0; \
  6643.   __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
  6644.   __ret; \
  6645. })
  6646. #else
  6647. #define vdup_n_f16(__p0) __extension__ ({ \
  6648.   float16x4_t __ret; \
  6649.   float16_t __s0 = __p0; \
  6650.   __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
  6651.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  6652.   __ret; \
  6653. })
  6654. #endif
  6655.  
  6656. #ifdef __LITTLE_ENDIAN__
  6657. __ai int32x2_t vdup_n_s32(int32_t __p0) {
  6658.   int32x2_t __ret;
  6659.   __ret = (int32x2_t) {__p0, __p0};
  6660.   return __ret;
  6661. }
  6662. #else
  6663. __ai int32x2_t vdup_n_s32(int32_t __p0) {
  6664.   int32x2_t __ret;
  6665.   __ret = (int32x2_t) {__p0, __p0};
  6666.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  6667.   return __ret;
  6668. }
  6669. #endif
  6670.  
  6671. __ai int64x1_t vdup_n_s64(int64_t __p0) {
  6672.   int64x1_t __ret;
  6673.   __ret = (int64x1_t) {__p0};
  6674.   return __ret;
  6675. }
  6676. #ifdef __LITTLE_ENDIAN__
  6677. __ai int16x4_t vdup_n_s16(int16_t __p0) {
  6678.   int16x4_t __ret;
  6679.   __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
  6680.   return __ret;
  6681. }
  6682. #else
  6683. __ai int16x4_t vdup_n_s16(int16_t __p0) {
  6684.   int16x4_t __ret;
  6685.   __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
  6686.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6687.   return __ret;
  6688. }
  6689. #endif
  6690.  
  6691. #ifdef __LITTLE_ENDIAN__
  6692. __ai uint8x16_t veorq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  6693.   uint8x16_t __ret;
  6694.   __ret = __p0 ^ __p1;
  6695.   return __ret;
  6696. }
  6697. #else
  6698. __ai uint8x16_t veorq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  6699.   uint8x16_t __ret;
  6700.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  6701.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  6702.   __ret = __rev0 ^ __rev1;
  6703.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  6704.   return __ret;
  6705. }
  6706. #endif
  6707.  
  6708. #ifdef __LITTLE_ENDIAN__
  6709. __ai uint32x4_t veorq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  6710.   uint32x4_t __ret;
  6711.   __ret = __p0 ^ __p1;
  6712.   return __ret;
  6713. }
  6714. #else
  6715. __ai uint32x4_t veorq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  6716.   uint32x4_t __ret;
  6717.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  6718.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  6719.   __ret = __rev0 ^ __rev1;
  6720.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6721.   return __ret;
  6722. }
  6723. #endif
  6724.  
  6725. #ifdef __LITTLE_ENDIAN__
  6726. __ai uint64x2_t veorq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  6727.   uint64x2_t __ret;
  6728.   __ret = __p0 ^ __p1;
  6729.   return __ret;
  6730. }
  6731. #else
  6732. __ai uint64x2_t veorq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  6733.   uint64x2_t __ret;
  6734.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  6735.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  6736.   __ret = __rev0 ^ __rev1;
  6737.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  6738.   return __ret;
  6739. }
  6740. #endif
  6741.  
  6742. #ifdef __LITTLE_ENDIAN__
  6743. __ai uint16x8_t veorq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  6744.   uint16x8_t __ret;
  6745.   __ret = __p0 ^ __p1;
  6746.   return __ret;
  6747. }
  6748. #else
  6749. __ai uint16x8_t veorq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  6750.   uint16x8_t __ret;
  6751.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  6752.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  6753.   __ret = __rev0 ^ __rev1;
  6754.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6755.   return __ret;
  6756. }
  6757. #endif
  6758.  
  6759. #ifdef __LITTLE_ENDIAN__
  6760. __ai int8x16_t veorq_s8(int8x16_t __p0, int8x16_t __p1) {
  6761.   int8x16_t __ret;
  6762.   __ret = __p0 ^ __p1;
  6763.   return __ret;
  6764. }
  6765. #else
  6766. __ai int8x16_t veorq_s8(int8x16_t __p0, int8x16_t __p1) {
  6767.   int8x16_t __ret;
  6768.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  6769.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  6770.   __ret = __rev0 ^ __rev1;
  6771.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  6772.   return __ret;
  6773. }
  6774. #endif
  6775.  
  6776. #ifdef __LITTLE_ENDIAN__
  6777. __ai int32x4_t veorq_s32(int32x4_t __p0, int32x4_t __p1) {
  6778.   int32x4_t __ret;
  6779.   __ret = __p0 ^ __p1;
  6780.   return __ret;
  6781. }
  6782. #else
  6783. __ai int32x4_t veorq_s32(int32x4_t __p0, int32x4_t __p1) {
  6784.   int32x4_t __ret;
  6785.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  6786.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  6787.   __ret = __rev0 ^ __rev1;
  6788.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6789.   return __ret;
  6790. }
  6791. #endif
  6792.  
  6793. #ifdef __LITTLE_ENDIAN__
  6794. __ai int64x2_t veorq_s64(int64x2_t __p0, int64x2_t __p1) {
  6795.   int64x2_t __ret;
  6796.   __ret = __p0 ^ __p1;
  6797.   return __ret;
  6798. }
  6799. #else
  6800. __ai int64x2_t veorq_s64(int64x2_t __p0, int64x2_t __p1) {
  6801.   int64x2_t __ret;
  6802.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  6803.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  6804.   __ret = __rev0 ^ __rev1;
  6805.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  6806.   return __ret;
  6807. }
  6808. #endif
  6809.  
  6810. #ifdef __LITTLE_ENDIAN__
  6811. __ai int16x8_t veorq_s16(int16x8_t __p0, int16x8_t __p1) {
  6812.   int16x8_t __ret;
  6813.   __ret = __p0 ^ __p1;
  6814.   return __ret;
  6815. }
  6816. #else
  6817. __ai int16x8_t veorq_s16(int16x8_t __p0, int16x8_t __p1) {
  6818.   int16x8_t __ret;
  6819.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  6820.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  6821.   __ret = __rev0 ^ __rev1;
  6822.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6823.   return __ret;
  6824. }
  6825. #endif
  6826.  
  6827. #ifdef __LITTLE_ENDIAN__
  6828. __ai uint8x8_t veor_u8(uint8x8_t __p0, uint8x8_t __p1) {
  6829.   uint8x8_t __ret;
  6830.   __ret = __p0 ^ __p1;
  6831.   return __ret;
  6832. }
  6833. #else
  6834. __ai uint8x8_t veor_u8(uint8x8_t __p0, uint8x8_t __p1) {
  6835.   uint8x8_t __ret;
  6836.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  6837.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  6838.   __ret = __rev0 ^ __rev1;
  6839.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6840.   return __ret;
  6841. }
  6842. #endif
  6843.  
  6844. #ifdef __LITTLE_ENDIAN__
  6845. __ai uint32x2_t veor_u32(uint32x2_t __p0, uint32x2_t __p1) {
  6846.   uint32x2_t __ret;
  6847.   __ret = __p0 ^ __p1;
  6848.   return __ret;
  6849. }
  6850. #else
  6851. __ai uint32x2_t veor_u32(uint32x2_t __p0, uint32x2_t __p1) {
  6852.   uint32x2_t __ret;
  6853.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  6854.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  6855.   __ret = __rev0 ^ __rev1;
  6856.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  6857.   return __ret;
  6858. }
  6859. #endif
  6860.  
  6861. __ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
  6862.   uint64x1_t __ret;
  6863.   __ret = __p0 ^ __p1;
  6864.   return __ret;
  6865. }
  6866. #ifdef __LITTLE_ENDIAN__
  6867. __ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
  6868.   uint16x4_t __ret;
  6869.   __ret = __p0 ^ __p1;
  6870.   return __ret;
  6871. }
  6872. #else
  6873. __ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
  6874.   uint16x4_t __ret;
  6875.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  6876.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  6877.   __ret = __rev0 ^ __rev1;
  6878.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6879.   return __ret;
  6880. }
  6881. #endif
  6882.  
  6883. #ifdef __LITTLE_ENDIAN__
  6884. __ai int8x8_t veor_s8(int8x8_t __p0, int8x8_t __p1) {
  6885.   int8x8_t __ret;
  6886.   __ret = __p0 ^ __p1;
  6887.   return __ret;
  6888. }
  6889. #else
  6890. __ai int8x8_t veor_s8(int8x8_t __p0, int8x8_t __p1) {
  6891.   int8x8_t __ret;
  6892.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  6893.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  6894.   __ret = __rev0 ^ __rev1;
  6895.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  6896.   return __ret;
  6897. }
  6898. #endif
  6899.  
  6900. #ifdef __LITTLE_ENDIAN__
  6901. __ai int32x2_t veor_s32(int32x2_t __p0, int32x2_t __p1) {
  6902.   int32x2_t __ret;
  6903.   __ret = __p0 ^ __p1;
  6904.   return __ret;
  6905. }
  6906. #else
  6907. __ai int32x2_t veor_s32(int32x2_t __p0, int32x2_t __p1) {
  6908.   int32x2_t __ret;
  6909.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  6910.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  6911.   __ret = __rev0 ^ __rev1;
  6912.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  6913.   return __ret;
  6914. }
  6915. #endif
  6916.  
  6917. __ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
  6918.   int64x1_t __ret;
  6919.   __ret = __p0 ^ __p1;
  6920.   return __ret;
  6921. }
  6922. #ifdef __LITTLE_ENDIAN__
  6923. __ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
  6924.   int16x4_t __ret;
  6925.   __ret = __p0 ^ __p1;
  6926.   return __ret;
  6927. }
  6928. #else
  6929. __ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
  6930.   int16x4_t __ret;
  6931.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  6932.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  6933.   __ret = __rev0 ^ __rev1;
  6934.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  6935.   return __ret;
  6936. }
  6937. #endif
  6938.  
  6939. #ifdef __LITTLE_ENDIAN__
  6940. #define vext_p8(__p0, __p1, __p2) __extension__ ({ \
  6941.   poly8x8_t __ret; \
  6942.   poly8x8_t __s0 = __p0; \
  6943.   poly8x8_t __s1 = __p1; \
  6944.   __ret = (poly8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
  6945.   __ret; \
  6946. })
  6947. #else
  6948. #define vext_p8(__p0, __p1, __p2) __extension__ ({ \
  6949.   poly8x8_t __ret; \
  6950.   poly8x8_t __s0 = __p0; \
  6951.   poly8x8_t __s1 = __p1; \
  6952.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  6953.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  6954.   __ret = (poly8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
  6955.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  6956.   __ret; \
  6957. })
  6958. #endif
  6959.  
  6960. #ifdef __LITTLE_ENDIAN__
  6961. #define vext_p16(__p0, __p1, __p2) __extension__ ({ \
  6962.   poly16x4_t __ret; \
  6963.   poly16x4_t __s0 = __p0; \
  6964.   poly16x4_t __s1 = __p1; \
  6965.   __ret = (poly16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
  6966.   __ret; \
  6967. })
  6968. #else
  6969. #define vext_p16(__p0, __p1, __p2) __extension__ ({ \
  6970.   poly16x4_t __ret; \
  6971.   poly16x4_t __s0 = __p0; \
  6972.   poly16x4_t __s1 = __p1; \
  6973.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  6974.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  6975.   __ret = (poly16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
  6976.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  6977.   __ret; \
  6978. })
  6979. #endif
  6980.  
  6981. #ifdef __LITTLE_ENDIAN__
  6982. #define vextq_p8(__p0, __p1, __p2) __extension__ ({ \
  6983.   poly8x16_t __ret; \
  6984.   poly8x16_t __s0 = __p0; \
  6985.   poly8x16_t __s1 = __p1; \
  6986.   __ret = (poly8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
  6987.   __ret; \
  6988. })
  6989. #else
  6990. #define vextq_p8(__p0, __p1, __p2) __extension__ ({ \
  6991.   poly8x16_t __ret; \
  6992.   poly8x16_t __s0 = __p0; \
  6993.   poly8x16_t __s1 = __p1; \
  6994.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  6995.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  6996.   __ret = (poly8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
  6997.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  6998.   __ret; \
  6999. })
  7000. #endif
  7001.  
  7002. #ifdef __LITTLE_ENDIAN__
  7003. #define vextq_p16(__p0, __p1, __p2) __extension__ ({ \
  7004.   poly16x8_t __ret; \
  7005.   poly16x8_t __s0 = __p0; \
  7006.   poly16x8_t __s1 = __p1; \
  7007.   __ret = (poly16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
  7008.   __ret; \
  7009. })
  7010. #else
  7011. #define vextq_p16(__p0, __p1, __p2) __extension__ ({ \
  7012.   poly16x8_t __ret; \
  7013.   poly16x8_t __s0 = __p0; \
  7014.   poly16x8_t __s1 = __p1; \
  7015.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7016.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  7017.   __ret = (poly16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
  7018.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  7019.   __ret; \
  7020. })
  7021. #endif
  7022.  
  7023. #ifdef __LITTLE_ENDIAN__
  7024. #define vextq_u8(__p0, __p1, __p2) __extension__ ({ \
  7025.   uint8x16_t __ret; \
  7026.   uint8x16_t __s0 = __p0; \
  7027.   uint8x16_t __s1 = __p1; \
  7028.   __ret = (uint8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
  7029.   __ret; \
  7030. })
  7031. #else
  7032. #define vextq_u8(__p0, __p1, __p2) __extension__ ({ \
  7033.   uint8x16_t __ret; \
  7034.   uint8x16_t __s0 = __p0; \
  7035.   uint8x16_t __s1 = __p1; \
  7036.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  7037.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  7038.   __ret = (uint8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
  7039.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  7040.   __ret; \
  7041. })
  7042. #endif
  7043.  
  7044. #ifdef __LITTLE_ENDIAN__
  7045. #define vextq_u32(__p0, __p1, __p2) __extension__ ({ \
  7046.   uint32x4_t __ret; \
  7047.   uint32x4_t __s0 = __p0; \
  7048.   uint32x4_t __s1 = __p1; \
  7049.   __ret = (uint32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
  7050.   __ret; \
  7051. })
  7052. #else
  7053. #define vextq_u32(__p0, __p1, __p2) __extension__ ({ \
  7054.   uint32x4_t __ret; \
  7055.   uint32x4_t __s0 = __p0; \
  7056.   uint32x4_t __s1 = __p1; \
  7057.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7058.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  7059.   __ret = (uint32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
  7060.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  7061.   __ret; \
  7062. })
  7063. #endif
  7064.  
  7065. #ifdef __LITTLE_ENDIAN__
  7066. #define vextq_u64(__p0, __p1, __p2) __extension__ ({ \
  7067.   uint64x2_t __ret; \
  7068.   uint64x2_t __s0 = __p0; \
  7069.   uint64x2_t __s1 = __p1; \
  7070.   __ret = (uint64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
  7071.   __ret; \
  7072. })
  7073. #else
  7074. #define vextq_u64(__p0, __p1, __p2) __extension__ ({ \
  7075.   uint64x2_t __ret; \
  7076.   uint64x2_t __s0 = __p0; \
  7077.   uint64x2_t __s1 = __p1; \
  7078.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  7079.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  7080.   __ret = (uint64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
  7081.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  7082.   __ret; \
  7083. })
  7084. #endif
  7085.  
  7086. #ifdef __LITTLE_ENDIAN__
  7087. #define vextq_u16(__p0, __p1, __p2) __extension__ ({ \
  7088.   uint16x8_t __ret; \
  7089.   uint16x8_t __s0 = __p0; \
  7090.   uint16x8_t __s1 = __p1; \
  7091.   __ret = (uint16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
  7092.   __ret; \
  7093. })
  7094. #else
  7095. #define vextq_u16(__p0, __p1, __p2) __extension__ ({ \
  7096.   uint16x8_t __ret; \
  7097.   uint16x8_t __s0 = __p0; \
  7098.   uint16x8_t __s1 = __p1; \
  7099.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7100.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  7101.   __ret = (uint16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
  7102.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  7103.   __ret; \
  7104. })
  7105. #endif
  7106.  
  7107. #ifdef __LITTLE_ENDIAN__
  7108. #define vextq_s8(__p0, __p1, __p2) __extension__ ({ \
  7109.   int8x16_t __ret; \
  7110.   int8x16_t __s0 = __p0; \
  7111.   int8x16_t __s1 = __p1; \
  7112.   __ret = (int8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
  7113.   __ret; \
  7114. })
  7115. #else
  7116. #define vextq_s8(__p0, __p1, __p2) __extension__ ({ \
  7117.   int8x16_t __ret; \
  7118.   int8x16_t __s0 = __p0; \
  7119.   int8x16_t __s1 = __p1; \
  7120.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  7121.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  7122.   __ret = (int8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
  7123.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  7124.   __ret; \
  7125. })
  7126. #endif
  7127.  
  7128. #ifdef __LITTLE_ENDIAN__
  7129. #define vextq_f32(__p0, __p1, __p2) __extension__ ({ \
  7130.   float32x4_t __ret; \
  7131.   float32x4_t __s0 = __p0; \
  7132.   float32x4_t __s1 = __p1; \
  7133.   __ret = (float32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 41); \
  7134.   __ret; \
  7135. })
  7136. #else
  7137. #define vextq_f32(__p0, __p1, __p2) __extension__ ({ \
  7138.   float32x4_t __ret; \
  7139.   float32x4_t __s0 = __p0; \
  7140.   float32x4_t __s1 = __p1; \
  7141.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7142.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  7143.   __ret = (float32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 41); \
  7144.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  7145.   __ret; \
  7146. })
  7147. #endif
  7148.  
  7149. #ifdef __LITTLE_ENDIAN__
  7150. #define vextq_s32(__p0, __p1, __p2) __extension__ ({ \
  7151.   int32x4_t __ret; \
  7152.   int32x4_t __s0 = __p0; \
  7153.   int32x4_t __s1 = __p1; \
  7154.   __ret = (int32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
  7155.   __ret; \
  7156. })
  7157. #else
  7158. #define vextq_s32(__p0, __p1, __p2) __extension__ ({ \
  7159.   int32x4_t __ret; \
  7160.   int32x4_t __s0 = __p0; \
  7161.   int32x4_t __s1 = __p1; \
  7162.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7163.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  7164.   __ret = (int32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
  7165.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  7166.   __ret; \
  7167. })
  7168. #endif
  7169.  
  7170. #ifdef __LITTLE_ENDIAN__
  7171. #define vextq_s64(__p0, __p1, __p2) __extension__ ({ \
  7172.   int64x2_t __ret; \
  7173.   int64x2_t __s0 = __p0; \
  7174.   int64x2_t __s1 = __p1; \
  7175.   __ret = (int64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
  7176.   __ret; \
  7177. })
  7178. #else
  7179. #define vextq_s64(__p0, __p1, __p2) __extension__ ({ \
  7180.   int64x2_t __ret; \
  7181.   int64x2_t __s0 = __p0; \
  7182.   int64x2_t __s1 = __p1; \
  7183.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  7184.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  7185.   __ret = (int64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
  7186.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  7187.   __ret; \
  7188. })
  7189. #endif
  7190.  
  7191. #ifdef __LITTLE_ENDIAN__
  7192. #define vextq_s16(__p0, __p1, __p2) __extension__ ({ \
  7193.   int16x8_t __ret; \
  7194.   int16x8_t __s0 = __p0; \
  7195.   int16x8_t __s1 = __p1; \
  7196.   __ret = (int16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
  7197.   __ret; \
  7198. })
  7199. #else
  7200. #define vextq_s16(__p0, __p1, __p2) __extension__ ({ \
  7201.   int16x8_t __ret; \
  7202.   int16x8_t __s0 = __p0; \
  7203.   int16x8_t __s1 = __p1; \
  7204.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7205.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  7206.   __ret = (int16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
  7207.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  7208.   __ret; \
  7209. })
  7210. #endif
  7211.  
  7212. #ifdef __LITTLE_ENDIAN__
  7213. #define vext_u8(__p0, __p1, __p2) __extension__ ({ \
  7214.   uint8x8_t __ret; \
  7215.   uint8x8_t __s0 = __p0; \
  7216.   uint8x8_t __s1 = __p1; \
  7217.   __ret = (uint8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
  7218.   __ret; \
  7219. })
  7220. #else
  7221. #define vext_u8(__p0, __p1, __p2) __extension__ ({ \
  7222.   uint8x8_t __ret; \
  7223.   uint8x8_t __s0 = __p0; \
  7224.   uint8x8_t __s1 = __p1; \
  7225.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7226.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  7227.   __ret = (uint8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
  7228.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  7229.   __ret; \
  7230. })
  7231. #endif
  7232.  
  7233. #ifdef __LITTLE_ENDIAN__
  7234. #define vext_u32(__p0, __p1, __p2) __extension__ ({ \
  7235.   uint32x2_t __ret; \
  7236.   uint32x2_t __s0 = __p0; \
  7237.   uint32x2_t __s1 = __p1; \
  7238.   __ret = (uint32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
  7239.   __ret; \
  7240. })
  7241. #else
  7242. #define vext_u32(__p0, __p1, __p2) __extension__ ({ \
  7243.   uint32x2_t __ret; \
  7244.   uint32x2_t __s0 = __p0; \
  7245.   uint32x2_t __s1 = __p1; \
  7246.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  7247.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  7248.   __ret = (uint32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
  7249.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  7250.   __ret; \
  7251. })
  7252. #endif
  7253.  
  7254. #define vext_u64(__p0, __p1, __p2) __extension__ ({ \
  7255.   uint64x1_t __ret; \
  7256.   uint64x1_t __s0 = __p0; \
  7257.   uint64x1_t __s1 = __p1; \
  7258.   __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
  7259.   __ret; \
  7260. })
  7261. #ifdef __LITTLE_ENDIAN__
  7262. #define vext_u16(__p0, __p1, __p2) __extension__ ({ \
  7263.   uint16x4_t __ret; \
  7264.   uint16x4_t __s0 = __p0; \
  7265.   uint16x4_t __s1 = __p1; \
  7266.   __ret = (uint16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
  7267.   __ret; \
  7268. })
  7269. #else
  7270. #define vext_u16(__p0, __p1, __p2) __extension__ ({ \
  7271.   uint16x4_t __ret; \
  7272.   uint16x4_t __s0 = __p0; \
  7273.   uint16x4_t __s1 = __p1; \
  7274.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7275.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  7276.   __ret = (uint16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
  7277.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  7278.   __ret; \
  7279. })
  7280. #endif
  7281.  
  7282. #ifdef __LITTLE_ENDIAN__
  7283. #define vext_s8(__p0, __p1, __p2) __extension__ ({ \
  7284.   int8x8_t __ret; \
  7285.   int8x8_t __s0 = __p0; \
  7286.   int8x8_t __s1 = __p1; \
  7287.   __ret = (int8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
  7288.   __ret; \
  7289. })
  7290. #else
  7291. #define vext_s8(__p0, __p1, __p2) __extension__ ({ \
  7292.   int8x8_t __ret; \
  7293.   int8x8_t __s0 = __p0; \
  7294.   int8x8_t __s1 = __p1; \
  7295.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7296.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  7297.   __ret = (int8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
  7298.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  7299.   __ret; \
  7300. })
  7301. #endif
  7302.  
  7303. #ifdef __LITTLE_ENDIAN__
  7304. #define vext_f32(__p0, __p1, __p2) __extension__ ({ \
  7305.   float32x2_t __ret; \
  7306.   float32x2_t __s0 = __p0; \
  7307.   float32x2_t __s1 = __p1; \
  7308.   __ret = (float32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 9); \
  7309.   __ret; \
  7310. })
  7311. #else
  7312. #define vext_f32(__p0, __p1, __p2) __extension__ ({ \
  7313.   float32x2_t __ret; \
  7314.   float32x2_t __s0 = __p0; \
  7315.   float32x2_t __s1 = __p1; \
  7316.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  7317.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  7318.   __ret = (float32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 9); \
  7319.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  7320.   __ret; \
  7321. })
  7322. #endif
  7323.  
  7324. #ifdef __LITTLE_ENDIAN__
  7325. #define vext_s32(__p0, __p1, __p2) __extension__ ({ \
  7326.   int32x2_t __ret; \
  7327.   int32x2_t __s0 = __p0; \
  7328.   int32x2_t __s1 = __p1; \
  7329.   __ret = (int32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
  7330.   __ret; \
  7331. })
  7332. #else
  7333. #define vext_s32(__p0, __p1, __p2) __extension__ ({ \
  7334.   int32x2_t __ret; \
  7335.   int32x2_t __s0 = __p0; \
  7336.   int32x2_t __s1 = __p1; \
  7337.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  7338.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  7339.   __ret = (int32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  7340.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  7341.   __ret; \
  7342. })
  7343. #endif
  7344.  
  7345. #define vext_s64(__p0, __p1, __p2) __extension__ ({ \
  7346.   int64x1_t __ret; \
  7347.   int64x1_t __s0 = __p0; \
  7348.   int64x1_t __s1 = __p1; \
  7349.   __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
  7350.   __ret; \
  7351. })
  7352. #ifdef __LITTLE_ENDIAN__
  7353. #define vext_s16(__p0, __p1, __p2) __extension__ ({ \
  7354.   int16x4_t __ret; \
  7355.   int16x4_t __s0 = __p0; \
  7356.   int16x4_t __s1 = __p1; \
  7357.   __ret = (int16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
  7358.   __ret; \
  7359. })
  7360. #else
  7361. #define vext_s16(__p0, __p1, __p2) __extension__ ({ \
  7362.   int16x4_t __ret; \
  7363.   int16x4_t __s0 = __p0; \
  7364.   int16x4_t __s1 = __p1; \
  7365.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7366.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  7367.   __ret = (int16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  7368.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  7369.   __ret; \
  7370. })
  7371. #endif
  7372.  
  7373. #ifdef __LITTLE_ENDIAN__
  7374. __ai poly8x8_t vget_high_p8(poly8x16_t __p0) {
  7375.   poly8x8_t __ret;
  7376.   __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  7377.   return __ret;
  7378. }
  7379. #else
  7380. __ai poly8x8_t vget_high_p8(poly8x16_t __p0) {
  7381.   poly8x8_t __ret;
  7382.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  7383.   __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
  7384.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  7385.   return __ret;
  7386. }
  7387. __ai poly8x8_t __noswap_vget_high_p8(poly8x16_t __p0) {
  7388.   poly8x8_t __ret;
  7389.   __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  7390.   return __ret;
  7391. }
  7392. #endif
  7393.  
  7394. #ifdef __LITTLE_ENDIAN__
  7395. __ai poly16x4_t vget_high_p16(poly16x8_t __p0) {
  7396.   poly16x4_t __ret;
  7397.   __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  7398.   return __ret;
  7399. }
  7400. #else
  7401. __ai poly16x4_t vget_high_p16(poly16x8_t __p0) {
  7402.   poly16x4_t __ret;
  7403.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  7404.   __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
  7405.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  7406.   return __ret;
  7407. }
  7408. #endif
  7409.  
  7410. #ifdef __LITTLE_ENDIAN__
  7411. __ai uint8x8_t vget_high_u8(uint8x16_t __p0) {
  7412.   uint8x8_t __ret;
  7413.   __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  7414.   return __ret;
  7415. }
  7416. #else
  7417. __ai uint8x8_t vget_high_u8(uint8x16_t __p0) {
  7418.   uint8x8_t __ret;
  7419.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  7420.   __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
  7421.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  7422.   return __ret;
  7423. }
  7424. __ai uint8x8_t __noswap_vget_high_u8(uint8x16_t __p0) {
  7425.   uint8x8_t __ret;
  7426.   __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  7427.   return __ret;
  7428. }
  7429. #endif
  7430.  
  7431. #ifdef __LITTLE_ENDIAN__
  7432. __ai uint32x2_t vget_high_u32(uint32x4_t __p0) {
  7433.   uint32x2_t __ret;
  7434.   __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  7435.   return __ret;
  7436. }
  7437. #else
  7438. __ai uint32x2_t vget_high_u32(uint32x4_t __p0) {
  7439.   uint32x2_t __ret;
  7440.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  7441.   __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
  7442.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  7443.   return __ret;
  7444. }
  7445. __ai uint32x2_t __noswap_vget_high_u32(uint32x4_t __p0) {
  7446.   uint32x2_t __ret;
  7447.   __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  7448.   return __ret;
  7449. }
  7450. #endif
  7451.  
  7452. #ifdef __LITTLE_ENDIAN__
  7453. __ai uint64x1_t vget_high_u64(uint64x2_t __p0) {
  7454.   uint64x1_t __ret;
  7455.   __ret = __builtin_shufflevector(__p0, __p0, 1);
  7456.   return __ret;
  7457. }
  7458. #else
  7459. __ai uint64x1_t vget_high_u64(uint64x2_t __p0) {
  7460.   uint64x1_t __ret;
  7461.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  7462.   __ret = __builtin_shufflevector(__rev0, __rev0, 1);
  7463.   return __ret;
  7464. }
  7465. #endif
  7466.  
  7467. #ifdef __LITTLE_ENDIAN__
  7468. __ai uint16x4_t vget_high_u16(uint16x8_t __p0) {
  7469.   uint16x4_t __ret;
  7470.   __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  7471.   return __ret;
  7472. }
  7473. #else
  7474. __ai uint16x4_t vget_high_u16(uint16x8_t __p0) {
  7475.   uint16x4_t __ret;
  7476.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  7477.   __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
  7478.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  7479.   return __ret;
  7480. }
  7481. __ai uint16x4_t __noswap_vget_high_u16(uint16x8_t __p0) {
  7482.   uint16x4_t __ret;
  7483.   __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  7484.   return __ret;
  7485. }
  7486. #endif
  7487.  
  7488. #ifdef __LITTLE_ENDIAN__
  7489. __ai int8x8_t vget_high_s8(int8x16_t __p0) {
  7490.   int8x8_t __ret;
  7491.   __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  7492.   return __ret;
  7493. }
  7494. #else
  7495. __ai int8x8_t vget_high_s8(int8x16_t __p0) {
  7496.   int8x8_t __ret;
  7497.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  7498.   __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
  7499.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  7500.   return __ret;
  7501. }
  7502. __ai int8x8_t __noswap_vget_high_s8(int8x16_t __p0) {
  7503.   int8x8_t __ret;
  7504.   __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
  7505.   return __ret;
  7506. }
  7507. #endif
  7508.  
  7509. #ifdef __LITTLE_ENDIAN__
  7510. __ai float32x2_t vget_high_f32(float32x4_t __p0) {
  7511.   float32x2_t __ret;
  7512.   __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  7513.   return __ret;
  7514. }
  7515. #else
  7516. __ai float32x2_t vget_high_f32(float32x4_t __p0) {
  7517.   float32x2_t __ret;
  7518.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  7519.   __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
  7520.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  7521.   return __ret;
  7522. }
  7523. __ai float32x2_t __noswap_vget_high_f32(float32x4_t __p0) {
  7524.   float32x2_t __ret;
  7525.   __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  7526.   return __ret;
  7527. }
  7528. #endif
  7529.  
  7530. #ifdef __LITTLE_ENDIAN__
  7531. __ai float16x4_t vget_high_f16(float16x8_t __p0) {
  7532.   float16x4_t __ret;
  7533.   __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  7534.   return __ret;
  7535. }
  7536. #else
  7537. __ai float16x4_t vget_high_f16(float16x8_t __p0) {
  7538.   float16x4_t __ret;
  7539.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  7540.   __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
  7541.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  7542.   return __ret;
  7543. }
  7544. __ai float16x4_t __noswap_vget_high_f16(float16x8_t __p0) {
  7545.   float16x4_t __ret;
  7546.   __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  7547.   return __ret;
  7548. }
  7549. #endif
  7550.  
  7551. #ifdef __LITTLE_ENDIAN__
  7552. __ai int32x2_t vget_high_s32(int32x4_t __p0) {
  7553.   int32x2_t __ret;
  7554.   __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  7555.   return __ret;
  7556. }
  7557. #else
  7558. __ai int32x2_t vget_high_s32(int32x4_t __p0) {
  7559.   int32x2_t __ret;
  7560.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  7561.   __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
  7562.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  7563.   return __ret;
  7564. }
  7565. __ai int32x2_t __noswap_vget_high_s32(int32x4_t __p0) {
  7566.   int32x2_t __ret;
  7567.   __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
  7568.   return __ret;
  7569. }
  7570. #endif
  7571.  
  7572. #ifdef __LITTLE_ENDIAN__
  7573. __ai int64x1_t vget_high_s64(int64x2_t __p0) {
  7574.   int64x1_t __ret;
  7575.   __ret = __builtin_shufflevector(__p0, __p0, 1);
  7576.   return __ret;
  7577. }
  7578. #else
  7579. __ai int64x1_t vget_high_s64(int64x2_t __p0) {
  7580.   int64x1_t __ret;
  7581.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  7582.   __ret = __builtin_shufflevector(__rev0, __rev0, 1);
  7583.   return __ret;
  7584. }
  7585. #endif
  7586.  
  7587. #ifdef __LITTLE_ENDIAN__
  7588. __ai int16x4_t vget_high_s16(int16x8_t __p0) {
  7589.   int16x4_t __ret;
  7590.   __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  7591.   return __ret;
  7592. }
  7593. #else
  7594. __ai int16x4_t vget_high_s16(int16x8_t __p0) {
  7595.   int16x4_t __ret;
  7596.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  7597.   __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
  7598.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  7599.   return __ret;
  7600. }
  7601. __ai int16x4_t __noswap_vget_high_s16(int16x8_t __p0) {
  7602.   int16x4_t __ret;
  7603.   __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  7604.   return __ret;
  7605. }
  7606. #endif
  7607.  
  7608. #ifdef __LITTLE_ENDIAN__
  7609. #define vget_lane_p8(__p0, __p1) __extension__ ({ \
  7610.   poly8_t __ret; \
  7611.   poly8x8_t __s0 = __p0; \
  7612.   __ret = (poly8_t) __builtin_neon_vget_lane_i8((poly8x8_t)__s0, __p1); \
  7613.   __ret; \
  7614. })
  7615. #else
  7616. #define vget_lane_p8(__p0, __p1) __extension__ ({ \
  7617.   poly8_t __ret; \
  7618.   poly8x8_t __s0 = __p0; \
  7619.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7620.   __ret = (poly8_t) __builtin_neon_vget_lane_i8((poly8x8_t)__rev0, __p1); \
  7621.   __ret; \
  7622. })
  7623. #define __noswap_vget_lane_p8(__p0, __p1) __extension__ ({ \
  7624.   poly8_t __ret; \
  7625.   poly8x8_t __s0 = __p0; \
  7626.   __ret = (poly8_t) __builtin_neon_vget_lane_i8((poly8x8_t)__s0, __p1); \
  7627.   __ret; \
  7628. })
  7629. #endif
  7630.  
  7631. #ifdef __LITTLE_ENDIAN__
  7632. #define vget_lane_p16(__p0, __p1) __extension__ ({ \
  7633.   poly16_t __ret; \
  7634.   poly16x4_t __s0 = __p0; \
  7635.   __ret = (poly16_t) __builtin_neon_vget_lane_i16((poly16x4_t)__s0, __p1); \
  7636.   __ret; \
  7637. })
  7638. #else
  7639. #define vget_lane_p16(__p0, __p1) __extension__ ({ \
  7640.   poly16_t __ret; \
  7641.   poly16x4_t __s0 = __p0; \
  7642.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7643.   __ret = (poly16_t) __builtin_neon_vget_lane_i16((poly16x4_t)__rev0, __p1); \
  7644.   __ret; \
  7645. })
  7646. #define __noswap_vget_lane_p16(__p0, __p1) __extension__ ({ \
  7647.   poly16_t __ret; \
  7648.   poly16x4_t __s0 = __p0; \
  7649.   __ret = (poly16_t) __builtin_neon_vget_lane_i16((poly16x4_t)__s0, __p1); \
  7650.   __ret; \
  7651. })
  7652. #endif
  7653.  
  7654. #ifdef __LITTLE_ENDIAN__
  7655. #define vgetq_lane_p8(__p0, __p1) __extension__ ({ \
  7656.   poly8_t __ret; \
  7657.   poly8x16_t __s0 = __p0; \
  7658.   __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((poly8x16_t)__s0, __p1); \
  7659.   __ret; \
  7660. })
  7661. #else
  7662. #define vgetq_lane_p8(__p0, __p1) __extension__ ({ \
  7663.   poly8_t __ret; \
  7664.   poly8x16_t __s0 = __p0; \
  7665.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  7666.   __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((poly8x16_t)__rev0, __p1); \
  7667.   __ret; \
  7668. })
  7669. #define __noswap_vgetq_lane_p8(__p0, __p1) __extension__ ({ \
  7670.   poly8_t __ret; \
  7671.   poly8x16_t __s0 = __p0; \
  7672.   __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((poly8x16_t)__s0, __p1); \
  7673.   __ret; \
  7674. })
  7675. #endif
  7676.  
  7677. #ifdef __LITTLE_ENDIAN__
  7678. #define vgetq_lane_p16(__p0, __p1) __extension__ ({ \
  7679.   poly16_t __ret; \
  7680.   poly16x8_t __s0 = __p0; \
  7681.   __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((poly16x8_t)__s0, __p1); \
  7682.   __ret; \
  7683. })
  7684. #else
  7685. #define vgetq_lane_p16(__p0, __p1) __extension__ ({ \
  7686.   poly16_t __ret; \
  7687.   poly16x8_t __s0 = __p0; \
  7688.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7689.   __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((poly16x8_t)__rev0, __p1); \
  7690.   __ret; \
  7691. })
  7692. #define __noswap_vgetq_lane_p16(__p0, __p1) __extension__ ({ \
  7693.   poly16_t __ret; \
  7694.   poly16x8_t __s0 = __p0; \
  7695.   __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((poly16x8_t)__s0, __p1); \
  7696.   __ret; \
  7697. })
  7698. #endif
  7699.  
  7700. #ifdef __LITTLE_ENDIAN__
  7701. #define vgetq_lane_u8(__p0, __p1) __extension__ ({ \
  7702.   uint8_t __ret; \
  7703.   uint8x16_t __s0 = __p0; \
  7704.   __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  7705.   __ret; \
  7706. })
  7707. #else
  7708. #define vgetq_lane_u8(__p0, __p1) __extension__ ({ \
  7709.   uint8_t __ret; \
  7710.   uint8x16_t __s0 = __p0; \
  7711.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  7712.   __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
  7713.   __ret; \
  7714. })
  7715. #define __noswap_vgetq_lane_u8(__p0, __p1) __extension__ ({ \
  7716.   uint8_t __ret; \
  7717.   uint8x16_t __s0 = __p0; \
  7718.   __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  7719.   __ret; \
  7720. })
  7721. #endif
  7722.  
  7723. #ifdef __LITTLE_ENDIAN__
  7724. #define vgetq_lane_u32(__p0, __p1) __extension__ ({ \
  7725.   uint32_t __ret; \
  7726.   uint32x4_t __s0 = __p0; \
  7727.   __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int32x4_t)__s0, __p1); \
  7728.   __ret; \
  7729. })
  7730. #else
  7731. #define vgetq_lane_u32(__p0, __p1) __extension__ ({ \
  7732.   uint32_t __ret; \
  7733.   uint32x4_t __s0 = __p0; \
  7734.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7735.   __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int32x4_t)__rev0, __p1); \
  7736.   __ret; \
  7737. })
  7738. #define __noswap_vgetq_lane_u32(__p0, __p1) __extension__ ({ \
  7739.   uint32_t __ret; \
  7740.   uint32x4_t __s0 = __p0; \
  7741.   __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int32x4_t)__s0, __p1); \
  7742.   __ret; \
  7743. })
  7744. #endif
  7745.  
  7746. #ifdef __LITTLE_ENDIAN__
  7747. #define vgetq_lane_u64(__p0, __p1) __extension__ ({ \
  7748.   uint64_t __ret; \
  7749.   uint64x2_t __s0 = __p0; \
  7750.   __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int64x2_t)__s0, __p1); \
  7751.   __ret; \
  7752. })
  7753. #else
  7754. #define vgetq_lane_u64(__p0, __p1) __extension__ ({ \
  7755.   uint64_t __ret; \
  7756.   uint64x2_t __s0 = __p0; \
  7757.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  7758.   __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int64x2_t)__rev0, __p1); \
  7759.   __ret; \
  7760. })
  7761. #define __noswap_vgetq_lane_u64(__p0, __p1) __extension__ ({ \
  7762.   uint64_t __ret; \
  7763.   uint64x2_t __s0 = __p0; \
  7764.   __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int64x2_t)__s0, __p1); \
  7765.   __ret; \
  7766. })
  7767. #endif
  7768.  
  7769. #ifdef __LITTLE_ENDIAN__
  7770. #define vgetq_lane_u16(__p0, __p1) __extension__ ({ \
  7771.   uint16_t __ret; \
  7772.   uint16x8_t __s0 = __p0; \
  7773.   __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int16x8_t)__s0, __p1); \
  7774.   __ret; \
  7775. })
  7776. #else
  7777. #define vgetq_lane_u16(__p0, __p1) __extension__ ({ \
  7778.   uint16_t __ret; \
  7779.   uint16x8_t __s0 = __p0; \
  7780.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7781.   __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int16x8_t)__rev0, __p1); \
  7782.   __ret; \
  7783. })
  7784. #define __noswap_vgetq_lane_u16(__p0, __p1) __extension__ ({ \
  7785.   uint16_t __ret; \
  7786.   uint16x8_t __s0 = __p0; \
  7787.   __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int16x8_t)__s0, __p1); \
  7788.   __ret; \
  7789. })
  7790. #endif
  7791.  
  7792. #ifdef __LITTLE_ENDIAN__
  7793. #define vgetq_lane_s8(__p0, __p1) __extension__ ({ \
  7794.   int8_t __ret; \
  7795.   int8x16_t __s0 = __p0; \
  7796.   __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  7797.   __ret; \
  7798. })
  7799. #else
  7800. #define vgetq_lane_s8(__p0, __p1) __extension__ ({ \
  7801.   int8_t __ret; \
  7802.   int8x16_t __s0 = __p0; \
  7803.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  7804.   __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
  7805.   __ret; \
  7806. })
  7807. #define __noswap_vgetq_lane_s8(__p0, __p1) __extension__ ({ \
  7808.   int8_t __ret; \
  7809.   int8x16_t __s0 = __p0; \
  7810.   __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
  7811.   __ret; \
  7812. })
  7813. #endif
  7814.  
  7815. #ifdef __LITTLE_ENDIAN__
  7816. #define vgetq_lane_f32(__p0, __p1) __extension__ ({ \
  7817.   float32_t __ret; \
  7818.   float32x4_t __s0 = __p0; \
  7819.   __ret = (float32_t) __builtin_neon_vgetq_lane_f32((float32x4_t)__s0, __p1); \
  7820.   __ret; \
  7821. })
  7822. #else
  7823. #define vgetq_lane_f32(__p0, __p1) __extension__ ({ \
  7824.   float32_t __ret; \
  7825.   float32x4_t __s0 = __p0; \
  7826.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7827.   __ret = (float32_t) __builtin_neon_vgetq_lane_f32((float32x4_t)__rev0, __p1); \
  7828.   __ret; \
  7829. })
  7830. #define __noswap_vgetq_lane_f32(__p0, __p1) __extension__ ({ \
  7831.   float32_t __ret; \
  7832.   float32x4_t __s0 = __p0; \
  7833.   __ret = (float32_t) __builtin_neon_vgetq_lane_f32((float32x4_t)__s0, __p1); \
  7834.   __ret; \
  7835. })
  7836. #endif
  7837.  
  7838. #ifdef __LITTLE_ENDIAN__
  7839. #define vgetq_lane_s32(__p0, __p1) __extension__ ({ \
  7840.   int32_t __ret; \
  7841.   int32x4_t __s0 = __p0; \
  7842.   __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int32x4_t)__s0, __p1); \
  7843.   __ret; \
  7844. })
  7845. #else
  7846. #define vgetq_lane_s32(__p0, __p1) __extension__ ({ \
  7847.   int32_t __ret; \
  7848.   int32x4_t __s0 = __p0; \
  7849.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7850.   __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int32x4_t)__rev0, __p1); \
  7851.   __ret; \
  7852. })
  7853. #define __noswap_vgetq_lane_s32(__p0, __p1) __extension__ ({ \
  7854.   int32_t __ret; \
  7855.   int32x4_t __s0 = __p0; \
  7856.   __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int32x4_t)__s0, __p1); \
  7857.   __ret; \
  7858. })
  7859. #endif
  7860.  
  7861. #ifdef __LITTLE_ENDIAN__
  7862. #define vgetq_lane_s64(__p0, __p1) __extension__ ({ \
  7863.   int64_t __ret; \
  7864.   int64x2_t __s0 = __p0; \
  7865.   __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int64x2_t)__s0, __p1); \
  7866.   __ret; \
  7867. })
  7868. #else
  7869. #define vgetq_lane_s64(__p0, __p1) __extension__ ({ \
  7870.   int64_t __ret; \
  7871.   int64x2_t __s0 = __p0; \
  7872.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  7873.   __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int64x2_t)__rev0, __p1); \
  7874.   __ret; \
  7875. })
  7876. #define __noswap_vgetq_lane_s64(__p0, __p1) __extension__ ({ \
  7877.   int64_t __ret; \
  7878.   int64x2_t __s0 = __p0; \
  7879.   __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int64x2_t)__s0, __p1); \
  7880.   __ret; \
  7881. })
  7882. #endif
  7883.  
  7884. #ifdef __LITTLE_ENDIAN__
  7885. #define vgetq_lane_s16(__p0, __p1) __extension__ ({ \
  7886.   int16_t __ret; \
  7887.   int16x8_t __s0 = __p0; \
  7888.   __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int16x8_t)__s0, __p1); \
  7889.   __ret; \
  7890. })
  7891. #else
  7892. #define vgetq_lane_s16(__p0, __p1) __extension__ ({ \
  7893.   int16_t __ret; \
  7894.   int16x8_t __s0 = __p0; \
  7895.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7896.   __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int16x8_t)__rev0, __p1); \
  7897.   __ret; \
  7898. })
  7899. #define __noswap_vgetq_lane_s16(__p0, __p1) __extension__ ({ \
  7900.   int16_t __ret; \
  7901.   int16x8_t __s0 = __p0; \
  7902.   __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int16x8_t)__s0, __p1); \
  7903.   __ret; \
  7904. })
  7905. #endif
  7906.  
  7907. #ifdef __LITTLE_ENDIAN__
  7908. #define vget_lane_u8(__p0, __p1) __extension__ ({ \
  7909.   uint8_t __ret; \
  7910.   uint8x8_t __s0 = __p0; \
  7911.   __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  7912.   __ret; \
  7913. })
  7914. #else
  7915. #define vget_lane_u8(__p0, __p1) __extension__ ({ \
  7916.   uint8_t __ret; \
  7917.   uint8x8_t __s0 = __p0; \
  7918.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7919.   __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
  7920.   __ret; \
  7921. })
  7922. #define __noswap_vget_lane_u8(__p0, __p1) __extension__ ({ \
  7923.   uint8_t __ret; \
  7924.   uint8x8_t __s0 = __p0; \
  7925.   __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  7926.   __ret; \
  7927. })
  7928. #endif
  7929.  
  7930. #ifdef __LITTLE_ENDIAN__
  7931. #define vget_lane_u32(__p0, __p1) __extension__ ({ \
  7932.   uint32_t __ret; \
  7933.   uint32x2_t __s0 = __p0; \
  7934.   __ret = (uint32_t) __builtin_neon_vget_lane_i32((int32x2_t)__s0, __p1); \
  7935.   __ret; \
  7936. })
  7937. #else
  7938. #define vget_lane_u32(__p0, __p1) __extension__ ({ \
  7939.   uint32_t __ret; \
  7940.   uint32x2_t __s0 = __p0; \
  7941.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  7942.   __ret = (uint32_t) __builtin_neon_vget_lane_i32((int32x2_t)__rev0, __p1); \
  7943.   __ret; \
  7944. })
  7945. #define __noswap_vget_lane_u32(__p0, __p1) __extension__ ({ \
  7946.   uint32_t __ret; \
  7947.   uint32x2_t __s0 = __p0; \
  7948.   __ret = (uint32_t) __builtin_neon_vget_lane_i32((int32x2_t)__s0, __p1); \
  7949.   __ret; \
  7950. })
  7951. #endif
  7952.  
  7953. #define vget_lane_u64(__p0, __p1) __extension__ ({ \
  7954.   uint64_t __ret; \
  7955.   uint64x1_t __s0 = __p0; \
  7956.   __ret = (uint64_t) __builtin_neon_vget_lane_i64((int64x1_t)__s0, __p1); \
  7957.   __ret; \
  7958. })
  7959. #ifdef __LITTLE_ENDIAN__
  7960. #define vget_lane_u16(__p0, __p1) __extension__ ({ \
  7961.   uint16_t __ret; \
  7962.   uint16x4_t __s0 = __p0; \
  7963.   __ret = (uint16_t) __builtin_neon_vget_lane_i16((int16x4_t)__s0, __p1); \
  7964.   __ret; \
  7965. })
  7966. #else
  7967. #define vget_lane_u16(__p0, __p1) __extension__ ({ \
  7968.   uint16_t __ret; \
  7969.   uint16x4_t __s0 = __p0; \
  7970.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  7971.   __ret = (uint16_t) __builtin_neon_vget_lane_i16((int16x4_t)__rev0, __p1); \
  7972.   __ret; \
  7973. })
  7974. #define __noswap_vget_lane_u16(__p0, __p1) __extension__ ({ \
  7975.   uint16_t __ret; \
  7976.   uint16x4_t __s0 = __p0; \
  7977.   __ret = (uint16_t) __builtin_neon_vget_lane_i16((int16x4_t)__s0, __p1); \
  7978.   __ret; \
  7979. })
  7980. #endif
  7981.  
  7982. #ifdef __LITTLE_ENDIAN__
  7983. #define vget_lane_s8(__p0, __p1) __extension__ ({ \
  7984.   int8_t __ret; \
  7985.   int8x8_t __s0 = __p0; \
  7986.   __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  7987.   __ret; \
  7988. })
  7989. #else
  7990. #define vget_lane_s8(__p0, __p1) __extension__ ({ \
  7991.   int8_t __ret; \
  7992.   int8x8_t __s0 = __p0; \
  7993.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  7994.   __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
  7995.   __ret; \
  7996. })
  7997. #define __noswap_vget_lane_s8(__p0, __p1) __extension__ ({ \
  7998.   int8_t __ret; \
  7999.   int8x8_t __s0 = __p0; \
  8000.   __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
  8001.   __ret; \
  8002. })
  8003. #endif
  8004.  
  8005. #ifdef __LITTLE_ENDIAN__
  8006. #define vget_lane_f32(__p0, __p1) __extension__ ({ \
  8007.   float32_t __ret; \
  8008.   float32x2_t __s0 = __p0; \
  8009.   __ret = (float32_t) __builtin_neon_vget_lane_f32((float32x2_t)__s0, __p1); \
  8010.   __ret; \
  8011. })
  8012. #else
  8013. #define vget_lane_f32(__p0, __p1) __extension__ ({ \
  8014.   float32_t __ret; \
  8015.   float32x2_t __s0 = __p0; \
  8016.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  8017.   __ret = (float32_t) __builtin_neon_vget_lane_f32((float32x2_t)__rev0, __p1); \
  8018.   __ret; \
  8019. })
  8020. #define __noswap_vget_lane_f32(__p0, __p1) __extension__ ({ \
  8021.   float32_t __ret; \
  8022.   float32x2_t __s0 = __p0; \
  8023.   __ret = (float32_t) __builtin_neon_vget_lane_f32((float32x2_t)__s0, __p1); \
  8024.   __ret; \
  8025. })
  8026. #endif
  8027.  
  8028. #ifdef __LITTLE_ENDIAN__
  8029. #define vget_lane_s32(__p0, __p1) __extension__ ({ \
  8030.   int32_t __ret; \
  8031.   int32x2_t __s0 = __p0; \
  8032.   __ret = (int32_t) __builtin_neon_vget_lane_i32((int32x2_t)__s0, __p1); \
  8033.   __ret; \
  8034. })
  8035. #else
  8036. #define vget_lane_s32(__p0, __p1) __extension__ ({ \
  8037.   int32_t __ret; \
  8038.   int32x2_t __s0 = __p0; \
  8039.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  8040.   __ret = (int32_t) __builtin_neon_vget_lane_i32((int32x2_t)__rev0, __p1); \
  8041.   __ret; \
  8042. })
  8043. #define __noswap_vget_lane_s32(__p0, __p1) __extension__ ({ \
  8044.   int32_t __ret; \
  8045.   int32x2_t __s0 = __p0; \
  8046.   __ret = (int32_t) __builtin_neon_vget_lane_i32((int32x2_t)__s0, __p1); \
  8047.   __ret; \
  8048. })
  8049. #endif
  8050.  
  8051. #define vget_lane_s64(__p0, __p1) __extension__ ({ \
  8052.   int64_t __ret; \
  8053.   int64x1_t __s0 = __p0; \
  8054.   __ret = (int64_t) __builtin_neon_vget_lane_i64((int64x1_t)__s0, __p1); \
  8055.   __ret; \
  8056. })
  8057. #ifdef __LITTLE_ENDIAN__
  8058. #define vget_lane_s16(__p0, __p1) __extension__ ({ \
  8059.   int16_t __ret; \
  8060.   int16x4_t __s0 = __p0; \
  8061.   __ret = (int16_t) __builtin_neon_vget_lane_i16((int16x4_t)__s0, __p1); \
  8062.   __ret; \
  8063. })
  8064. #else
  8065. #define vget_lane_s16(__p0, __p1) __extension__ ({ \
  8066.   int16_t __ret; \
  8067.   int16x4_t __s0 = __p0; \
  8068.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  8069.   __ret = (int16_t) __builtin_neon_vget_lane_i16((int16x4_t)__rev0, __p1); \
  8070.   __ret; \
  8071. })
  8072. #define __noswap_vget_lane_s16(__p0, __p1) __extension__ ({ \
  8073.   int16_t __ret; \
  8074.   int16x4_t __s0 = __p0; \
  8075.   __ret = (int16_t) __builtin_neon_vget_lane_i16((int16x4_t)__s0, __p1); \
  8076.   __ret; \
  8077. })
  8078. #endif
  8079.  
  8080. #ifdef __LITTLE_ENDIAN__
  8081. __ai poly8x8_t vget_low_p8(poly8x16_t __p0) {
  8082.   poly8x8_t __ret;
  8083.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
  8084.   return __ret;
  8085. }
  8086. #else
  8087. __ai poly8x8_t vget_low_p8(poly8x16_t __p0) {
  8088.   poly8x8_t __ret;
  8089.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8090.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
  8091.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8092.   return __ret;
  8093. }
  8094. #endif
  8095.  
  8096. #ifdef __LITTLE_ENDIAN__
  8097. __ai poly16x4_t vget_low_p16(poly16x8_t __p0) {
  8098.   poly16x4_t __ret;
  8099.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  8100.   return __ret;
  8101. }
  8102. #else
  8103. __ai poly16x4_t vget_low_p16(poly16x8_t __p0) {
  8104.   poly16x4_t __ret;
  8105.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8106.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
  8107.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8108.   return __ret;
  8109. }
  8110. #endif
  8111.  
  8112. #ifdef __LITTLE_ENDIAN__
  8113. __ai uint8x8_t vget_low_u8(uint8x16_t __p0) {
  8114.   uint8x8_t __ret;
  8115.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
  8116.   return __ret;
  8117. }
  8118. #else
  8119. __ai uint8x8_t vget_low_u8(uint8x16_t __p0) {
  8120.   uint8x8_t __ret;
  8121.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8122.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
  8123.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8124.   return __ret;
  8125. }
  8126. #endif
  8127.  
  8128. #ifdef __LITTLE_ENDIAN__
  8129. __ai uint32x2_t vget_low_u32(uint32x4_t __p0) {
  8130.   uint32x2_t __ret;
  8131.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
  8132.   return __ret;
  8133. }
  8134. #else
  8135. __ai uint32x2_t vget_low_u32(uint32x4_t __p0) {
  8136.   uint32x2_t __ret;
  8137.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8138.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
  8139.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  8140.   return __ret;
  8141. }
  8142. #endif
  8143.  
  8144. #ifdef __LITTLE_ENDIAN__
  8145. __ai uint64x1_t vget_low_u64(uint64x2_t __p0) {
  8146.   uint64x1_t __ret;
  8147.   __ret = __builtin_shufflevector(__p0, __p0, 0);
  8148.   return __ret;
  8149. }
  8150. #else
  8151. __ai uint64x1_t vget_low_u64(uint64x2_t __p0) {
  8152.   uint64x1_t __ret;
  8153.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  8154.   __ret = __builtin_shufflevector(__rev0, __rev0, 0);
  8155.   return __ret;
  8156. }
  8157. #endif
  8158.  
  8159. #ifdef __LITTLE_ENDIAN__
  8160. __ai uint16x4_t vget_low_u16(uint16x8_t __p0) {
  8161.   uint16x4_t __ret;
  8162.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  8163.   return __ret;
  8164. }
  8165. #else
  8166. __ai uint16x4_t vget_low_u16(uint16x8_t __p0) {
  8167.   uint16x4_t __ret;
  8168.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8169.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
  8170.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8171.   return __ret;
  8172. }
  8173. #endif
  8174.  
  8175. #ifdef __LITTLE_ENDIAN__
  8176. __ai int8x8_t vget_low_s8(int8x16_t __p0) {
  8177.   int8x8_t __ret;
  8178.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
  8179.   return __ret;
  8180. }
  8181. #else
  8182. __ai int8x8_t vget_low_s8(int8x16_t __p0) {
  8183.   int8x8_t __ret;
  8184.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8185.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
  8186.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8187.   return __ret;
  8188. }
  8189. #endif
  8190.  
  8191. #ifdef __LITTLE_ENDIAN__
  8192. __ai float32x2_t vget_low_f32(float32x4_t __p0) {
  8193.   float32x2_t __ret;
  8194.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
  8195.   return __ret;
  8196. }
  8197. #else
  8198. __ai float32x2_t vget_low_f32(float32x4_t __p0) {
  8199.   float32x2_t __ret;
  8200.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8201.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
  8202.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  8203.   return __ret;
  8204. }
  8205. #endif
  8206.  
  8207. #ifdef __LITTLE_ENDIAN__
  8208. __ai float16x4_t vget_low_f16(float16x8_t __p0) {
  8209.   float16x4_t __ret;
  8210.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  8211.   return __ret;
  8212. }
  8213. #else
  8214. __ai float16x4_t vget_low_f16(float16x8_t __p0) {
  8215.   float16x4_t __ret;
  8216.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8217.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
  8218.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8219.   return __ret;
  8220. }
  8221. #endif
  8222.  
  8223. #ifdef __LITTLE_ENDIAN__
  8224. __ai int32x2_t vget_low_s32(int32x4_t __p0) {
  8225.   int32x2_t __ret;
  8226.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
  8227.   return __ret;
  8228. }
  8229. #else
  8230. __ai int32x2_t vget_low_s32(int32x4_t __p0) {
  8231.   int32x2_t __ret;
  8232.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8233.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
  8234.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  8235.   return __ret;
  8236. }
  8237. #endif
  8238.  
  8239. #ifdef __LITTLE_ENDIAN__
  8240. __ai int64x1_t vget_low_s64(int64x2_t __p0) {
  8241.   int64x1_t __ret;
  8242.   __ret = __builtin_shufflevector(__p0, __p0, 0);
  8243.   return __ret;
  8244. }
  8245. #else
  8246. __ai int64x1_t vget_low_s64(int64x2_t __p0) {
  8247.   int64x1_t __ret;
  8248.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  8249.   __ret = __builtin_shufflevector(__rev0, __rev0, 0);
  8250.   return __ret;
  8251. }
  8252. #endif
  8253.  
  8254. #ifdef __LITTLE_ENDIAN__
  8255. __ai int16x4_t vget_low_s16(int16x8_t __p0) {
  8256.   int16x4_t __ret;
  8257.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  8258.   return __ret;
  8259. }
  8260. #else
  8261. __ai int16x4_t vget_low_s16(int16x8_t __p0) {
  8262.   int16x4_t __ret;
  8263.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8264.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
  8265.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8266.   return __ret;
  8267. }
  8268. #endif
  8269.  
  8270. #ifdef __LITTLE_ENDIAN__
  8271. __ai uint8x16_t vhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  8272.   uint8x16_t __ret;
  8273.   __ret = (uint8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  8274.   return __ret;
  8275. }
  8276. #else
  8277. __ai uint8x16_t vhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  8278.   uint8x16_t __ret;
  8279.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8280.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8281.   __ret = (uint8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  8282.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8283.   return __ret;
  8284. }
  8285. #endif
  8286.  
  8287. #ifdef __LITTLE_ENDIAN__
  8288. __ai uint32x4_t vhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  8289.   uint32x4_t __ret;
  8290.   __ret = (uint32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  8291.   return __ret;
  8292. }
  8293. #else
  8294. __ai uint32x4_t vhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  8295.   uint32x4_t __ret;
  8296.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8297.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  8298.   __ret = (uint32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  8299.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8300.   return __ret;
  8301. }
  8302. #endif
  8303.  
  8304. #ifdef __LITTLE_ENDIAN__
  8305. __ai uint16x8_t vhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  8306.   uint16x8_t __ret;
  8307.   __ret = (uint16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  8308.   return __ret;
  8309. }
  8310. #else
  8311. __ai uint16x8_t vhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  8312.   uint16x8_t __ret;
  8313.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8314.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  8315.   __ret = (uint16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  8316.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8317.   return __ret;
  8318. }
  8319. #endif
  8320.  
  8321. #ifdef __LITTLE_ENDIAN__
  8322. __ai int8x16_t vhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  8323.   int8x16_t __ret;
  8324.   __ret = (int8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  8325.   return __ret;
  8326. }
  8327. #else
  8328. __ai int8x16_t vhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  8329.   int8x16_t __ret;
  8330.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8331.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8332.   __ret = (int8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  8333.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8334.   return __ret;
  8335. }
  8336. #endif
  8337.  
  8338. #ifdef __LITTLE_ENDIAN__
  8339. __ai int32x4_t vhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  8340.   int32x4_t __ret;
  8341.   __ret = (int32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  8342.   return __ret;
  8343. }
  8344. #else
  8345. __ai int32x4_t vhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  8346.   int32x4_t __ret;
  8347.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8348.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  8349.   __ret = (int32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  8350.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8351.   return __ret;
  8352. }
  8353. #endif
  8354.  
  8355. #ifdef __LITTLE_ENDIAN__
  8356. __ai int16x8_t vhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  8357.   int16x8_t __ret;
  8358.   __ret = (int16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  8359.   return __ret;
  8360. }
  8361. #else
  8362. __ai int16x8_t vhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  8363.   int16x8_t __ret;
  8364.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8365.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  8366.   __ret = (int16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  8367.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8368.   return __ret;
  8369. }
  8370. #endif
  8371.  
  8372. #ifdef __LITTLE_ENDIAN__
  8373. __ai uint8x8_t vhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  8374.   uint8x8_t __ret;
  8375.   __ret = (uint8x8_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  8376.   return __ret;
  8377. }
  8378. #else
  8379. __ai uint8x8_t vhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  8380.   uint8x8_t __ret;
  8381.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8382.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  8383.   __ret = (uint8x8_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  8384.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8385.   return __ret;
  8386. }
  8387. #endif
  8388.  
  8389. #ifdef __LITTLE_ENDIAN__
  8390. __ai uint32x2_t vhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  8391.   uint32x2_t __ret;
  8392.   __ret = (uint32x2_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  8393.   return __ret;
  8394. }
  8395. #else
  8396. __ai uint32x2_t vhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  8397.   uint32x2_t __ret;
  8398.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  8399.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  8400.   __ret = (uint32x2_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  8401.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  8402.   return __ret;
  8403. }
  8404. #endif
  8405.  
  8406. #ifdef __LITTLE_ENDIAN__
  8407. __ai uint16x4_t vhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  8408.   uint16x4_t __ret;
  8409.   __ret = (uint16x4_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  8410.   return __ret;
  8411. }
  8412. #else
  8413. __ai uint16x4_t vhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  8414.   uint16x4_t __ret;
  8415.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8416.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  8417.   __ret = (uint16x4_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  8418.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8419.   return __ret;
  8420. }
  8421. #endif
  8422.  
  8423. #ifdef __LITTLE_ENDIAN__
  8424. __ai int8x8_t vhadd_s8(int8x8_t __p0, int8x8_t __p1) {
  8425.   int8x8_t __ret;
  8426.   __ret = (int8x8_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  8427.   return __ret;
  8428. }
  8429. #else
  8430. __ai int8x8_t vhadd_s8(int8x8_t __p0, int8x8_t __p1) {
  8431.   int8x8_t __ret;
  8432.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8433.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  8434.   __ret = (int8x8_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  8435.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8436.   return __ret;
  8437. }
  8438. #endif
  8439.  
  8440. #ifdef __LITTLE_ENDIAN__
  8441. __ai int32x2_t vhadd_s32(int32x2_t __p0, int32x2_t __p1) {
  8442.   int32x2_t __ret;
  8443.   __ret = (int32x2_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  8444.   return __ret;
  8445. }
  8446. #else
  8447. __ai int32x2_t vhadd_s32(int32x2_t __p0, int32x2_t __p1) {
  8448.   int32x2_t __ret;
  8449.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  8450.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  8451.   __ret = (int32x2_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  8452.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  8453.   return __ret;
  8454. }
  8455. #endif
  8456.  
  8457. #ifdef __LITTLE_ENDIAN__
  8458. __ai int16x4_t vhadd_s16(int16x4_t __p0, int16x4_t __p1) {
  8459.   int16x4_t __ret;
  8460.   __ret = (int16x4_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  8461.   return __ret;
  8462. }
  8463. #else
  8464. __ai int16x4_t vhadd_s16(int16x4_t __p0, int16x4_t __p1) {
  8465.   int16x4_t __ret;
  8466.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8467.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  8468.   __ret = (int16x4_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  8469.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8470.   return __ret;
  8471. }
  8472. #endif
  8473.  
  8474. #ifdef __LITTLE_ENDIAN__
  8475. __ai uint8x16_t vhsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  8476.   uint8x16_t __ret;
  8477.   __ret = (uint8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  8478.   return __ret;
  8479. }
  8480. #else
  8481. __ai uint8x16_t vhsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  8482.   uint8x16_t __ret;
  8483.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8484.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8485.   __ret = (uint8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  8486.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8487.   return __ret;
  8488. }
  8489. #endif
  8490.  
  8491. #ifdef __LITTLE_ENDIAN__
  8492. __ai uint32x4_t vhsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  8493.   uint32x4_t __ret;
  8494.   __ret = (uint32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  8495.   return __ret;
  8496. }
  8497. #else
  8498. __ai uint32x4_t vhsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  8499.   uint32x4_t __ret;
  8500.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8501.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  8502.   __ret = (uint32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  8503.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8504.   return __ret;
  8505. }
  8506. #endif
  8507.  
  8508. #ifdef __LITTLE_ENDIAN__
  8509. __ai uint16x8_t vhsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  8510.   uint16x8_t __ret;
  8511.   __ret = (uint16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  8512.   return __ret;
  8513. }
  8514. #else
  8515. __ai uint16x8_t vhsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  8516.   uint16x8_t __ret;
  8517.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8518.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  8519.   __ret = (uint16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  8520.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8521.   return __ret;
  8522. }
  8523. #endif
  8524.  
  8525. #ifdef __LITTLE_ENDIAN__
  8526. __ai int8x16_t vhsubq_s8(int8x16_t __p0, int8x16_t __p1) {
  8527.   int8x16_t __ret;
  8528.   __ret = (int8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  8529.   return __ret;
  8530. }
  8531. #else
  8532. __ai int8x16_t vhsubq_s8(int8x16_t __p0, int8x16_t __p1) {
  8533.   int8x16_t __ret;
  8534.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8535.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8536.   __ret = (int8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  8537.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  8538.   return __ret;
  8539. }
  8540. #endif
  8541.  
  8542. #ifdef __LITTLE_ENDIAN__
  8543. __ai int32x4_t vhsubq_s32(int32x4_t __p0, int32x4_t __p1) {
  8544.   int32x4_t __ret;
  8545.   __ret = (int32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  8546.   return __ret;
  8547. }
  8548. #else
  8549. __ai int32x4_t vhsubq_s32(int32x4_t __p0, int32x4_t __p1) {
  8550.   int32x4_t __ret;
  8551.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8552.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  8553.   __ret = (int32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  8554.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8555.   return __ret;
  8556. }
  8557. #endif
  8558.  
  8559. #ifdef __LITTLE_ENDIAN__
  8560. __ai int16x8_t vhsubq_s16(int16x8_t __p0, int16x8_t __p1) {
  8561.   int16x8_t __ret;
  8562.   __ret = (int16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  8563.   return __ret;
  8564. }
  8565. #else
  8566. __ai int16x8_t vhsubq_s16(int16x8_t __p0, int16x8_t __p1) {
  8567.   int16x8_t __ret;
  8568.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8569.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  8570.   __ret = (int16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  8571.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8572.   return __ret;
  8573. }
  8574. #endif
  8575.  
  8576. #ifdef __LITTLE_ENDIAN__
  8577. __ai uint8x8_t vhsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
  8578.   uint8x8_t __ret;
  8579.   __ret = (uint8x8_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  8580.   return __ret;
  8581. }
  8582. #else
  8583. __ai uint8x8_t vhsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
  8584.   uint8x8_t __ret;
  8585.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8586.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  8587.   __ret = (uint8x8_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  8588.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8589.   return __ret;
  8590. }
  8591. #endif
  8592.  
  8593. #ifdef __LITTLE_ENDIAN__
  8594. __ai uint32x2_t vhsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
  8595.   uint32x2_t __ret;
  8596.   __ret = (uint32x2_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  8597.   return __ret;
  8598. }
  8599. #else
  8600. __ai uint32x2_t vhsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
  8601.   uint32x2_t __ret;
  8602.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  8603.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  8604.   __ret = (uint32x2_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  8605.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  8606.   return __ret;
  8607. }
  8608. #endif
  8609.  
  8610. #ifdef __LITTLE_ENDIAN__
  8611. __ai uint16x4_t vhsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
  8612.   uint16x4_t __ret;
  8613.   __ret = (uint16x4_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  8614.   return __ret;
  8615. }
  8616. #else
  8617. __ai uint16x4_t vhsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
  8618.   uint16x4_t __ret;
  8619.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8620.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  8621.   __ret = (uint16x4_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  8622.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8623.   return __ret;
  8624. }
  8625. #endif
  8626.  
  8627. #ifdef __LITTLE_ENDIAN__
  8628. __ai int8x8_t vhsub_s8(int8x8_t __p0, int8x8_t __p1) {
  8629.   int8x8_t __ret;
  8630.   __ret = (int8x8_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  8631.   return __ret;
  8632. }
  8633. #else
  8634. __ai int8x8_t vhsub_s8(int8x8_t __p0, int8x8_t __p1) {
  8635.   int8x8_t __ret;
  8636.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  8637.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  8638.   __ret = (int8x8_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  8639.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  8640.   return __ret;
  8641. }
  8642. #endif
  8643.  
  8644. #ifdef __LITTLE_ENDIAN__
  8645. __ai int32x2_t vhsub_s32(int32x2_t __p0, int32x2_t __p1) {
  8646.   int32x2_t __ret;
  8647.   __ret = (int32x2_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  8648.   return __ret;
  8649. }
  8650. #else
  8651. __ai int32x2_t vhsub_s32(int32x2_t __p0, int32x2_t __p1) {
  8652.   int32x2_t __ret;
  8653.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  8654.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  8655.   __ret = (int32x2_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  8656.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  8657.   return __ret;
  8658. }
  8659. #endif
  8660.  
  8661. #ifdef __LITTLE_ENDIAN__
  8662. __ai int16x4_t vhsub_s16(int16x4_t __p0, int16x4_t __p1) {
  8663.   int16x4_t __ret;
  8664.   __ret = (int16x4_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  8665.   return __ret;
  8666. }
  8667. #else
  8668. __ai int16x4_t vhsub_s16(int16x4_t __p0, int16x4_t __p1) {
  8669.   int16x4_t __ret;
  8670.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  8671.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  8672.   __ret = (int16x4_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  8673.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  8674.   return __ret;
  8675. }
  8676. #endif
  8677.  
  8678. #ifdef __LITTLE_ENDIAN__
  8679. #define vld1_p8(__p0) __extension__ ({ \
  8680.   poly8x8_t __ret; \
  8681.   __ret = (poly8x8_t) __builtin_neon_vld1_v(__p0, 4); \
  8682.   __ret; \
  8683. })
  8684. #else
  8685. #define vld1_p8(__p0) __extension__ ({ \
  8686.   poly8x8_t __ret; \
  8687.   __ret = (poly8x8_t) __builtin_neon_vld1_v(__p0, 4); \
  8688.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  8689.   __ret; \
  8690. })
  8691. #endif
  8692.  
  8693. #ifdef __LITTLE_ENDIAN__
  8694. #define vld1_p16(__p0) __extension__ ({ \
  8695.   poly16x4_t __ret; \
  8696.   __ret = (poly16x4_t) __builtin_neon_vld1_v(__p0, 5); \
  8697.   __ret; \
  8698. })
  8699. #else
  8700. #define vld1_p16(__p0) __extension__ ({ \
  8701.   poly16x4_t __ret; \
  8702.   __ret = (poly16x4_t) __builtin_neon_vld1_v(__p0, 5); \
  8703.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  8704.   __ret; \
  8705. })
  8706. #endif
  8707.  
  8708. #ifdef __LITTLE_ENDIAN__
  8709. #define vld1q_p8(__p0) __extension__ ({ \
  8710.   poly8x16_t __ret; \
  8711.   __ret = (poly8x16_t) __builtin_neon_vld1q_v(__p0, 36); \
  8712.   __ret; \
  8713. })
  8714. #else
  8715. #define vld1q_p8(__p0) __extension__ ({ \
  8716.   poly8x16_t __ret; \
  8717.   __ret = (poly8x16_t) __builtin_neon_vld1q_v(__p0, 36); \
  8718.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  8719.   __ret; \
  8720. })
  8721. #endif
  8722.  
  8723. #ifdef __LITTLE_ENDIAN__
  8724. #define vld1q_p16(__p0) __extension__ ({ \
  8725.   poly16x8_t __ret; \
  8726.   __ret = (poly16x8_t) __builtin_neon_vld1q_v(__p0, 37); \
  8727.   __ret; \
  8728. })
  8729. #else
  8730. #define vld1q_p16(__p0) __extension__ ({ \
  8731.   poly16x8_t __ret; \
  8732.   __ret = (poly16x8_t) __builtin_neon_vld1q_v(__p0, 37); \
  8733.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  8734.   __ret; \
  8735. })
  8736. #endif
  8737.  
  8738. #ifdef __LITTLE_ENDIAN__
  8739. #define vld1q_u8(__p0) __extension__ ({ \
  8740.   uint8x16_t __ret; \
  8741.   __ret = (uint8x16_t) __builtin_neon_vld1q_v(__p0, 48); \
  8742.   __ret; \
  8743. })
  8744. #else
  8745. #define vld1q_u8(__p0) __extension__ ({ \
  8746.   uint8x16_t __ret; \
  8747.   __ret = (uint8x16_t) __builtin_neon_vld1q_v(__p0, 48); \
  8748.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  8749.   __ret; \
  8750. })
  8751. #endif
  8752.  
  8753. #ifdef __LITTLE_ENDIAN__
  8754. #define vld1q_u32(__p0) __extension__ ({ \
  8755.   uint32x4_t __ret; \
  8756.   __ret = (uint32x4_t) __builtin_neon_vld1q_v(__p0, 50); \
  8757.   __ret; \
  8758. })
  8759. #else
  8760. #define vld1q_u32(__p0) __extension__ ({ \
  8761.   uint32x4_t __ret; \
  8762.   __ret = (uint32x4_t) __builtin_neon_vld1q_v(__p0, 50); \
  8763.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  8764.   __ret; \
  8765. })
  8766. #endif
  8767.  
  8768. #ifdef __LITTLE_ENDIAN__
  8769. #define vld1q_u64(__p0) __extension__ ({ \
  8770.   uint64x2_t __ret; \
  8771.   __ret = (uint64x2_t) __builtin_neon_vld1q_v(__p0, 51); \
  8772.   __ret; \
  8773. })
  8774. #else
  8775. #define vld1q_u64(__p0) __extension__ ({ \
  8776.   uint64x2_t __ret; \
  8777.   __ret = (uint64x2_t) __builtin_neon_vld1q_v(__p0, 51); \
  8778.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  8779.   __ret; \
  8780. })
  8781. #endif
  8782.  
  8783. #ifdef __LITTLE_ENDIAN__
  8784. #define vld1q_u16(__p0) __extension__ ({ \
  8785.   uint16x8_t __ret; \
  8786.   __ret = (uint16x8_t) __builtin_neon_vld1q_v(__p0, 49); \
  8787.   __ret; \
  8788. })
  8789. #else
  8790. #define vld1q_u16(__p0) __extension__ ({ \
  8791.   uint16x8_t __ret; \
  8792.   __ret = (uint16x8_t) __builtin_neon_vld1q_v(__p0, 49); \
  8793.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  8794.   __ret; \
  8795. })
  8796. #endif
  8797.  
  8798. #ifdef __LITTLE_ENDIAN__
  8799. #define vld1q_s8(__p0) __extension__ ({ \
  8800.   int8x16_t __ret; \
  8801.   __ret = (int8x16_t) __builtin_neon_vld1q_v(__p0, 32); \
  8802.   __ret; \
  8803. })
  8804. #else
  8805. #define vld1q_s8(__p0) __extension__ ({ \
  8806.   int8x16_t __ret; \
  8807.   __ret = (int8x16_t) __builtin_neon_vld1q_v(__p0, 32); \
  8808.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  8809.   __ret; \
  8810. })
  8811. #endif
  8812.  
  8813. #ifdef __LITTLE_ENDIAN__
  8814. #define vld1q_f32(__p0) __extension__ ({ \
  8815.   float32x4_t __ret; \
  8816.   __ret = (float32x4_t) __builtin_neon_vld1q_v(__p0, 41); \
  8817.   __ret; \
  8818. })
  8819. #else
  8820. #define vld1q_f32(__p0) __extension__ ({ \
  8821.   float32x4_t __ret; \
  8822.   __ret = (float32x4_t) __builtin_neon_vld1q_v(__p0, 41); \
  8823.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  8824.   __ret; \
  8825. })
  8826. #endif
  8827.  
  8828. #ifdef __LITTLE_ENDIAN__
  8829. #define vld1q_s32(__p0) __extension__ ({ \
  8830.   int32x4_t __ret; \
  8831.   __ret = (int32x4_t) __builtin_neon_vld1q_v(__p0, 34); \
  8832.   __ret; \
  8833. })
  8834. #else
  8835. #define vld1q_s32(__p0) __extension__ ({ \
  8836.   int32x4_t __ret; \
  8837.   __ret = (int32x4_t) __builtin_neon_vld1q_v(__p0, 34); \
  8838.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  8839.   __ret; \
  8840. })
  8841. #endif
  8842.  
  8843. #ifdef __LITTLE_ENDIAN__
  8844. #define vld1q_s64(__p0) __extension__ ({ \
  8845.   int64x2_t __ret; \
  8846.   __ret = (int64x2_t) __builtin_neon_vld1q_v(__p0, 35); \
  8847.   __ret; \
  8848. })
  8849. #else
  8850. #define vld1q_s64(__p0) __extension__ ({ \
  8851.   int64x2_t __ret; \
  8852.   __ret = (int64x2_t) __builtin_neon_vld1q_v(__p0, 35); \
  8853.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  8854.   __ret; \
  8855. })
  8856. #endif
  8857.  
  8858. #ifdef __LITTLE_ENDIAN__
  8859. #define vld1q_s16(__p0) __extension__ ({ \
  8860.   int16x8_t __ret; \
  8861.   __ret = (int16x8_t) __builtin_neon_vld1q_v(__p0, 33); \
  8862.   __ret; \
  8863. })
  8864. #else
  8865. #define vld1q_s16(__p0) __extension__ ({ \
  8866.   int16x8_t __ret; \
  8867.   __ret = (int16x8_t) __builtin_neon_vld1q_v(__p0, 33); \
  8868.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  8869.   __ret; \
  8870. })
  8871. #endif
  8872.  
  8873. #ifdef __LITTLE_ENDIAN__
  8874. #define vld1_u8(__p0) __extension__ ({ \
  8875.   uint8x8_t __ret; \
  8876.   __ret = (uint8x8_t) __builtin_neon_vld1_v(__p0, 16); \
  8877.   __ret; \
  8878. })
  8879. #else
  8880. #define vld1_u8(__p0) __extension__ ({ \
  8881.   uint8x8_t __ret; \
  8882.   __ret = (uint8x8_t) __builtin_neon_vld1_v(__p0, 16); \
  8883.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  8884.   __ret; \
  8885. })
  8886. #endif
  8887.  
  8888. #ifdef __LITTLE_ENDIAN__
  8889. #define vld1_u32(__p0) __extension__ ({ \
  8890.   uint32x2_t __ret; \
  8891.   __ret = (uint32x2_t) __builtin_neon_vld1_v(__p0, 18); \
  8892.   __ret; \
  8893. })
  8894. #else
  8895. #define vld1_u32(__p0) __extension__ ({ \
  8896.   uint32x2_t __ret; \
  8897.   __ret = (uint32x2_t) __builtin_neon_vld1_v(__p0, 18); \
  8898.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  8899.   __ret; \
  8900. })
  8901. #endif
  8902.  
  8903. #define vld1_u64(__p0) __extension__ ({ \
  8904.   uint64x1_t __ret; \
  8905.   __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
  8906.   __ret; \
  8907. })
  8908. #ifdef __LITTLE_ENDIAN__
  8909. #define vld1_u16(__p0) __extension__ ({ \
  8910.   uint16x4_t __ret; \
  8911.   __ret = (uint16x4_t) __builtin_neon_vld1_v(__p0, 17); \
  8912.   __ret; \
  8913. })
  8914. #else
  8915. #define vld1_u16(__p0) __extension__ ({ \
  8916.   uint16x4_t __ret; \
  8917.   __ret = (uint16x4_t) __builtin_neon_vld1_v(__p0, 17); \
  8918.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  8919.   __ret; \
  8920. })
  8921. #endif
  8922.  
  8923. #ifdef __LITTLE_ENDIAN__
  8924. #define vld1_s8(__p0) __extension__ ({ \
  8925.   int8x8_t __ret; \
  8926.   __ret = (int8x8_t) __builtin_neon_vld1_v(__p0, 0); \
  8927.   __ret; \
  8928. })
  8929. #else
  8930. #define vld1_s8(__p0) __extension__ ({ \
  8931.   int8x8_t __ret; \
  8932.   __ret = (int8x8_t) __builtin_neon_vld1_v(__p0, 0); \
  8933.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  8934.   __ret; \
  8935. })
  8936. #endif
  8937.  
  8938. #ifdef __LITTLE_ENDIAN__
  8939. #define vld1_f32(__p0) __extension__ ({ \
  8940.   float32x2_t __ret; \
  8941.   __ret = (float32x2_t) __builtin_neon_vld1_v(__p0, 9); \
  8942.   __ret; \
  8943. })
  8944. #else
  8945. #define vld1_f32(__p0) __extension__ ({ \
  8946.   float32x2_t __ret; \
  8947.   __ret = (float32x2_t) __builtin_neon_vld1_v(__p0, 9); \
  8948.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  8949.   __ret; \
  8950. })
  8951. #endif
  8952.  
  8953. #ifdef __LITTLE_ENDIAN__
  8954. #define vld1_s32(__p0) __extension__ ({ \
  8955.   int32x2_t __ret; \
  8956.   __ret = (int32x2_t) __builtin_neon_vld1_v(__p0, 2); \
  8957.   __ret; \
  8958. })
  8959. #else
  8960. #define vld1_s32(__p0) __extension__ ({ \
  8961.   int32x2_t __ret; \
  8962.   __ret = (int32x2_t) __builtin_neon_vld1_v(__p0, 2); \
  8963.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  8964.   __ret; \
  8965. })
  8966. #endif
  8967.  
  8968. #define vld1_s64(__p0) __extension__ ({ \
  8969.   int64x1_t __ret; \
  8970.   __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
  8971.   __ret; \
  8972. })
  8973. #ifdef __LITTLE_ENDIAN__
  8974. #define vld1_s16(__p0) __extension__ ({ \
  8975.   int16x4_t __ret; \
  8976.   __ret = (int16x4_t) __builtin_neon_vld1_v(__p0, 1); \
  8977.   __ret; \
  8978. })
  8979. #else
  8980. #define vld1_s16(__p0) __extension__ ({ \
  8981.   int16x4_t __ret; \
  8982.   __ret = (int16x4_t) __builtin_neon_vld1_v(__p0, 1); \
  8983.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  8984.   __ret; \
  8985. })
  8986. #endif
  8987.  
  8988. #ifdef __LITTLE_ENDIAN__
  8989. #define vld1_dup_p8(__p0) __extension__ ({ \
  8990.   poly8x8_t __ret; \
  8991.   __ret = (poly8x8_t) __builtin_neon_vld1_dup_v(__p0, 4); \
  8992.   __ret; \
  8993. })
  8994. #else
  8995. #define vld1_dup_p8(__p0) __extension__ ({ \
  8996.   poly8x8_t __ret; \
  8997.   __ret = (poly8x8_t) __builtin_neon_vld1_dup_v(__p0, 4); \
  8998.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  8999.   __ret; \
  9000. })
  9001. #endif
  9002.  
  9003. #ifdef __LITTLE_ENDIAN__
  9004. #define vld1_dup_p16(__p0) __extension__ ({ \
  9005.   poly16x4_t __ret; \
  9006.   __ret = (poly16x4_t) __builtin_neon_vld1_dup_v(__p0, 5); \
  9007.   __ret; \
  9008. })
  9009. #else
  9010. #define vld1_dup_p16(__p0) __extension__ ({ \
  9011.   poly16x4_t __ret; \
  9012.   __ret = (poly16x4_t) __builtin_neon_vld1_dup_v(__p0, 5); \
  9013.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9014.   __ret; \
  9015. })
  9016. #endif
  9017.  
  9018. #ifdef __LITTLE_ENDIAN__
  9019. #define vld1q_dup_p8(__p0) __extension__ ({ \
  9020.   poly8x16_t __ret; \
  9021.   __ret = (poly8x16_t) __builtin_neon_vld1q_dup_v(__p0, 36); \
  9022.   __ret; \
  9023. })
  9024. #else
  9025. #define vld1q_dup_p8(__p0) __extension__ ({ \
  9026.   poly8x16_t __ret; \
  9027.   __ret = (poly8x16_t) __builtin_neon_vld1q_dup_v(__p0, 36); \
  9028.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9029.   __ret; \
  9030. })
  9031. #endif
  9032.  
  9033. #ifdef __LITTLE_ENDIAN__
  9034. #define vld1q_dup_p16(__p0) __extension__ ({ \
  9035.   poly16x8_t __ret; \
  9036.   __ret = (poly16x8_t) __builtin_neon_vld1q_dup_v(__p0, 37); \
  9037.   __ret; \
  9038. })
  9039. #else
  9040. #define vld1q_dup_p16(__p0) __extension__ ({ \
  9041.   poly16x8_t __ret; \
  9042.   __ret = (poly16x8_t) __builtin_neon_vld1q_dup_v(__p0, 37); \
  9043.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9044.   __ret; \
  9045. })
  9046. #endif
  9047.  
  9048. #ifdef __LITTLE_ENDIAN__
  9049. #define vld1q_dup_u8(__p0) __extension__ ({ \
  9050.   uint8x16_t __ret; \
  9051.   __ret = (uint8x16_t) __builtin_neon_vld1q_dup_v(__p0, 48); \
  9052.   __ret; \
  9053. })
  9054. #else
  9055. #define vld1q_dup_u8(__p0) __extension__ ({ \
  9056.   uint8x16_t __ret; \
  9057.   __ret = (uint8x16_t) __builtin_neon_vld1q_dup_v(__p0, 48); \
  9058.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9059.   __ret; \
  9060. })
  9061. #endif
  9062.  
  9063. #ifdef __LITTLE_ENDIAN__
  9064. #define vld1q_dup_u32(__p0) __extension__ ({ \
  9065.   uint32x4_t __ret; \
  9066.   __ret = (uint32x4_t) __builtin_neon_vld1q_dup_v(__p0, 50); \
  9067.   __ret; \
  9068. })
  9069. #else
  9070. #define vld1q_dup_u32(__p0) __extension__ ({ \
  9071.   uint32x4_t __ret; \
  9072.   __ret = (uint32x4_t) __builtin_neon_vld1q_dup_v(__p0, 50); \
  9073.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9074.   __ret; \
  9075. })
  9076. #endif
  9077.  
  9078. #ifdef __LITTLE_ENDIAN__
  9079. #define vld1q_dup_u64(__p0) __extension__ ({ \
  9080.   uint64x2_t __ret; \
  9081.   __ret = (uint64x2_t) __builtin_neon_vld1q_dup_v(__p0, 51); \
  9082.   __ret; \
  9083. })
  9084. #else
  9085. #define vld1q_dup_u64(__p0) __extension__ ({ \
  9086.   uint64x2_t __ret; \
  9087.   __ret = (uint64x2_t) __builtin_neon_vld1q_dup_v(__p0, 51); \
  9088.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9089.   __ret; \
  9090. })
  9091. #endif
  9092.  
  9093. #ifdef __LITTLE_ENDIAN__
  9094. #define vld1q_dup_u16(__p0) __extension__ ({ \
  9095.   uint16x8_t __ret; \
  9096.   __ret = (uint16x8_t) __builtin_neon_vld1q_dup_v(__p0, 49); \
  9097.   __ret; \
  9098. })
  9099. #else
  9100. #define vld1q_dup_u16(__p0) __extension__ ({ \
  9101.   uint16x8_t __ret; \
  9102.   __ret = (uint16x8_t) __builtin_neon_vld1q_dup_v(__p0, 49); \
  9103.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9104.   __ret; \
  9105. })
  9106. #endif
  9107.  
  9108. #ifdef __LITTLE_ENDIAN__
  9109. #define vld1q_dup_s8(__p0) __extension__ ({ \
  9110.   int8x16_t __ret; \
  9111.   __ret = (int8x16_t) __builtin_neon_vld1q_dup_v(__p0, 32); \
  9112.   __ret; \
  9113. })
  9114. #else
  9115. #define vld1q_dup_s8(__p0) __extension__ ({ \
  9116.   int8x16_t __ret; \
  9117.   __ret = (int8x16_t) __builtin_neon_vld1q_dup_v(__p0, 32); \
  9118.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9119.   __ret; \
  9120. })
  9121. #endif
  9122.  
  9123. #ifdef __LITTLE_ENDIAN__
  9124. #define vld1q_dup_f32(__p0) __extension__ ({ \
  9125.   float32x4_t __ret; \
  9126.   __ret = (float32x4_t) __builtin_neon_vld1q_dup_v(__p0, 41); \
  9127.   __ret; \
  9128. })
  9129. #else
  9130. #define vld1q_dup_f32(__p0) __extension__ ({ \
  9131.   float32x4_t __ret; \
  9132.   __ret = (float32x4_t) __builtin_neon_vld1q_dup_v(__p0, 41); \
  9133.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9134.   __ret; \
  9135. })
  9136. #endif
  9137.  
  9138. #ifdef __LITTLE_ENDIAN__
  9139. #define vld1q_dup_s32(__p0) __extension__ ({ \
  9140.   int32x4_t __ret; \
  9141.   __ret = (int32x4_t) __builtin_neon_vld1q_dup_v(__p0, 34); \
  9142.   __ret; \
  9143. })
  9144. #else
  9145. #define vld1q_dup_s32(__p0) __extension__ ({ \
  9146.   int32x4_t __ret; \
  9147.   __ret = (int32x4_t) __builtin_neon_vld1q_dup_v(__p0, 34); \
  9148.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9149.   __ret; \
  9150. })
  9151. #endif
  9152.  
  9153. #ifdef __LITTLE_ENDIAN__
  9154. #define vld1q_dup_s64(__p0) __extension__ ({ \
  9155.   int64x2_t __ret; \
  9156.   __ret = (int64x2_t) __builtin_neon_vld1q_dup_v(__p0, 35); \
  9157.   __ret; \
  9158. })
  9159. #else
  9160. #define vld1q_dup_s64(__p0) __extension__ ({ \
  9161.   int64x2_t __ret; \
  9162.   __ret = (int64x2_t) __builtin_neon_vld1q_dup_v(__p0, 35); \
  9163.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9164.   __ret; \
  9165. })
  9166. #endif
  9167.  
  9168. #ifdef __LITTLE_ENDIAN__
  9169. #define vld1q_dup_s16(__p0) __extension__ ({ \
  9170.   int16x8_t __ret; \
  9171.   __ret = (int16x8_t) __builtin_neon_vld1q_dup_v(__p0, 33); \
  9172.   __ret; \
  9173. })
  9174. #else
  9175. #define vld1q_dup_s16(__p0) __extension__ ({ \
  9176.   int16x8_t __ret; \
  9177.   __ret = (int16x8_t) __builtin_neon_vld1q_dup_v(__p0, 33); \
  9178.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9179.   __ret; \
  9180. })
  9181. #endif
  9182.  
  9183. #ifdef __LITTLE_ENDIAN__
  9184. #define vld1_dup_u8(__p0) __extension__ ({ \
  9185.   uint8x8_t __ret; \
  9186.   __ret = (uint8x8_t) __builtin_neon_vld1_dup_v(__p0, 16); \
  9187.   __ret; \
  9188. })
  9189. #else
  9190. #define vld1_dup_u8(__p0) __extension__ ({ \
  9191.   uint8x8_t __ret; \
  9192.   __ret = (uint8x8_t) __builtin_neon_vld1_dup_v(__p0, 16); \
  9193.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9194.   __ret; \
  9195. })
  9196. #endif
  9197.  
  9198. #ifdef __LITTLE_ENDIAN__
  9199. #define vld1_dup_u32(__p0) __extension__ ({ \
  9200.   uint32x2_t __ret; \
  9201.   __ret = (uint32x2_t) __builtin_neon_vld1_dup_v(__p0, 18); \
  9202.   __ret; \
  9203. })
  9204. #else
  9205. #define vld1_dup_u32(__p0) __extension__ ({ \
  9206.   uint32x2_t __ret; \
  9207.   __ret = (uint32x2_t) __builtin_neon_vld1_dup_v(__p0, 18); \
  9208.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9209.   __ret; \
  9210. })
  9211. #endif
  9212.  
  9213. #define vld1_dup_u64(__p0) __extension__ ({ \
  9214.   uint64x1_t __ret; \
  9215.   __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
  9216.   __ret; \
  9217. })
  9218. #ifdef __LITTLE_ENDIAN__
  9219. #define vld1_dup_u16(__p0) __extension__ ({ \
  9220.   uint16x4_t __ret; \
  9221.   __ret = (uint16x4_t) __builtin_neon_vld1_dup_v(__p0, 17); \
  9222.   __ret; \
  9223. })
  9224. #else
  9225. #define vld1_dup_u16(__p0) __extension__ ({ \
  9226.   uint16x4_t __ret; \
  9227.   __ret = (uint16x4_t) __builtin_neon_vld1_dup_v(__p0, 17); \
  9228.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9229.   __ret; \
  9230. })
  9231. #endif
  9232.  
  9233. #ifdef __LITTLE_ENDIAN__
  9234. #define vld1_dup_s8(__p0) __extension__ ({ \
  9235.   int8x8_t __ret; \
  9236.   __ret = (int8x8_t) __builtin_neon_vld1_dup_v(__p0, 0); \
  9237.   __ret; \
  9238. })
  9239. #else
  9240. #define vld1_dup_s8(__p0) __extension__ ({ \
  9241.   int8x8_t __ret; \
  9242.   __ret = (int8x8_t) __builtin_neon_vld1_dup_v(__p0, 0); \
  9243.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9244.   __ret; \
  9245. })
  9246. #endif
  9247.  
  9248. #ifdef __LITTLE_ENDIAN__
  9249. #define vld1_dup_f32(__p0) __extension__ ({ \
  9250.   float32x2_t __ret; \
  9251.   __ret = (float32x2_t) __builtin_neon_vld1_dup_v(__p0, 9); \
  9252.   __ret; \
  9253. })
  9254. #else
  9255. #define vld1_dup_f32(__p0) __extension__ ({ \
  9256.   float32x2_t __ret; \
  9257.   __ret = (float32x2_t) __builtin_neon_vld1_dup_v(__p0, 9); \
  9258.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9259.   __ret; \
  9260. })
  9261. #endif
  9262.  
  9263. #ifdef __LITTLE_ENDIAN__
  9264. #define vld1_dup_s32(__p0) __extension__ ({ \
  9265.   int32x2_t __ret; \
  9266.   __ret = (int32x2_t) __builtin_neon_vld1_dup_v(__p0, 2); \
  9267.   __ret; \
  9268. })
  9269. #else
  9270. #define vld1_dup_s32(__p0) __extension__ ({ \
  9271.   int32x2_t __ret; \
  9272.   __ret = (int32x2_t) __builtin_neon_vld1_dup_v(__p0, 2); \
  9273.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9274.   __ret; \
  9275. })
  9276. #endif
  9277.  
  9278. #define vld1_dup_s64(__p0) __extension__ ({ \
  9279.   int64x1_t __ret; \
  9280.   __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
  9281.   __ret; \
  9282. })
  9283. #ifdef __LITTLE_ENDIAN__
  9284. #define vld1_dup_s16(__p0) __extension__ ({ \
  9285.   int16x4_t __ret; \
  9286.   __ret = (int16x4_t) __builtin_neon_vld1_dup_v(__p0, 1); \
  9287.   __ret; \
  9288. })
  9289. #else
  9290. #define vld1_dup_s16(__p0) __extension__ ({ \
  9291.   int16x4_t __ret; \
  9292.   __ret = (int16x4_t) __builtin_neon_vld1_dup_v(__p0, 1); \
  9293.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9294.   __ret; \
  9295. })
  9296. #endif
  9297.  
  9298. #ifdef __LITTLE_ENDIAN__
  9299. #define vld1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  9300.   poly8x8_t __ret; \
  9301.   poly8x8_t __s1 = __p1; \
  9302.   __ret = (poly8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 4); \
  9303.   __ret; \
  9304. })
  9305. #else
  9306. #define vld1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  9307.   poly8x8_t __ret; \
  9308.   poly8x8_t __s1 = __p1; \
  9309.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  9310.   __ret = (poly8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 4); \
  9311.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9312.   __ret; \
  9313. })
  9314. #endif
  9315.  
  9316. #ifdef __LITTLE_ENDIAN__
  9317. #define vld1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  9318.   poly16x4_t __ret; \
  9319.   poly16x4_t __s1 = __p1; \
  9320.   __ret = (poly16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 5); \
  9321.   __ret; \
  9322. })
  9323. #else
  9324. #define vld1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  9325.   poly16x4_t __ret; \
  9326.   poly16x4_t __s1 = __p1; \
  9327.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  9328.   __ret = (poly16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 5); \
  9329.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9330.   __ret; \
  9331. })
  9332. #endif
  9333.  
  9334. #ifdef __LITTLE_ENDIAN__
  9335. #define vld1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  9336.   poly8x16_t __ret; \
  9337.   poly8x16_t __s1 = __p1; \
  9338.   __ret = (poly8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 36); \
  9339.   __ret; \
  9340. })
  9341. #else
  9342. #define vld1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  9343.   poly8x16_t __ret; \
  9344.   poly8x16_t __s1 = __p1; \
  9345.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9346.   __ret = (poly8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 36); \
  9347.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9348.   __ret; \
  9349. })
  9350. #endif
  9351.  
  9352. #ifdef __LITTLE_ENDIAN__
  9353. #define vld1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  9354.   poly16x8_t __ret; \
  9355.   poly16x8_t __s1 = __p1; \
  9356.   __ret = (poly16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 37); \
  9357.   __ret; \
  9358. })
  9359. #else
  9360. #define vld1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  9361.   poly16x8_t __ret; \
  9362.   poly16x8_t __s1 = __p1; \
  9363.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  9364.   __ret = (poly16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 37); \
  9365.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9366.   __ret; \
  9367. })
  9368. #endif
  9369.  
  9370. #ifdef __LITTLE_ENDIAN__
  9371. #define vld1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  9372.   uint8x16_t __ret; \
  9373.   uint8x16_t __s1 = __p1; \
  9374.   __ret = (uint8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 48); \
  9375.   __ret; \
  9376. })
  9377. #else
  9378. #define vld1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  9379.   uint8x16_t __ret; \
  9380.   uint8x16_t __s1 = __p1; \
  9381.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9382.   __ret = (uint8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 48); \
  9383.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9384.   __ret; \
  9385. })
  9386. #endif
  9387.  
  9388. #ifdef __LITTLE_ENDIAN__
  9389. #define vld1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  9390.   uint32x4_t __ret; \
  9391.   uint32x4_t __s1 = __p1; \
  9392.   __ret = (uint32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 50); \
  9393.   __ret; \
  9394. })
  9395. #else
  9396. #define vld1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  9397.   uint32x4_t __ret; \
  9398.   uint32x4_t __s1 = __p1; \
  9399.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  9400.   __ret = (uint32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 50); \
  9401.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9402.   __ret; \
  9403. })
  9404. #endif
  9405.  
  9406. #ifdef __LITTLE_ENDIAN__
  9407. #define vld1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  9408.   uint64x2_t __ret; \
  9409.   uint64x2_t __s1 = __p1; \
  9410.   __ret = (uint64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 51); \
  9411.   __ret; \
  9412. })
  9413. #else
  9414. #define vld1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  9415.   uint64x2_t __ret; \
  9416.   uint64x2_t __s1 = __p1; \
  9417.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  9418.   __ret = (uint64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 51); \
  9419.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9420.   __ret; \
  9421. })
  9422. #endif
  9423.  
  9424. #ifdef __LITTLE_ENDIAN__
  9425. #define vld1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  9426.   uint16x8_t __ret; \
  9427.   uint16x8_t __s1 = __p1; \
  9428.   __ret = (uint16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 49); \
  9429.   __ret; \
  9430. })
  9431. #else
  9432. #define vld1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  9433.   uint16x8_t __ret; \
  9434.   uint16x8_t __s1 = __p1; \
  9435.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  9436.   __ret = (uint16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 49); \
  9437.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9438.   __ret; \
  9439. })
  9440. #endif
  9441.  
  9442. #ifdef __LITTLE_ENDIAN__
  9443. #define vld1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  9444.   int8x16_t __ret; \
  9445.   int8x16_t __s1 = __p1; \
  9446.   __ret = (int8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 32); \
  9447.   __ret; \
  9448. })
  9449. #else
  9450. #define vld1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  9451.   int8x16_t __ret; \
  9452.   int8x16_t __s1 = __p1; \
  9453.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9454.   __ret = (int8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 32); \
  9455.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9456.   __ret; \
  9457. })
  9458. #endif
  9459.  
  9460. #ifdef __LITTLE_ENDIAN__
  9461. #define vld1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  9462.   float32x4_t __ret; \
  9463.   float32x4_t __s1 = __p1; \
  9464.   __ret = (float32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 41); \
  9465.   __ret; \
  9466. })
  9467. #else
  9468. #define vld1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  9469.   float32x4_t __ret; \
  9470.   float32x4_t __s1 = __p1; \
  9471.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  9472.   __ret = (float32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 41); \
  9473.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9474.   __ret; \
  9475. })
  9476. #endif
  9477.  
  9478. #ifdef __LITTLE_ENDIAN__
  9479. #define vld1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  9480.   int32x4_t __ret; \
  9481.   int32x4_t __s1 = __p1; \
  9482.   __ret = (int32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 34); \
  9483.   __ret; \
  9484. })
  9485. #else
  9486. #define vld1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  9487.   int32x4_t __ret; \
  9488.   int32x4_t __s1 = __p1; \
  9489.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  9490.   __ret = (int32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 34); \
  9491.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9492.   __ret; \
  9493. })
  9494. #endif
  9495.  
  9496. #ifdef __LITTLE_ENDIAN__
  9497. #define vld1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  9498.   int64x2_t __ret; \
  9499.   int64x2_t __s1 = __p1; \
  9500.   __ret = (int64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 35); \
  9501.   __ret; \
  9502. })
  9503. #else
  9504. #define vld1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  9505.   int64x2_t __ret; \
  9506.   int64x2_t __s1 = __p1; \
  9507.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  9508.   __ret = (int64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 35); \
  9509.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9510.   __ret; \
  9511. })
  9512. #endif
  9513.  
  9514. #ifdef __LITTLE_ENDIAN__
  9515. #define vld1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  9516.   int16x8_t __ret; \
  9517.   int16x8_t __s1 = __p1; \
  9518.   __ret = (int16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 33); \
  9519.   __ret; \
  9520. })
  9521. #else
  9522. #define vld1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  9523.   int16x8_t __ret; \
  9524.   int16x8_t __s1 = __p1; \
  9525.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  9526.   __ret = (int16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 33); \
  9527.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9528.   __ret; \
  9529. })
  9530. #endif
  9531.  
  9532. #ifdef __LITTLE_ENDIAN__
  9533. #define vld1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  9534.   uint8x8_t __ret; \
  9535.   uint8x8_t __s1 = __p1; \
  9536.   __ret = (uint8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 16); \
  9537.   __ret; \
  9538. })
  9539. #else
  9540. #define vld1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  9541.   uint8x8_t __ret; \
  9542.   uint8x8_t __s1 = __p1; \
  9543.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  9544.   __ret = (uint8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 16); \
  9545.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9546.   __ret; \
  9547. })
  9548. #endif
  9549.  
  9550. #ifdef __LITTLE_ENDIAN__
  9551. #define vld1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  9552.   uint32x2_t __ret; \
  9553.   uint32x2_t __s1 = __p1; \
  9554.   __ret = (uint32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 18); \
  9555.   __ret; \
  9556. })
  9557. #else
  9558. #define vld1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  9559.   uint32x2_t __ret; \
  9560.   uint32x2_t __s1 = __p1; \
  9561.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  9562.   __ret = (uint32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 18); \
  9563.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9564.   __ret; \
  9565. })
  9566. #endif
  9567.  
  9568. #define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  9569.   uint64x1_t __ret; \
  9570.   uint64x1_t __s1 = __p1; \
  9571.   __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
  9572.   __ret; \
  9573. })
  9574. #ifdef __LITTLE_ENDIAN__
  9575. #define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  9576.   uint16x4_t __ret; \
  9577.   uint16x4_t __s1 = __p1; \
  9578.   __ret = (uint16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 17); \
  9579.   __ret; \
  9580. })
  9581. #else
  9582. #define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  9583.   uint16x4_t __ret; \
  9584.   uint16x4_t __s1 = __p1; \
  9585.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  9586.   __ret = (uint16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 17); \
  9587.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9588.   __ret; \
  9589. })
  9590. #endif
  9591.  
  9592. #ifdef __LITTLE_ENDIAN__
  9593. #define vld1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  9594.   int8x8_t __ret; \
  9595.   int8x8_t __s1 = __p1; \
  9596.   __ret = (int8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 0); \
  9597.   __ret; \
  9598. })
  9599. #else
  9600. #define vld1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  9601.   int8x8_t __ret; \
  9602.   int8x8_t __s1 = __p1; \
  9603.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  9604.   __ret = (int8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 0); \
  9605.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  9606.   __ret; \
  9607. })
  9608. #endif
  9609.  
  9610. #ifdef __LITTLE_ENDIAN__
  9611. #define vld1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  9612.   float32x2_t __ret; \
  9613.   float32x2_t __s1 = __p1; \
  9614.   __ret = (float32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 9); \
  9615.   __ret; \
  9616. })
  9617. #else
  9618. #define vld1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  9619.   float32x2_t __ret; \
  9620.   float32x2_t __s1 = __p1; \
  9621.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  9622.   __ret = (float32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 9); \
  9623.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9624.   __ret; \
  9625. })
  9626. #endif
  9627.  
  9628. #ifdef __LITTLE_ENDIAN__
  9629. #define vld1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  9630.   int32x2_t __ret; \
  9631.   int32x2_t __s1 = __p1; \
  9632.   __ret = (int32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 2); \
  9633.   __ret; \
  9634. })
  9635. #else
  9636. #define vld1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  9637.   int32x2_t __ret; \
  9638.   int32x2_t __s1 = __p1; \
  9639.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  9640.   __ret = (int32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 2); \
  9641.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  9642.   __ret; \
  9643. })
  9644. #endif
  9645.  
  9646. #define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  9647.   int64x1_t __ret; \
  9648.   int64x1_t __s1 = __p1; \
  9649.   __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
  9650.   __ret; \
  9651. })
  9652. #ifdef __LITTLE_ENDIAN__
  9653. #define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  9654.   int16x4_t __ret; \
  9655.   int16x4_t __s1 = __p1; \
  9656.   __ret = (int16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 1); \
  9657.   __ret; \
  9658. })
  9659. #else
  9660. #define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  9661.   int16x4_t __ret; \
  9662.   int16x4_t __s1 = __p1; \
  9663.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  9664.   __ret = (int16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 1); \
  9665.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  9666.   __ret; \
  9667. })
  9668. #endif
  9669.  
  9670. #ifdef __LITTLE_ENDIAN__
  9671. #define vld1_p8_x2(__p0) __extension__ ({ \
  9672.   poly8x8x2_t __ret; \
  9673.   __builtin_neon_vld1_x2_v(&__ret, __p0, 4); \
  9674.   __ret; \
  9675. })
  9676. #else
  9677. #define vld1_p8_x2(__p0) __extension__ ({ \
  9678.   poly8x8x2_t __ret; \
  9679.   __builtin_neon_vld1_x2_v(&__ret, __p0, 4); \
  9680.  \
  9681.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  9682.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  9683.   __ret; \
  9684. })
  9685. #endif
  9686.  
  9687. #ifdef __LITTLE_ENDIAN__
  9688. #define vld1_p16_x2(__p0) __extension__ ({ \
  9689.   poly16x4x2_t __ret; \
  9690.   __builtin_neon_vld1_x2_v(&__ret, __p0, 5); \
  9691.   __ret; \
  9692. })
  9693. #else
  9694. #define vld1_p16_x2(__p0) __extension__ ({ \
  9695.   poly16x4x2_t __ret; \
  9696.   __builtin_neon_vld1_x2_v(&__ret, __p0, 5); \
  9697.  \
  9698.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  9699.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  9700.   __ret; \
  9701. })
  9702. #endif
  9703.  
  9704. #ifdef __LITTLE_ENDIAN__
  9705. #define vld1q_p8_x2(__p0) __extension__ ({ \
  9706.   poly8x16x2_t __ret; \
  9707.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 36); \
  9708.   __ret; \
  9709. })
  9710. #else
  9711. #define vld1q_p8_x2(__p0) __extension__ ({ \
  9712.   poly8x16x2_t __ret; \
  9713.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 36); \
  9714.  \
  9715.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9716.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9717.   __ret; \
  9718. })
  9719. #endif
  9720.  
  9721. #ifdef __LITTLE_ENDIAN__
  9722. #define vld1q_p16_x2(__p0) __extension__ ({ \
  9723.   poly16x8x2_t __ret; \
  9724.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 37); \
  9725.   __ret; \
  9726. })
  9727. #else
  9728. #define vld1q_p16_x2(__p0) __extension__ ({ \
  9729.   poly16x8x2_t __ret; \
  9730.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 37); \
  9731.  \
  9732.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  9733.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  9734.   __ret; \
  9735. })
  9736. #endif
  9737.  
  9738. #ifdef __LITTLE_ENDIAN__
  9739. #define vld1q_u8_x2(__p0) __extension__ ({ \
  9740.   uint8x16x2_t __ret; \
  9741.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 48); \
  9742.   __ret; \
  9743. })
  9744. #else
  9745. #define vld1q_u8_x2(__p0) __extension__ ({ \
  9746.   uint8x16x2_t __ret; \
  9747.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 48); \
  9748.  \
  9749.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9750.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9751.   __ret; \
  9752. })
  9753. #endif
  9754.  
  9755. #ifdef __LITTLE_ENDIAN__
  9756. #define vld1q_u32_x2(__p0) __extension__ ({ \
  9757.   uint32x4x2_t __ret; \
  9758.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 50); \
  9759.   __ret; \
  9760. })
  9761. #else
  9762. #define vld1q_u32_x2(__p0) __extension__ ({ \
  9763.   uint32x4x2_t __ret; \
  9764.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 50); \
  9765.  \
  9766.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  9767.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  9768.   __ret; \
  9769. })
  9770. #endif
  9771.  
  9772. #ifdef __LITTLE_ENDIAN__
  9773. #define vld1q_u64_x2(__p0) __extension__ ({ \
  9774.   uint64x2x2_t __ret; \
  9775.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 51); \
  9776.   __ret; \
  9777. })
  9778. #else
  9779. #define vld1q_u64_x2(__p0) __extension__ ({ \
  9780.   uint64x2x2_t __ret; \
  9781.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 51); \
  9782.  \
  9783.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  9784.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  9785.   __ret; \
  9786. })
  9787. #endif
  9788.  
  9789. #ifdef __LITTLE_ENDIAN__
  9790. #define vld1q_u16_x2(__p0) __extension__ ({ \
  9791.   uint16x8x2_t __ret; \
  9792.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 49); \
  9793.   __ret; \
  9794. })
  9795. #else
  9796. #define vld1q_u16_x2(__p0) __extension__ ({ \
  9797.   uint16x8x2_t __ret; \
  9798.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 49); \
  9799.  \
  9800.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  9801.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  9802.   __ret; \
  9803. })
  9804. #endif
  9805.  
  9806. #ifdef __LITTLE_ENDIAN__
  9807. #define vld1q_s8_x2(__p0) __extension__ ({ \
  9808.   int8x16x2_t __ret; \
  9809.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 32); \
  9810.   __ret; \
  9811. })
  9812. #else
  9813. #define vld1q_s8_x2(__p0) __extension__ ({ \
  9814.   int8x16x2_t __ret; \
  9815.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 32); \
  9816.  \
  9817.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9818.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  9819.   __ret; \
  9820. })
  9821. #endif
  9822.  
  9823. #ifdef __LITTLE_ENDIAN__
  9824. #define vld1q_f32_x2(__p0) __extension__ ({ \
  9825.   float32x4x2_t __ret; \
  9826.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 41); \
  9827.   __ret; \
  9828. })
  9829. #else
  9830. #define vld1q_f32_x2(__p0) __extension__ ({ \
  9831.   float32x4x2_t __ret; \
  9832.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 41); \
  9833.  \
  9834.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  9835.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  9836.   __ret; \
  9837. })
  9838. #endif
  9839.  
  9840. #ifdef __LITTLE_ENDIAN__
  9841. #define vld1q_s32_x2(__p0) __extension__ ({ \
  9842.   int32x4x2_t __ret; \
  9843.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 34); \
  9844.   __ret; \
  9845. })
  9846. #else
  9847. #define vld1q_s32_x2(__p0) __extension__ ({ \
  9848.   int32x4x2_t __ret; \
  9849.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 34); \
  9850.  \
  9851.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  9852.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  9853.   __ret; \
  9854. })
  9855. #endif
  9856.  
  9857. #ifdef __LITTLE_ENDIAN__
  9858. #define vld1q_s64_x2(__p0) __extension__ ({ \
  9859.   int64x2x2_t __ret; \
  9860.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 35); \
  9861.   __ret; \
  9862. })
  9863. #else
  9864. #define vld1q_s64_x2(__p0) __extension__ ({ \
  9865.   int64x2x2_t __ret; \
  9866.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 35); \
  9867.  \
  9868.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  9869.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  9870.   __ret; \
  9871. })
  9872. #endif
  9873.  
  9874. #ifdef __LITTLE_ENDIAN__
  9875. #define vld1q_s16_x2(__p0) __extension__ ({ \
  9876.   int16x8x2_t __ret; \
  9877.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 33); \
  9878.   __ret; \
  9879. })
  9880. #else
  9881. #define vld1q_s16_x2(__p0) __extension__ ({ \
  9882.   int16x8x2_t __ret; \
  9883.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 33); \
  9884.  \
  9885.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  9886.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  9887.   __ret; \
  9888. })
  9889. #endif
  9890.  
  9891. #ifdef __LITTLE_ENDIAN__
  9892. #define vld1_u8_x2(__p0) __extension__ ({ \
  9893.   uint8x8x2_t __ret; \
  9894.   __builtin_neon_vld1_x2_v(&__ret, __p0, 16); \
  9895.   __ret; \
  9896. })
  9897. #else
  9898. #define vld1_u8_x2(__p0) __extension__ ({ \
  9899.   uint8x8x2_t __ret; \
  9900.   __builtin_neon_vld1_x2_v(&__ret, __p0, 16); \
  9901.  \
  9902.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  9903.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  9904.   __ret; \
  9905. })
  9906. #endif
  9907.  
  9908. #ifdef __LITTLE_ENDIAN__
  9909. #define vld1_u32_x2(__p0) __extension__ ({ \
  9910.   uint32x2x2_t __ret; \
  9911.   __builtin_neon_vld1_x2_v(&__ret, __p0, 18); \
  9912.   __ret; \
  9913. })
  9914. #else
  9915. #define vld1_u32_x2(__p0) __extension__ ({ \
  9916.   uint32x2x2_t __ret; \
  9917.   __builtin_neon_vld1_x2_v(&__ret, __p0, 18); \
  9918.  \
  9919.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  9920.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  9921.   __ret; \
  9922. })
  9923. #endif
  9924.  
  9925. #define vld1_u64_x2(__p0) __extension__ ({ \
  9926.   uint64x1x2_t __ret; \
  9927.   __builtin_neon_vld1_x2_v(&__ret, __p0, 19); \
  9928.   __ret; \
  9929. })
  9930. #ifdef __LITTLE_ENDIAN__
  9931. #define vld1_u16_x2(__p0) __extension__ ({ \
  9932.   uint16x4x2_t __ret; \
  9933.   __builtin_neon_vld1_x2_v(&__ret, __p0, 17); \
  9934.   __ret; \
  9935. })
  9936. #else
  9937. #define vld1_u16_x2(__p0) __extension__ ({ \
  9938.   uint16x4x2_t __ret; \
  9939.   __builtin_neon_vld1_x2_v(&__ret, __p0, 17); \
  9940.  \
  9941.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  9942.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  9943.   __ret; \
  9944. })
  9945. #endif
  9946.  
  9947. #ifdef __LITTLE_ENDIAN__
  9948. #define vld1_s8_x2(__p0) __extension__ ({ \
  9949.   int8x8x2_t __ret; \
  9950.   __builtin_neon_vld1_x2_v(&__ret, __p0, 0); \
  9951.   __ret; \
  9952. })
  9953. #else
  9954. #define vld1_s8_x2(__p0) __extension__ ({ \
  9955.   int8x8x2_t __ret; \
  9956.   __builtin_neon_vld1_x2_v(&__ret, __p0, 0); \
  9957.  \
  9958.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  9959.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  9960.   __ret; \
  9961. })
  9962. #endif
  9963.  
  9964. #ifdef __LITTLE_ENDIAN__
  9965. #define vld1_f32_x2(__p0) __extension__ ({ \
  9966.   float32x2x2_t __ret; \
  9967.   __builtin_neon_vld1_x2_v(&__ret, __p0, 9); \
  9968.   __ret; \
  9969. })
  9970. #else
  9971. #define vld1_f32_x2(__p0) __extension__ ({ \
  9972.   float32x2x2_t __ret; \
  9973.   __builtin_neon_vld1_x2_v(&__ret, __p0, 9); \
  9974.  \
  9975.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  9976.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  9977.   __ret; \
  9978. })
  9979. #endif
  9980.  
  9981. #ifdef __LITTLE_ENDIAN__
  9982. #define vld1_s32_x2(__p0) __extension__ ({ \
  9983.   int32x2x2_t __ret; \
  9984.   __builtin_neon_vld1_x2_v(&__ret, __p0, 2); \
  9985.   __ret; \
  9986. })
  9987. #else
  9988. #define vld1_s32_x2(__p0) __extension__ ({ \
  9989.   int32x2x2_t __ret; \
  9990.   __builtin_neon_vld1_x2_v(&__ret, __p0, 2); \
  9991.  \
  9992.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  9993.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  9994.   __ret; \
  9995. })
  9996. #endif
  9997.  
  9998. #define vld1_s64_x2(__p0) __extension__ ({ \
  9999.   int64x1x2_t __ret; \
  10000.   __builtin_neon_vld1_x2_v(&__ret, __p0, 3); \
  10001.   __ret; \
  10002. })
  10003. #ifdef __LITTLE_ENDIAN__
  10004. #define vld1_s16_x2(__p0) __extension__ ({ \
  10005.   int16x4x2_t __ret; \
  10006.   __builtin_neon_vld1_x2_v(&__ret, __p0, 1); \
  10007.   __ret; \
  10008. })
  10009. #else
  10010. #define vld1_s16_x2(__p0) __extension__ ({ \
  10011.   int16x4x2_t __ret; \
  10012.   __builtin_neon_vld1_x2_v(&__ret, __p0, 1); \
  10013.  \
  10014.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10015.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10016.   __ret; \
  10017. })
  10018. #endif
  10019.  
  10020. #ifdef __LITTLE_ENDIAN__
  10021. #define vld1_p8_x3(__p0) __extension__ ({ \
  10022.   poly8x8x3_t __ret; \
  10023.   __builtin_neon_vld1_x3_v(&__ret, __p0, 4); \
  10024.   __ret; \
  10025. })
  10026. #else
  10027. #define vld1_p8_x3(__p0) __extension__ ({ \
  10028.   poly8x8x3_t __ret; \
  10029.   __builtin_neon_vld1_x3_v(&__ret, __p0, 4); \
  10030.  \
  10031.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10032.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10033.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10034.   __ret; \
  10035. })
  10036. #endif
  10037.  
  10038. #ifdef __LITTLE_ENDIAN__
  10039. #define vld1_p16_x3(__p0) __extension__ ({ \
  10040.   poly16x4x3_t __ret; \
  10041.   __builtin_neon_vld1_x3_v(&__ret, __p0, 5); \
  10042.   __ret; \
  10043. })
  10044. #else
  10045. #define vld1_p16_x3(__p0) __extension__ ({ \
  10046.   poly16x4x3_t __ret; \
  10047.   __builtin_neon_vld1_x3_v(&__ret, __p0, 5); \
  10048.  \
  10049.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10050.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10051.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10052.   __ret; \
  10053. })
  10054. #endif
  10055.  
  10056. #ifdef __LITTLE_ENDIAN__
  10057. #define vld1q_p8_x3(__p0) __extension__ ({ \
  10058.   poly8x16x3_t __ret; \
  10059.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 36); \
  10060.   __ret; \
  10061. })
  10062. #else
  10063. #define vld1q_p8_x3(__p0) __extension__ ({ \
  10064.   poly8x16x3_t __ret; \
  10065.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 36); \
  10066.  \
  10067.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10068.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10069.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10070.   __ret; \
  10071. })
  10072. #endif
  10073.  
  10074. #ifdef __LITTLE_ENDIAN__
  10075. #define vld1q_p16_x3(__p0) __extension__ ({ \
  10076.   poly16x8x3_t __ret; \
  10077.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 37); \
  10078.   __ret; \
  10079. })
  10080. #else
  10081. #define vld1q_p16_x3(__p0) __extension__ ({ \
  10082.   poly16x8x3_t __ret; \
  10083.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 37); \
  10084.  \
  10085.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10086.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10087.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10088.   __ret; \
  10089. })
  10090. #endif
  10091.  
  10092. #ifdef __LITTLE_ENDIAN__
  10093. #define vld1q_u8_x3(__p0) __extension__ ({ \
  10094.   uint8x16x3_t __ret; \
  10095.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 48); \
  10096.   __ret; \
  10097. })
  10098. #else
  10099. #define vld1q_u8_x3(__p0) __extension__ ({ \
  10100.   uint8x16x3_t __ret; \
  10101.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 48); \
  10102.  \
  10103.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10104.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10105.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10106.   __ret; \
  10107. })
  10108. #endif
  10109.  
  10110. #ifdef __LITTLE_ENDIAN__
  10111. #define vld1q_u32_x3(__p0) __extension__ ({ \
  10112.   uint32x4x3_t __ret; \
  10113.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 50); \
  10114.   __ret; \
  10115. })
  10116. #else
  10117. #define vld1q_u32_x3(__p0) __extension__ ({ \
  10118.   uint32x4x3_t __ret; \
  10119.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 50); \
  10120.  \
  10121.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10122.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10123.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10124.   __ret; \
  10125. })
  10126. #endif
  10127.  
  10128. #ifdef __LITTLE_ENDIAN__
  10129. #define vld1q_u64_x3(__p0) __extension__ ({ \
  10130.   uint64x2x3_t __ret; \
  10131.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 51); \
  10132.   __ret; \
  10133. })
  10134. #else
  10135. #define vld1q_u64_x3(__p0) __extension__ ({ \
  10136.   uint64x2x3_t __ret; \
  10137.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 51); \
  10138.  \
  10139.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10140.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10141.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10142.   __ret; \
  10143. })
  10144. #endif
  10145.  
  10146. #ifdef __LITTLE_ENDIAN__
  10147. #define vld1q_u16_x3(__p0) __extension__ ({ \
  10148.   uint16x8x3_t __ret; \
  10149.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 49); \
  10150.   __ret; \
  10151. })
  10152. #else
  10153. #define vld1q_u16_x3(__p0) __extension__ ({ \
  10154.   uint16x8x3_t __ret; \
  10155.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 49); \
  10156.  \
  10157.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10158.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10159.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10160.   __ret; \
  10161. })
  10162. #endif
  10163.  
  10164. #ifdef __LITTLE_ENDIAN__
  10165. #define vld1q_s8_x3(__p0) __extension__ ({ \
  10166.   int8x16x3_t __ret; \
  10167.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 32); \
  10168.   __ret; \
  10169. })
  10170. #else
  10171. #define vld1q_s8_x3(__p0) __extension__ ({ \
  10172.   int8x16x3_t __ret; \
  10173.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 32); \
  10174.  \
  10175.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10176.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10177.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10178.   __ret; \
  10179. })
  10180. #endif
  10181.  
  10182. #ifdef __LITTLE_ENDIAN__
  10183. #define vld1q_f32_x3(__p0) __extension__ ({ \
  10184.   float32x4x3_t __ret; \
  10185.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 41); \
  10186.   __ret; \
  10187. })
  10188. #else
  10189. #define vld1q_f32_x3(__p0) __extension__ ({ \
  10190.   float32x4x3_t __ret; \
  10191.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 41); \
  10192.  \
  10193.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10194.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10195.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10196.   __ret; \
  10197. })
  10198. #endif
  10199.  
  10200. #ifdef __LITTLE_ENDIAN__
  10201. #define vld1q_s32_x3(__p0) __extension__ ({ \
  10202.   int32x4x3_t __ret; \
  10203.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 34); \
  10204.   __ret; \
  10205. })
  10206. #else
  10207. #define vld1q_s32_x3(__p0) __extension__ ({ \
  10208.   int32x4x3_t __ret; \
  10209.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 34); \
  10210.  \
  10211.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10212.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10213.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10214.   __ret; \
  10215. })
  10216. #endif
  10217.  
  10218. #ifdef __LITTLE_ENDIAN__
  10219. #define vld1q_s64_x3(__p0) __extension__ ({ \
  10220.   int64x2x3_t __ret; \
  10221.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 35); \
  10222.   __ret; \
  10223. })
  10224. #else
  10225. #define vld1q_s64_x3(__p0) __extension__ ({ \
  10226.   int64x2x3_t __ret; \
  10227.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 35); \
  10228.  \
  10229.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10230.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10231.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10232.   __ret; \
  10233. })
  10234. #endif
  10235.  
  10236. #ifdef __LITTLE_ENDIAN__
  10237. #define vld1q_s16_x3(__p0) __extension__ ({ \
  10238.   int16x8x3_t __ret; \
  10239.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 33); \
  10240.   __ret; \
  10241. })
  10242. #else
  10243. #define vld1q_s16_x3(__p0) __extension__ ({ \
  10244.   int16x8x3_t __ret; \
  10245.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 33); \
  10246.  \
  10247.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10248.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10249.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10250.   __ret; \
  10251. })
  10252. #endif
  10253.  
  10254. #ifdef __LITTLE_ENDIAN__
  10255. #define vld1_u8_x3(__p0) __extension__ ({ \
  10256.   uint8x8x3_t __ret; \
  10257.   __builtin_neon_vld1_x3_v(&__ret, __p0, 16); \
  10258.   __ret; \
  10259. })
  10260. #else
  10261. #define vld1_u8_x3(__p0) __extension__ ({ \
  10262.   uint8x8x3_t __ret; \
  10263.   __builtin_neon_vld1_x3_v(&__ret, __p0, 16); \
  10264.  \
  10265.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10266.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10267.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10268.   __ret; \
  10269. })
  10270. #endif
  10271.  
  10272. #ifdef __LITTLE_ENDIAN__
  10273. #define vld1_u32_x3(__p0) __extension__ ({ \
  10274.   uint32x2x3_t __ret; \
  10275.   __builtin_neon_vld1_x3_v(&__ret, __p0, 18); \
  10276.   __ret; \
  10277. })
  10278. #else
  10279. #define vld1_u32_x3(__p0) __extension__ ({ \
  10280.   uint32x2x3_t __ret; \
  10281.   __builtin_neon_vld1_x3_v(&__ret, __p0, 18); \
  10282.  \
  10283.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10284.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10285.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10286.   __ret; \
  10287. })
  10288. #endif
  10289.  
  10290. #define vld1_u64_x3(__p0) __extension__ ({ \
  10291.   uint64x1x3_t __ret; \
  10292.   __builtin_neon_vld1_x3_v(&__ret, __p0, 19); \
  10293.   __ret; \
  10294. })
  10295. #ifdef __LITTLE_ENDIAN__
  10296. #define vld1_u16_x3(__p0) __extension__ ({ \
  10297.   uint16x4x3_t __ret; \
  10298.   __builtin_neon_vld1_x3_v(&__ret, __p0, 17); \
  10299.   __ret; \
  10300. })
  10301. #else
  10302. #define vld1_u16_x3(__p0) __extension__ ({ \
  10303.   uint16x4x3_t __ret; \
  10304.   __builtin_neon_vld1_x3_v(&__ret, __p0, 17); \
  10305.  \
  10306.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10307.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10308.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10309.   __ret; \
  10310. })
  10311. #endif
  10312.  
  10313. #ifdef __LITTLE_ENDIAN__
  10314. #define vld1_s8_x3(__p0) __extension__ ({ \
  10315.   int8x8x3_t __ret; \
  10316.   __builtin_neon_vld1_x3_v(&__ret, __p0, 0); \
  10317.   __ret; \
  10318. })
  10319. #else
  10320. #define vld1_s8_x3(__p0) __extension__ ({ \
  10321.   int8x8x3_t __ret; \
  10322.   __builtin_neon_vld1_x3_v(&__ret, __p0, 0); \
  10323.  \
  10324.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10325.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10326.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10327.   __ret; \
  10328. })
  10329. #endif
  10330.  
  10331. #ifdef __LITTLE_ENDIAN__
  10332. #define vld1_f32_x3(__p0) __extension__ ({ \
  10333.   float32x2x3_t __ret; \
  10334.   __builtin_neon_vld1_x3_v(&__ret, __p0, 9); \
  10335.   __ret; \
  10336. })
  10337. #else
  10338. #define vld1_f32_x3(__p0) __extension__ ({ \
  10339.   float32x2x3_t __ret; \
  10340.   __builtin_neon_vld1_x3_v(&__ret, __p0, 9); \
  10341.  \
  10342.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10343.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10344.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10345.   __ret; \
  10346. })
  10347. #endif
  10348.  
  10349. #ifdef __LITTLE_ENDIAN__
  10350. #define vld1_s32_x3(__p0) __extension__ ({ \
  10351.   int32x2x3_t __ret; \
  10352.   __builtin_neon_vld1_x3_v(&__ret, __p0, 2); \
  10353.   __ret; \
  10354. })
  10355. #else
  10356. #define vld1_s32_x3(__p0) __extension__ ({ \
  10357.   int32x2x3_t __ret; \
  10358.   __builtin_neon_vld1_x3_v(&__ret, __p0, 2); \
  10359.  \
  10360.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10361.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10362.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10363.   __ret; \
  10364. })
  10365. #endif
  10366.  
  10367. #define vld1_s64_x3(__p0) __extension__ ({ \
  10368.   int64x1x3_t __ret; \
  10369.   __builtin_neon_vld1_x3_v(&__ret, __p0, 3); \
  10370.   __ret; \
  10371. })
  10372. #ifdef __LITTLE_ENDIAN__
  10373. #define vld1_s16_x3(__p0) __extension__ ({ \
  10374.   int16x4x3_t __ret; \
  10375.   __builtin_neon_vld1_x3_v(&__ret, __p0, 1); \
  10376.   __ret; \
  10377. })
  10378. #else
  10379. #define vld1_s16_x3(__p0) __extension__ ({ \
  10380.   int16x4x3_t __ret; \
  10381.   __builtin_neon_vld1_x3_v(&__ret, __p0, 1); \
  10382.  \
  10383.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10384.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10385.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10386.   __ret; \
  10387. })
  10388. #endif
  10389.  
  10390. #ifdef __LITTLE_ENDIAN__
  10391. #define vld1_p8_x4(__p0) __extension__ ({ \
  10392.   poly8x8x4_t __ret; \
  10393.   __builtin_neon_vld1_x4_v(&__ret, __p0, 4); \
  10394.   __ret; \
  10395. })
  10396. #else
  10397. #define vld1_p8_x4(__p0) __extension__ ({ \
  10398.   poly8x8x4_t __ret; \
  10399.   __builtin_neon_vld1_x4_v(&__ret, __p0, 4); \
  10400.  \
  10401.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10402.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10403.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10404.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  10405.   __ret; \
  10406. })
  10407. #endif
  10408.  
  10409. #ifdef __LITTLE_ENDIAN__
  10410. #define vld1_p16_x4(__p0) __extension__ ({ \
  10411.   poly16x4x4_t __ret; \
  10412.   __builtin_neon_vld1_x4_v(&__ret, __p0, 5); \
  10413.   __ret; \
  10414. })
  10415. #else
  10416. #define vld1_p16_x4(__p0) __extension__ ({ \
  10417.   poly16x4x4_t __ret; \
  10418.   __builtin_neon_vld1_x4_v(&__ret, __p0, 5); \
  10419.  \
  10420.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10421.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10422.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10423.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  10424.   __ret; \
  10425. })
  10426. #endif
  10427.  
  10428. #ifdef __LITTLE_ENDIAN__
  10429. #define vld1q_p8_x4(__p0) __extension__ ({ \
  10430.   poly8x16x4_t __ret; \
  10431.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 36); \
  10432.   __ret; \
  10433. })
  10434. #else
  10435. #define vld1q_p8_x4(__p0) __extension__ ({ \
  10436.   poly8x16x4_t __ret; \
  10437.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 36); \
  10438.  \
  10439.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10440.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10441.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10442.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10443.   __ret; \
  10444. })
  10445. #endif
  10446.  
  10447. #ifdef __LITTLE_ENDIAN__
  10448. #define vld1q_p16_x4(__p0) __extension__ ({ \
  10449.   poly16x8x4_t __ret; \
  10450.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 37); \
  10451.   __ret; \
  10452. })
  10453. #else
  10454. #define vld1q_p16_x4(__p0) __extension__ ({ \
  10455.   poly16x8x4_t __ret; \
  10456.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 37); \
  10457.  \
  10458.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10459.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10460.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10461.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  10462.   __ret; \
  10463. })
  10464. #endif
  10465.  
  10466. #ifdef __LITTLE_ENDIAN__
  10467. #define vld1q_u8_x4(__p0) __extension__ ({ \
  10468.   uint8x16x4_t __ret; \
  10469.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 48); \
  10470.   __ret; \
  10471. })
  10472. #else
  10473. #define vld1q_u8_x4(__p0) __extension__ ({ \
  10474.   uint8x16x4_t __ret; \
  10475.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 48); \
  10476.  \
  10477.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10478.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10479.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10480.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10481.   __ret; \
  10482. })
  10483. #endif
  10484.  
  10485. #ifdef __LITTLE_ENDIAN__
  10486. #define vld1q_u32_x4(__p0) __extension__ ({ \
  10487.   uint32x4x4_t __ret; \
  10488.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 50); \
  10489.   __ret; \
  10490. })
  10491. #else
  10492. #define vld1q_u32_x4(__p0) __extension__ ({ \
  10493.   uint32x4x4_t __ret; \
  10494.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 50); \
  10495.  \
  10496.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10497.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10498.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10499.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  10500.   __ret; \
  10501. })
  10502. #endif
  10503.  
  10504. #ifdef __LITTLE_ENDIAN__
  10505. #define vld1q_u64_x4(__p0) __extension__ ({ \
  10506.   uint64x2x4_t __ret; \
  10507.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 51); \
  10508.   __ret; \
  10509. })
  10510. #else
  10511. #define vld1q_u64_x4(__p0) __extension__ ({ \
  10512.   uint64x2x4_t __ret; \
  10513.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 51); \
  10514.  \
  10515.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10516.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10517.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10518.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  10519.   __ret; \
  10520. })
  10521. #endif
  10522.  
  10523. #ifdef __LITTLE_ENDIAN__
  10524. #define vld1q_u16_x4(__p0) __extension__ ({ \
  10525.   uint16x8x4_t __ret; \
  10526.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 49); \
  10527.   __ret; \
  10528. })
  10529. #else
  10530. #define vld1q_u16_x4(__p0) __extension__ ({ \
  10531.   uint16x8x4_t __ret; \
  10532.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 49); \
  10533.  \
  10534.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10535.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10536.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10537.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  10538.   __ret; \
  10539. })
  10540. #endif
  10541.  
  10542. #ifdef __LITTLE_ENDIAN__
  10543. #define vld1q_s8_x4(__p0) __extension__ ({ \
  10544.   int8x16x4_t __ret; \
  10545.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 32); \
  10546.   __ret; \
  10547. })
  10548. #else
  10549. #define vld1q_s8_x4(__p0) __extension__ ({ \
  10550.   int8x16x4_t __ret; \
  10551.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 32); \
  10552.  \
  10553.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10554.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10555.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10556.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10557.   __ret; \
  10558. })
  10559. #endif
  10560.  
  10561. #ifdef __LITTLE_ENDIAN__
  10562. #define vld1q_f32_x4(__p0) __extension__ ({ \
  10563.   float32x4x4_t __ret; \
  10564.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 41); \
  10565.   __ret; \
  10566. })
  10567. #else
  10568. #define vld1q_f32_x4(__p0) __extension__ ({ \
  10569.   float32x4x4_t __ret; \
  10570.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 41); \
  10571.  \
  10572.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10573.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10574.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10575.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  10576.   __ret; \
  10577. })
  10578. #endif
  10579.  
  10580. #ifdef __LITTLE_ENDIAN__
  10581. #define vld1q_s32_x4(__p0) __extension__ ({ \
  10582.   int32x4x4_t __ret; \
  10583.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 34); \
  10584.   __ret; \
  10585. })
  10586. #else
  10587. #define vld1q_s32_x4(__p0) __extension__ ({ \
  10588.   int32x4x4_t __ret; \
  10589.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 34); \
  10590.  \
  10591.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10592.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10593.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10594.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  10595.   __ret; \
  10596. })
  10597. #endif
  10598.  
  10599. #ifdef __LITTLE_ENDIAN__
  10600. #define vld1q_s64_x4(__p0) __extension__ ({ \
  10601.   int64x2x4_t __ret; \
  10602.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 35); \
  10603.   __ret; \
  10604. })
  10605. #else
  10606. #define vld1q_s64_x4(__p0) __extension__ ({ \
  10607.   int64x2x4_t __ret; \
  10608.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 35); \
  10609.  \
  10610.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10611.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10612.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10613.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  10614.   __ret; \
  10615. })
  10616. #endif
  10617.  
  10618. #ifdef __LITTLE_ENDIAN__
  10619. #define vld1q_s16_x4(__p0) __extension__ ({ \
  10620.   int16x8x4_t __ret; \
  10621.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 33); \
  10622.   __ret; \
  10623. })
  10624. #else
  10625. #define vld1q_s16_x4(__p0) __extension__ ({ \
  10626.   int16x8x4_t __ret; \
  10627.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 33); \
  10628.  \
  10629.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10630.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10631.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10632.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  10633.   __ret; \
  10634. })
  10635. #endif
  10636.  
  10637. #ifdef __LITTLE_ENDIAN__
  10638. #define vld1_u8_x4(__p0) __extension__ ({ \
  10639.   uint8x8x4_t __ret; \
  10640.   __builtin_neon_vld1_x4_v(&__ret, __p0, 16); \
  10641.   __ret; \
  10642. })
  10643. #else
  10644. #define vld1_u8_x4(__p0) __extension__ ({ \
  10645.   uint8x8x4_t __ret; \
  10646.   __builtin_neon_vld1_x4_v(&__ret, __p0, 16); \
  10647.  \
  10648.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10649.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10650.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10651.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  10652.   __ret; \
  10653. })
  10654. #endif
  10655.  
  10656. #ifdef __LITTLE_ENDIAN__
  10657. #define vld1_u32_x4(__p0) __extension__ ({ \
  10658.   uint32x2x4_t __ret; \
  10659.   __builtin_neon_vld1_x4_v(&__ret, __p0, 18); \
  10660.   __ret; \
  10661. })
  10662. #else
  10663. #define vld1_u32_x4(__p0) __extension__ ({ \
  10664.   uint32x2x4_t __ret; \
  10665.   __builtin_neon_vld1_x4_v(&__ret, __p0, 18); \
  10666.  \
  10667.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10668.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10669.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10670.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  10671.   __ret; \
  10672. })
  10673. #endif
  10674.  
  10675. #define vld1_u64_x4(__p0) __extension__ ({ \
  10676.   uint64x1x4_t __ret; \
  10677.   __builtin_neon_vld1_x4_v(&__ret, __p0, 19); \
  10678.   __ret; \
  10679. })
  10680. #ifdef __LITTLE_ENDIAN__
  10681. #define vld1_u16_x4(__p0) __extension__ ({ \
  10682.   uint16x4x4_t __ret; \
  10683.   __builtin_neon_vld1_x4_v(&__ret, __p0, 17); \
  10684.   __ret; \
  10685. })
  10686. #else
  10687. #define vld1_u16_x4(__p0) __extension__ ({ \
  10688.   uint16x4x4_t __ret; \
  10689.   __builtin_neon_vld1_x4_v(&__ret, __p0, 17); \
  10690.  \
  10691.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10692.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10693.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10694.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  10695.   __ret; \
  10696. })
  10697. #endif
  10698.  
  10699. #ifdef __LITTLE_ENDIAN__
  10700. #define vld1_s8_x4(__p0) __extension__ ({ \
  10701.   int8x8x4_t __ret; \
  10702.   __builtin_neon_vld1_x4_v(&__ret, __p0, 0); \
  10703.   __ret; \
  10704. })
  10705. #else
  10706. #define vld1_s8_x4(__p0) __extension__ ({ \
  10707.   int8x8x4_t __ret; \
  10708.   __builtin_neon_vld1_x4_v(&__ret, __p0, 0); \
  10709.  \
  10710.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10711.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10712.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  10713.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  10714.   __ret; \
  10715. })
  10716. #endif
  10717.  
  10718. #ifdef __LITTLE_ENDIAN__
  10719. #define vld1_f32_x4(__p0) __extension__ ({ \
  10720.   float32x2x4_t __ret; \
  10721.   __builtin_neon_vld1_x4_v(&__ret, __p0, 9); \
  10722.   __ret; \
  10723. })
  10724. #else
  10725. #define vld1_f32_x4(__p0) __extension__ ({ \
  10726.   float32x2x4_t __ret; \
  10727.   __builtin_neon_vld1_x4_v(&__ret, __p0, 9); \
  10728.  \
  10729.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10730.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10731.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10732.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  10733.   __ret; \
  10734. })
  10735. #endif
  10736.  
  10737. #ifdef __LITTLE_ENDIAN__
  10738. #define vld1_s32_x4(__p0) __extension__ ({ \
  10739.   int32x2x4_t __ret; \
  10740.   __builtin_neon_vld1_x4_v(&__ret, __p0, 2); \
  10741.   __ret; \
  10742. })
  10743. #else
  10744. #define vld1_s32_x4(__p0) __extension__ ({ \
  10745.   int32x2x4_t __ret; \
  10746.   __builtin_neon_vld1_x4_v(&__ret, __p0, 2); \
  10747.  \
  10748.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10749.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10750.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  10751.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  10752.   __ret; \
  10753. })
  10754. #endif
  10755.  
  10756. #define vld1_s64_x4(__p0) __extension__ ({ \
  10757.   int64x1x4_t __ret; \
  10758.   __builtin_neon_vld1_x4_v(&__ret, __p0, 3); \
  10759.   __ret; \
  10760. })
  10761. #ifdef __LITTLE_ENDIAN__
  10762. #define vld1_s16_x4(__p0) __extension__ ({ \
  10763.   int16x4x4_t __ret; \
  10764.   __builtin_neon_vld1_x4_v(&__ret, __p0, 1); \
  10765.   __ret; \
  10766. })
  10767. #else
  10768. #define vld1_s16_x4(__p0) __extension__ ({ \
  10769.   int16x4x4_t __ret; \
  10770.   __builtin_neon_vld1_x4_v(&__ret, __p0, 1); \
  10771.  \
  10772.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10773.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10774.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  10775.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  10776.   __ret; \
  10777. })
  10778. #endif
  10779.  
  10780. #ifdef __LITTLE_ENDIAN__
  10781. #define vld2_p8(__p0) __extension__ ({ \
  10782.   poly8x8x2_t __ret; \
  10783.   __builtin_neon_vld2_v(&__ret, __p0, 4); \
  10784.   __ret; \
  10785. })
  10786. #else
  10787. #define vld2_p8(__p0) __extension__ ({ \
  10788.   poly8x8x2_t __ret; \
  10789.   __builtin_neon_vld2_v(&__ret, __p0, 4); \
  10790.  \
  10791.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10792.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10793.   __ret; \
  10794. })
  10795. #endif
  10796.  
  10797. #ifdef __LITTLE_ENDIAN__
  10798. #define vld2_p16(__p0) __extension__ ({ \
  10799.   poly16x4x2_t __ret; \
  10800.   __builtin_neon_vld2_v(&__ret, __p0, 5); \
  10801.   __ret; \
  10802. })
  10803. #else
  10804. #define vld2_p16(__p0) __extension__ ({ \
  10805.   poly16x4x2_t __ret; \
  10806.   __builtin_neon_vld2_v(&__ret, __p0, 5); \
  10807.  \
  10808.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10809.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10810.   __ret; \
  10811. })
  10812. #endif
  10813.  
  10814. #ifdef __LITTLE_ENDIAN__
  10815. #define vld2q_p8(__p0) __extension__ ({ \
  10816.   poly8x16x2_t __ret; \
  10817.   __builtin_neon_vld2q_v(&__ret, __p0, 36); \
  10818.   __ret; \
  10819. })
  10820. #else
  10821. #define vld2q_p8(__p0) __extension__ ({ \
  10822.   poly8x16x2_t __ret; \
  10823.   __builtin_neon_vld2q_v(&__ret, __p0, 36); \
  10824.  \
  10825.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10826.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10827.   __ret; \
  10828. })
  10829. #endif
  10830.  
  10831. #ifdef __LITTLE_ENDIAN__
  10832. #define vld2q_p16(__p0) __extension__ ({ \
  10833.   poly16x8x2_t __ret; \
  10834.   __builtin_neon_vld2q_v(&__ret, __p0, 37); \
  10835.   __ret; \
  10836. })
  10837. #else
  10838. #define vld2q_p16(__p0) __extension__ ({ \
  10839.   poly16x8x2_t __ret; \
  10840.   __builtin_neon_vld2q_v(&__ret, __p0, 37); \
  10841.  \
  10842.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10843.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10844.   __ret; \
  10845. })
  10846. #endif
  10847.  
  10848. #ifdef __LITTLE_ENDIAN__
  10849. #define vld2q_u8(__p0) __extension__ ({ \
  10850.   uint8x16x2_t __ret; \
  10851.   __builtin_neon_vld2q_v(&__ret, __p0, 48); \
  10852.   __ret; \
  10853. })
  10854. #else
  10855. #define vld2q_u8(__p0) __extension__ ({ \
  10856.   uint8x16x2_t __ret; \
  10857.   __builtin_neon_vld2q_v(&__ret, __p0, 48); \
  10858.  \
  10859.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10860.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10861.   __ret; \
  10862. })
  10863. #endif
  10864.  
  10865. #ifdef __LITTLE_ENDIAN__
  10866. #define vld2q_u32(__p0) __extension__ ({ \
  10867.   uint32x4x2_t __ret; \
  10868.   __builtin_neon_vld2q_v(&__ret, __p0, 50); \
  10869.   __ret; \
  10870. })
  10871. #else
  10872. #define vld2q_u32(__p0) __extension__ ({ \
  10873.   uint32x4x2_t __ret; \
  10874.   __builtin_neon_vld2q_v(&__ret, __p0, 50); \
  10875.  \
  10876.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10877.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10878.   __ret; \
  10879. })
  10880. #endif
  10881.  
  10882. #ifdef __LITTLE_ENDIAN__
  10883. #define vld2q_u16(__p0) __extension__ ({ \
  10884.   uint16x8x2_t __ret; \
  10885.   __builtin_neon_vld2q_v(&__ret, __p0, 49); \
  10886.   __ret; \
  10887. })
  10888. #else
  10889. #define vld2q_u16(__p0) __extension__ ({ \
  10890.   uint16x8x2_t __ret; \
  10891.   __builtin_neon_vld2q_v(&__ret, __p0, 49); \
  10892.  \
  10893.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10894.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10895.   __ret; \
  10896. })
  10897. #endif
  10898.  
  10899. #ifdef __LITTLE_ENDIAN__
  10900. #define vld2q_s8(__p0) __extension__ ({ \
  10901.   int8x16x2_t __ret; \
  10902.   __builtin_neon_vld2q_v(&__ret, __p0, 32); \
  10903.   __ret; \
  10904. })
  10905. #else
  10906. #define vld2q_s8(__p0) __extension__ ({ \
  10907.   int8x16x2_t __ret; \
  10908.   __builtin_neon_vld2q_v(&__ret, __p0, 32); \
  10909.  \
  10910.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10911.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  10912.   __ret; \
  10913. })
  10914. #endif
  10915.  
  10916. #ifdef __LITTLE_ENDIAN__
  10917. #define vld2q_f32(__p0) __extension__ ({ \
  10918.   float32x4x2_t __ret; \
  10919.   __builtin_neon_vld2q_v(&__ret, __p0, 41); \
  10920.   __ret; \
  10921. })
  10922. #else
  10923. #define vld2q_f32(__p0) __extension__ ({ \
  10924.   float32x4x2_t __ret; \
  10925.   __builtin_neon_vld2q_v(&__ret, __p0, 41); \
  10926.  \
  10927.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10928.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10929.   __ret; \
  10930. })
  10931. #endif
  10932.  
  10933. #ifdef __LITTLE_ENDIAN__
  10934. #define vld2q_s32(__p0) __extension__ ({ \
  10935.   int32x4x2_t __ret; \
  10936.   __builtin_neon_vld2q_v(&__ret, __p0, 34); \
  10937.   __ret; \
  10938. })
  10939. #else
  10940. #define vld2q_s32(__p0) __extension__ ({ \
  10941.   int32x4x2_t __ret; \
  10942.   __builtin_neon_vld2q_v(&__ret, __p0, 34); \
  10943.  \
  10944.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  10945.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  10946.   __ret; \
  10947. })
  10948. #endif
  10949.  
  10950. #ifdef __LITTLE_ENDIAN__
  10951. #define vld2q_s16(__p0) __extension__ ({ \
  10952.   int16x8x2_t __ret; \
  10953.   __builtin_neon_vld2q_v(&__ret, __p0, 33); \
  10954.   __ret; \
  10955. })
  10956. #else
  10957. #define vld2q_s16(__p0) __extension__ ({ \
  10958.   int16x8x2_t __ret; \
  10959.   __builtin_neon_vld2q_v(&__ret, __p0, 33); \
  10960.  \
  10961.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10962.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10963.   __ret; \
  10964. })
  10965. #endif
  10966.  
  10967. #ifdef __LITTLE_ENDIAN__
  10968. #define vld2_u8(__p0) __extension__ ({ \
  10969.   uint8x8x2_t __ret; \
  10970.   __builtin_neon_vld2_v(&__ret, __p0, 16); \
  10971.   __ret; \
  10972. })
  10973. #else
  10974. #define vld2_u8(__p0) __extension__ ({ \
  10975.   uint8x8x2_t __ret; \
  10976.   __builtin_neon_vld2_v(&__ret, __p0, 16); \
  10977.  \
  10978.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  10979.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  10980.   __ret; \
  10981. })
  10982. #endif
  10983.  
  10984. #ifdef __LITTLE_ENDIAN__
  10985. #define vld2_u32(__p0) __extension__ ({ \
  10986.   uint32x2x2_t __ret; \
  10987.   __builtin_neon_vld2_v(&__ret, __p0, 18); \
  10988.   __ret; \
  10989. })
  10990. #else
  10991. #define vld2_u32(__p0) __extension__ ({ \
  10992.   uint32x2x2_t __ret; \
  10993.   __builtin_neon_vld2_v(&__ret, __p0, 18); \
  10994.  \
  10995.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  10996.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  10997.   __ret; \
  10998. })
  10999. #endif
  11000.  
  11001. #define vld2_u64(__p0) __extension__ ({ \
  11002.   uint64x1x2_t __ret; \
  11003.   __builtin_neon_vld2_v(&__ret, __p0, 19); \
  11004.   __ret; \
  11005. })
  11006. #ifdef __LITTLE_ENDIAN__
  11007. #define vld2_u16(__p0) __extension__ ({ \
  11008.   uint16x4x2_t __ret; \
  11009.   __builtin_neon_vld2_v(&__ret, __p0, 17); \
  11010.   __ret; \
  11011. })
  11012. #else
  11013. #define vld2_u16(__p0) __extension__ ({ \
  11014.   uint16x4x2_t __ret; \
  11015.   __builtin_neon_vld2_v(&__ret, __p0, 17); \
  11016.  \
  11017.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11018.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11019.   __ret; \
  11020. })
  11021. #endif
  11022.  
  11023. #ifdef __LITTLE_ENDIAN__
  11024. #define vld2_s8(__p0) __extension__ ({ \
  11025.   int8x8x2_t __ret; \
  11026.   __builtin_neon_vld2_v(&__ret, __p0, 0); \
  11027.   __ret; \
  11028. })
  11029. #else
  11030. #define vld2_s8(__p0) __extension__ ({ \
  11031.   int8x8x2_t __ret; \
  11032.   __builtin_neon_vld2_v(&__ret, __p0, 0); \
  11033.  \
  11034.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11035.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11036.   __ret; \
  11037. })
  11038. #endif
  11039.  
  11040. #ifdef __LITTLE_ENDIAN__
  11041. #define vld2_f32(__p0) __extension__ ({ \
  11042.   float32x2x2_t __ret; \
  11043.   __builtin_neon_vld2_v(&__ret, __p0, 9); \
  11044.   __ret; \
  11045. })
  11046. #else
  11047. #define vld2_f32(__p0) __extension__ ({ \
  11048.   float32x2x2_t __ret; \
  11049.   __builtin_neon_vld2_v(&__ret, __p0, 9); \
  11050.  \
  11051.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11052.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11053.   __ret; \
  11054. })
  11055. #endif
  11056.  
  11057. #ifdef __LITTLE_ENDIAN__
  11058. #define vld2_s32(__p0) __extension__ ({ \
  11059.   int32x2x2_t __ret; \
  11060.   __builtin_neon_vld2_v(&__ret, __p0, 2); \
  11061.   __ret; \
  11062. })
  11063. #else
  11064. #define vld2_s32(__p0) __extension__ ({ \
  11065.   int32x2x2_t __ret; \
  11066.   __builtin_neon_vld2_v(&__ret, __p0, 2); \
  11067.  \
  11068.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11069.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11070.   __ret; \
  11071. })
  11072. #endif
  11073.  
  11074. #define vld2_s64(__p0) __extension__ ({ \
  11075.   int64x1x2_t __ret; \
  11076.   __builtin_neon_vld2_v(&__ret, __p0, 3); \
  11077.   __ret; \
  11078. })
  11079. #ifdef __LITTLE_ENDIAN__
  11080. #define vld2_s16(__p0) __extension__ ({ \
  11081.   int16x4x2_t __ret; \
  11082.   __builtin_neon_vld2_v(&__ret, __p0, 1); \
  11083.   __ret; \
  11084. })
  11085. #else
  11086. #define vld2_s16(__p0) __extension__ ({ \
  11087.   int16x4x2_t __ret; \
  11088.   __builtin_neon_vld2_v(&__ret, __p0, 1); \
  11089.  \
  11090.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11091.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11092.   __ret; \
  11093. })
  11094. #endif
  11095.  
  11096. #ifdef __LITTLE_ENDIAN__
  11097. #define vld2_dup_p8(__p0) __extension__ ({ \
  11098.   poly8x8x2_t __ret; \
  11099.   __builtin_neon_vld2_dup_v(&__ret, __p0, 4); \
  11100.   __ret; \
  11101. })
  11102. #else
  11103. #define vld2_dup_p8(__p0) __extension__ ({ \
  11104.   poly8x8x2_t __ret; \
  11105.   __builtin_neon_vld2_dup_v(&__ret, __p0, 4); \
  11106.  \
  11107.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11108.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11109.   __ret; \
  11110. })
  11111. #endif
  11112.  
  11113. #ifdef __LITTLE_ENDIAN__
  11114. #define vld2_dup_p16(__p0) __extension__ ({ \
  11115.   poly16x4x2_t __ret; \
  11116.   __builtin_neon_vld2_dup_v(&__ret, __p0, 5); \
  11117.   __ret; \
  11118. })
  11119. #else
  11120. #define vld2_dup_p16(__p0) __extension__ ({ \
  11121.   poly16x4x2_t __ret; \
  11122.   __builtin_neon_vld2_dup_v(&__ret, __p0, 5); \
  11123.  \
  11124.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11125.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11126.   __ret; \
  11127. })
  11128. #endif
  11129.  
  11130. #ifdef __LITTLE_ENDIAN__
  11131. #define vld2q_dup_p8(__p0) __extension__ ({ \
  11132.   poly8x16x2_t __ret; \
  11133.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 36); \
  11134.   __ret; \
  11135. })
  11136. #else
  11137. #define vld2q_dup_p8(__p0) __extension__ ({ \
  11138.   poly8x16x2_t __ret; \
  11139.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 36); \
  11140.  \
  11141.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11142.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11143.   __ret; \
  11144. })
  11145. #endif
  11146.  
  11147. #ifdef __LITTLE_ENDIAN__
  11148. #define vld2q_dup_p16(__p0) __extension__ ({ \
  11149.   poly16x8x2_t __ret; \
  11150.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 37); \
  11151.   __ret; \
  11152. })
  11153. #else
  11154. #define vld2q_dup_p16(__p0) __extension__ ({ \
  11155.   poly16x8x2_t __ret; \
  11156.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 37); \
  11157.  \
  11158.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11159.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11160.   __ret; \
  11161. })
  11162. #endif
  11163.  
  11164. #ifdef __LITTLE_ENDIAN__
  11165. #define vld2q_dup_u8(__p0) __extension__ ({ \
  11166.   uint8x16x2_t __ret; \
  11167.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 48); \
  11168.   __ret; \
  11169. })
  11170. #else
  11171. #define vld2q_dup_u8(__p0) __extension__ ({ \
  11172.   uint8x16x2_t __ret; \
  11173.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 48); \
  11174.  \
  11175.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11176.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11177.   __ret; \
  11178. })
  11179. #endif
  11180.  
  11181. #ifdef __LITTLE_ENDIAN__
  11182. #define vld2q_dup_u32(__p0) __extension__ ({ \
  11183.   uint32x4x2_t __ret; \
  11184.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 50); \
  11185.   __ret; \
  11186. })
  11187. #else
  11188. #define vld2q_dup_u32(__p0) __extension__ ({ \
  11189.   uint32x4x2_t __ret; \
  11190.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 50); \
  11191.  \
  11192.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11193.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11194.   __ret; \
  11195. })
  11196. #endif
  11197.  
  11198. #ifdef __LITTLE_ENDIAN__
  11199. #define vld2q_dup_u64(__p0) __extension__ ({ \
  11200.   uint64x2x2_t __ret; \
  11201.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 51); \
  11202.   __ret; \
  11203. })
  11204. #else
  11205. #define vld2q_dup_u64(__p0) __extension__ ({ \
  11206.   uint64x2x2_t __ret; \
  11207.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 51); \
  11208.  \
  11209.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11210.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11211.   __ret; \
  11212. })
  11213. #endif
  11214.  
  11215. #ifdef __LITTLE_ENDIAN__
  11216. #define vld2q_dup_u16(__p0) __extension__ ({ \
  11217.   uint16x8x2_t __ret; \
  11218.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 49); \
  11219.   __ret; \
  11220. })
  11221. #else
  11222. #define vld2q_dup_u16(__p0) __extension__ ({ \
  11223.   uint16x8x2_t __ret; \
  11224.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 49); \
  11225.  \
  11226.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11227.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11228.   __ret; \
  11229. })
  11230. #endif
  11231.  
  11232. #ifdef __LITTLE_ENDIAN__
  11233. #define vld2q_dup_s8(__p0) __extension__ ({ \
  11234.   int8x16x2_t __ret; \
  11235.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 32); \
  11236.   __ret; \
  11237. })
  11238. #else
  11239. #define vld2q_dup_s8(__p0) __extension__ ({ \
  11240.   int8x16x2_t __ret; \
  11241.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 32); \
  11242.  \
  11243.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11244.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11245.   __ret; \
  11246. })
  11247. #endif
  11248.  
  11249. #ifdef __LITTLE_ENDIAN__
  11250. #define vld2q_dup_f32(__p0) __extension__ ({ \
  11251.   float32x4x2_t __ret; \
  11252.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 41); \
  11253.   __ret; \
  11254. })
  11255. #else
  11256. #define vld2q_dup_f32(__p0) __extension__ ({ \
  11257.   float32x4x2_t __ret; \
  11258.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 41); \
  11259.  \
  11260.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11261.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11262.   __ret; \
  11263. })
  11264. #endif
  11265.  
  11266. #ifdef __LITTLE_ENDIAN__
  11267. #define vld2q_dup_s32(__p0) __extension__ ({ \
  11268.   int32x4x2_t __ret; \
  11269.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 34); \
  11270.   __ret; \
  11271. })
  11272. #else
  11273. #define vld2q_dup_s32(__p0) __extension__ ({ \
  11274.   int32x4x2_t __ret; \
  11275.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 34); \
  11276.  \
  11277.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11278.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11279.   __ret; \
  11280. })
  11281. #endif
  11282.  
  11283. #ifdef __LITTLE_ENDIAN__
  11284. #define vld2q_dup_s64(__p0) __extension__ ({ \
  11285.   int64x2x2_t __ret; \
  11286.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 35); \
  11287.   __ret; \
  11288. })
  11289. #else
  11290. #define vld2q_dup_s64(__p0) __extension__ ({ \
  11291.   int64x2x2_t __ret; \
  11292.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 35); \
  11293.  \
  11294.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11295.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11296.   __ret; \
  11297. })
  11298. #endif
  11299.  
  11300. #ifdef __LITTLE_ENDIAN__
  11301. #define vld2q_dup_s16(__p0) __extension__ ({ \
  11302.   int16x8x2_t __ret; \
  11303.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 33); \
  11304.   __ret; \
  11305. })
  11306. #else
  11307. #define vld2q_dup_s16(__p0) __extension__ ({ \
  11308.   int16x8x2_t __ret; \
  11309.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 33); \
  11310.  \
  11311.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11312.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11313.   __ret; \
  11314. })
  11315. #endif
  11316.  
  11317. #ifdef __LITTLE_ENDIAN__
  11318. #define vld2_dup_u8(__p0) __extension__ ({ \
  11319.   uint8x8x2_t __ret; \
  11320.   __builtin_neon_vld2_dup_v(&__ret, __p0, 16); \
  11321.   __ret; \
  11322. })
  11323. #else
  11324. #define vld2_dup_u8(__p0) __extension__ ({ \
  11325.   uint8x8x2_t __ret; \
  11326.   __builtin_neon_vld2_dup_v(&__ret, __p0, 16); \
  11327.  \
  11328.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11329.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11330.   __ret; \
  11331. })
  11332. #endif
  11333.  
  11334. #ifdef __LITTLE_ENDIAN__
  11335. #define vld2_dup_u32(__p0) __extension__ ({ \
  11336.   uint32x2x2_t __ret; \
  11337.   __builtin_neon_vld2_dup_v(&__ret, __p0, 18); \
  11338.   __ret; \
  11339. })
  11340. #else
  11341. #define vld2_dup_u32(__p0) __extension__ ({ \
  11342.   uint32x2x2_t __ret; \
  11343.   __builtin_neon_vld2_dup_v(&__ret, __p0, 18); \
  11344.  \
  11345.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11346.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11347.   __ret; \
  11348. })
  11349. #endif
  11350.  
  11351. #define vld2_dup_u64(__p0) __extension__ ({ \
  11352.   uint64x1x2_t __ret; \
  11353.   __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
  11354.   __ret; \
  11355. })
  11356. #ifdef __LITTLE_ENDIAN__
  11357. #define vld2_dup_u16(__p0) __extension__ ({ \
  11358.   uint16x4x2_t __ret; \
  11359.   __builtin_neon_vld2_dup_v(&__ret, __p0, 17); \
  11360.   __ret; \
  11361. })
  11362. #else
  11363. #define vld2_dup_u16(__p0) __extension__ ({ \
  11364.   uint16x4x2_t __ret; \
  11365.   __builtin_neon_vld2_dup_v(&__ret, __p0, 17); \
  11366.  \
  11367.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11368.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11369.   __ret; \
  11370. })
  11371. #endif
  11372.  
  11373. #ifdef __LITTLE_ENDIAN__
  11374. #define vld2_dup_s8(__p0) __extension__ ({ \
  11375.   int8x8x2_t __ret; \
  11376.   __builtin_neon_vld2_dup_v(&__ret, __p0, 0); \
  11377.   __ret; \
  11378. })
  11379. #else
  11380. #define vld2_dup_s8(__p0) __extension__ ({ \
  11381.   int8x8x2_t __ret; \
  11382.   __builtin_neon_vld2_dup_v(&__ret, __p0, 0); \
  11383.  \
  11384.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11385.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11386.   __ret; \
  11387. })
  11388. #endif
  11389.  
  11390. #ifdef __LITTLE_ENDIAN__
  11391. #define vld2_dup_f32(__p0) __extension__ ({ \
  11392.   float32x2x2_t __ret; \
  11393.   __builtin_neon_vld2_dup_v(&__ret, __p0, 9); \
  11394.   __ret; \
  11395. })
  11396. #else
  11397. #define vld2_dup_f32(__p0) __extension__ ({ \
  11398.   float32x2x2_t __ret; \
  11399.   __builtin_neon_vld2_dup_v(&__ret, __p0, 9); \
  11400.  \
  11401.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11402.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11403.   __ret; \
  11404. })
  11405. #endif
  11406.  
  11407. #ifdef __LITTLE_ENDIAN__
  11408. #define vld2_dup_s32(__p0) __extension__ ({ \
  11409.   int32x2x2_t __ret; \
  11410.   __builtin_neon_vld2_dup_v(&__ret, __p0, 2); \
  11411.   __ret; \
  11412. })
  11413. #else
  11414. #define vld2_dup_s32(__p0) __extension__ ({ \
  11415.   int32x2x2_t __ret; \
  11416.   __builtin_neon_vld2_dup_v(&__ret, __p0, 2); \
  11417.  \
  11418.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11419.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11420.   __ret; \
  11421. })
  11422. #endif
  11423.  
  11424. #define vld2_dup_s64(__p0) __extension__ ({ \
  11425.   int64x1x2_t __ret; \
  11426.   __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
  11427.   __ret; \
  11428. })
  11429. #ifdef __LITTLE_ENDIAN__
  11430. #define vld2_dup_s16(__p0) __extension__ ({ \
  11431.   int16x4x2_t __ret; \
  11432.   __builtin_neon_vld2_dup_v(&__ret, __p0, 1); \
  11433.   __ret; \
  11434. })
  11435. #else
  11436. #define vld2_dup_s16(__p0) __extension__ ({ \
  11437.   int16x4x2_t __ret; \
  11438.   __builtin_neon_vld2_dup_v(&__ret, __p0, 1); \
  11439.  \
  11440.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11441.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11442.   __ret; \
  11443. })
  11444. #endif
  11445.  
  11446. #ifdef __LITTLE_ENDIAN__
  11447. #define vld2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  11448.   poly8x8x2_t __ret; \
  11449.   poly8x8x2_t __s1 = __p1; \
  11450.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 4); \
  11451.   __ret; \
  11452. })
  11453. #else
  11454. #define vld2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  11455.   poly8x8x2_t __ret; \
  11456.   poly8x8x2_t __s1 = __p1; \
  11457.   poly8x8x2_t __rev1; \
  11458.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11459.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11460.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 4); \
  11461.  \
  11462.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11463.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11464.   __ret; \
  11465. })
  11466. #endif
  11467.  
  11468. #ifdef __LITTLE_ENDIAN__
  11469. #define vld2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  11470.   poly16x4x2_t __ret; \
  11471.   poly16x4x2_t __s1 = __p1; \
  11472.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 5); \
  11473.   __ret; \
  11474. })
  11475. #else
  11476. #define vld2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  11477.   poly16x4x2_t __ret; \
  11478.   poly16x4x2_t __s1 = __p1; \
  11479.   poly16x4x2_t __rev1; \
  11480.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  11481.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  11482.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 5); \
  11483.  \
  11484.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11485.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11486.   __ret; \
  11487. })
  11488. #endif
  11489.  
  11490. #ifdef __LITTLE_ENDIAN__
  11491. #define vld2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  11492.   poly16x8x2_t __ret; \
  11493.   poly16x8x2_t __s1 = __p1; \
  11494.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 37); \
  11495.   __ret; \
  11496. })
  11497. #else
  11498. #define vld2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  11499.   poly16x8x2_t __ret; \
  11500.   poly16x8x2_t __s1 = __p1; \
  11501.   poly16x8x2_t __rev1; \
  11502.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11503.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11504.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 37); \
  11505.  \
  11506.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11507.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11508.   __ret; \
  11509. })
  11510. #endif
  11511.  
  11512. #ifdef __LITTLE_ENDIAN__
  11513. #define vld2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  11514.   uint32x4x2_t __ret; \
  11515.   uint32x4x2_t __s1 = __p1; \
  11516.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 50); \
  11517.   __ret; \
  11518. })
  11519. #else
  11520. #define vld2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  11521.   uint32x4x2_t __ret; \
  11522.   uint32x4x2_t __s1 = __p1; \
  11523.   uint32x4x2_t __rev1; \
  11524.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  11525.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  11526.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 50); \
  11527.  \
  11528.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11529.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11530.   __ret; \
  11531. })
  11532. #endif
  11533.  
  11534. #ifdef __LITTLE_ENDIAN__
  11535. #define vld2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  11536.   uint16x8x2_t __ret; \
  11537.   uint16x8x2_t __s1 = __p1; \
  11538.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 49); \
  11539.   __ret; \
  11540. })
  11541. #else
  11542. #define vld2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  11543.   uint16x8x2_t __ret; \
  11544.   uint16x8x2_t __s1 = __p1; \
  11545.   uint16x8x2_t __rev1; \
  11546.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11547.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11548.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 49); \
  11549.  \
  11550.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11551.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11552.   __ret; \
  11553. })
  11554. #endif
  11555.  
  11556. #ifdef __LITTLE_ENDIAN__
  11557. #define vld2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  11558.   float32x4x2_t __ret; \
  11559.   float32x4x2_t __s1 = __p1; \
  11560.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 41); \
  11561.   __ret; \
  11562. })
  11563. #else
  11564. #define vld2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  11565.   float32x4x2_t __ret; \
  11566.   float32x4x2_t __s1 = __p1; \
  11567.   float32x4x2_t __rev1; \
  11568.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  11569.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  11570.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 41); \
  11571.  \
  11572.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11573.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11574.   __ret; \
  11575. })
  11576. #endif
  11577.  
  11578. #ifdef __LITTLE_ENDIAN__
  11579. #define vld2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  11580.   int32x4x2_t __ret; \
  11581.   int32x4x2_t __s1 = __p1; \
  11582.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 34); \
  11583.   __ret; \
  11584. })
  11585. #else
  11586. #define vld2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  11587.   int32x4x2_t __ret; \
  11588.   int32x4x2_t __s1 = __p1; \
  11589.   int32x4x2_t __rev1; \
  11590.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  11591.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  11592.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 34); \
  11593.  \
  11594.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11595.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11596.   __ret; \
  11597. })
  11598. #endif
  11599.  
  11600. #ifdef __LITTLE_ENDIAN__
  11601. #define vld2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  11602.   int16x8x2_t __ret; \
  11603.   int16x8x2_t __s1 = __p1; \
  11604.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 33); \
  11605.   __ret; \
  11606. })
  11607. #else
  11608. #define vld2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  11609.   int16x8x2_t __ret; \
  11610.   int16x8x2_t __s1 = __p1; \
  11611.   int16x8x2_t __rev1; \
  11612.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11613.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11614.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 33); \
  11615.  \
  11616.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11617.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11618.   __ret; \
  11619. })
  11620. #endif
  11621.  
  11622. #ifdef __LITTLE_ENDIAN__
  11623. #define vld2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  11624.   uint8x8x2_t __ret; \
  11625.   uint8x8x2_t __s1 = __p1; \
  11626.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 16); \
  11627.   __ret; \
  11628. })
  11629. #else
  11630. #define vld2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  11631.   uint8x8x2_t __ret; \
  11632.   uint8x8x2_t __s1 = __p1; \
  11633.   uint8x8x2_t __rev1; \
  11634.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11635.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11636.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 16); \
  11637.  \
  11638.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11639.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11640.   __ret; \
  11641. })
  11642. #endif
  11643.  
  11644. #ifdef __LITTLE_ENDIAN__
  11645. #define vld2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  11646.   uint32x2x2_t __ret; \
  11647.   uint32x2x2_t __s1 = __p1; \
  11648.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 18); \
  11649.   __ret; \
  11650. })
  11651. #else
  11652. #define vld2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  11653.   uint32x2x2_t __ret; \
  11654.   uint32x2x2_t __s1 = __p1; \
  11655.   uint32x2x2_t __rev1; \
  11656.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  11657.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  11658.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 18); \
  11659.  \
  11660.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11661.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11662.   __ret; \
  11663. })
  11664. #endif
  11665.  
  11666. #ifdef __LITTLE_ENDIAN__
  11667. #define vld2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  11668.   uint16x4x2_t __ret; \
  11669.   uint16x4x2_t __s1 = __p1; \
  11670.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 17); \
  11671.   __ret; \
  11672. })
  11673. #else
  11674. #define vld2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  11675.   uint16x4x2_t __ret; \
  11676.   uint16x4x2_t __s1 = __p1; \
  11677.   uint16x4x2_t __rev1; \
  11678.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  11679.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  11680.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 17); \
  11681.  \
  11682.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11683.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11684.   __ret; \
  11685. })
  11686. #endif
  11687.  
  11688. #ifdef __LITTLE_ENDIAN__
  11689. #define vld2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  11690.   int8x8x2_t __ret; \
  11691.   int8x8x2_t __s1 = __p1; \
  11692.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 0); \
  11693.   __ret; \
  11694. })
  11695. #else
  11696. #define vld2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  11697.   int8x8x2_t __ret; \
  11698.   int8x8x2_t __s1 = __p1; \
  11699.   int8x8x2_t __rev1; \
  11700.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11701.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11702.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 0); \
  11703.  \
  11704.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11705.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11706.   __ret; \
  11707. })
  11708. #endif
  11709.  
  11710. #ifdef __LITTLE_ENDIAN__
  11711. #define vld2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  11712.   float32x2x2_t __ret; \
  11713.   float32x2x2_t __s1 = __p1; \
  11714.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 9); \
  11715.   __ret; \
  11716. })
  11717. #else
  11718. #define vld2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  11719.   float32x2x2_t __ret; \
  11720.   float32x2x2_t __s1 = __p1; \
  11721.   float32x2x2_t __rev1; \
  11722.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  11723.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  11724.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 9); \
  11725.  \
  11726.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11727.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11728.   __ret; \
  11729. })
  11730. #endif
  11731.  
  11732. #ifdef __LITTLE_ENDIAN__
  11733. #define vld2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  11734.   int32x2x2_t __ret; \
  11735.   int32x2x2_t __s1 = __p1; \
  11736.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 2); \
  11737.   __ret; \
  11738. })
  11739. #else
  11740. #define vld2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  11741.   int32x2x2_t __ret; \
  11742.   int32x2x2_t __s1 = __p1; \
  11743.   int32x2x2_t __rev1; \
  11744.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  11745.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  11746.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 2); \
  11747.  \
  11748.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  11749.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  11750.   __ret; \
  11751. })
  11752. #endif
  11753.  
  11754. #ifdef __LITTLE_ENDIAN__
  11755. #define vld2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  11756.   int16x4x2_t __ret; \
  11757.   int16x4x2_t __s1 = __p1; \
  11758.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 1); \
  11759.   __ret; \
  11760. })
  11761. #else
  11762. #define vld2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  11763.   int16x4x2_t __ret; \
  11764.   int16x4x2_t __s1 = __p1; \
  11765.   int16x4x2_t __rev1; \
  11766.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  11767.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  11768.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 1); \
  11769.  \
  11770.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11771.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11772.   __ret; \
  11773. })
  11774. #endif
  11775.  
  11776. #ifdef __LITTLE_ENDIAN__
  11777. #define vld3_p8(__p0) __extension__ ({ \
  11778.   poly8x8x3_t __ret; \
  11779.   __builtin_neon_vld3_v(&__ret, __p0, 4); \
  11780.   __ret; \
  11781. })
  11782. #else
  11783. #define vld3_p8(__p0) __extension__ ({ \
  11784.   poly8x8x3_t __ret; \
  11785.   __builtin_neon_vld3_v(&__ret, __p0, 4); \
  11786.  \
  11787.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11788.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11789.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  11790.   __ret; \
  11791. })
  11792. #endif
  11793.  
  11794. #ifdef __LITTLE_ENDIAN__
  11795. #define vld3_p16(__p0) __extension__ ({ \
  11796.   poly16x4x3_t __ret; \
  11797.   __builtin_neon_vld3_v(&__ret, __p0, 5); \
  11798.   __ret; \
  11799. })
  11800. #else
  11801. #define vld3_p16(__p0) __extension__ ({ \
  11802.   poly16x4x3_t __ret; \
  11803.   __builtin_neon_vld3_v(&__ret, __p0, 5); \
  11804.  \
  11805.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11806.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11807.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  11808.   __ret; \
  11809. })
  11810. #endif
  11811.  
  11812. #ifdef __LITTLE_ENDIAN__
  11813. #define vld3q_p8(__p0) __extension__ ({ \
  11814.   poly8x16x3_t __ret; \
  11815.   __builtin_neon_vld3q_v(&__ret, __p0, 36); \
  11816.   __ret; \
  11817. })
  11818. #else
  11819. #define vld3q_p8(__p0) __extension__ ({ \
  11820.   poly8x16x3_t __ret; \
  11821.   __builtin_neon_vld3q_v(&__ret, __p0, 36); \
  11822.  \
  11823.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11824.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11825.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11826.   __ret; \
  11827. })
  11828. #endif
  11829.  
  11830. #ifdef __LITTLE_ENDIAN__
  11831. #define vld3q_p16(__p0) __extension__ ({ \
  11832.   poly16x8x3_t __ret; \
  11833.   __builtin_neon_vld3q_v(&__ret, __p0, 37); \
  11834.   __ret; \
  11835. })
  11836. #else
  11837. #define vld3q_p16(__p0) __extension__ ({ \
  11838.   poly16x8x3_t __ret; \
  11839.   __builtin_neon_vld3q_v(&__ret, __p0, 37); \
  11840.  \
  11841.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11842.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11843.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  11844.   __ret; \
  11845. })
  11846. #endif
  11847.  
  11848. #ifdef __LITTLE_ENDIAN__
  11849. #define vld3q_u8(__p0) __extension__ ({ \
  11850.   uint8x16x3_t __ret; \
  11851.   __builtin_neon_vld3q_v(&__ret, __p0, 48); \
  11852.   __ret; \
  11853. })
  11854. #else
  11855. #define vld3q_u8(__p0) __extension__ ({ \
  11856.   uint8x16x3_t __ret; \
  11857.   __builtin_neon_vld3q_v(&__ret, __p0, 48); \
  11858.  \
  11859.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11860.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11861.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11862.   __ret; \
  11863. })
  11864. #endif
  11865.  
  11866. #ifdef __LITTLE_ENDIAN__
  11867. #define vld3q_u32(__p0) __extension__ ({ \
  11868.   uint32x4x3_t __ret; \
  11869.   __builtin_neon_vld3q_v(&__ret, __p0, 50); \
  11870.   __ret; \
  11871. })
  11872. #else
  11873. #define vld3q_u32(__p0) __extension__ ({ \
  11874.   uint32x4x3_t __ret; \
  11875.   __builtin_neon_vld3q_v(&__ret, __p0, 50); \
  11876.  \
  11877.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11878.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11879.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  11880.   __ret; \
  11881. })
  11882. #endif
  11883.  
  11884. #ifdef __LITTLE_ENDIAN__
  11885. #define vld3q_u16(__p0) __extension__ ({ \
  11886.   uint16x8x3_t __ret; \
  11887.   __builtin_neon_vld3q_v(&__ret, __p0, 49); \
  11888.   __ret; \
  11889. })
  11890. #else
  11891. #define vld3q_u16(__p0) __extension__ ({ \
  11892.   uint16x8x3_t __ret; \
  11893.   __builtin_neon_vld3q_v(&__ret, __p0, 49); \
  11894.  \
  11895.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11896.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11897.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  11898.   __ret; \
  11899. })
  11900. #endif
  11901.  
  11902. #ifdef __LITTLE_ENDIAN__
  11903. #define vld3q_s8(__p0) __extension__ ({ \
  11904.   int8x16x3_t __ret; \
  11905.   __builtin_neon_vld3q_v(&__ret, __p0, 32); \
  11906.   __ret; \
  11907. })
  11908. #else
  11909. #define vld3q_s8(__p0) __extension__ ({ \
  11910.   int8x16x3_t __ret; \
  11911.   __builtin_neon_vld3q_v(&__ret, __p0, 32); \
  11912.  \
  11913.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11914.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11915.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  11916.   __ret; \
  11917. })
  11918. #endif
  11919.  
  11920. #ifdef __LITTLE_ENDIAN__
  11921. #define vld3q_f32(__p0) __extension__ ({ \
  11922.   float32x4x3_t __ret; \
  11923.   __builtin_neon_vld3q_v(&__ret, __p0, 41); \
  11924.   __ret; \
  11925. })
  11926. #else
  11927. #define vld3q_f32(__p0) __extension__ ({ \
  11928.   float32x4x3_t __ret; \
  11929.   __builtin_neon_vld3q_v(&__ret, __p0, 41); \
  11930.  \
  11931.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11932.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11933.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  11934.   __ret; \
  11935. })
  11936. #endif
  11937.  
  11938. #ifdef __LITTLE_ENDIAN__
  11939. #define vld3q_s32(__p0) __extension__ ({ \
  11940.   int32x4x3_t __ret; \
  11941.   __builtin_neon_vld3q_v(&__ret, __p0, 34); \
  11942.   __ret; \
  11943. })
  11944. #else
  11945. #define vld3q_s32(__p0) __extension__ ({ \
  11946.   int32x4x3_t __ret; \
  11947.   __builtin_neon_vld3q_v(&__ret, __p0, 34); \
  11948.  \
  11949.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  11950.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  11951.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  11952.   __ret; \
  11953. })
  11954. #endif
  11955.  
  11956. #ifdef __LITTLE_ENDIAN__
  11957. #define vld3q_s16(__p0) __extension__ ({ \
  11958.   int16x8x3_t __ret; \
  11959.   __builtin_neon_vld3q_v(&__ret, __p0, 33); \
  11960.   __ret; \
  11961. })
  11962. #else
  11963. #define vld3q_s16(__p0) __extension__ ({ \
  11964.   int16x8x3_t __ret; \
  11965.   __builtin_neon_vld3q_v(&__ret, __p0, 33); \
  11966.  \
  11967.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11968.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11969.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  11970.   __ret; \
  11971. })
  11972. #endif
  11973.  
  11974. #ifdef __LITTLE_ENDIAN__
  11975. #define vld3_u8(__p0) __extension__ ({ \
  11976.   uint8x8x3_t __ret; \
  11977.   __builtin_neon_vld3_v(&__ret, __p0, 16); \
  11978.   __ret; \
  11979. })
  11980. #else
  11981. #define vld3_u8(__p0) __extension__ ({ \
  11982.   uint8x8x3_t __ret; \
  11983.   __builtin_neon_vld3_v(&__ret, __p0, 16); \
  11984.  \
  11985.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  11986.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  11987.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  11988.   __ret; \
  11989. })
  11990. #endif
  11991.  
  11992. #ifdef __LITTLE_ENDIAN__
  11993. #define vld3_u32(__p0) __extension__ ({ \
  11994.   uint32x2x3_t __ret; \
  11995.   __builtin_neon_vld3_v(&__ret, __p0, 18); \
  11996.   __ret; \
  11997. })
  11998. #else
  11999. #define vld3_u32(__p0) __extension__ ({ \
  12000.   uint32x2x3_t __ret; \
  12001.   __builtin_neon_vld3_v(&__ret, __p0, 18); \
  12002.  \
  12003.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12004.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12005.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12006.   __ret; \
  12007. })
  12008. #endif
  12009.  
  12010. #define vld3_u64(__p0) __extension__ ({ \
  12011.   uint64x1x3_t __ret; \
  12012.   __builtin_neon_vld3_v(&__ret, __p0, 19); \
  12013.   __ret; \
  12014. })
  12015. #ifdef __LITTLE_ENDIAN__
  12016. #define vld3_u16(__p0) __extension__ ({ \
  12017.   uint16x4x3_t __ret; \
  12018.   __builtin_neon_vld3_v(&__ret, __p0, 17); \
  12019.   __ret; \
  12020. })
  12021. #else
  12022. #define vld3_u16(__p0) __extension__ ({ \
  12023.   uint16x4x3_t __ret; \
  12024.   __builtin_neon_vld3_v(&__ret, __p0, 17); \
  12025.  \
  12026.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12027.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12028.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12029.   __ret; \
  12030. })
  12031. #endif
  12032.  
  12033. #ifdef __LITTLE_ENDIAN__
  12034. #define vld3_s8(__p0) __extension__ ({ \
  12035.   int8x8x3_t __ret; \
  12036.   __builtin_neon_vld3_v(&__ret, __p0, 0); \
  12037.   __ret; \
  12038. })
  12039. #else
  12040. #define vld3_s8(__p0) __extension__ ({ \
  12041.   int8x8x3_t __ret; \
  12042.   __builtin_neon_vld3_v(&__ret, __p0, 0); \
  12043.  \
  12044.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12045.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12046.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12047.   __ret; \
  12048. })
  12049. #endif
  12050.  
  12051. #ifdef __LITTLE_ENDIAN__
  12052. #define vld3_f32(__p0) __extension__ ({ \
  12053.   float32x2x3_t __ret; \
  12054.   __builtin_neon_vld3_v(&__ret, __p0, 9); \
  12055.   __ret; \
  12056. })
  12057. #else
  12058. #define vld3_f32(__p0) __extension__ ({ \
  12059.   float32x2x3_t __ret; \
  12060.   __builtin_neon_vld3_v(&__ret, __p0, 9); \
  12061.  \
  12062.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12063.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12064.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12065.   __ret; \
  12066. })
  12067. #endif
  12068.  
  12069. #ifdef __LITTLE_ENDIAN__
  12070. #define vld3_s32(__p0) __extension__ ({ \
  12071.   int32x2x3_t __ret; \
  12072.   __builtin_neon_vld3_v(&__ret, __p0, 2); \
  12073.   __ret; \
  12074. })
  12075. #else
  12076. #define vld3_s32(__p0) __extension__ ({ \
  12077.   int32x2x3_t __ret; \
  12078.   __builtin_neon_vld3_v(&__ret, __p0, 2); \
  12079.  \
  12080.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12081.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12082.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12083.   __ret; \
  12084. })
  12085. #endif
  12086.  
  12087. #define vld3_s64(__p0) __extension__ ({ \
  12088.   int64x1x3_t __ret; \
  12089.   __builtin_neon_vld3_v(&__ret, __p0, 3); \
  12090.   __ret; \
  12091. })
  12092. #ifdef __LITTLE_ENDIAN__
  12093. #define vld3_s16(__p0) __extension__ ({ \
  12094.   int16x4x3_t __ret; \
  12095.   __builtin_neon_vld3_v(&__ret, __p0, 1); \
  12096.   __ret; \
  12097. })
  12098. #else
  12099. #define vld3_s16(__p0) __extension__ ({ \
  12100.   int16x4x3_t __ret; \
  12101.   __builtin_neon_vld3_v(&__ret, __p0, 1); \
  12102.  \
  12103.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12104.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12105.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12106.   __ret; \
  12107. })
  12108. #endif
  12109.  
  12110. #ifdef __LITTLE_ENDIAN__
  12111. #define vld3_dup_p8(__p0) __extension__ ({ \
  12112.   poly8x8x3_t __ret; \
  12113.   __builtin_neon_vld3_dup_v(&__ret, __p0, 4); \
  12114.   __ret; \
  12115. })
  12116. #else
  12117. #define vld3_dup_p8(__p0) __extension__ ({ \
  12118.   poly8x8x3_t __ret; \
  12119.   __builtin_neon_vld3_dup_v(&__ret, __p0, 4); \
  12120.  \
  12121.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12122.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12123.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12124.   __ret; \
  12125. })
  12126. #endif
  12127.  
  12128. #ifdef __LITTLE_ENDIAN__
  12129. #define vld3_dup_p16(__p0) __extension__ ({ \
  12130.   poly16x4x3_t __ret; \
  12131.   __builtin_neon_vld3_dup_v(&__ret, __p0, 5); \
  12132.   __ret; \
  12133. })
  12134. #else
  12135. #define vld3_dup_p16(__p0) __extension__ ({ \
  12136.   poly16x4x3_t __ret; \
  12137.   __builtin_neon_vld3_dup_v(&__ret, __p0, 5); \
  12138.  \
  12139.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12140.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12141.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12142.   __ret; \
  12143. })
  12144. #endif
  12145.  
  12146. #ifdef __LITTLE_ENDIAN__
  12147. #define vld3q_dup_p8(__p0) __extension__ ({ \
  12148.   poly8x16x3_t __ret; \
  12149.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 36); \
  12150.   __ret; \
  12151. })
  12152. #else
  12153. #define vld3q_dup_p8(__p0) __extension__ ({ \
  12154.   poly8x16x3_t __ret; \
  12155.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 36); \
  12156.  \
  12157.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12158.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12159.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12160.   __ret; \
  12161. })
  12162. #endif
  12163.  
  12164. #ifdef __LITTLE_ENDIAN__
  12165. #define vld3q_dup_p16(__p0) __extension__ ({ \
  12166.   poly16x8x3_t __ret; \
  12167.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 37); \
  12168.   __ret; \
  12169. })
  12170. #else
  12171. #define vld3q_dup_p16(__p0) __extension__ ({ \
  12172.   poly16x8x3_t __ret; \
  12173.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 37); \
  12174.  \
  12175.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12176.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12177.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12178.   __ret; \
  12179. })
  12180. #endif
  12181.  
  12182. #ifdef __LITTLE_ENDIAN__
  12183. #define vld3q_dup_u8(__p0) __extension__ ({ \
  12184.   uint8x16x3_t __ret; \
  12185.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 48); \
  12186.   __ret; \
  12187. })
  12188. #else
  12189. #define vld3q_dup_u8(__p0) __extension__ ({ \
  12190.   uint8x16x3_t __ret; \
  12191.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 48); \
  12192.  \
  12193.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12194.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12195.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12196.   __ret; \
  12197. })
  12198. #endif
  12199.  
  12200. #ifdef __LITTLE_ENDIAN__
  12201. #define vld3q_dup_u32(__p0) __extension__ ({ \
  12202.   uint32x4x3_t __ret; \
  12203.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 50); \
  12204.   __ret; \
  12205. })
  12206. #else
  12207. #define vld3q_dup_u32(__p0) __extension__ ({ \
  12208.   uint32x4x3_t __ret; \
  12209.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 50); \
  12210.  \
  12211.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12212.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12213.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12214.   __ret; \
  12215. })
  12216. #endif
  12217.  
  12218. #ifdef __LITTLE_ENDIAN__
  12219. #define vld3q_dup_u64(__p0) __extension__ ({ \
  12220.   uint64x2x3_t __ret; \
  12221.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 51); \
  12222.   __ret; \
  12223. })
  12224. #else
  12225. #define vld3q_dup_u64(__p0) __extension__ ({ \
  12226.   uint64x2x3_t __ret; \
  12227.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 51); \
  12228.  \
  12229.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12230.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12231.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12232.   __ret; \
  12233. })
  12234. #endif
  12235.  
  12236. #ifdef __LITTLE_ENDIAN__
  12237. #define vld3q_dup_u16(__p0) __extension__ ({ \
  12238.   uint16x8x3_t __ret; \
  12239.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 49); \
  12240.   __ret; \
  12241. })
  12242. #else
  12243. #define vld3q_dup_u16(__p0) __extension__ ({ \
  12244.   uint16x8x3_t __ret; \
  12245.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 49); \
  12246.  \
  12247.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12248.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12249.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12250.   __ret; \
  12251. })
  12252. #endif
  12253.  
  12254. #ifdef __LITTLE_ENDIAN__
  12255. #define vld3q_dup_s8(__p0) __extension__ ({ \
  12256.   int8x16x3_t __ret; \
  12257.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 32); \
  12258.   __ret; \
  12259. })
  12260. #else
  12261. #define vld3q_dup_s8(__p0) __extension__ ({ \
  12262.   int8x16x3_t __ret; \
  12263.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 32); \
  12264.  \
  12265.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12266.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12267.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12268.   __ret; \
  12269. })
  12270. #endif
  12271.  
  12272. #ifdef __LITTLE_ENDIAN__
  12273. #define vld3q_dup_f32(__p0) __extension__ ({ \
  12274.   float32x4x3_t __ret; \
  12275.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 41); \
  12276.   __ret; \
  12277. })
  12278. #else
  12279. #define vld3q_dup_f32(__p0) __extension__ ({ \
  12280.   float32x4x3_t __ret; \
  12281.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 41); \
  12282.  \
  12283.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12284.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12285.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12286.   __ret; \
  12287. })
  12288. #endif
  12289.  
  12290. #ifdef __LITTLE_ENDIAN__
  12291. #define vld3q_dup_s32(__p0) __extension__ ({ \
  12292.   int32x4x3_t __ret; \
  12293.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 34); \
  12294.   __ret; \
  12295. })
  12296. #else
  12297. #define vld3q_dup_s32(__p0) __extension__ ({ \
  12298.   int32x4x3_t __ret; \
  12299.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 34); \
  12300.  \
  12301.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12302.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12303.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12304.   __ret; \
  12305. })
  12306. #endif
  12307.  
  12308. #ifdef __LITTLE_ENDIAN__
  12309. #define vld3q_dup_s64(__p0) __extension__ ({ \
  12310.   int64x2x3_t __ret; \
  12311.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 35); \
  12312.   __ret; \
  12313. })
  12314. #else
  12315. #define vld3q_dup_s64(__p0) __extension__ ({ \
  12316.   int64x2x3_t __ret; \
  12317.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 35); \
  12318.  \
  12319.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12320.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12321.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12322.   __ret; \
  12323. })
  12324. #endif
  12325.  
  12326. #ifdef __LITTLE_ENDIAN__
  12327. #define vld3q_dup_s16(__p0) __extension__ ({ \
  12328.   int16x8x3_t __ret; \
  12329.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 33); \
  12330.   __ret; \
  12331. })
  12332. #else
  12333. #define vld3q_dup_s16(__p0) __extension__ ({ \
  12334.   int16x8x3_t __ret; \
  12335.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 33); \
  12336.  \
  12337.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12338.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12339.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12340.   __ret; \
  12341. })
  12342. #endif
  12343.  
  12344. #ifdef __LITTLE_ENDIAN__
  12345. #define vld3_dup_u8(__p0) __extension__ ({ \
  12346.   uint8x8x3_t __ret; \
  12347.   __builtin_neon_vld3_dup_v(&__ret, __p0, 16); \
  12348.   __ret; \
  12349. })
  12350. #else
  12351. #define vld3_dup_u8(__p0) __extension__ ({ \
  12352.   uint8x8x3_t __ret; \
  12353.   __builtin_neon_vld3_dup_v(&__ret, __p0, 16); \
  12354.  \
  12355.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12356.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12357.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12358.   __ret; \
  12359. })
  12360. #endif
  12361.  
  12362. #ifdef __LITTLE_ENDIAN__
  12363. #define vld3_dup_u32(__p0) __extension__ ({ \
  12364.   uint32x2x3_t __ret; \
  12365.   __builtin_neon_vld3_dup_v(&__ret, __p0, 18); \
  12366.   __ret; \
  12367. })
  12368. #else
  12369. #define vld3_dup_u32(__p0) __extension__ ({ \
  12370.   uint32x2x3_t __ret; \
  12371.   __builtin_neon_vld3_dup_v(&__ret, __p0, 18); \
  12372.  \
  12373.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12374.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12375.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12376.   __ret; \
  12377. })
  12378. #endif
  12379.  
  12380. #define vld3_dup_u64(__p0) __extension__ ({ \
  12381.   uint64x1x3_t __ret; \
  12382.   __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
  12383.   __ret; \
  12384. })
  12385. #ifdef __LITTLE_ENDIAN__
  12386. #define vld3_dup_u16(__p0) __extension__ ({ \
  12387.   uint16x4x3_t __ret; \
  12388.   __builtin_neon_vld3_dup_v(&__ret, __p0, 17); \
  12389.   __ret; \
  12390. })
  12391. #else
  12392. #define vld3_dup_u16(__p0) __extension__ ({ \
  12393.   uint16x4x3_t __ret; \
  12394.   __builtin_neon_vld3_dup_v(&__ret, __p0, 17); \
  12395.  \
  12396.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12397.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12398.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12399.   __ret; \
  12400. })
  12401. #endif
  12402.  
  12403. #ifdef __LITTLE_ENDIAN__
  12404. #define vld3_dup_s8(__p0) __extension__ ({ \
  12405.   int8x8x3_t __ret; \
  12406.   __builtin_neon_vld3_dup_v(&__ret, __p0, 0); \
  12407.   __ret; \
  12408. })
  12409. #else
  12410. #define vld3_dup_s8(__p0) __extension__ ({ \
  12411.   int8x8x3_t __ret; \
  12412.   __builtin_neon_vld3_dup_v(&__ret, __p0, 0); \
  12413.  \
  12414.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12415.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12416.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12417.   __ret; \
  12418. })
  12419. #endif
  12420.  
  12421. #ifdef __LITTLE_ENDIAN__
  12422. #define vld3_dup_f32(__p0) __extension__ ({ \
  12423.   float32x2x3_t __ret; \
  12424.   __builtin_neon_vld3_dup_v(&__ret, __p0, 9); \
  12425.   __ret; \
  12426. })
  12427. #else
  12428. #define vld3_dup_f32(__p0) __extension__ ({ \
  12429.   float32x2x3_t __ret; \
  12430.   __builtin_neon_vld3_dup_v(&__ret, __p0, 9); \
  12431.  \
  12432.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12433.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12434.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12435.   __ret; \
  12436. })
  12437. #endif
  12438.  
  12439. #ifdef __LITTLE_ENDIAN__
  12440. #define vld3_dup_s32(__p0) __extension__ ({ \
  12441.   int32x2x3_t __ret; \
  12442.   __builtin_neon_vld3_dup_v(&__ret, __p0, 2); \
  12443.   __ret; \
  12444. })
  12445. #else
  12446. #define vld3_dup_s32(__p0) __extension__ ({ \
  12447.   int32x2x3_t __ret; \
  12448.   __builtin_neon_vld3_dup_v(&__ret, __p0, 2); \
  12449.  \
  12450.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12451.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12452.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12453.   __ret; \
  12454. })
  12455. #endif
  12456.  
  12457. #define vld3_dup_s64(__p0) __extension__ ({ \
  12458.   int64x1x3_t __ret; \
  12459.   __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
  12460.   __ret; \
  12461. })
  12462. #ifdef __LITTLE_ENDIAN__
  12463. #define vld3_dup_s16(__p0) __extension__ ({ \
  12464.   int16x4x3_t __ret; \
  12465.   __builtin_neon_vld3_dup_v(&__ret, __p0, 1); \
  12466.   __ret; \
  12467. })
  12468. #else
  12469. #define vld3_dup_s16(__p0) __extension__ ({ \
  12470.   int16x4x3_t __ret; \
  12471.   __builtin_neon_vld3_dup_v(&__ret, __p0, 1); \
  12472.  \
  12473.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12474.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12475.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12476.   __ret; \
  12477. })
  12478. #endif
  12479.  
  12480. #ifdef __LITTLE_ENDIAN__
  12481. #define vld3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  12482.   poly8x8x3_t __ret; \
  12483.   poly8x8x3_t __s1 = __p1; \
  12484.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 4); \
  12485.   __ret; \
  12486. })
  12487. #else
  12488. #define vld3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  12489.   poly8x8x3_t __ret; \
  12490.   poly8x8x3_t __s1 = __p1; \
  12491.   poly8x8x3_t __rev1; \
  12492.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12493.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12494.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12495.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 4); \
  12496.  \
  12497.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12498.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12499.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12500.   __ret; \
  12501. })
  12502. #endif
  12503.  
  12504. #ifdef __LITTLE_ENDIAN__
  12505. #define vld3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  12506.   poly16x4x3_t __ret; \
  12507.   poly16x4x3_t __s1 = __p1; \
  12508.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 5); \
  12509.   __ret; \
  12510. })
  12511. #else
  12512. #define vld3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  12513.   poly16x4x3_t __ret; \
  12514.   poly16x4x3_t __s1 = __p1; \
  12515.   poly16x4x3_t __rev1; \
  12516.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  12517.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  12518.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  12519.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 5); \
  12520.  \
  12521.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12522.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12523.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12524.   __ret; \
  12525. })
  12526. #endif
  12527.  
  12528. #ifdef __LITTLE_ENDIAN__
  12529. #define vld3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  12530.   poly16x8x3_t __ret; \
  12531.   poly16x8x3_t __s1 = __p1; \
  12532.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 37); \
  12533.   __ret; \
  12534. })
  12535. #else
  12536. #define vld3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  12537.   poly16x8x3_t __ret; \
  12538.   poly16x8x3_t __s1 = __p1; \
  12539.   poly16x8x3_t __rev1; \
  12540.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12541.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12542.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12543.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 37); \
  12544.  \
  12545.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12546.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12547.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12548.   __ret; \
  12549. })
  12550. #endif
  12551.  
  12552. #ifdef __LITTLE_ENDIAN__
  12553. #define vld3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  12554.   uint32x4x3_t __ret; \
  12555.   uint32x4x3_t __s1 = __p1; \
  12556.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 50); \
  12557.   __ret; \
  12558. })
  12559. #else
  12560. #define vld3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  12561.   uint32x4x3_t __ret; \
  12562.   uint32x4x3_t __s1 = __p1; \
  12563.   uint32x4x3_t __rev1; \
  12564.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  12565.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  12566.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  12567.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 50); \
  12568.  \
  12569.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12570.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12571.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12572.   __ret; \
  12573. })
  12574. #endif
  12575.  
  12576. #ifdef __LITTLE_ENDIAN__
  12577. #define vld3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  12578.   uint16x8x3_t __ret; \
  12579.   uint16x8x3_t __s1 = __p1; \
  12580.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 49); \
  12581.   __ret; \
  12582. })
  12583. #else
  12584. #define vld3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  12585.   uint16x8x3_t __ret; \
  12586.   uint16x8x3_t __s1 = __p1; \
  12587.   uint16x8x3_t __rev1; \
  12588.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12589.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12590.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12591.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 49); \
  12592.  \
  12593.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12594.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12595.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12596.   __ret; \
  12597. })
  12598. #endif
  12599.  
  12600. #ifdef __LITTLE_ENDIAN__
  12601. #define vld3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  12602.   float32x4x3_t __ret; \
  12603.   float32x4x3_t __s1 = __p1; \
  12604.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 41); \
  12605.   __ret; \
  12606. })
  12607. #else
  12608. #define vld3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  12609.   float32x4x3_t __ret; \
  12610.   float32x4x3_t __s1 = __p1; \
  12611.   float32x4x3_t __rev1; \
  12612.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  12613.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  12614.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  12615.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 41); \
  12616.  \
  12617.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12618.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12619.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12620.   __ret; \
  12621. })
  12622. #endif
  12623.  
  12624. #ifdef __LITTLE_ENDIAN__
  12625. #define vld3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  12626.   int32x4x3_t __ret; \
  12627.   int32x4x3_t __s1 = __p1; \
  12628.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 34); \
  12629.   __ret; \
  12630. })
  12631. #else
  12632. #define vld3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  12633.   int32x4x3_t __ret; \
  12634.   int32x4x3_t __s1 = __p1; \
  12635.   int32x4x3_t __rev1; \
  12636.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  12637.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  12638.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  12639.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 34); \
  12640.  \
  12641.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12642.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12643.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12644.   __ret; \
  12645. })
  12646. #endif
  12647.  
  12648. #ifdef __LITTLE_ENDIAN__
  12649. #define vld3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  12650.   int16x8x3_t __ret; \
  12651.   int16x8x3_t __s1 = __p1; \
  12652.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 33); \
  12653.   __ret; \
  12654. })
  12655. #else
  12656. #define vld3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  12657.   int16x8x3_t __ret; \
  12658.   int16x8x3_t __s1 = __p1; \
  12659.   int16x8x3_t __rev1; \
  12660.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12661.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12662.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12663.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 33); \
  12664.  \
  12665.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12666.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12667.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12668.   __ret; \
  12669. })
  12670. #endif
  12671.  
  12672. #ifdef __LITTLE_ENDIAN__
  12673. #define vld3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  12674.   uint8x8x3_t __ret; \
  12675.   uint8x8x3_t __s1 = __p1; \
  12676.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 16); \
  12677.   __ret; \
  12678. })
  12679. #else
  12680. #define vld3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  12681.   uint8x8x3_t __ret; \
  12682.   uint8x8x3_t __s1 = __p1; \
  12683.   uint8x8x3_t __rev1; \
  12684.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12685.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12686.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12687.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 16); \
  12688.  \
  12689.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12690.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12691.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12692.   __ret; \
  12693. })
  12694. #endif
  12695.  
  12696. #ifdef __LITTLE_ENDIAN__
  12697. #define vld3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  12698.   uint32x2x3_t __ret; \
  12699.   uint32x2x3_t __s1 = __p1; \
  12700.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 18); \
  12701.   __ret; \
  12702. })
  12703. #else
  12704. #define vld3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  12705.   uint32x2x3_t __ret; \
  12706.   uint32x2x3_t __s1 = __p1; \
  12707.   uint32x2x3_t __rev1; \
  12708.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  12709.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  12710.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  12711.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 18); \
  12712.  \
  12713.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12714.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12715.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12716.   __ret; \
  12717. })
  12718. #endif
  12719.  
  12720. #ifdef __LITTLE_ENDIAN__
  12721. #define vld3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  12722.   uint16x4x3_t __ret; \
  12723.   uint16x4x3_t __s1 = __p1; \
  12724.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 17); \
  12725.   __ret; \
  12726. })
  12727. #else
  12728. #define vld3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  12729.   uint16x4x3_t __ret; \
  12730.   uint16x4x3_t __s1 = __p1; \
  12731.   uint16x4x3_t __rev1; \
  12732.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  12733.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  12734.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  12735.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 17); \
  12736.  \
  12737.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12738.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12739.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12740.   __ret; \
  12741. })
  12742. #endif
  12743.  
  12744. #ifdef __LITTLE_ENDIAN__
  12745. #define vld3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  12746.   int8x8x3_t __ret; \
  12747.   int8x8x3_t __s1 = __p1; \
  12748.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 0); \
  12749.   __ret; \
  12750. })
  12751. #else
  12752. #define vld3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  12753.   int8x8x3_t __ret; \
  12754.   int8x8x3_t __s1 = __p1; \
  12755.   int8x8x3_t __rev1; \
  12756.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12757.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12758.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12759.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 0); \
  12760.  \
  12761.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12762.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12763.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12764.   __ret; \
  12765. })
  12766. #endif
  12767.  
  12768. #ifdef __LITTLE_ENDIAN__
  12769. #define vld3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  12770.   float32x2x3_t __ret; \
  12771.   float32x2x3_t __s1 = __p1; \
  12772.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 9); \
  12773.   __ret; \
  12774. })
  12775. #else
  12776. #define vld3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  12777.   float32x2x3_t __ret; \
  12778.   float32x2x3_t __s1 = __p1; \
  12779.   float32x2x3_t __rev1; \
  12780.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  12781.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  12782.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  12783.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 9); \
  12784.  \
  12785.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12786.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12787.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12788.   __ret; \
  12789. })
  12790. #endif
  12791.  
  12792. #ifdef __LITTLE_ENDIAN__
  12793. #define vld3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  12794.   int32x2x3_t __ret; \
  12795.   int32x2x3_t __s1 = __p1; \
  12796.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 2); \
  12797.   __ret; \
  12798. })
  12799. #else
  12800. #define vld3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  12801.   int32x2x3_t __ret; \
  12802.   int32x2x3_t __s1 = __p1; \
  12803.   int32x2x3_t __rev1; \
  12804.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  12805.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  12806.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  12807.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 2); \
  12808.  \
  12809.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  12810.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  12811.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  12812.   __ret; \
  12813. })
  12814. #endif
  12815.  
  12816. #ifdef __LITTLE_ENDIAN__
  12817. #define vld3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  12818.   int16x4x3_t __ret; \
  12819.   int16x4x3_t __s1 = __p1; \
  12820.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 1); \
  12821.   __ret; \
  12822. })
  12823. #else
  12824. #define vld3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  12825.   int16x4x3_t __ret; \
  12826.   int16x4x3_t __s1 = __p1; \
  12827.   int16x4x3_t __rev1; \
  12828.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  12829.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  12830.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  12831.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 1); \
  12832.  \
  12833.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12834.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12835.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12836.   __ret; \
  12837. })
  12838. #endif
  12839.  
  12840. #ifdef __LITTLE_ENDIAN__
  12841. #define vld4_p8(__p0) __extension__ ({ \
  12842.   poly8x8x4_t __ret; \
  12843.   __builtin_neon_vld4_v(&__ret, __p0, 4); \
  12844.   __ret; \
  12845. })
  12846. #else
  12847. #define vld4_p8(__p0) __extension__ ({ \
  12848.   poly8x8x4_t __ret; \
  12849.   __builtin_neon_vld4_v(&__ret, __p0, 4); \
  12850.  \
  12851.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12852.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12853.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12854.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  12855.   __ret; \
  12856. })
  12857. #endif
  12858.  
  12859. #ifdef __LITTLE_ENDIAN__
  12860. #define vld4_p16(__p0) __extension__ ({ \
  12861.   poly16x4x4_t __ret; \
  12862.   __builtin_neon_vld4_v(&__ret, __p0, 5); \
  12863.   __ret; \
  12864. })
  12865. #else
  12866. #define vld4_p16(__p0) __extension__ ({ \
  12867.   poly16x4x4_t __ret; \
  12868.   __builtin_neon_vld4_v(&__ret, __p0, 5); \
  12869.  \
  12870.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12871.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12872.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12873.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  12874.   __ret; \
  12875. })
  12876. #endif
  12877.  
  12878. #ifdef __LITTLE_ENDIAN__
  12879. #define vld4q_p8(__p0) __extension__ ({ \
  12880.   poly8x16x4_t __ret; \
  12881.   __builtin_neon_vld4q_v(&__ret, __p0, 36); \
  12882.   __ret; \
  12883. })
  12884. #else
  12885. #define vld4q_p8(__p0) __extension__ ({ \
  12886.   poly8x16x4_t __ret; \
  12887.   __builtin_neon_vld4q_v(&__ret, __p0, 36); \
  12888.  \
  12889.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12890.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12891.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12892.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12893.   __ret; \
  12894. })
  12895. #endif
  12896.  
  12897. #ifdef __LITTLE_ENDIAN__
  12898. #define vld4q_p16(__p0) __extension__ ({ \
  12899.   poly16x8x4_t __ret; \
  12900.   __builtin_neon_vld4q_v(&__ret, __p0, 37); \
  12901.   __ret; \
  12902. })
  12903. #else
  12904. #define vld4q_p16(__p0) __extension__ ({ \
  12905.   poly16x8x4_t __ret; \
  12906.   __builtin_neon_vld4q_v(&__ret, __p0, 37); \
  12907.  \
  12908.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12909.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12910.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12911.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  12912.   __ret; \
  12913. })
  12914. #endif
  12915.  
  12916. #ifdef __LITTLE_ENDIAN__
  12917. #define vld4q_u8(__p0) __extension__ ({ \
  12918.   uint8x16x4_t __ret; \
  12919.   __builtin_neon_vld4q_v(&__ret, __p0, 48); \
  12920.   __ret; \
  12921. })
  12922. #else
  12923. #define vld4q_u8(__p0) __extension__ ({ \
  12924.   uint8x16x4_t __ret; \
  12925.   __builtin_neon_vld4q_v(&__ret, __p0, 48); \
  12926.  \
  12927.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12928.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12929.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12930.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12931.   __ret; \
  12932. })
  12933. #endif
  12934.  
  12935. #ifdef __LITTLE_ENDIAN__
  12936. #define vld4q_u32(__p0) __extension__ ({ \
  12937.   uint32x4x4_t __ret; \
  12938.   __builtin_neon_vld4q_v(&__ret, __p0, 50); \
  12939.   __ret; \
  12940. })
  12941. #else
  12942. #define vld4q_u32(__p0) __extension__ ({ \
  12943.   uint32x4x4_t __ret; \
  12944.   __builtin_neon_vld4q_v(&__ret, __p0, 50); \
  12945.  \
  12946.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  12947.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  12948.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  12949.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  12950.   __ret; \
  12951. })
  12952. #endif
  12953.  
  12954. #ifdef __LITTLE_ENDIAN__
  12955. #define vld4q_u16(__p0) __extension__ ({ \
  12956.   uint16x8x4_t __ret; \
  12957.   __builtin_neon_vld4q_v(&__ret, __p0, 49); \
  12958.   __ret; \
  12959. })
  12960. #else
  12961. #define vld4q_u16(__p0) __extension__ ({ \
  12962.   uint16x8x4_t __ret; \
  12963.   __builtin_neon_vld4q_v(&__ret, __p0, 49); \
  12964.  \
  12965.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  12966.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  12967.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  12968.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  12969.   __ret; \
  12970. })
  12971. #endif
  12972.  
  12973. #ifdef __LITTLE_ENDIAN__
  12974. #define vld4q_s8(__p0) __extension__ ({ \
  12975.   int8x16x4_t __ret; \
  12976.   __builtin_neon_vld4q_v(&__ret, __p0, 32); \
  12977.   __ret; \
  12978. })
  12979. #else
  12980. #define vld4q_s8(__p0) __extension__ ({ \
  12981.   int8x16x4_t __ret; \
  12982.   __builtin_neon_vld4q_v(&__ret, __p0, 32); \
  12983.  \
  12984.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12985.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12986.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12987.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  12988.   __ret; \
  12989. })
  12990. #endif
  12991.  
  12992. #ifdef __LITTLE_ENDIAN__
  12993. #define vld4q_f32(__p0) __extension__ ({ \
  12994.   float32x4x4_t __ret; \
  12995.   __builtin_neon_vld4q_v(&__ret, __p0, 41); \
  12996.   __ret; \
  12997. })
  12998. #else
  12999. #define vld4q_f32(__p0) __extension__ ({ \
  13000.   float32x4x4_t __ret; \
  13001.   __builtin_neon_vld4q_v(&__ret, __p0, 41); \
  13002.  \
  13003.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13004.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13005.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13006.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13007.   __ret; \
  13008. })
  13009. #endif
  13010.  
  13011. #ifdef __LITTLE_ENDIAN__
  13012. #define vld4q_s32(__p0) __extension__ ({ \
  13013.   int32x4x4_t __ret; \
  13014.   __builtin_neon_vld4q_v(&__ret, __p0, 34); \
  13015.   __ret; \
  13016. })
  13017. #else
  13018. #define vld4q_s32(__p0) __extension__ ({ \
  13019.   int32x4x4_t __ret; \
  13020.   __builtin_neon_vld4q_v(&__ret, __p0, 34); \
  13021.  \
  13022.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13023.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13024.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13025.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13026.   __ret; \
  13027. })
  13028. #endif
  13029.  
  13030. #ifdef __LITTLE_ENDIAN__
  13031. #define vld4q_s16(__p0) __extension__ ({ \
  13032.   int16x8x4_t __ret; \
  13033.   __builtin_neon_vld4q_v(&__ret, __p0, 33); \
  13034.   __ret; \
  13035. })
  13036. #else
  13037. #define vld4q_s16(__p0) __extension__ ({ \
  13038.   int16x8x4_t __ret; \
  13039.   __builtin_neon_vld4q_v(&__ret, __p0, 33); \
  13040.  \
  13041.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13042.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13043.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13044.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13045.   __ret; \
  13046. })
  13047. #endif
  13048.  
  13049. #ifdef __LITTLE_ENDIAN__
  13050. #define vld4_u8(__p0) __extension__ ({ \
  13051.   uint8x8x4_t __ret; \
  13052.   __builtin_neon_vld4_v(&__ret, __p0, 16); \
  13053.   __ret; \
  13054. })
  13055. #else
  13056. #define vld4_u8(__p0) __extension__ ({ \
  13057.   uint8x8x4_t __ret; \
  13058.   __builtin_neon_vld4_v(&__ret, __p0, 16); \
  13059.  \
  13060.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13061.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13062.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13063.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13064.   __ret; \
  13065. })
  13066. #endif
  13067.  
  13068. #ifdef __LITTLE_ENDIAN__
  13069. #define vld4_u32(__p0) __extension__ ({ \
  13070.   uint32x2x4_t __ret; \
  13071.   __builtin_neon_vld4_v(&__ret, __p0, 18); \
  13072.   __ret; \
  13073. })
  13074. #else
  13075. #define vld4_u32(__p0) __extension__ ({ \
  13076.   uint32x2x4_t __ret; \
  13077.   __builtin_neon_vld4_v(&__ret, __p0, 18); \
  13078.  \
  13079.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13080.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13081.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13082.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13083.   __ret; \
  13084. })
  13085. #endif
  13086.  
  13087. #define vld4_u64(__p0) __extension__ ({ \
  13088.   uint64x1x4_t __ret; \
  13089.   __builtin_neon_vld4_v(&__ret, __p0, 19); \
  13090.   __ret; \
  13091. })
  13092. #ifdef __LITTLE_ENDIAN__
  13093. #define vld4_u16(__p0) __extension__ ({ \
  13094.   uint16x4x4_t __ret; \
  13095.   __builtin_neon_vld4_v(&__ret, __p0, 17); \
  13096.   __ret; \
  13097. })
  13098. #else
  13099. #define vld4_u16(__p0) __extension__ ({ \
  13100.   uint16x4x4_t __ret; \
  13101.   __builtin_neon_vld4_v(&__ret, __p0, 17); \
  13102.  \
  13103.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13104.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13105.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13106.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13107.   __ret; \
  13108. })
  13109. #endif
  13110.  
  13111. #ifdef __LITTLE_ENDIAN__
  13112. #define vld4_s8(__p0) __extension__ ({ \
  13113.   int8x8x4_t __ret; \
  13114.   __builtin_neon_vld4_v(&__ret, __p0, 0); \
  13115.   __ret; \
  13116. })
  13117. #else
  13118. #define vld4_s8(__p0) __extension__ ({ \
  13119.   int8x8x4_t __ret; \
  13120.   __builtin_neon_vld4_v(&__ret, __p0, 0); \
  13121.  \
  13122.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13123.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13124.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13125.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13126.   __ret; \
  13127. })
  13128. #endif
  13129.  
  13130. #ifdef __LITTLE_ENDIAN__
  13131. #define vld4_f32(__p0) __extension__ ({ \
  13132.   float32x2x4_t __ret; \
  13133.   __builtin_neon_vld4_v(&__ret, __p0, 9); \
  13134.   __ret; \
  13135. })
  13136. #else
  13137. #define vld4_f32(__p0) __extension__ ({ \
  13138.   float32x2x4_t __ret; \
  13139.   __builtin_neon_vld4_v(&__ret, __p0, 9); \
  13140.  \
  13141.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13142.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13143.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13144.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13145.   __ret; \
  13146. })
  13147. #endif
  13148.  
  13149. #ifdef __LITTLE_ENDIAN__
  13150. #define vld4_s32(__p0) __extension__ ({ \
  13151.   int32x2x4_t __ret; \
  13152.   __builtin_neon_vld4_v(&__ret, __p0, 2); \
  13153.   __ret; \
  13154. })
  13155. #else
  13156. #define vld4_s32(__p0) __extension__ ({ \
  13157.   int32x2x4_t __ret; \
  13158.   __builtin_neon_vld4_v(&__ret, __p0, 2); \
  13159.  \
  13160.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13161.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13162.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13163.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13164.   __ret; \
  13165. })
  13166. #endif
  13167.  
  13168. #define vld4_s64(__p0) __extension__ ({ \
  13169.   int64x1x4_t __ret; \
  13170.   __builtin_neon_vld4_v(&__ret, __p0, 3); \
  13171.   __ret; \
  13172. })
  13173. #ifdef __LITTLE_ENDIAN__
  13174. #define vld4_s16(__p0) __extension__ ({ \
  13175.   int16x4x4_t __ret; \
  13176.   __builtin_neon_vld4_v(&__ret, __p0, 1); \
  13177.   __ret; \
  13178. })
  13179. #else
  13180. #define vld4_s16(__p0) __extension__ ({ \
  13181.   int16x4x4_t __ret; \
  13182.   __builtin_neon_vld4_v(&__ret, __p0, 1); \
  13183.  \
  13184.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13185.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13186.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13187.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13188.   __ret; \
  13189. })
  13190. #endif
  13191.  
  13192. #ifdef __LITTLE_ENDIAN__
  13193. #define vld4_dup_p8(__p0) __extension__ ({ \
  13194.   poly8x8x4_t __ret; \
  13195.   __builtin_neon_vld4_dup_v(&__ret, __p0, 4); \
  13196.   __ret; \
  13197. })
  13198. #else
  13199. #define vld4_dup_p8(__p0) __extension__ ({ \
  13200.   poly8x8x4_t __ret; \
  13201.   __builtin_neon_vld4_dup_v(&__ret, __p0, 4); \
  13202.  \
  13203.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13204.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13205.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13206.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13207.   __ret; \
  13208. })
  13209. #endif
  13210.  
  13211. #ifdef __LITTLE_ENDIAN__
  13212. #define vld4_dup_p16(__p0) __extension__ ({ \
  13213.   poly16x4x4_t __ret; \
  13214.   __builtin_neon_vld4_dup_v(&__ret, __p0, 5); \
  13215.   __ret; \
  13216. })
  13217. #else
  13218. #define vld4_dup_p16(__p0) __extension__ ({ \
  13219.   poly16x4x4_t __ret; \
  13220.   __builtin_neon_vld4_dup_v(&__ret, __p0, 5); \
  13221.  \
  13222.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13223.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13224.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13225.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13226.   __ret; \
  13227. })
  13228. #endif
  13229.  
  13230. #ifdef __LITTLE_ENDIAN__
  13231. #define vld4q_dup_p8(__p0) __extension__ ({ \
  13232.   poly8x16x4_t __ret; \
  13233.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 36); \
  13234.   __ret; \
  13235. })
  13236. #else
  13237. #define vld4q_dup_p8(__p0) __extension__ ({ \
  13238.   poly8x16x4_t __ret; \
  13239.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 36); \
  13240.  \
  13241.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13242.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13243.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13244.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13245.   __ret; \
  13246. })
  13247. #endif
  13248.  
  13249. #ifdef __LITTLE_ENDIAN__
  13250. #define vld4q_dup_p16(__p0) __extension__ ({ \
  13251.   poly16x8x4_t __ret; \
  13252.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 37); \
  13253.   __ret; \
  13254. })
  13255. #else
  13256. #define vld4q_dup_p16(__p0) __extension__ ({ \
  13257.   poly16x8x4_t __ret; \
  13258.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 37); \
  13259.  \
  13260.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13261.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13262.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13263.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13264.   __ret; \
  13265. })
  13266. #endif
  13267.  
  13268. #ifdef __LITTLE_ENDIAN__
  13269. #define vld4q_dup_u8(__p0) __extension__ ({ \
  13270.   uint8x16x4_t __ret; \
  13271.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 48); \
  13272.   __ret; \
  13273. })
  13274. #else
  13275. #define vld4q_dup_u8(__p0) __extension__ ({ \
  13276.   uint8x16x4_t __ret; \
  13277.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 48); \
  13278.  \
  13279.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13280.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13281.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13282.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13283.   __ret; \
  13284. })
  13285. #endif
  13286.  
  13287. #ifdef __LITTLE_ENDIAN__
  13288. #define vld4q_dup_u32(__p0) __extension__ ({ \
  13289.   uint32x4x4_t __ret; \
  13290.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 50); \
  13291.   __ret; \
  13292. })
  13293. #else
  13294. #define vld4q_dup_u32(__p0) __extension__ ({ \
  13295.   uint32x4x4_t __ret; \
  13296.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 50); \
  13297.  \
  13298.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13299.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13300.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13301.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13302.   __ret; \
  13303. })
  13304. #endif
  13305.  
  13306. #ifdef __LITTLE_ENDIAN__
  13307. #define vld4q_dup_u64(__p0) __extension__ ({ \
  13308.   uint64x2x4_t __ret; \
  13309.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 51); \
  13310.   __ret; \
  13311. })
  13312. #else
  13313. #define vld4q_dup_u64(__p0) __extension__ ({ \
  13314.   uint64x2x4_t __ret; \
  13315.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 51); \
  13316.  \
  13317.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13318.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13319.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13320.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13321.   __ret; \
  13322. })
  13323. #endif
  13324.  
  13325. #ifdef __LITTLE_ENDIAN__
  13326. #define vld4q_dup_u16(__p0) __extension__ ({ \
  13327.   uint16x8x4_t __ret; \
  13328.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 49); \
  13329.   __ret; \
  13330. })
  13331. #else
  13332. #define vld4q_dup_u16(__p0) __extension__ ({ \
  13333.   uint16x8x4_t __ret; \
  13334.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 49); \
  13335.  \
  13336.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13337.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13338.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13339.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13340.   __ret; \
  13341. })
  13342. #endif
  13343.  
  13344. #ifdef __LITTLE_ENDIAN__
  13345. #define vld4q_dup_s8(__p0) __extension__ ({ \
  13346.   int8x16x4_t __ret; \
  13347.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 32); \
  13348.   __ret; \
  13349. })
  13350. #else
  13351. #define vld4q_dup_s8(__p0) __extension__ ({ \
  13352.   int8x16x4_t __ret; \
  13353.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 32); \
  13354.  \
  13355.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13356.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13357.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13358.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  13359.   __ret; \
  13360. })
  13361. #endif
  13362.  
  13363. #ifdef __LITTLE_ENDIAN__
  13364. #define vld4q_dup_f32(__p0) __extension__ ({ \
  13365.   float32x4x4_t __ret; \
  13366.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 41); \
  13367.   __ret; \
  13368. })
  13369. #else
  13370. #define vld4q_dup_f32(__p0) __extension__ ({ \
  13371.   float32x4x4_t __ret; \
  13372.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 41); \
  13373.  \
  13374.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13375.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13376.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13377.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13378.   __ret; \
  13379. })
  13380. #endif
  13381.  
  13382. #ifdef __LITTLE_ENDIAN__
  13383. #define vld4q_dup_s32(__p0) __extension__ ({ \
  13384.   int32x4x4_t __ret; \
  13385.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 34); \
  13386.   __ret; \
  13387. })
  13388. #else
  13389. #define vld4q_dup_s32(__p0) __extension__ ({ \
  13390.   int32x4x4_t __ret; \
  13391.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 34); \
  13392.  \
  13393.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13394.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13395.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13396.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13397.   __ret; \
  13398. })
  13399. #endif
  13400.  
  13401. #ifdef __LITTLE_ENDIAN__
  13402. #define vld4q_dup_s64(__p0) __extension__ ({ \
  13403.   int64x2x4_t __ret; \
  13404.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 35); \
  13405.   __ret; \
  13406. })
  13407. #else
  13408. #define vld4q_dup_s64(__p0) __extension__ ({ \
  13409.   int64x2x4_t __ret; \
  13410.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 35); \
  13411.  \
  13412.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13413.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13414.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13415.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13416.   __ret; \
  13417. })
  13418. #endif
  13419.  
  13420. #ifdef __LITTLE_ENDIAN__
  13421. #define vld4q_dup_s16(__p0) __extension__ ({ \
  13422.   int16x8x4_t __ret; \
  13423.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 33); \
  13424.   __ret; \
  13425. })
  13426. #else
  13427. #define vld4q_dup_s16(__p0) __extension__ ({ \
  13428.   int16x8x4_t __ret; \
  13429.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 33); \
  13430.  \
  13431.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13432.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13433.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13434.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13435.   __ret; \
  13436. })
  13437. #endif
  13438.  
  13439. #ifdef __LITTLE_ENDIAN__
  13440. #define vld4_dup_u8(__p0) __extension__ ({ \
  13441.   uint8x8x4_t __ret; \
  13442.   __builtin_neon_vld4_dup_v(&__ret, __p0, 16); \
  13443.   __ret; \
  13444. })
  13445. #else
  13446. #define vld4_dup_u8(__p0) __extension__ ({ \
  13447.   uint8x8x4_t __ret; \
  13448.   __builtin_neon_vld4_dup_v(&__ret, __p0, 16); \
  13449.  \
  13450.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13451.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13452.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13453.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13454.   __ret; \
  13455. })
  13456. #endif
  13457.  
  13458. #ifdef __LITTLE_ENDIAN__
  13459. #define vld4_dup_u32(__p0) __extension__ ({ \
  13460.   uint32x2x4_t __ret; \
  13461.   __builtin_neon_vld4_dup_v(&__ret, __p0, 18); \
  13462.   __ret; \
  13463. })
  13464. #else
  13465. #define vld4_dup_u32(__p0) __extension__ ({ \
  13466.   uint32x2x4_t __ret; \
  13467.   __builtin_neon_vld4_dup_v(&__ret, __p0, 18); \
  13468.  \
  13469.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13470.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13471.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13472.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13473.   __ret; \
  13474. })
  13475. #endif
  13476.  
  13477. #define vld4_dup_u64(__p0) __extension__ ({ \
  13478.   uint64x1x4_t __ret; \
  13479.   __builtin_neon_vld4_dup_v(&__ret, __p0, 19); \
  13480.   __ret; \
  13481. })
  13482. #ifdef __LITTLE_ENDIAN__
  13483. #define vld4_dup_u16(__p0) __extension__ ({ \
  13484.   uint16x4x4_t __ret; \
  13485.   __builtin_neon_vld4_dup_v(&__ret, __p0, 17); \
  13486.   __ret; \
  13487. })
  13488. #else
  13489. #define vld4_dup_u16(__p0) __extension__ ({ \
  13490.   uint16x4x4_t __ret; \
  13491.   __builtin_neon_vld4_dup_v(&__ret, __p0, 17); \
  13492.  \
  13493.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13494.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13495.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13496.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13497.   __ret; \
  13498. })
  13499. #endif
  13500.  
  13501. #ifdef __LITTLE_ENDIAN__
  13502. #define vld4_dup_s8(__p0) __extension__ ({ \
  13503.   int8x8x4_t __ret; \
  13504.   __builtin_neon_vld4_dup_v(&__ret, __p0, 0); \
  13505.   __ret; \
  13506. })
  13507. #else
  13508. #define vld4_dup_s8(__p0) __extension__ ({ \
  13509.   int8x8x4_t __ret; \
  13510.   __builtin_neon_vld4_dup_v(&__ret, __p0, 0); \
  13511.  \
  13512.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13513.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13514.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13515.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13516.   __ret; \
  13517. })
  13518. #endif
  13519.  
  13520. #ifdef __LITTLE_ENDIAN__
  13521. #define vld4_dup_f32(__p0) __extension__ ({ \
  13522.   float32x2x4_t __ret; \
  13523.   __builtin_neon_vld4_dup_v(&__ret, __p0, 9); \
  13524.   __ret; \
  13525. })
  13526. #else
  13527. #define vld4_dup_f32(__p0) __extension__ ({ \
  13528.   float32x2x4_t __ret; \
  13529.   __builtin_neon_vld4_dup_v(&__ret, __p0, 9); \
  13530.  \
  13531.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13532.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13533.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13534.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13535.   __ret; \
  13536. })
  13537. #endif
  13538.  
  13539. #ifdef __LITTLE_ENDIAN__
  13540. #define vld4_dup_s32(__p0) __extension__ ({ \
  13541.   int32x2x4_t __ret; \
  13542.   __builtin_neon_vld4_dup_v(&__ret, __p0, 2); \
  13543.   __ret; \
  13544. })
  13545. #else
  13546. #define vld4_dup_s32(__p0) __extension__ ({ \
  13547.   int32x2x4_t __ret; \
  13548.   __builtin_neon_vld4_dup_v(&__ret, __p0, 2); \
  13549.  \
  13550.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13551.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13552.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13553.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13554.   __ret; \
  13555. })
  13556. #endif
  13557.  
  13558. #define vld4_dup_s64(__p0) __extension__ ({ \
  13559.   int64x1x4_t __ret; \
  13560.   __builtin_neon_vld4_dup_v(&__ret, __p0, 3); \
  13561.   __ret; \
  13562. })
  13563. #ifdef __LITTLE_ENDIAN__
  13564. #define vld4_dup_s16(__p0) __extension__ ({ \
  13565.   int16x4x4_t __ret; \
  13566.   __builtin_neon_vld4_dup_v(&__ret, __p0, 1); \
  13567.   __ret; \
  13568. })
  13569. #else
  13570. #define vld4_dup_s16(__p0) __extension__ ({ \
  13571.   int16x4x4_t __ret; \
  13572.   __builtin_neon_vld4_dup_v(&__ret, __p0, 1); \
  13573.  \
  13574.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13575.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13576.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13577.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13578.   __ret; \
  13579. })
  13580. #endif
  13581.  
  13582. #ifdef __LITTLE_ENDIAN__
  13583. #define vld4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  13584.   poly8x8x4_t __ret; \
  13585.   poly8x8x4_t __s1 = __p1; \
  13586.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 4); \
  13587.   __ret; \
  13588. })
  13589. #else
  13590. #define vld4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  13591.   poly8x8x4_t __ret; \
  13592.   poly8x8x4_t __s1 = __p1; \
  13593.   poly8x8x4_t __rev1; \
  13594.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13595.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13596.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13597.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13598.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 4); \
  13599.  \
  13600.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13601.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13602.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13603.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13604.   __ret; \
  13605. })
  13606. #endif
  13607.  
  13608. #ifdef __LITTLE_ENDIAN__
  13609. #define vld4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  13610.   poly16x4x4_t __ret; \
  13611.   poly16x4x4_t __s1 = __p1; \
  13612.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 5); \
  13613.   __ret; \
  13614. })
  13615. #else
  13616. #define vld4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  13617.   poly16x4x4_t __ret; \
  13618.   poly16x4x4_t __s1 = __p1; \
  13619.   poly16x4x4_t __rev1; \
  13620.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  13621.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  13622.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  13623.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  13624.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 5); \
  13625.  \
  13626.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13627.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13628.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13629.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13630.   __ret; \
  13631. })
  13632. #endif
  13633.  
  13634. #ifdef __LITTLE_ENDIAN__
  13635. #define vld4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  13636.   poly16x8x4_t __ret; \
  13637.   poly16x8x4_t __s1 = __p1; \
  13638.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 37); \
  13639.   __ret; \
  13640. })
  13641. #else
  13642. #define vld4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  13643.   poly16x8x4_t __ret; \
  13644.   poly16x8x4_t __s1 = __p1; \
  13645.   poly16x8x4_t __rev1; \
  13646.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13647.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13648.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13649.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13650.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 37); \
  13651.  \
  13652.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13653.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13654.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13655.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13656.   __ret; \
  13657. })
  13658. #endif
  13659.  
  13660. #ifdef __LITTLE_ENDIAN__
  13661. #define vld4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  13662.   uint32x4x4_t __ret; \
  13663.   uint32x4x4_t __s1 = __p1; \
  13664.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 50); \
  13665.   __ret; \
  13666. })
  13667. #else
  13668. #define vld4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  13669.   uint32x4x4_t __ret; \
  13670.   uint32x4x4_t __s1 = __p1; \
  13671.   uint32x4x4_t __rev1; \
  13672.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  13673.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  13674.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  13675.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  13676.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 50); \
  13677.  \
  13678.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13679.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13680.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13681.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13682.   __ret; \
  13683. })
  13684. #endif
  13685.  
  13686. #ifdef __LITTLE_ENDIAN__
  13687. #define vld4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  13688.   uint16x8x4_t __ret; \
  13689.   uint16x8x4_t __s1 = __p1; \
  13690.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 49); \
  13691.   __ret; \
  13692. })
  13693. #else
  13694. #define vld4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  13695.   uint16x8x4_t __ret; \
  13696.   uint16x8x4_t __s1 = __p1; \
  13697.   uint16x8x4_t __rev1; \
  13698.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13699.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13700.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13701.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13702.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 49); \
  13703.  \
  13704.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13705.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13706.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13707.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13708.   __ret; \
  13709. })
  13710. #endif
  13711.  
  13712. #ifdef __LITTLE_ENDIAN__
  13713. #define vld4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  13714.   float32x4x4_t __ret; \
  13715.   float32x4x4_t __s1 = __p1; \
  13716.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 41); \
  13717.   __ret; \
  13718. })
  13719. #else
  13720. #define vld4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  13721.   float32x4x4_t __ret; \
  13722.   float32x4x4_t __s1 = __p1; \
  13723.   float32x4x4_t __rev1; \
  13724.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  13725.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  13726.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  13727.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  13728.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 41); \
  13729.  \
  13730.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13731.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13732.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13733.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13734.   __ret; \
  13735. })
  13736. #endif
  13737.  
  13738. #ifdef __LITTLE_ENDIAN__
  13739. #define vld4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  13740.   int32x4x4_t __ret; \
  13741.   int32x4x4_t __s1 = __p1; \
  13742.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 34); \
  13743.   __ret; \
  13744. })
  13745. #else
  13746. #define vld4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  13747.   int32x4x4_t __ret; \
  13748.   int32x4x4_t __s1 = __p1; \
  13749.   int32x4x4_t __rev1; \
  13750.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  13751.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  13752.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  13753.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  13754.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 34); \
  13755.  \
  13756.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13757.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13758.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13759.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13760.   __ret; \
  13761. })
  13762. #endif
  13763.  
  13764. #ifdef __LITTLE_ENDIAN__
  13765. #define vld4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  13766.   int16x8x4_t __ret; \
  13767.   int16x8x4_t __s1 = __p1; \
  13768.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 33); \
  13769.   __ret; \
  13770. })
  13771. #else
  13772. #define vld4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  13773.   int16x8x4_t __ret; \
  13774.   int16x8x4_t __s1 = __p1; \
  13775.   int16x8x4_t __rev1; \
  13776.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13777.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13778.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13779.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13780.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 33); \
  13781.  \
  13782.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13783.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13784.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13785.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13786.   __ret; \
  13787. })
  13788. #endif
  13789.  
  13790. #ifdef __LITTLE_ENDIAN__
  13791. #define vld4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  13792.   uint8x8x4_t __ret; \
  13793.   uint8x8x4_t __s1 = __p1; \
  13794.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 16); \
  13795.   __ret; \
  13796. })
  13797. #else
  13798. #define vld4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  13799.   uint8x8x4_t __ret; \
  13800.   uint8x8x4_t __s1 = __p1; \
  13801.   uint8x8x4_t __rev1; \
  13802.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13803.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13804.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13805.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13806.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 16); \
  13807.  \
  13808.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13809.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13810.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13811.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13812.   __ret; \
  13813. })
  13814. #endif
  13815.  
  13816. #ifdef __LITTLE_ENDIAN__
  13817. #define vld4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  13818.   uint32x2x4_t __ret; \
  13819.   uint32x2x4_t __s1 = __p1; \
  13820.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 18); \
  13821.   __ret; \
  13822. })
  13823. #else
  13824. #define vld4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  13825.   uint32x2x4_t __ret; \
  13826.   uint32x2x4_t __s1 = __p1; \
  13827.   uint32x2x4_t __rev1; \
  13828.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  13829.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  13830.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  13831.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  13832.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 18); \
  13833.  \
  13834.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13835.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13836.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13837.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13838.   __ret; \
  13839. })
  13840. #endif
  13841.  
  13842. #ifdef __LITTLE_ENDIAN__
  13843. #define vld4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  13844.   uint16x4x4_t __ret; \
  13845.   uint16x4x4_t __s1 = __p1; \
  13846.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 17); \
  13847.   __ret; \
  13848. })
  13849. #else
  13850. #define vld4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  13851.   uint16x4x4_t __ret; \
  13852.   uint16x4x4_t __s1 = __p1; \
  13853.   uint16x4x4_t __rev1; \
  13854.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  13855.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  13856.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  13857.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  13858.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 17); \
  13859.  \
  13860.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13861.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13862.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13863.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13864.   __ret; \
  13865. })
  13866. #endif
  13867.  
  13868. #ifdef __LITTLE_ENDIAN__
  13869. #define vld4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  13870.   int8x8x4_t __ret; \
  13871.   int8x8x4_t __s1 = __p1; \
  13872.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 0); \
  13873.   __ret; \
  13874. })
  13875. #else
  13876. #define vld4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  13877.   int8x8x4_t __ret; \
  13878.   int8x8x4_t __s1 = __p1; \
  13879.   int8x8x4_t __rev1; \
  13880.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13881.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13882.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13883.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13884.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 0); \
  13885.  \
  13886.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  13887.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  13888.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  13889.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  13890.   __ret; \
  13891. })
  13892. #endif
  13893.  
  13894. #ifdef __LITTLE_ENDIAN__
  13895. #define vld4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  13896.   float32x2x4_t __ret; \
  13897.   float32x2x4_t __s1 = __p1; \
  13898.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 9); \
  13899.   __ret; \
  13900. })
  13901. #else
  13902. #define vld4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  13903.   float32x2x4_t __ret; \
  13904.   float32x2x4_t __s1 = __p1; \
  13905.   float32x2x4_t __rev1; \
  13906.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  13907.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  13908.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  13909.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  13910.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 9); \
  13911.  \
  13912.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13913.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13914.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13915.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13916.   __ret; \
  13917. })
  13918. #endif
  13919.  
  13920. #ifdef __LITTLE_ENDIAN__
  13921. #define vld4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  13922.   int32x2x4_t __ret; \
  13923.   int32x2x4_t __s1 = __p1; \
  13924.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 2); \
  13925.   __ret; \
  13926. })
  13927. #else
  13928. #define vld4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  13929.   int32x2x4_t __ret; \
  13930.   int32x2x4_t __s1 = __p1; \
  13931.   int32x2x4_t __rev1; \
  13932.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  13933.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  13934.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  13935.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  13936.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 2); \
  13937.  \
  13938.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  13939.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  13940.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  13941.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  13942.   __ret; \
  13943. })
  13944. #endif
  13945.  
  13946. #ifdef __LITTLE_ENDIAN__
  13947. #define vld4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  13948.   int16x4x4_t __ret; \
  13949.   int16x4x4_t __s1 = __p1; \
  13950.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 1); \
  13951.   __ret; \
  13952. })
  13953. #else
  13954. #define vld4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  13955.   int16x4x4_t __ret; \
  13956.   int16x4x4_t __s1 = __p1; \
  13957.   int16x4x4_t __rev1; \
  13958.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  13959.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  13960.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  13961.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  13962.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 1); \
  13963.  \
  13964.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  13965.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  13966.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  13967.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  13968.   __ret; \
  13969. })
  13970. #endif
  13971.  
  13972. #ifdef __LITTLE_ENDIAN__
  13973. __ai uint8x16_t vmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  13974.   uint8x16_t __ret;
  13975.   __ret = (uint8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  13976.   return __ret;
  13977. }
  13978. #else
  13979. __ai uint8x16_t vmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  13980.   uint8x16_t __ret;
  13981.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  13982.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  13983.   __ret = (uint8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  13984.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  13985.   return __ret;
  13986. }
  13987. #endif
  13988.  
  13989. #ifdef __LITTLE_ENDIAN__
  13990. __ai uint32x4_t vmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  13991.   uint32x4_t __ret;
  13992.   __ret = (uint32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  13993.   return __ret;
  13994. }
  13995. #else
  13996. __ai uint32x4_t vmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  13997.   uint32x4_t __ret;
  13998.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  13999.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14000.   __ret = (uint32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  14001.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14002.   return __ret;
  14003. }
  14004. #endif
  14005.  
  14006. #ifdef __LITTLE_ENDIAN__
  14007. __ai uint16x8_t vmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  14008.   uint16x8_t __ret;
  14009.   __ret = (uint16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  14010.   return __ret;
  14011. }
  14012. #else
  14013. __ai uint16x8_t vmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  14014.   uint16x8_t __ret;
  14015.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14016.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14017.   __ret = (uint16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  14018.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14019.   return __ret;
  14020. }
  14021. #endif
  14022.  
  14023. #ifdef __LITTLE_ENDIAN__
  14024. __ai int8x16_t vmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
  14025.   int8x16_t __ret;
  14026.   __ret = (int8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  14027.   return __ret;
  14028. }
  14029. #else
  14030. __ai int8x16_t vmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
  14031.   int8x16_t __ret;
  14032.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14033.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14034.   __ret = (int8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  14035.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14036.   return __ret;
  14037. }
  14038. #endif
  14039.  
  14040. #ifdef __LITTLE_ENDIAN__
  14041. __ai float32x4_t vmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
  14042.   float32x4_t __ret;
  14043.   __ret = (float32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  14044.   return __ret;
  14045. }
  14046. #else
  14047. __ai float32x4_t vmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
  14048.   float32x4_t __ret;
  14049.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14050.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14051.   __ret = (float32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  14052.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14053.   return __ret;
  14054. }
  14055. #endif
  14056.  
  14057. #ifdef __LITTLE_ENDIAN__
  14058. __ai int32x4_t vmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
  14059.   int32x4_t __ret;
  14060.   __ret = (int32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  14061.   return __ret;
  14062. }
  14063. #else
  14064. __ai int32x4_t vmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
  14065.   int32x4_t __ret;
  14066.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14067.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14068.   __ret = (int32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  14069.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14070.   return __ret;
  14071. }
  14072. #endif
  14073.  
  14074. #ifdef __LITTLE_ENDIAN__
  14075. __ai int16x8_t vmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
  14076.   int16x8_t __ret;
  14077.   __ret = (int16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  14078.   return __ret;
  14079. }
  14080. #else
  14081. __ai int16x8_t vmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
  14082.   int16x8_t __ret;
  14083.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14084.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14085.   __ret = (int16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  14086.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14087.   return __ret;
  14088. }
  14089. #endif
  14090.  
  14091. #ifdef __LITTLE_ENDIAN__
  14092. __ai uint8x8_t vmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
  14093.   uint8x8_t __ret;
  14094.   __ret = (uint8x8_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  14095.   return __ret;
  14096. }
  14097. #else
  14098. __ai uint8x8_t vmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
  14099.   uint8x8_t __ret;
  14100.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14101.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14102.   __ret = (uint8x8_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  14103.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14104.   return __ret;
  14105. }
  14106. #endif
  14107.  
  14108. #ifdef __LITTLE_ENDIAN__
  14109. __ai uint32x2_t vmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
  14110.   uint32x2_t __ret;
  14111.   __ret = (uint32x2_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  14112.   return __ret;
  14113. }
  14114. #else
  14115. __ai uint32x2_t vmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
  14116.   uint32x2_t __ret;
  14117.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  14118.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  14119.   __ret = (uint32x2_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  14120.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  14121.   return __ret;
  14122. }
  14123. #endif
  14124.  
  14125. #ifdef __LITTLE_ENDIAN__
  14126. __ai uint16x4_t vmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
  14127.   uint16x4_t __ret;
  14128.   __ret = (uint16x4_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  14129.   return __ret;
  14130. }
  14131. #else
  14132. __ai uint16x4_t vmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
  14133.   uint16x4_t __ret;
  14134.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14135.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14136.   __ret = (uint16x4_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  14137.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14138.   return __ret;
  14139. }
  14140. #endif
  14141.  
  14142. #ifdef __LITTLE_ENDIAN__
  14143. __ai int8x8_t vmax_s8(int8x8_t __p0, int8x8_t __p1) {
  14144.   int8x8_t __ret;
  14145.   __ret = (int8x8_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  14146.   return __ret;
  14147. }
  14148. #else
  14149. __ai int8x8_t vmax_s8(int8x8_t __p0, int8x8_t __p1) {
  14150.   int8x8_t __ret;
  14151.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14152.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14153.   __ret = (int8x8_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  14154.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14155.   return __ret;
  14156. }
  14157. #endif
  14158.  
  14159. #ifdef __LITTLE_ENDIAN__
  14160. __ai float32x2_t vmax_f32(float32x2_t __p0, float32x2_t __p1) {
  14161.   float32x2_t __ret;
  14162.   __ret = (float32x2_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  14163.   return __ret;
  14164. }
  14165. #else
  14166. __ai float32x2_t vmax_f32(float32x2_t __p0, float32x2_t __p1) {
  14167.   float32x2_t __ret;
  14168.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  14169.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  14170.   __ret = (float32x2_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  14171.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  14172.   return __ret;
  14173. }
  14174. #endif
  14175.  
  14176. #ifdef __LITTLE_ENDIAN__
  14177. __ai int32x2_t vmax_s32(int32x2_t __p0, int32x2_t __p1) {
  14178.   int32x2_t __ret;
  14179.   __ret = (int32x2_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  14180.   return __ret;
  14181. }
  14182. #else
  14183. __ai int32x2_t vmax_s32(int32x2_t __p0, int32x2_t __p1) {
  14184.   int32x2_t __ret;
  14185.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  14186.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  14187.   __ret = (int32x2_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  14188.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  14189.   return __ret;
  14190. }
  14191. #endif
  14192.  
  14193. #ifdef __LITTLE_ENDIAN__
  14194. __ai int16x4_t vmax_s16(int16x4_t __p0, int16x4_t __p1) {
  14195.   int16x4_t __ret;
  14196.   __ret = (int16x4_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  14197.   return __ret;
  14198. }
  14199. #else
  14200. __ai int16x4_t vmax_s16(int16x4_t __p0, int16x4_t __p1) {
  14201.   int16x4_t __ret;
  14202.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14203.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14204.   __ret = (int16x4_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  14205.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14206.   return __ret;
  14207. }
  14208. #endif
  14209.  
  14210. #ifdef __LITTLE_ENDIAN__
  14211. __ai uint8x16_t vminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  14212.   uint8x16_t __ret;
  14213.   __ret = (uint8x16_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  14214.   return __ret;
  14215. }
  14216. #else
  14217. __ai uint8x16_t vminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  14218.   uint8x16_t __ret;
  14219.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14220.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14221.   __ret = (uint8x16_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  14222.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14223.   return __ret;
  14224. }
  14225. #endif
  14226.  
  14227. #ifdef __LITTLE_ENDIAN__
  14228. __ai uint32x4_t vminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  14229.   uint32x4_t __ret;
  14230.   __ret = (uint32x4_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  14231.   return __ret;
  14232. }
  14233. #else
  14234. __ai uint32x4_t vminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  14235.   uint32x4_t __ret;
  14236.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14237.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14238.   __ret = (uint32x4_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  14239.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14240.   return __ret;
  14241. }
  14242. #endif
  14243.  
  14244. #ifdef __LITTLE_ENDIAN__
  14245. __ai uint16x8_t vminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  14246.   uint16x8_t __ret;
  14247.   __ret = (uint16x8_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  14248.   return __ret;
  14249. }
  14250. #else
  14251. __ai uint16x8_t vminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  14252.   uint16x8_t __ret;
  14253.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14254.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14255.   __ret = (uint16x8_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  14256.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14257.   return __ret;
  14258. }
  14259. #endif
  14260.  
  14261. #ifdef __LITTLE_ENDIAN__
  14262. __ai int8x16_t vminq_s8(int8x16_t __p0, int8x16_t __p1) {
  14263.   int8x16_t __ret;
  14264.   __ret = (int8x16_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  14265.   return __ret;
  14266. }
  14267. #else
  14268. __ai int8x16_t vminq_s8(int8x16_t __p0, int8x16_t __p1) {
  14269.   int8x16_t __ret;
  14270.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14271.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14272.   __ret = (int8x16_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  14273.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14274.   return __ret;
  14275. }
  14276. #endif
  14277.  
  14278. #ifdef __LITTLE_ENDIAN__
  14279. __ai float32x4_t vminq_f32(float32x4_t __p0, float32x4_t __p1) {
  14280.   float32x4_t __ret;
  14281.   __ret = (float32x4_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  14282.   return __ret;
  14283. }
  14284. #else
  14285. __ai float32x4_t vminq_f32(float32x4_t __p0, float32x4_t __p1) {
  14286.   float32x4_t __ret;
  14287.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14288.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14289.   __ret = (float32x4_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  14290.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14291.   return __ret;
  14292. }
  14293. #endif
  14294.  
  14295. #ifdef __LITTLE_ENDIAN__
  14296. __ai int32x4_t vminq_s32(int32x4_t __p0, int32x4_t __p1) {
  14297.   int32x4_t __ret;
  14298.   __ret = (int32x4_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  14299.   return __ret;
  14300. }
  14301. #else
  14302. __ai int32x4_t vminq_s32(int32x4_t __p0, int32x4_t __p1) {
  14303.   int32x4_t __ret;
  14304.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14305.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14306.   __ret = (int32x4_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  14307.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14308.   return __ret;
  14309. }
  14310. #endif
  14311.  
  14312. #ifdef __LITTLE_ENDIAN__
  14313. __ai int16x8_t vminq_s16(int16x8_t __p0, int16x8_t __p1) {
  14314.   int16x8_t __ret;
  14315.   __ret = (int16x8_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  14316.   return __ret;
  14317. }
  14318. #else
  14319. __ai int16x8_t vminq_s16(int16x8_t __p0, int16x8_t __p1) {
  14320.   int16x8_t __ret;
  14321.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14322.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14323.   __ret = (int16x8_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  14324.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14325.   return __ret;
  14326. }
  14327. #endif
  14328.  
  14329. #ifdef __LITTLE_ENDIAN__
  14330. __ai uint8x8_t vmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
  14331.   uint8x8_t __ret;
  14332.   __ret = (uint8x8_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  14333.   return __ret;
  14334. }
  14335. #else
  14336. __ai uint8x8_t vmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
  14337.   uint8x8_t __ret;
  14338.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14339.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14340.   __ret = (uint8x8_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  14341.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14342.   return __ret;
  14343. }
  14344. #endif
  14345.  
  14346. #ifdef __LITTLE_ENDIAN__
  14347. __ai uint32x2_t vmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
  14348.   uint32x2_t __ret;
  14349.   __ret = (uint32x2_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  14350.   return __ret;
  14351. }
  14352. #else
  14353. __ai uint32x2_t vmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
  14354.   uint32x2_t __ret;
  14355.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  14356.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  14357.   __ret = (uint32x2_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  14358.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  14359.   return __ret;
  14360. }
  14361. #endif
  14362.  
  14363. #ifdef __LITTLE_ENDIAN__
  14364. __ai uint16x4_t vmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
  14365.   uint16x4_t __ret;
  14366.   __ret = (uint16x4_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  14367.   return __ret;
  14368. }
  14369. #else
  14370. __ai uint16x4_t vmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
  14371.   uint16x4_t __ret;
  14372.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14373.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14374.   __ret = (uint16x4_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  14375.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14376.   return __ret;
  14377. }
  14378. #endif
  14379.  
  14380. #ifdef __LITTLE_ENDIAN__
  14381. __ai int8x8_t vmin_s8(int8x8_t __p0, int8x8_t __p1) {
  14382.   int8x8_t __ret;
  14383.   __ret = (int8x8_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  14384.   return __ret;
  14385. }
  14386. #else
  14387. __ai int8x8_t vmin_s8(int8x8_t __p0, int8x8_t __p1) {
  14388.   int8x8_t __ret;
  14389.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14390.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14391.   __ret = (int8x8_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  14392.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14393.   return __ret;
  14394. }
  14395. #endif
  14396.  
  14397. #ifdef __LITTLE_ENDIAN__
  14398. __ai float32x2_t vmin_f32(float32x2_t __p0, float32x2_t __p1) {
  14399.   float32x2_t __ret;
  14400.   __ret = (float32x2_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  14401.   return __ret;
  14402. }
  14403. #else
  14404. __ai float32x2_t vmin_f32(float32x2_t __p0, float32x2_t __p1) {
  14405.   float32x2_t __ret;
  14406.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  14407.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  14408.   __ret = (float32x2_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  14409.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  14410.   return __ret;
  14411. }
  14412. #endif
  14413.  
  14414. #ifdef __LITTLE_ENDIAN__
  14415. __ai int32x2_t vmin_s32(int32x2_t __p0, int32x2_t __p1) {
  14416.   int32x2_t __ret;
  14417.   __ret = (int32x2_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  14418.   return __ret;
  14419. }
  14420. #else
  14421. __ai int32x2_t vmin_s32(int32x2_t __p0, int32x2_t __p1) {
  14422.   int32x2_t __ret;
  14423.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  14424.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  14425.   __ret = (int32x2_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  14426.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  14427.   return __ret;
  14428. }
  14429. #endif
  14430.  
  14431. #ifdef __LITTLE_ENDIAN__
  14432. __ai int16x4_t vmin_s16(int16x4_t __p0, int16x4_t __p1) {
  14433.   int16x4_t __ret;
  14434.   __ret = (int16x4_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  14435.   return __ret;
  14436. }
  14437. #else
  14438. __ai int16x4_t vmin_s16(int16x4_t __p0, int16x4_t __p1) {
  14439.   int16x4_t __ret;
  14440.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14441.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14442.   __ret = (int16x4_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  14443.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14444.   return __ret;
  14445. }
  14446. #endif
  14447.  
  14448. #ifdef __LITTLE_ENDIAN__
  14449. __ai uint8x16_t vmlaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  14450.   uint8x16_t __ret;
  14451.   __ret = __p0 + __p1 * __p2;
  14452.   return __ret;
  14453. }
  14454. #else
  14455. __ai uint8x16_t vmlaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  14456.   uint8x16_t __ret;
  14457.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14458.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14459.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14460.   __ret = __rev0 + __rev1 * __rev2;
  14461.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14462.   return __ret;
  14463. }
  14464. #endif
  14465.  
  14466. #ifdef __LITTLE_ENDIAN__
  14467. __ai uint32x4_t vmlaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  14468.   uint32x4_t __ret;
  14469.   __ret = __p0 + __p1 * __p2;
  14470.   return __ret;
  14471. }
  14472. #else
  14473. __ai uint32x4_t vmlaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  14474.   uint32x4_t __ret;
  14475.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14476.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14477.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  14478.   __ret = __rev0 + __rev1 * __rev2;
  14479.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14480.   return __ret;
  14481. }
  14482. #endif
  14483.  
  14484. #ifdef __LITTLE_ENDIAN__
  14485. __ai uint16x8_t vmlaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  14486.   uint16x8_t __ret;
  14487.   __ret = __p0 + __p1 * __p2;
  14488.   return __ret;
  14489. }
  14490. #else
  14491. __ai uint16x8_t vmlaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  14492.   uint16x8_t __ret;
  14493.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14494.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14495.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  14496.   __ret = __rev0 + __rev1 * __rev2;
  14497.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14498.   return __ret;
  14499. }
  14500. #endif
  14501.  
  14502. #ifdef __LITTLE_ENDIAN__
  14503. __ai int8x16_t vmlaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  14504.   int8x16_t __ret;
  14505.   __ret = __p0 + __p1 * __p2;
  14506.   return __ret;
  14507. }
  14508. #else
  14509. __ai int8x16_t vmlaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  14510.   int8x16_t __ret;
  14511.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14512.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14513.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14514.   __ret = __rev0 + __rev1 * __rev2;
  14515.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  14516.   return __ret;
  14517. }
  14518. #endif
  14519.  
  14520. #ifdef __LITTLE_ENDIAN__
  14521. __ai float32x4_t vmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  14522.   float32x4_t __ret;
  14523.   __ret = __p0 + __p1 * __p2;
  14524.   return __ret;
  14525. }
  14526. #else
  14527. __ai float32x4_t vmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  14528.   float32x4_t __ret;
  14529.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14530.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14531.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  14532.   __ret = __rev0 + __rev1 * __rev2;
  14533.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14534.   return __ret;
  14535. }
  14536. #endif
  14537.  
  14538. #ifdef __LITTLE_ENDIAN__
  14539. __ai int32x4_t vmlaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  14540.   int32x4_t __ret;
  14541.   __ret = __p0 + __p1 * __p2;
  14542.   return __ret;
  14543. }
  14544. #else
  14545. __ai int32x4_t vmlaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  14546.   int32x4_t __ret;
  14547.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14548.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14549.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  14550.   __ret = __rev0 + __rev1 * __rev2;
  14551.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14552.   return __ret;
  14553. }
  14554. #endif
  14555.  
  14556. #ifdef __LITTLE_ENDIAN__
  14557. __ai int16x8_t vmlaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  14558.   int16x8_t __ret;
  14559.   __ret = __p0 + __p1 * __p2;
  14560.   return __ret;
  14561. }
  14562. #else
  14563. __ai int16x8_t vmlaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  14564.   int16x8_t __ret;
  14565.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14566.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14567.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  14568.   __ret = __rev0 + __rev1 * __rev2;
  14569.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14570.   return __ret;
  14571. }
  14572. #endif
  14573.  
  14574. #ifdef __LITTLE_ENDIAN__
  14575. __ai uint8x8_t vmla_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  14576.   uint8x8_t __ret;
  14577.   __ret = __p0 + __p1 * __p2;
  14578.   return __ret;
  14579. }
  14580. #else
  14581. __ai uint8x8_t vmla_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  14582.   uint8x8_t __ret;
  14583.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14584.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14585.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  14586.   __ret = __rev0 + __rev1 * __rev2;
  14587.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14588.   return __ret;
  14589. }
  14590. #endif
  14591.  
  14592. #ifdef __LITTLE_ENDIAN__
  14593. __ai uint32x2_t vmla_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  14594.   uint32x2_t __ret;
  14595.   __ret = __p0 + __p1 * __p2;
  14596.   return __ret;
  14597. }
  14598. #else
  14599. __ai uint32x2_t vmla_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  14600.   uint32x2_t __ret;
  14601.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  14602.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  14603.   uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  14604.   __ret = __rev0 + __rev1 * __rev2;
  14605.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  14606.   return __ret;
  14607. }
  14608. #endif
  14609.  
  14610. #ifdef __LITTLE_ENDIAN__
  14611. __ai uint16x4_t vmla_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  14612.   uint16x4_t __ret;
  14613.   __ret = __p0 + __p1 * __p2;
  14614.   return __ret;
  14615. }
  14616. #else
  14617. __ai uint16x4_t vmla_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  14618.   uint16x4_t __ret;
  14619.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14620.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14621.   uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  14622.   __ret = __rev0 + __rev1 * __rev2;
  14623.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14624.   return __ret;
  14625. }
  14626. #endif
  14627.  
  14628. #ifdef __LITTLE_ENDIAN__
  14629. __ai int8x8_t vmla_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  14630.   int8x8_t __ret;
  14631.   __ret = __p0 + __p1 * __p2;
  14632.   return __ret;
  14633. }
  14634. #else
  14635. __ai int8x8_t vmla_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  14636.   int8x8_t __ret;
  14637.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14638.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14639.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  14640.   __ret = __rev0 + __rev1 * __rev2;
  14641.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14642.   return __ret;
  14643. }
  14644. #endif
  14645.  
  14646. #ifdef __LITTLE_ENDIAN__
  14647. __ai float32x2_t vmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  14648.   float32x2_t __ret;
  14649.   __ret = __p0 + __p1 * __p2;
  14650.   return __ret;
  14651. }
  14652. #else
  14653. __ai float32x2_t vmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  14654.   float32x2_t __ret;
  14655.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  14656.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  14657.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  14658.   __ret = __rev0 + __rev1 * __rev2;
  14659.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  14660.   return __ret;
  14661. }
  14662. #endif
  14663.  
  14664. #ifdef __LITTLE_ENDIAN__
  14665. __ai int32x2_t vmla_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  14666.   int32x2_t __ret;
  14667.   __ret = __p0 + __p1 * __p2;
  14668.   return __ret;
  14669. }
  14670. #else
  14671. __ai int32x2_t vmla_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  14672.   int32x2_t __ret;
  14673.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  14674.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  14675.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  14676.   __ret = __rev0 + __rev1 * __rev2;
  14677.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  14678.   return __ret;
  14679. }
  14680. #endif
  14681.  
  14682. #ifdef __LITTLE_ENDIAN__
  14683. __ai int16x4_t vmla_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  14684.   int16x4_t __ret;
  14685.   __ret = __p0 + __p1 * __p2;
  14686.   return __ret;
  14687. }
  14688. #else
  14689. __ai int16x4_t vmla_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  14690.   int16x4_t __ret;
  14691.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14692.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14693.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  14694.   __ret = __rev0 + __rev1 * __rev2;
  14695.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14696.   return __ret;
  14697. }
  14698. #endif
  14699.  
  14700. #ifdef __LITTLE_ENDIAN__
  14701. #define vmlaq_lane_u32(__p0_46, __p1_46, __p2_46, __p3_46) __extension__ ({ \
  14702.   uint32x4_t __ret_46; \
  14703.   uint32x4_t __s0_46 = __p0_46; \
  14704.   uint32x4_t __s1_46 = __p1_46; \
  14705.   uint32x2_t __s2_46 = __p2_46; \
  14706.   __ret_46 = __s0_46 + __s1_46 * splatq_lane_u32(__s2_46, __p3_46); \
  14707.   __ret_46; \
  14708. })
  14709. #else
  14710. #define vmlaq_lane_u32(__p0_47, __p1_47, __p2_47, __p3_47) __extension__ ({ \
  14711.   uint32x4_t __ret_47; \
  14712.   uint32x4_t __s0_47 = __p0_47; \
  14713.   uint32x4_t __s1_47 = __p1_47; \
  14714.   uint32x2_t __s2_47 = __p2_47; \
  14715.   uint32x4_t __rev0_47;  __rev0_47 = __builtin_shufflevector(__s0_47, __s0_47, 3, 2, 1, 0); \
  14716.   uint32x4_t __rev1_47;  __rev1_47 = __builtin_shufflevector(__s1_47, __s1_47, 3, 2, 1, 0); \
  14717.   uint32x2_t __rev2_47;  __rev2_47 = __builtin_shufflevector(__s2_47, __s2_47, 1, 0); \
  14718.   __ret_47 = __rev0_47 + __rev1_47 * __noswap_splatq_lane_u32(__rev2_47, __p3_47); \
  14719.   __ret_47 = __builtin_shufflevector(__ret_47, __ret_47, 3, 2, 1, 0); \
  14720.   __ret_47; \
  14721. })
  14722. #endif
  14723.  
  14724. #ifdef __LITTLE_ENDIAN__
  14725. #define vmlaq_lane_u16(__p0_48, __p1_48, __p2_48, __p3_48) __extension__ ({ \
  14726.   uint16x8_t __ret_48; \
  14727.   uint16x8_t __s0_48 = __p0_48; \
  14728.   uint16x8_t __s1_48 = __p1_48; \
  14729.   uint16x4_t __s2_48 = __p2_48; \
  14730.   __ret_48 = __s0_48 + __s1_48 * splatq_lane_u16(__s2_48, __p3_48); \
  14731.   __ret_48; \
  14732. })
  14733. #else
  14734. #define vmlaq_lane_u16(__p0_49, __p1_49, __p2_49, __p3_49) __extension__ ({ \
  14735.   uint16x8_t __ret_49; \
  14736.   uint16x8_t __s0_49 = __p0_49; \
  14737.   uint16x8_t __s1_49 = __p1_49; \
  14738.   uint16x4_t __s2_49 = __p2_49; \
  14739.   uint16x8_t __rev0_49;  __rev0_49 = __builtin_shufflevector(__s0_49, __s0_49, 7, 6, 5, 4, 3, 2, 1, 0); \
  14740.   uint16x8_t __rev1_49;  __rev1_49 = __builtin_shufflevector(__s1_49, __s1_49, 7, 6, 5, 4, 3, 2, 1, 0); \
  14741.   uint16x4_t __rev2_49;  __rev2_49 = __builtin_shufflevector(__s2_49, __s2_49, 3, 2, 1, 0); \
  14742.   __ret_49 = __rev0_49 + __rev1_49 * __noswap_splatq_lane_u16(__rev2_49, __p3_49); \
  14743.   __ret_49 = __builtin_shufflevector(__ret_49, __ret_49, 7, 6, 5, 4, 3, 2, 1, 0); \
  14744.   __ret_49; \
  14745. })
  14746. #endif
  14747.  
  14748. #ifdef __LITTLE_ENDIAN__
  14749. #define vmlaq_lane_f32(__p0_50, __p1_50, __p2_50, __p3_50) __extension__ ({ \
  14750.   float32x4_t __ret_50; \
  14751.   float32x4_t __s0_50 = __p0_50; \
  14752.   float32x4_t __s1_50 = __p1_50; \
  14753.   float32x2_t __s2_50 = __p2_50; \
  14754.   __ret_50 = __s0_50 + __s1_50 * splatq_lane_f32(__s2_50, __p3_50); \
  14755.   __ret_50; \
  14756. })
  14757. #else
  14758. #define vmlaq_lane_f32(__p0_51, __p1_51, __p2_51, __p3_51) __extension__ ({ \
  14759.   float32x4_t __ret_51; \
  14760.   float32x4_t __s0_51 = __p0_51; \
  14761.   float32x4_t __s1_51 = __p1_51; \
  14762.   float32x2_t __s2_51 = __p2_51; \
  14763.   float32x4_t __rev0_51;  __rev0_51 = __builtin_shufflevector(__s0_51, __s0_51, 3, 2, 1, 0); \
  14764.   float32x4_t __rev1_51;  __rev1_51 = __builtin_shufflevector(__s1_51, __s1_51, 3, 2, 1, 0); \
  14765.   float32x2_t __rev2_51;  __rev2_51 = __builtin_shufflevector(__s2_51, __s2_51, 1, 0); \
  14766.   __ret_51 = __rev0_51 + __rev1_51 * __noswap_splatq_lane_f32(__rev2_51, __p3_51); \
  14767.   __ret_51 = __builtin_shufflevector(__ret_51, __ret_51, 3, 2, 1, 0); \
  14768.   __ret_51; \
  14769. })
  14770. #endif
  14771.  
  14772. #ifdef __LITTLE_ENDIAN__
  14773. #define vmlaq_lane_s32(__p0_52, __p1_52, __p2_52, __p3_52) __extension__ ({ \
  14774.   int32x4_t __ret_52; \
  14775.   int32x4_t __s0_52 = __p0_52; \
  14776.   int32x4_t __s1_52 = __p1_52; \
  14777.   int32x2_t __s2_52 = __p2_52; \
  14778.   __ret_52 = __s0_52 + __s1_52 * splatq_lane_s32(__s2_52, __p3_52); \
  14779.   __ret_52; \
  14780. })
  14781. #else
  14782. #define vmlaq_lane_s32(__p0_53, __p1_53, __p2_53, __p3_53) __extension__ ({ \
  14783.   int32x4_t __ret_53; \
  14784.   int32x4_t __s0_53 = __p0_53; \
  14785.   int32x4_t __s1_53 = __p1_53; \
  14786.   int32x2_t __s2_53 = __p2_53; \
  14787.   int32x4_t __rev0_53;  __rev0_53 = __builtin_shufflevector(__s0_53, __s0_53, 3, 2, 1, 0); \
  14788.   int32x4_t __rev1_53;  __rev1_53 = __builtin_shufflevector(__s1_53, __s1_53, 3, 2, 1, 0); \
  14789.   int32x2_t __rev2_53;  __rev2_53 = __builtin_shufflevector(__s2_53, __s2_53, 1, 0); \
  14790.   __ret_53 = __rev0_53 + __rev1_53 * __noswap_splatq_lane_s32(__rev2_53, __p3_53); \
  14791.   __ret_53 = __builtin_shufflevector(__ret_53, __ret_53, 3, 2, 1, 0); \
  14792.   __ret_53; \
  14793. })
  14794. #endif
  14795.  
  14796. #ifdef __LITTLE_ENDIAN__
  14797. #define vmlaq_lane_s16(__p0_54, __p1_54, __p2_54, __p3_54) __extension__ ({ \
  14798.   int16x8_t __ret_54; \
  14799.   int16x8_t __s0_54 = __p0_54; \
  14800.   int16x8_t __s1_54 = __p1_54; \
  14801.   int16x4_t __s2_54 = __p2_54; \
  14802.   __ret_54 = __s0_54 + __s1_54 * splatq_lane_s16(__s2_54, __p3_54); \
  14803.   __ret_54; \
  14804. })
  14805. #else
  14806. #define vmlaq_lane_s16(__p0_55, __p1_55, __p2_55, __p3_55) __extension__ ({ \
  14807.   int16x8_t __ret_55; \
  14808.   int16x8_t __s0_55 = __p0_55; \
  14809.   int16x8_t __s1_55 = __p1_55; \
  14810.   int16x4_t __s2_55 = __p2_55; \
  14811.   int16x8_t __rev0_55;  __rev0_55 = __builtin_shufflevector(__s0_55, __s0_55, 7, 6, 5, 4, 3, 2, 1, 0); \
  14812.   int16x8_t __rev1_55;  __rev1_55 = __builtin_shufflevector(__s1_55, __s1_55, 7, 6, 5, 4, 3, 2, 1, 0); \
  14813.   int16x4_t __rev2_55;  __rev2_55 = __builtin_shufflevector(__s2_55, __s2_55, 3, 2, 1, 0); \
  14814.   __ret_55 = __rev0_55 + __rev1_55 * __noswap_splatq_lane_s16(__rev2_55, __p3_55); \
  14815.   __ret_55 = __builtin_shufflevector(__ret_55, __ret_55, 7, 6, 5, 4, 3, 2, 1, 0); \
  14816.   __ret_55; \
  14817. })
  14818. #endif
  14819.  
  14820. #ifdef __LITTLE_ENDIAN__
  14821. #define vmla_lane_u32(__p0_56, __p1_56, __p2_56, __p3_56) __extension__ ({ \
  14822.   uint32x2_t __ret_56; \
  14823.   uint32x2_t __s0_56 = __p0_56; \
  14824.   uint32x2_t __s1_56 = __p1_56; \
  14825.   uint32x2_t __s2_56 = __p2_56; \
  14826.   __ret_56 = __s0_56 + __s1_56 * splat_lane_u32(__s2_56, __p3_56); \
  14827.   __ret_56; \
  14828. })
  14829. #else
  14830. #define vmla_lane_u32(__p0_57, __p1_57, __p2_57, __p3_57) __extension__ ({ \
  14831.   uint32x2_t __ret_57; \
  14832.   uint32x2_t __s0_57 = __p0_57; \
  14833.   uint32x2_t __s1_57 = __p1_57; \
  14834.   uint32x2_t __s2_57 = __p2_57; \
  14835.   uint32x2_t __rev0_57;  __rev0_57 = __builtin_shufflevector(__s0_57, __s0_57, 1, 0); \
  14836.   uint32x2_t __rev1_57;  __rev1_57 = __builtin_shufflevector(__s1_57, __s1_57, 1, 0); \
  14837.   uint32x2_t __rev2_57;  __rev2_57 = __builtin_shufflevector(__s2_57, __s2_57, 1, 0); \
  14838.   __ret_57 = __rev0_57 + __rev1_57 * __noswap_splat_lane_u32(__rev2_57, __p3_57); \
  14839.   __ret_57 = __builtin_shufflevector(__ret_57, __ret_57, 1, 0); \
  14840.   __ret_57; \
  14841. })
  14842. #endif
  14843.  
  14844. #ifdef __LITTLE_ENDIAN__
  14845. #define vmla_lane_u16(__p0_58, __p1_58, __p2_58, __p3_58) __extension__ ({ \
  14846.   uint16x4_t __ret_58; \
  14847.   uint16x4_t __s0_58 = __p0_58; \
  14848.   uint16x4_t __s1_58 = __p1_58; \
  14849.   uint16x4_t __s2_58 = __p2_58; \
  14850.   __ret_58 = __s0_58 + __s1_58 * splat_lane_u16(__s2_58, __p3_58); \
  14851.   __ret_58; \
  14852. })
  14853. #else
  14854. #define vmla_lane_u16(__p0_59, __p1_59, __p2_59, __p3_59) __extension__ ({ \
  14855.   uint16x4_t __ret_59; \
  14856.   uint16x4_t __s0_59 = __p0_59; \
  14857.   uint16x4_t __s1_59 = __p1_59; \
  14858.   uint16x4_t __s2_59 = __p2_59; \
  14859.   uint16x4_t __rev0_59;  __rev0_59 = __builtin_shufflevector(__s0_59, __s0_59, 3, 2, 1, 0); \
  14860.   uint16x4_t __rev1_59;  __rev1_59 = __builtin_shufflevector(__s1_59, __s1_59, 3, 2, 1, 0); \
  14861.   uint16x4_t __rev2_59;  __rev2_59 = __builtin_shufflevector(__s2_59, __s2_59, 3, 2, 1, 0); \
  14862.   __ret_59 = __rev0_59 + __rev1_59 * __noswap_splat_lane_u16(__rev2_59, __p3_59); \
  14863.   __ret_59 = __builtin_shufflevector(__ret_59, __ret_59, 3, 2, 1, 0); \
  14864.   __ret_59; \
  14865. })
  14866. #endif
  14867.  
  14868. #ifdef __LITTLE_ENDIAN__
  14869. #define vmla_lane_f32(__p0_60, __p1_60, __p2_60, __p3_60) __extension__ ({ \
  14870.   float32x2_t __ret_60; \
  14871.   float32x2_t __s0_60 = __p0_60; \
  14872.   float32x2_t __s1_60 = __p1_60; \
  14873.   float32x2_t __s2_60 = __p2_60; \
  14874.   __ret_60 = __s0_60 + __s1_60 * splat_lane_f32(__s2_60, __p3_60); \
  14875.   __ret_60; \
  14876. })
  14877. #else
  14878. #define vmla_lane_f32(__p0_61, __p1_61, __p2_61, __p3_61) __extension__ ({ \
  14879.   float32x2_t __ret_61; \
  14880.   float32x2_t __s0_61 = __p0_61; \
  14881.   float32x2_t __s1_61 = __p1_61; \
  14882.   float32x2_t __s2_61 = __p2_61; \
  14883.   float32x2_t __rev0_61;  __rev0_61 = __builtin_shufflevector(__s0_61, __s0_61, 1, 0); \
  14884.   float32x2_t __rev1_61;  __rev1_61 = __builtin_shufflevector(__s1_61, __s1_61, 1, 0); \
  14885.   float32x2_t __rev2_61;  __rev2_61 = __builtin_shufflevector(__s2_61, __s2_61, 1, 0); \
  14886.   __ret_61 = __rev0_61 + __rev1_61 * __noswap_splat_lane_f32(__rev2_61, __p3_61); \
  14887.   __ret_61 = __builtin_shufflevector(__ret_61, __ret_61, 1, 0); \
  14888.   __ret_61; \
  14889. })
  14890. #endif
  14891.  
  14892. #ifdef __LITTLE_ENDIAN__
  14893. #define vmla_lane_s32(__p0_62, __p1_62, __p2_62, __p3_62) __extension__ ({ \
  14894.   int32x2_t __ret_62; \
  14895.   int32x2_t __s0_62 = __p0_62; \
  14896.   int32x2_t __s1_62 = __p1_62; \
  14897.   int32x2_t __s2_62 = __p2_62; \
  14898.   __ret_62 = __s0_62 + __s1_62 * splat_lane_s32(__s2_62, __p3_62); \
  14899.   __ret_62; \
  14900. })
  14901. #else
  14902. #define vmla_lane_s32(__p0_63, __p1_63, __p2_63, __p3_63) __extension__ ({ \
  14903.   int32x2_t __ret_63; \
  14904.   int32x2_t __s0_63 = __p0_63; \
  14905.   int32x2_t __s1_63 = __p1_63; \
  14906.   int32x2_t __s2_63 = __p2_63; \
  14907.   int32x2_t __rev0_63;  __rev0_63 = __builtin_shufflevector(__s0_63, __s0_63, 1, 0); \
  14908.   int32x2_t __rev1_63;  __rev1_63 = __builtin_shufflevector(__s1_63, __s1_63, 1, 0); \
  14909.   int32x2_t __rev2_63;  __rev2_63 = __builtin_shufflevector(__s2_63, __s2_63, 1, 0); \
  14910.   __ret_63 = __rev0_63 + __rev1_63 * __noswap_splat_lane_s32(__rev2_63, __p3_63); \
  14911.   __ret_63 = __builtin_shufflevector(__ret_63, __ret_63, 1, 0); \
  14912.   __ret_63; \
  14913. })
  14914. #endif
  14915.  
  14916. #ifdef __LITTLE_ENDIAN__
  14917. #define vmla_lane_s16(__p0_64, __p1_64, __p2_64, __p3_64) __extension__ ({ \
  14918.   int16x4_t __ret_64; \
  14919.   int16x4_t __s0_64 = __p0_64; \
  14920.   int16x4_t __s1_64 = __p1_64; \
  14921.   int16x4_t __s2_64 = __p2_64; \
  14922.   __ret_64 = __s0_64 + __s1_64 * splat_lane_s16(__s2_64, __p3_64); \
  14923.   __ret_64; \
  14924. })
  14925. #else
  14926. #define vmla_lane_s16(__p0_65, __p1_65, __p2_65, __p3_65) __extension__ ({ \
  14927.   int16x4_t __ret_65; \
  14928.   int16x4_t __s0_65 = __p0_65; \
  14929.   int16x4_t __s1_65 = __p1_65; \
  14930.   int16x4_t __s2_65 = __p2_65; \
  14931.   int16x4_t __rev0_65;  __rev0_65 = __builtin_shufflevector(__s0_65, __s0_65, 3, 2, 1, 0); \
  14932.   int16x4_t __rev1_65;  __rev1_65 = __builtin_shufflevector(__s1_65, __s1_65, 3, 2, 1, 0); \
  14933.   int16x4_t __rev2_65;  __rev2_65 = __builtin_shufflevector(__s2_65, __s2_65, 3, 2, 1, 0); \
  14934.   __ret_65 = __rev0_65 + __rev1_65 * __noswap_splat_lane_s16(__rev2_65, __p3_65); \
  14935.   __ret_65 = __builtin_shufflevector(__ret_65, __ret_65, 3, 2, 1, 0); \
  14936.   __ret_65; \
  14937. })
  14938. #endif
  14939.  
  14940. #ifdef __LITTLE_ENDIAN__
  14941. __ai uint32x4_t vmlaq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
  14942.   uint32x4_t __ret;
  14943.   __ret = __p0 + __p1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
  14944.   return __ret;
  14945. }
  14946. #else
  14947. __ai uint32x4_t vmlaq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
  14948.   uint32x4_t __ret;
  14949.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14950.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14951.   __ret = __rev0 + __rev1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
  14952.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14953.   return __ret;
  14954. }
  14955. #endif
  14956.  
  14957. #ifdef __LITTLE_ENDIAN__
  14958. __ai uint16x8_t vmlaq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
  14959.   uint16x8_t __ret;
  14960.   __ret = __p0 + __p1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
  14961.   return __ret;
  14962. }
  14963. #else
  14964. __ai uint16x8_t vmlaq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
  14965.   uint16x8_t __ret;
  14966.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  14967.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  14968.   __ret = __rev0 + __rev1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
  14969.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  14970.   return __ret;
  14971. }
  14972. #endif
  14973.  
  14974. #ifdef __LITTLE_ENDIAN__
  14975. __ai float32x4_t vmlaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
  14976.   float32x4_t __ret;
  14977.   __ret = __p0 + __p1 * (float32x4_t) {__p2, __p2, __p2, __p2};
  14978.   return __ret;
  14979. }
  14980. #else
  14981. __ai float32x4_t vmlaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
  14982.   float32x4_t __ret;
  14983.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  14984.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  14985.   __ret = __rev0 + __rev1 * (float32x4_t) {__p2, __p2, __p2, __p2};
  14986.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  14987.   return __ret;
  14988. }
  14989. #endif
  14990.  
  14991. #ifdef __LITTLE_ENDIAN__
  14992. __ai int32x4_t vmlaq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
  14993.   int32x4_t __ret;
  14994.   __ret = __p0 + __p1 * (int32x4_t) {__p2, __p2, __p2, __p2};
  14995.   return __ret;
  14996. }
  14997. #else
  14998. __ai int32x4_t vmlaq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
  14999.   int32x4_t __ret;
  15000.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15001.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15002.   __ret = __rev0 + __rev1 * (int32x4_t) {__p2, __p2, __p2, __p2};
  15003.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15004.   return __ret;
  15005. }
  15006. #endif
  15007.  
  15008. #ifdef __LITTLE_ENDIAN__
  15009. __ai int16x8_t vmlaq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
  15010.   int16x8_t __ret;
  15011.   __ret = __p0 + __p1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
  15012.   return __ret;
  15013. }
  15014. #else
  15015. __ai int16x8_t vmlaq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
  15016.   int16x8_t __ret;
  15017.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  15018.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  15019.   __ret = __rev0 + __rev1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
  15020.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15021.   return __ret;
  15022. }
  15023. #endif
  15024.  
  15025. #ifdef __LITTLE_ENDIAN__
  15026. __ai uint32x2_t vmla_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  15027.   uint32x2_t __ret;
  15028.   __ret = __p0 + __p1 * (uint32x2_t) {__p2, __p2};
  15029.   return __ret;
  15030. }
  15031. #else
  15032. __ai uint32x2_t vmla_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  15033.   uint32x2_t __ret;
  15034.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  15035.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  15036.   __ret = __rev0 + __rev1 * (uint32x2_t) {__p2, __p2};
  15037.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15038.   return __ret;
  15039. }
  15040. #endif
  15041.  
  15042. #ifdef __LITTLE_ENDIAN__
  15043. __ai uint16x4_t vmla_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  15044.   uint16x4_t __ret;
  15045.   __ret = __p0 + __p1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
  15046.   return __ret;
  15047. }
  15048. #else
  15049. __ai uint16x4_t vmla_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  15050.   uint16x4_t __ret;
  15051.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15052.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15053.   __ret = __rev0 + __rev1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
  15054.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15055.   return __ret;
  15056. }
  15057. #endif
  15058.  
  15059. #ifdef __LITTLE_ENDIAN__
  15060. __ai float32x2_t vmla_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
  15061.   float32x2_t __ret;
  15062.   __ret = __p0 + __p1 * (float32x2_t) {__p2, __p2};
  15063.   return __ret;
  15064. }
  15065. #else
  15066. __ai float32x2_t vmla_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
  15067.   float32x2_t __ret;
  15068.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  15069.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  15070.   __ret = __rev0 + __rev1 * (float32x2_t) {__p2, __p2};
  15071.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15072.   return __ret;
  15073. }
  15074. #endif
  15075.  
  15076. #ifdef __LITTLE_ENDIAN__
  15077. __ai int32x2_t vmla_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
  15078.   int32x2_t __ret;
  15079.   __ret = __p0 + __p1 * (int32x2_t) {__p2, __p2};
  15080.   return __ret;
  15081. }
  15082. #else
  15083. __ai int32x2_t vmla_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
  15084.   int32x2_t __ret;
  15085.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  15086.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  15087.   __ret = __rev0 + __rev1 * (int32x2_t) {__p2, __p2};
  15088.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15089.   return __ret;
  15090. }
  15091. #endif
  15092.  
  15093. #ifdef __LITTLE_ENDIAN__
  15094. __ai int16x4_t vmla_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
  15095.   int16x4_t __ret;
  15096.   __ret = __p0 + __p1 * (int16x4_t) {__p2, __p2, __p2, __p2};
  15097.   return __ret;
  15098. }
  15099. #else
  15100. __ai int16x4_t vmla_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
  15101.   int16x4_t __ret;
  15102.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15103.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15104.   __ret = __rev0 + __rev1 * (int16x4_t) {__p2, __p2, __p2, __p2};
  15105.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15106.   return __ret;
  15107. }
  15108. #endif
  15109.  
  15110. #ifdef __LITTLE_ENDIAN__
  15111. __ai uint8x16_t vmlsq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  15112.   uint8x16_t __ret;
  15113.   __ret = __p0 - __p1 * __p2;
  15114.   return __ret;
  15115. }
  15116. #else
  15117. __ai uint8x16_t vmlsq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  15118.   uint8x16_t __ret;
  15119.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15120.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15121.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15122.   __ret = __rev0 - __rev1 * __rev2;
  15123.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15124.   return __ret;
  15125. }
  15126. #endif
  15127.  
  15128. #ifdef __LITTLE_ENDIAN__
  15129. __ai uint32x4_t vmlsq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  15130.   uint32x4_t __ret;
  15131.   __ret = __p0 - __p1 * __p2;
  15132.   return __ret;
  15133. }
  15134. #else
  15135. __ai uint32x4_t vmlsq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  15136.   uint32x4_t __ret;
  15137.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15138.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15139.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  15140.   __ret = __rev0 - __rev1 * __rev2;
  15141.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15142.   return __ret;
  15143. }
  15144. #endif
  15145.  
  15146. #ifdef __LITTLE_ENDIAN__
  15147. __ai uint16x8_t vmlsq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  15148.   uint16x8_t __ret;
  15149.   __ret = __p0 - __p1 * __p2;
  15150.   return __ret;
  15151. }
  15152. #else
  15153. __ai uint16x8_t vmlsq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  15154.   uint16x8_t __ret;
  15155.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  15156.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  15157.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  15158.   __ret = __rev0 - __rev1 * __rev2;
  15159.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15160.   return __ret;
  15161. }
  15162. #endif
  15163.  
  15164. #ifdef __LITTLE_ENDIAN__
  15165. __ai int8x16_t vmlsq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  15166.   int8x16_t __ret;
  15167.   __ret = __p0 - __p1 * __p2;
  15168.   return __ret;
  15169. }
  15170. #else
  15171. __ai int8x16_t vmlsq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  15172.   int8x16_t __ret;
  15173.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15174.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15175.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15176.   __ret = __rev0 - __rev1 * __rev2;
  15177.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15178.   return __ret;
  15179. }
  15180. #endif
  15181.  
  15182. #ifdef __LITTLE_ENDIAN__
  15183. __ai float32x4_t vmlsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  15184.   float32x4_t __ret;
  15185.   __ret = __p0 - __p1 * __p2;
  15186.   return __ret;
  15187. }
  15188. #else
  15189. __ai float32x4_t vmlsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  15190.   float32x4_t __ret;
  15191.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15192.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15193.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  15194.   __ret = __rev0 - __rev1 * __rev2;
  15195.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15196.   return __ret;
  15197. }
  15198. #endif
  15199.  
  15200. #ifdef __LITTLE_ENDIAN__
  15201. __ai int32x4_t vmlsq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  15202.   int32x4_t __ret;
  15203.   __ret = __p0 - __p1 * __p2;
  15204.   return __ret;
  15205. }
  15206. #else
  15207. __ai int32x4_t vmlsq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  15208.   int32x4_t __ret;
  15209.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15210.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15211.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  15212.   __ret = __rev0 - __rev1 * __rev2;
  15213.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15214.   return __ret;
  15215. }
  15216. #endif
  15217.  
  15218. #ifdef __LITTLE_ENDIAN__
  15219. __ai int16x8_t vmlsq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  15220.   int16x8_t __ret;
  15221.   __ret = __p0 - __p1 * __p2;
  15222.   return __ret;
  15223. }
  15224. #else
  15225. __ai int16x8_t vmlsq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  15226.   int16x8_t __ret;
  15227.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  15228.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  15229.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  15230.   __ret = __rev0 - __rev1 * __rev2;
  15231.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15232.   return __ret;
  15233. }
  15234. #endif
  15235.  
  15236. #ifdef __LITTLE_ENDIAN__
  15237. __ai uint8x8_t vmls_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  15238.   uint8x8_t __ret;
  15239.   __ret = __p0 - __p1 * __p2;
  15240.   return __ret;
  15241. }
  15242. #else
  15243. __ai uint8x8_t vmls_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  15244.   uint8x8_t __ret;
  15245.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  15246.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  15247.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  15248.   __ret = __rev0 - __rev1 * __rev2;
  15249.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15250.   return __ret;
  15251. }
  15252. #endif
  15253.  
  15254. #ifdef __LITTLE_ENDIAN__
  15255. __ai uint32x2_t vmls_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  15256.   uint32x2_t __ret;
  15257.   __ret = __p0 - __p1 * __p2;
  15258.   return __ret;
  15259. }
  15260. #else
  15261. __ai uint32x2_t vmls_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  15262.   uint32x2_t __ret;
  15263.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  15264.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  15265.   uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  15266.   __ret = __rev0 - __rev1 * __rev2;
  15267.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15268.   return __ret;
  15269. }
  15270. #endif
  15271.  
  15272. #ifdef __LITTLE_ENDIAN__
  15273. __ai uint16x4_t vmls_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  15274.   uint16x4_t __ret;
  15275.   __ret = __p0 - __p1 * __p2;
  15276.   return __ret;
  15277. }
  15278. #else
  15279. __ai uint16x4_t vmls_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  15280.   uint16x4_t __ret;
  15281.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15282.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15283.   uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  15284.   __ret = __rev0 - __rev1 * __rev2;
  15285.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15286.   return __ret;
  15287. }
  15288. #endif
  15289.  
  15290. #ifdef __LITTLE_ENDIAN__
  15291. __ai int8x8_t vmls_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  15292.   int8x8_t __ret;
  15293.   __ret = __p0 - __p1 * __p2;
  15294.   return __ret;
  15295. }
  15296. #else
  15297. __ai int8x8_t vmls_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  15298.   int8x8_t __ret;
  15299.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  15300.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  15301.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  15302.   __ret = __rev0 - __rev1 * __rev2;
  15303.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15304.   return __ret;
  15305. }
  15306. #endif
  15307.  
  15308. #ifdef __LITTLE_ENDIAN__
  15309. __ai float32x2_t vmls_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  15310.   float32x2_t __ret;
  15311.   __ret = __p0 - __p1 * __p2;
  15312.   return __ret;
  15313. }
  15314. #else
  15315. __ai float32x2_t vmls_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  15316.   float32x2_t __ret;
  15317.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  15318.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  15319.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  15320.   __ret = __rev0 - __rev1 * __rev2;
  15321.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15322.   return __ret;
  15323. }
  15324. #endif
  15325.  
  15326. #ifdef __LITTLE_ENDIAN__
  15327. __ai int32x2_t vmls_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  15328.   int32x2_t __ret;
  15329.   __ret = __p0 - __p1 * __p2;
  15330.   return __ret;
  15331. }
  15332. #else
  15333. __ai int32x2_t vmls_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  15334.   int32x2_t __ret;
  15335.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  15336.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  15337.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  15338.   __ret = __rev0 - __rev1 * __rev2;
  15339.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15340.   return __ret;
  15341. }
  15342. #endif
  15343.  
  15344. #ifdef __LITTLE_ENDIAN__
  15345. __ai int16x4_t vmls_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  15346.   int16x4_t __ret;
  15347.   __ret = __p0 - __p1 * __p2;
  15348.   return __ret;
  15349. }
  15350. #else
  15351. __ai int16x4_t vmls_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  15352.   int16x4_t __ret;
  15353.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15354.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15355.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  15356.   __ret = __rev0 - __rev1 * __rev2;
  15357.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15358.   return __ret;
  15359. }
  15360. #endif
  15361.  
  15362. #ifdef __LITTLE_ENDIAN__
  15363. #define vmlsq_lane_u32(__p0_66, __p1_66, __p2_66, __p3_66) __extension__ ({ \
  15364.   uint32x4_t __ret_66; \
  15365.   uint32x4_t __s0_66 = __p0_66; \
  15366.   uint32x4_t __s1_66 = __p1_66; \
  15367.   uint32x2_t __s2_66 = __p2_66; \
  15368.   __ret_66 = __s0_66 - __s1_66 * splatq_lane_u32(__s2_66, __p3_66); \
  15369.   __ret_66; \
  15370. })
  15371. #else
  15372. #define vmlsq_lane_u32(__p0_67, __p1_67, __p2_67, __p3_67) __extension__ ({ \
  15373.   uint32x4_t __ret_67; \
  15374.   uint32x4_t __s0_67 = __p0_67; \
  15375.   uint32x4_t __s1_67 = __p1_67; \
  15376.   uint32x2_t __s2_67 = __p2_67; \
  15377.   uint32x4_t __rev0_67;  __rev0_67 = __builtin_shufflevector(__s0_67, __s0_67, 3, 2, 1, 0); \
  15378.   uint32x4_t __rev1_67;  __rev1_67 = __builtin_shufflevector(__s1_67, __s1_67, 3, 2, 1, 0); \
  15379.   uint32x2_t __rev2_67;  __rev2_67 = __builtin_shufflevector(__s2_67, __s2_67, 1, 0); \
  15380.   __ret_67 = __rev0_67 - __rev1_67 * __noswap_splatq_lane_u32(__rev2_67, __p3_67); \
  15381.   __ret_67 = __builtin_shufflevector(__ret_67, __ret_67, 3, 2, 1, 0); \
  15382.   __ret_67; \
  15383. })
  15384. #endif
  15385.  
  15386. #ifdef __LITTLE_ENDIAN__
  15387. #define vmlsq_lane_u16(__p0_68, __p1_68, __p2_68, __p3_68) __extension__ ({ \
  15388.   uint16x8_t __ret_68; \
  15389.   uint16x8_t __s0_68 = __p0_68; \
  15390.   uint16x8_t __s1_68 = __p1_68; \
  15391.   uint16x4_t __s2_68 = __p2_68; \
  15392.   __ret_68 = __s0_68 - __s1_68 * splatq_lane_u16(__s2_68, __p3_68); \
  15393.   __ret_68; \
  15394. })
  15395. #else
  15396. #define vmlsq_lane_u16(__p0_69, __p1_69, __p2_69, __p3_69) __extension__ ({ \
  15397.   uint16x8_t __ret_69; \
  15398.   uint16x8_t __s0_69 = __p0_69; \
  15399.   uint16x8_t __s1_69 = __p1_69; \
  15400.   uint16x4_t __s2_69 = __p2_69; \
  15401.   uint16x8_t __rev0_69;  __rev0_69 = __builtin_shufflevector(__s0_69, __s0_69, 7, 6, 5, 4, 3, 2, 1, 0); \
  15402.   uint16x8_t __rev1_69;  __rev1_69 = __builtin_shufflevector(__s1_69, __s1_69, 7, 6, 5, 4, 3, 2, 1, 0); \
  15403.   uint16x4_t __rev2_69;  __rev2_69 = __builtin_shufflevector(__s2_69, __s2_69, 3, 2, 1, 0); \
  15404.   __ret_69 = __rev0_69 - __rev1_69 * __noswap_splatq_lane_u16(__rev2_69, __p3_69); \
  15405.   __ret_69 = __builtin_shufflevector(__ret_69, __ret_69, 7, 6, 5, 4, 3, 2, 1, 0); \
  15406.   __ret_69; \
  15407. })
  15408. #endif
  15409.  
  15410. #ifdef __LITTLE_ENDIAN__
  15411. #define vmlsq_lane_f32(__p0_70, __p1_70, __p2_70, __p3_70) __extension__ ({ \
  15412.   float32x4_t __ret_70; \
  15413.   float32x4_t __s0_70 = __p0_70; \
  15414.   float32x4_t __s1_70 = __p1_70; \
  15415.   float32x2_t __s2_70 = __p2_70; \
  15416.   __ret_70 = __s0_70 - __s1_70 * splatq_lane_f32(__s2_70, __p3_70); \
  15417.   __ret_70; \
  15418. })
  15419. #else
  15420. #define vmlsq_lane_f32(__p0_71, __p1_71, __p2_71, __p3_71) __extension__ ({ \
  15421.   float32x4_t __ret_71; \
  15422.   float32x4_t __s0_71 = __p0_71; \
  15423.   float32x4_t __s1_71 = __p1_71; \
  15424.   float32x2_t __s2_71 = __p2_71; \
  15425.   float32x4_t __rev0_71;  __rev0_71 = __builtin_shufflevector(__s0_71, __s0_71, 3, 2, 1, 0); \
  15426.   float32x4_t __rev1_71;  __rev1_71 = __builtin_shufflevector(__s1_71, __s1_71, 3, 2, 1, 0); \
  15427.   float32x2_t __rev2_71;  __rev2_71 = __builtin_shufflevector(__s2_71, __s2_71, 1, 0); \
  15428.   __ret_71 = __rev0_71 - __rev1_71 * __noswap_splatq_lane_f32(__rev2_71, __p3_71); \
  15429.   __ret_71 = __builtin_shufflevector(__ret_71, __ret_71, 3, 2, 1, 0); \
  15430.   __ret_71; \
  15431. })
  15432. #endif
  15433.  
  15434. #ifdef __LITTLE_ENDIAN__
  15435. #define vmlsq_lane_s32(__p0_72, __p1_72, __p2_72, __p3_72) __extension__ ({ \
  15436.   int32x4_t __ret_72; \
  15437.   int32x4_t __s0_72 = __p0_72; \
  15438.   int32x4_t __s1_72 = __p1_72; \
  15439.   int32x2_t __s2_72 = __p2_72; \
  15440.   __ret_72 = __s0_72 - __s1_72 * splatq_lane_s32(__s2_72, __p3_72); \
  15441.   __ret_72; \
  15442. })
  15443. #else
  15444. #define vmlsq_lane_s32(__p0_73, __p1_73, __p2_73, __p3_73) __extension__ ({ \
  15445.   int32x4_t __ret_73; \
  15446.   int32x4_t __s0_73 = __p0_73; \
  15447.   int32x4_t __s1_73 = __p1_73; \
  15448.   int32x2_t __s2_73 = __p2_73; \
  15449.   int32x4_t __rev0_73;  __rev0_73 = __builtin_shufflevector(__s0_73, __s0_73, 3, 2, 1, 0); \
  15450.   int32x4_t __rev1_73;  __rev1_73 = __builtin_shufflevector(__s1_73, __s1_73, 3, 2, 1, 0); \
  15451.   int32x2_t __rev2_73;  __rev2_73 = __builtin_shufflevector(__s2_73, __s2_73, 1, 0); \
  15452.   __ret_73 = __rev0_73 - __rev1_73 * __noswap_splatq_lane_s32(__rev2_73, __p3_73); \
  15453.   __ret_73 = __builtin_shufflevector(__ret_73, __ret_73, 3, 2, 1, 0); \
  15454.   __ret_73; \
  15455. })
  15456. #endif
  15457.  
  15458. #ifdef __LITTLE_ENDIAN__
  15459. #define vmlsq_lane_s16(__p0_74, __p1_74, __p2_74, __p3_74) __extension__ ({ \
  15460.   int16x8_t __ret_74; \
  15461.   int16x8_t __s0_74 = __p0_74; \
  15462.   int16x8_t __s1_74 = __p1_74; \
  15463.   int16x4_t __s2_74 = __p2_74; \
  15464.   __ret_74 = __s0_74 - __s1_74 * splatq_lane_s16(__s2_74, __p3_74); \
  15465.   __ret_74; \
  15466. })
  15467. #else
  15468. #define vmlsq_lane_s16(__p0_75, __p1_75, __p2_75, __p3_75) __extension__ ({ \
  15469.   int16x8_t __ret_75; \
  15470.   int16x8_t __s0_75 = __p0_75; \
  15471.   int16x8_t __s1_75 = __p1_75; \
  15472.   int16x4_t __s2_75 = __p2_75; \
  15473.   int16x8_t __rev0_75;  __rev0_75 = __builtin_shufflevector(__s0_75, __s0_75, 7, 6, 5, 4, 3, 2, 1, 0); \
  15474.   int16x8_t __rev1_75;  __rev1_75 = __builtin_shufflevector(__s1_75, __s1_75, 7, 6, 5, 4, 3, 2, 1, 0); \
  15475.   int16x4_t __rev2_75;  __rev2_75 = __builtin_shufflevector(__s2_75, __s2_75, 3, 2, 1, 0); \
  15476.   __ret_75 = __rev0_75 - __rev1_75 * __noswap_splatq_lane_s16(__rev2_75, __p3_75); \
  15477.   __ret_75 = __builtin_shufflevector(__ret_75, __ret_75, 7, 6, 5, 4, 3, 2, 1, 0); \
  15478.   __ret_75; \
  15479. })
  15480. #endif
  15481.  
  15482. #ifdef __LITTLE_ENDIAN__
  15483. #define vmls_lane_u32(__p0_76, __p1_76, __p2_76, __p3_76) __extension__ ({ \
  15484.   uint32x2_t __ret_76; \
  15485.   uint32x2_t __s0_76 = __p0_76; \
  15486.   uint32x2_t __s1_76 = __p1_76; \
  15487.   uint32x2_t __s2_76 = __p2_76; \
  15488.   __ret_76 = __s0_76 - __s1_76 * splat_lane_u32(__s2_76, __p3_76); \
  15489.   __ret_76; \
  15490. })
  15491. #else
  15492. #define vmls_lane_u32(__p0_77, __p1_77, __p2_77, __p3_77) __extension__ ({ \
  15493.   uint32x2_t __ret_77; \
  15494.   uint32x2_t __s0_77 = __p0_77; \
  15495.   uint32x2_t __s1_77 = __p1_77; \
  15496.   uint32x2_t __s2_77 = __p2_77; \
  15497.   uint32x2_t __rev0_77;  __rev0_77 = __builtin_shufflevector(__s0_77, __s0_77, 1, 0); \
  15498.   uint32x2_t __rev1_77;  __rev1_77 = __builtin_shufflevector(__s1_77, __s1_77, 1, 0); \
  15499.   uint32x2_t __rev2_77;  __rev2_77 = __builtin_shufflevector(__s2_77, __s2_77, 1, 0); \
  15500.   __ret_77 = __rev0_77 - __rev1_77 * __noswap_splat_lane_u32(__rev2_77, __p3_77); \
  15501.   __ret_77 = __builtin_shufflevector(__ret_77, __ret_77, 1, 0); \
  15502.   __ret_77; \
  15503. })
  15504. #endif
  15505.  
  15506. #ifdef __LITTLE_ENDIAN__
  15507. #define vmls_lane_u16(__p0_78, __p1_78, __p2_78, __p3_78) __extension__ ({ \
  15508.   uint16x4_t __ret_78; \
  15509.   uint16x4_t __s0_78 = __p0_78; \
  15510.   uint16x4_t __s1_78 = __p1_78; \
  15511.   uint16x4_t __s2_78 = __p2_78; \
  15512.   __ret_78 = __s0_78 - __s1_78 * splat_lane_u16(__s2_78, __p3_78); \
  15513.   __ret_78; \
  15514. })
  15515. #else
  15516. #define vmls_lane_u16(__p0_79, __p1_79, __p2_79, __p3_79) __extension__ ({ \
  15517.   uint16x4_t __ret_79; \
  15518.   uint16x4_t __s0_79 = __p0_79; \
  15519.   uint16x4_t __s1_79 = __p1_79; \
  15520.   uint16x4_t __s2_79 = __p2_79; \
  15521.   uint16x4_t __rev0_79;  __rev0_79 = __builtin_shufflevector(__s0_79, __s0_79, 3, 2, 1, 0); \
  15522.   uint16x4_t __rev1_79;  __rev1_79 = __builtin_shufflevector(__s1_79, __s1_79, 3, 2, 1, 0); \
  15523.   uint16x4_t __rev2_79;  __rev2_79 = __builtin_shufflevector(__s2_79, __s2_79, 3, 2, 1, 0); \
  15524.   __ret_79 = __rev0_79 - __rev1_79 * __noswap_splat_lane_u16(__rev2_79, __p3_79); \
  15525.   __ret_79 = __builtin_shufflevector(__ret_79, __ret_79, 3, 2, 1, 0); \
  15526.   __ret_79; \
  15527. })
  15528. #endif
  15529.  
  15530. #ifdef __LITTLE_ENDIAN__
  15531. #define vmls_lane_f32(__p0_80, __p1_80, __p2_80, __p3_80) __extension__ ({ \
  15532.   float32x2_t __ret_80; \
  15533.   float32x2_t __s0_80 = __p0_80; \
  15534.   float32x2_t __s1_80 = __p1_80; \
  15535.   float32x2_t __s2_80 = __p2_80; \
  15536.   __ret_80 = __s0_80 - __s1_80 * splat_lane_f32(__s2_80, __p3_80); \
  15537.   __ret_80; \
  15538. })
  15539. #else
  15540. #define vmls_lane_f32(__p0_81, __p1_81, __p2_81, __p3_81) __extension__ ({ \
  15541.   float32x2_t __ret_81; \
  15542.   float32x2_t __s0_81 = __p0_81; \
  15543.   float32x2_t __s1_81 = __p1_81; \
  15544.   float32x2_t __s2_81 = __p2_81; \
  15545.   float32x2_t __rev0_81;  __rev0_81 = __builtin_shufflevector(__s0_81, __s0_81, 1, 0); \
  15546.   float32x2_t __rev1_81;  __rev1_81 = __builtin_shufflevector(__s1_81, __s1_81, 1, 0); \
  15547.   float32x2_t __rev2_81;  __rev2_81 = __builtin_shufflevector(__s2_81, __s2_81, 1, 0); \
  15548.   __ret_81 = __rev0_81 - __rev1_81 * __noswap_splat_lane_f32(__rev2_81, __p3_81); \
  15549.   __ret_81 = __builtin_shufflevector(__ret_81, __ret_81, 1, 0); \
  15550.   __ret_81; \
  15551. })
  15552. #endif
  15553.  
  15554. #ifdef __LITTLE_ENDIAN__
  15555. #define vmls_lane_s32(__p0_82, __p1_82, __p2_82, __p3_82) __extension__ ({ \
  15556.   int32x2_t __ret_82; \
  15557.   int32x2_t __s0_82 = __p0_82; \
  15558.   int32x2_t __s1_82 = __p1_82; \
  15559.   int32x2_t __s2_82 = __p2_82; \
  15560.   __ret_82 = __s0_82 - __s1_82 * splat_lane_s32(__s2_82, __p3_82); \
  15561.   __ret_82; \
  15562. })
  15563. #else
  15564. #define vmls_lane_s32(__p0_83, __p1_83, __p2_83, __p3_83) __extension__ ({ \
  15565.   int32x2_t __ret_83; \
  15566.   int32x2_t __s0_83 = __p0_83; \
  15567.   int32x2_t __s1_83 = __p1_83; \
  15568.   int32x2_t __s2_83 = __p2_83; \
  15569.   int32x2_t __rev0_83;  __rev0_83 = __builtin_shufflevector(__s0_83, __s0_83, 1, 0); \
  15570.   int32x2_t __rev1_83;  __rev1_83 = __builtin_shufflevector(__s1_83, __s1_83, 1, 0); \
  15571.   int32x2_t __rev2_83;  __rev2_83 = __builtin_shufflevector(__s2_83, __s2_83, 1, 0); \
  15572.   __ret_83 = __rev0_83 - __rev1_83 * __noswap_splat_lane_s32(__rev2_83, __p3_83); \
  15573.   __ret_83 = __builtin_shufflevector(__ret_83, __ret_83, 1, 0); \
  15574.   __ret_83; \
  15575. })
  15576. #endif
  15577.  
  15578. #ifdef __LITTLE_ENDIAN__
  15579. #define vmls_lane_s16(__p0_84, __p1_84, __p2_84, __p3_84) __extension__ ({ \
  15580.   int16x4_t __ret_84; \
  15581.   int16x4_t __s0_84 = __p0_84; \
  15582.   int16x4_t __s1_84 = __p1_84; \
  15583.   int16x4_t __s2_84 = __p2_84; \
  15584.   __ret_84 = __s0_84 - __s1_84 * splat_lane_s16(__s2_84, __p3_84); \
  15585.   __ret_84; \
  15586. })
  15587. #else
  15588. #define vmls_lane_s16(__p0_85, __p1_85, __p2_85, __p3_85) __extension__ ({ \
  15589.   int16x4_t __ret_85; \
  15590.   int16x4_t __s0_85 = __p0_85; \
  15591.   int16x4_t __s1_85 = __p1_85; \
  15592.   int16x4_t __s2_85 = __p2_85; \
  15593.   int16x4_t __rev0_85;  __rev0_85 = __builtin_shufflevector(__s0_85, __s0_85, 3, 2, 1, 0); \
  15594.   int16x4_t __rev1_85;  __rev1_85 = __builtin_shufflevector(__s1_85, __s1_85, 3, 2, 1, 0); \
  15595.   int16x4_t __rev2_85;  __rev2_85 = __builtin_shufflevector(__s2_85, __s2_85, 3, 2, 1, 0); \
  15596.   __ret_85 = __rev0_85 - __rev1_85 * __noswap_splat_lane_s16(__rev2_85, __p3_85); \
  15597.   __ret_85 = __builtin_shufflevector(__ret_85, __ret_85, 3, 2, 1, 0); \
  15598.   __ret_85; \
  15599. })
  15600. #endif
  15601.  
  15602. #ifdef __LITTLE_ENDIAN__
  15603. __ai uint32x4_t vmlsq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
  15604.   uint32x4_t __ret;
  15605.   __ret = __p0 - __p1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
  15606.   return __ret;
  15607. }
  15608. #else
  15609. __ai uint32x4_t vmlsq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
  15610.   uint32x4_t __ret;
  15611.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15612.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15613.   __ret = __rev0 - __rev1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
  15614.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15615.   return __ret;
  15616. }
  15617. #endif
  15618.  
  15619. #ifdef __LITTLE_ENDIAN__
  15620. __ai uint16x8_t vmlsq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
  15621.   uint16x8_t __ret;
  15622.   __ret = __p0 - __p1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
  15623.   return __ret;
  15624. }
  15625. #else
  15626. __ai uint16x8_t vmlsq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
  15627.   uint16x8_t __ret;
  15628.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  15629.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  15630.   __ret = __rev0 - __rev1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
  15631.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15632.   return __ret;
  15633. }
  15634. #endif
  15635.  
  15636. #ifdef __LITTLE_ENDIAN__
  15637. __ai float32x4_t vmlsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
  15638.   float32x4_t __ret;
  15639.   __ret = __p0 - __p1 * (float32x4_t) {__p2, __p2, __p2, __p2};
  15640.   return __ret;
  15641. }
  15642. #else
  15643. __ai float32x4_t vmlsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
  15644.   float32x4_t __ret;
  15645.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15646.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15647.   __ret = __rev0 - __rev1 * (float32x4_t) {__p2, __p2, __p2, __p2};
  15648.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15649.   return __ret;
  15650. }
  15651. #endif
  15652.  
  15653. #ifdef __LITTLE_ENDIAN__
  15654. __ai int32x4_t vmlsq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
  15655.   int32x4_t __ret;
  15656.   __ret = __p0 - __p1 * (int32x4_t) {__p2, __p2, __p2, __p2};
  15657.   return __ret;
  15658. }
  15659. #else
  15660. __ai int32x4_t vmlsq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
  15661.   int32x4_t __ret;
  15662.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15663.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15664.   __ret = __rev0 - __rev1 * (int32x4_t) {__p2, __p2, __p2, __p2};
  15665.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15666.   return __ret;
  15667. }
  15668. #endif
  15669.  
  15670. #ifdef __LITTLE_ENDIAN__
  15671. __ai int16x8_t vmlsq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
  15672.   int16x8_t __ret;
  15673.   __ret = __p0 - __p1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
  15674.   return __ret;
  15675. }
  15676. #else
  15677. __ai int16x8_t vmlsq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
  15678.   int16x8_t __ret;
  15679.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  15680.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  15681.   __ret = __rev0 - __rev1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
  15682.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15683.   return __ret;
  15684. }
  15685. #endif
  15686.  
  15687. #ifdef __LITTLE_ENDIAN__
  15688. __ai uint32x2_t vmls_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  15689.   uint32x2_t __ret;
  15690.   __ret = __p0 - __p1 * (uint32x2_t) {__p2, __p2};
  15691.   return __ret;
  15692. }
  15693. #else
  15694. __ai uint32x2_t vmls_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  15695.   uint32x2_t __ret;
  15696.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  15697.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  15698.   __ret = __rev0 - __rev1 * (uint32x2_t) {__p2, __p2};
  15699.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15700.   return __ret;
  15701. }
  15702. #endif
  15703.  
  15704. #ifdef __LITTLE_ENDIAN__
  15705. __ai uint16x4_t vmls_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  15706.   uint16x4_t __ret;
  15707.   __ret = __p0 - __p1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
  15708.   return __ret;
  15709. }
  15710. #else
  15711. __ai uint16x4_t vmls_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  15712.   uint16x4_t __ret;
  15713.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15714.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15715.   __ret = __rev0 - __rev1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
  15716.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15717.   return __ret;
  15718. }
  15719. #endif
  15720.  
  15721. #ifdef __LITTLE_ENDIAN__
  15722. __ai float32x2_t vmls_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
  15723.   float32x2_t __ret;
  15724.   __ret = __p0 - __p1 * (float32x2_t) {__p2, __p2};
  15725.   return __ret;
  15726. }
  15727. #else
  15728. __ai float32x2_t vmls_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
  15729.   float32x2_t __ret;
  15730.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  15731.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  15732.   __ret = __rev0 - __rev1 * (float32x2_t) {__p2, __p2};
  15733.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15734.   return __ret;
  15735. }
  15736. #endif
  15737.  
  15738. #ifdef __LITTLE_ENDIAN__
  15739. __ai int32x2_t vmls_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
  15740.   int32x2_t __ret;
  15741.   __ret = __p0 - __p1 * (int32x2_t) {__p2, __p2};
  15742.   return __ret;
  15743. }
  15744. #else
  15745. __ai int32x2_t vmls_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
  15746.   int32x2_t __ret;
  15747.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  15748.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  15749.   __ret = __rev0 - __rev1 * (int32x2_t) {__p2, __p2};
  15750.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15751.   return __ret;
  15752. }
  15753. #endif
  15754.  
  15755. #ifdef __LITTLE_ENDIAN__
  15756. __ai int16x4_t vmls_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
  15757.   int16x4_t __ret;
  15758.   __ret = __p0 - __p1 * (int16x4_t) {__p2, __p2, __p2, __p2};
  15759.   return __ret;
  15760. }
  15761. #else
  15762. __ai int16x4_t vmls_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
  15763.   int16x4_t __ret;
  15764.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  15765.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  15766.   __ret = __rev0 - __rev1 * (int16x4_t) {__p2, __p2, __p2, __p2};
  15767.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15768.   return __ret;
  15769. }
  15770. #endif
  15771.  
  15772. #ifdef __LITTLE_ENDIAN__
  15773. __ai poly8x8_t vmov_n_p8(poly8_t __p0) {
  15774.   poly8x8_t __ret;
  15775.   __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15776.   return __ret;
  15777. }
  15778. #else
  15779. __ai poly8x8_t vmov_n_p8(poly8_t __p0) {
  15780.   poly8x8_t __ret;
  15781.   __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15782.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15783.   return __ret;
  15784. }
  15785. #endif
  15786.  
  15787. #ifdef __LITTLE_ENDIAN__
  15788. __ai poly16x4_t vmov_n_p16(poly16_t __p0) {
  15789.   poly16x4_t __ret;
  15790.   __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
  15791.   return __ret;
  15792. }
  15793. #else
  15794. __ai poly16x4_t vmov_n_p16(poly16_t __p0) {
  15795.   poly16x4_t __ret;
  15796.   __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
  15797.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15798.   return __ret;
  15799. }
  15800. #endif
  15801.  
  15802. #ifdef __LITTLE_ENDIAN__
  15803. __ai poly8x16_t vmovq_n_p8(poly8_t __p0) {
  15804.   poly8x16_t __ret;
  15805.   __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15806.   return __ret;
  15807. }
  15808. #else
  15809. __ai poly8x16_t vmovq_n_p8(poly8_t __p0) {
  15810.   poly8x16_t __ret;
  15811.   __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15812.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15813.   return __ret;
  15814. }
  15815. #endif
  15816.  
  15817. #ifdef __LITTLE_ENDIAN__
  15818. __ai poly16x8_t vmovq_n_p16(poly16_t __p0) {
  15819.   poly16x8_t __ret;
  15820.   __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15821.   return __ret;
  15822. }
  15823. #else
  15824. __ai poly16x8_t vmovq_n_p16(poly16_t __p0) {
  15825.   poly16x8_t __ret;
  15826.   __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15827.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15828.   return __ret;
  15829. }
  15830. #endif
  15831.  
  15832. #ifdef __LITTLE_ENDIAN__
  15833. __ai uint8x16_t vmovq_n_u8(uint8_t __p0) {
  15834.   uint8x16_t __ret;
  15835.   __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15836.   return __ret;
  15837. }
  15838. #else
  15839. __ai uint8x16_t vmovq_n_u8(uint8_t __p0) {
  15840.   uint8x16_t __ret;
  15841.   __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15842.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15843.   return __ret;
  15844. }
  15845. #endif
  15846.  
  15847. #ifdef __LITTLE_ENDIAN__
  15848. __ai uint32x4_t vmovq_n_u32(uint32_t __p0) {
  15849.   uint32x4_t __ret;
  15850.   __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
  15851.   return __ret;
  15852. }
  15853. #else
  15854. __ai uint32x4_t vmovq_n_u32(uint32_t __p0) {
  15855.   uint32x4_t __ret;
  15856.   __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
  15857.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15858.   return __ret;
  15859. }
  15860. #endif
  15861.  
  15862. #ifdef __LITTLE_ENDIAN__
  15863. __ai uint64x2_t vmovq_n_u64(uint64_t __p0) {
  15864.   uint64x2_t __ret;
  15865.   __ret = (uint64x2_t) {__p0, __p0};
  15866.   return __ret;
  15867. }
  15868. #else
  15869. __ai uint64x2_t vmovq_n_u64(uint64_t __p0) {
  15870.   uint64x2_t __ret;
  15871.   __ret = (uint64x2_t) {__p0, __p0};
  15872.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15873.   return __ret;
  15874. }
  15875. #endif
  15876.  
  15877. #ifdef __LITTLE_ENDIAN__
  15878. __ai uint16x8_t vmovq_n_u16(uint16_t __p0) {
  15879.   uint16x8_t __ret;
  15880.   __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15881.   return __ret;
  15882. }
  15883. #else
  15884. __ai uint16x8_t vmovq_n_u16(uint16_t __p0) {
  15885.   uint16x8_t __ret;
  15886.   __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15887.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15888.   return __ret;
  15889. }
  15890. #endif
  15891.  
  15892. #ifdef __LITTLE_ENDIAN__
  15893. __ai int8x16_t vmovq_n_s8(int8_t __p0) {
  15894.   int8x16_t __ret;
  15895.   __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15896.   return __ret;
  15897. }
  15898. #else
  15899. __ai int8x16_t vmovq_n_s8(int8_t __p0) {
  15900.   int8x16_t __ret;
  15901.   __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15902.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  15903.   return __ret;
  15904. }
  15905. #endif
  15906.  
  15907. #ifdef __LITTLE_ENDIAN__
  15908. __ai float32x4_t vmovq_n_f32(float32_t __p0) {
  15909.   float32x4_t __ret;
  15910.   __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
  15911.   return __ret;
  15912. }
  15913. #else
  15914. __ai float32x4_t vmovq_n_f32(float32_t __p0) {
  15915.   float32x4_t __ret;
  15916.   __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
  15917.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15918.   return __ret;
  15919. }
  15920. #endif
  15921.  
  15922. #ifdef __LITTLE_ENDIAN__
  15923. #define vmovq_n_f16(__p0) __extension__ ({ \
  15924.   float16x8_t __ret; \
  15925.   float16_t __s0 = __p0; \
  15926.   __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
  15927.   __ret; \
  15928. })
  15929. #else
  15930. #define vmovq_n_f16(__p0) __extension__ ({ \
  15931.   float16x8_t __ret; \
  15932.   float16_t __s0 = __p0; \
  15933.   __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
  15934.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  15935.   __ret; \
  15936. })
  15937. #endif
  15938.  
  15939. #ifdef __LITTLE_ENDIAN__
  15940. __ai int32x4_t vmovq_n_s32(int32_t __p0) {
  15941.   int32x4_t __ret;
  15942.   __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
  15943.   return __ret;
  15944. }
  15945. #else
  15946. __ai int32x4_t vmovq_n_s32(int32_t __p0) {
  15947.   int32x4_t __ret;
  15948.   __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
  15949.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  15950.   return __ret;
  15951. }
  15952. #endif
  15953.  
  15954. #ifdef __LITTLE_ENDIAN__
  15955. __ai int64x2_t vmovq_n_s64(int64_t __p0) {
  15956.   int64x2_t __ret;
  15957.   __ret = (int64x2_t) {__p0, __p0};
  15958.   return __ret;
  15959. }
  15960. #else
  15961. __ai int64x2_t vmovq_n_s64(int64_t __p0) {
  15962.   int64x2_t __ret;
  15963.   __ret = (int64x2_t) {__p0, __p0};
  15964.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  15965.   return __ret;
  15966. }
  15967. #endif
  15968.  
  15969. #ifdef __LITTLE_ENDIAN__
  15970. __ai int16x8_t vmovq_n_s16(int16_t __p0) {
  15971.   int16x8_t __ret;
  15972.   __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15973.   return __ret;
  15974. }
  15975. #else
  15976. __ai int16x8_t vmovq_n_s16(int16_t __p0) {
  15977.   int16x8_t __ret;
  15978.   __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15979.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15980.   return __ret;
  15981. }
  15982. #endif
  15983.  
  15984. #ifdef __LITTLE_ENDIAN__
  15985. __ai uint8x8_t vmov_n_u8(uint8_t __p0) {
  15986.   uint8x8_t __ret;
  15987.   __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15988.   return __ret;
  15989. }
  15990. #else
  15991. __ai uint8x8_t vmov_n_u8(uint8_t __p0) {
  15992.   uint8x8_t __ret;
  15993.   __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  15994.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  15995.   return __ret;
  15996. }
  15997. #endif
  15998.  
  15999. #ifdef __LITTLE_ENDIAN__
  16000. __ai uint32x2_t vmov_n_u32(uint32_t __p0) {
  16001.   uint32x2_t __ret;
  16002.   __ret = (uint32x2_t) {__p0, __p0};
  16003.   return __ret;
  16004. }
  16005. #else
  16006. __ai uint32x2_t vmov_n_u32(uint32_t __p0) {
  16007.   uint32x2_t __ret;
  16008.   __ret = (uint32x2_t) {__p0, __p0};
  16009.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16010.   return __ret;
  16011. }
  16012. #endif
  16013.  
  16014. __ai uint64x1_t vmov_n_u64(uint64_t __p0) {
  16015.   uint64x1_t __ret;
  16016.   __ret = (uint64x1_t) {__p0};
  16017.   return __ret;
  16018. }
  16019. #ifdef __LITTLE_ENDIAN__
  16020. __ai uint16x4_t vmov_n_u16(uint16_t __p0) {
  16021.   uint16x4_t __ret;
  16022.   __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
  16023.   return __ret;
  16024. }
  16025. #else
  16026. __ai uint16x4_t vmov_n_u16(uint16_t __p0) {
  16027.   uint16x4_t __ret;
  16028.   __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
  16029.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16030.   return __ret;
  16031. }
  16032. #endif
  16033.  
  16034. #ifdef __LITTLE_ENDIAN__
  16035. __ai int8x8_t vmov_n_s8(int8_t __p0) {
  16036.   int8x8_t __ret;
  16037.   __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  16038.   return __ret;
  16039. }
  16040. #else
  16041. __ai int8x8_t vmov_n_s8(int8_t __p0) {
  16042.   int8x8_t __ret;
  16043.   __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  16044.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16045.   return __ret;
  16046. }
  16047. #endif
  16048.  
  16049. #ifdef __LITTLE_ENDIAN__
  16050. __ai float32x2_t vmov_n_f32(float32_t __p0) {
  16051.   float32x2_t __ret;
  16052.   __ret = (float32x2_t) {__p0, __p0};
  16053.   return __ret;
  16054. }
  16055. #else
  16056. __ai float32x2_t vmov_n_f32(float32_t __p0) {
  16057.   float32x2_t __ret;
  16058.   __ret = (float32x2_t) {__p0, __p0};
  16059.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16060.   return __ret;
  16061. }
  16062. #endif
  16063.  
  16064. #ifdef __LITTLE_ENDIAN__
  16065. #define vmov_n_f16(__p0) __extension__ ({ \
  16066.   float16x4_t __ret; \
  16067.   float16_t __s0 = __p0; \
  16068.   __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
  16069.   __ret; \
  16070. })
  16071. #else
  16072. #define vmov_n_f16(__p0) __extension__ ({ \
  16073.   float16x4_t __ret; \
  16074.   float16_t __s0 = __p0; \
  16075.   __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
  16076.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  16077.   __ret; \
  16078. })
  16079. #endif
  16080.  
  16081. #ifdef __LITTLE_ENDIAN__
  16082. __ai int32x2_t vmov_n_s32(int32_t __p0) {
  16083.   int32x2_t __ret;
  16084.   __ret = (int32x2_t) {__p0, __p0};
  16085.   return __ret;
  16086. }
  16087. #else
  16088. __ai int32x2_t vmov_n_s32(int32_t __p0) {
  16089.   int32x2_t __ret;
  16090.   __ret = (int32x2_t) {__p0, __p0};
  16091.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16092.   return __ret;
  16093. }
  16094. #endif
  16095.  
  16096. __ai int64x1_t vmov_n_s64(int64_t __p0) {
  16097.   int64x1_t __ret;
  16098.   __ret = (int64x1_t) {__p0};
  16099.   return __ret;
  16100. }
  16101. #ifdef __LITTLE_ENDIAN__
  16102. __ai int16x4_t vmov_n_s16(int16_t __p0) {
  16103.   int16x4_t __ret;
  16104.   __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
  16105.   return __ret;
  16106. }
  16107. #else
  16108. __ai int16x4_t vmov_n_s16(int16_t __p0) {
  16109.   int16x4_t __ret;
  16110.   __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
  16111.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16112.   return __ret;
  16113. }
  16114. #endif
  16115.  
  16116. #ifdef __LITTLE_ENDIAN__
  16117. __ai uint16x8_t vmovl_u8(uint8x8_t __p0) {
  16118.   uint16x8_t __ret;
  16119.   __ret = (uint16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 49);
  16120.   return __ret;
  16121. }
  16122. #else
  16123. __ai uint16x8_t vmovl_u8(uint8x8_t __p0) {
  16124.   uint16x8_t __ret;
  16125.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16126.   __ret = (uint16x8_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 49);
  16127.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16128.   return __ret;
  16129. }
  16130. __ai uint16x8_t __noswap_vmovl_u8(uint8x8_t __p0) {
  16131.   uint16x8_t __ret;
  16132.   __ret = (uint16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 49);
  16133.   return __ret;
  16134. }
  16135. #endif
  16136.  
  16137. #ifdef __LITTLE_ENDIAN__
  16138. __ai uint64x2_t vmovl_u32(uint32x2_t __p0) {
  16139.   uint64x2_t __ret;
  16140.   __ret = (uint64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 51);
  16141.   return __ret;
  16142. }
  16143. #else
  16144. __ai uint64x2_t vmovl_u32(uint32x2_t __p0) {
  16145.   uint64x2_t __ret;
  16146.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16147.   __ret = (uint64x2_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 51);
  16148.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16149.   return __ret;
  16150. }
  16151. __ai uint64x2_t __noswap_vmovl_u32(uint32x2_t __p0) {
  16152.   uint64x2_t __ret;
  16153.   __ret = (uint64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 51);
  16154.   return __ret;
  16155. }
  16156. #endif
  16157.  
  16158. #ifdef __LITTLE_ENDIAN__
  16159. __ai uint32x4_t vmovl_u16(uint16x4_t __p0) {
  16160.   uint32x4_t __ret;
  16161.   __ret = (uint32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 50);
  16162.   return __ret;
  16163. }
  16164. #else
  16165. __ai uint32x4_t vmovl_u16(uint16x4_t __p0) {
  16166.   uint32x4_t __ret;
  16167.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16168.   __ret = (uint32x4_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 50);
  16169.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16170.   return __ret;
  16171. }
  16172. __ai uint32x4_t __noswap_vmovl_u16(uint16x4_t __p0) {
  16173.   uint32x4_t __ret;
  16174.   __ret = (uint32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 50);
  16175.   return __ret;
  16176. }
  16177. #endif
  16178.  
  16179. #ifdef __LITTLE_ENDIAN__
  16180. __ai int16x8_t vmovl_s8(int8x8_t __p0) {
  16181.   int16x8_t __ret;
  16182.   __ret = (int16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 33);
  16183.   return __ret;
  16184. }
  16185. #else
  16186. __ai int16x8_t vmovl_s8(int8x8_t __p0) {
  16187.   int16x8_t __ret;
  16188.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16189.   __ret = (int16x8_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 33);
  16190.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16191.   return __ret;
  16192. }
  16193. __ai int16x8_t __noswap_vmovl_s8(int8x8_t __p0) {
  16194.   int16x8_t __ret;
  16195.   __ret = (int16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 33);
  16196.   return __ret;
  16197. }
  16198. #endif
  16199.  
  16200. #ifdef __LITTLE_ENDIAN__
  16201. __ai int64x2_t vmovl_s32(int32x2_t __p0) {
  16202.   int64x2_t __ret;
  16203.   __ret = (int64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 35);
  16204.   return __ret;
  16205. }
  16206. #else
  16207. __ai int64x2_t vmovl_s32(int32x2_t __p0) {
  16208.   int64x2_t __ret;
  16209.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16210.   __ret = (int64x2_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 35);
  16211.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16212.   return __ret;
  16213. }
  16214. __ai int64x2_t __noswap_vmovl_s32(int32x2_t __p0) {
  16215.   int64x2_t __ret;
  16216.   __ret = (int64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 35);
  16217.   return __ret;
  16218. }
  16219. #endif
  16220.  
  16221. #ifdef __LITTLE_ENDIAN__
  16222. __ai int32x4_t vmovl_s16(int16x4_t __p0) {
  16223.   int32x4_t __ret;
  16224.   __ret = (int32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 34);
  16225.   return __ret;
  16226. }
  16227. #else
  16228. __ai int32x4_t vmovl_s16(int16x4_t __p0) {
  16229.   int32x4_t __ret;
  16230.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16231.   __ret = (int32x4_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 34);
  16232.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16233.   return __ret;
  16234. }
  16235. __ai int32x4_t __noswap_vmovl_s16(int16x4_t __p0) {
  16236.   int32x4_t __ret;
  16237.   __ret = (int32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 34);
  16238.   return __ret;
  16239. }
  16240. #endif
  16241.  
  16242. #ifdef __LITTLE_ENDIAN__
  16243. __ai uint16x4_t vmovn_u32(uint32x4_t __p0) {
  16244.   uint16x4_t __ret;
  16245.   __ret = (uint16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 17);
  16246.   return __ret;
  16247. }
  16248. #else
  16249. __ai uint16x4_t vmovn_u32(uint32x4_t __p0) {
  16250.   uint16x4_t __ret;
  16251.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16252.   __ret = (uint16x4_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 17);
  16253.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16254.   return __ret;
  16255. }
  16256. __ai uint16x4_t __noswap_vmovn_u32(uint32x4_t __p0) {
  16257.   uint16x4_t __ret;
  16258.   __ret = (uint16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 17);
  16259.   return __ret;
  16260. }
  16261. #endif
  16262.  
  16263. #ifdef __LITTLE_ENDIAN__
  16264. __ai uint32x2_t vmovn_u64(uint64x2_t __p0) {
  16265.   uint32x2_t __ret;
  16266.   __ret = (uint32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 18);
  16267.   return __ret;
  16268. }
  16269. #else
  16270. __ai uint32x2_t vmovn_u64(uint64x2_t __p0) {
  16271.   uint32x2_t __ret;
  16272.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16273.   __ret = (uint32x2_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 18);
  16274.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16275.   return __ret;
  16276. }
  16277. __ai uint32x2_t __noswap_vmovn_u64(uint64x2_t __p0) {
  16278.   uint32x2_t __ret;
  16279.   __ret = (uint32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 18);
  16280.   return __ret;
  16281. }
  16282. #endif
  16283.  
  16284. #ifdef __LITTLE_ENDIAN__
  16285. __ai uint8x8_t vmovn_u16(uint16x8_t __p0) {
  16286.   uint8x8_t __ret;
  16287.   __ret = (uint8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 16);
  16288.   return __ret;
  16289. }
  16290. #else
  16291. __ai uint8x8_t vmovn_u16(uint16x8_t __p0) {
  16292.   uint8x8_t __ret;
  16293.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16294.   __ret = (uint8x8_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 16);
  16295.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16296.   return __ret;
  16297. }
  16298. __ai uint8x8_t __noswap_vmovn_u16(uint16x8_t __p0) {
  16299.   uint8x8_t __ret;
  16300.   __ret = (uint8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 16);
  16301.   return __ret;
  16302. }
  16303. #endif
  16304.  
  16305. #ifdef __LITTLE_ENDIAN__
  16306. __ai int16x4_t vmovn_s32(int32x4_t __p0) {
  16307.   int16x4_t __ret;
  16308.   __ret = (int16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 1);
  16309.   return __ret;
  16310. }
  16311. #else
  16312. __ai int16x4_t vmovn_s32(int32x4_t __p0) {
  16313.   int16x4_t __ret;
  16314.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16315.   __ret = (int16x4_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 1);
  16316.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16317.   return __ret;
  16318. }
  16319. __ai int16x4_t __noswap_vmovn_s32(int32x4_t __p0) {
  16320.   int16x4_t __ret;
  16321.   __ret = (int16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 1);
  16322.   return __ret;
  16323. }
  16324. #endif
  16325.  
  16326. #ifdef __LITTLE_ENDIAN__
  16327. __ai int32x2_t vmovn_s64(int64x2_t __p0) {
  16328.   int32x2_t __ret;
  16329.   __ret = (int32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 2);
  16330.   return __ret;
  16331. }
  16332. #else
  16333. __ai int32x2_t vmovn_s64(int64x2_t __p0) {
  16334.   int32x2_t __ret;
  16335.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16336.   __ret = (int32x2_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 2);
  16337.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16338.   return __ret;
  16339. }
  16340. __ai int32x2_t __noswap_vmovn_s64(int64x2_t __p0) {
  16341.   int32x2_t __ret;
  16342.   __ret = (int32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 2);
  16343.   return __ret;
  16344. }
  16345. #endif
  16346.  
  16347. #ifdef __LITTLE_ENDIAN__
  16348. __ai int8x8_t vmovn_s16(int16x8_t __p0) {
  16349.   int8x8_t __ret;
  16350.   __ret = (int8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 0);
  16351.   return __ret;
  16352. }
  16353. #else
  16354. __ai int8x8_t vmovn_s16(int16x8_t __p0) {
  16355.   int8x8_t __ret;
  16356.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16357.   __ret = (int8x8_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 0);
  16358.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16359.   return __ret;
  16360. }
  16361. __ai int8x8_t __noswap_vmovn_s16(int16x8_t __p0) {
  16362.   int8x8_t __ret;
  16363.   __ret = (int8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 0);
  16364.   return __ret;
  16365. }
  16366. #endif
  16367.  
  16368. #ifdef __LITTLE_ENDIAN__
  16369. __ai uint8x16_t vmulq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  16370.   uint8x16_t __ret;
  16371.   __ret = __p0 * __p1;
  16372.   return __ret;
  16373. }
  16374. #else
  16375. __ai uint8x16_t vmulq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  16376.   uint8x16_t __ret;
  16377.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  16378.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  16379.   __ret = __rev0 * __rev1;
  16380.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  16381.   return __ret;
  16382. }
  16383. #endif
  16384.  
  16385. #ifdef __LITTLE_ENDIAN__
  16386. __ai uint32x4_t vmulq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  16387.   uint32x4_t __ret;
  16388.   __ret = __p0 * __p1;
  16389.   return __ret;
  16390. }
  16391. #else
  16392. __ai uint32x4_t vmulq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  16393.   uint32x4_t __ret;
  16394.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16395.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  16396.   __ret = __rev0 * __rev1;
  16397.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16398.   return __ret;
  16399. }
  16400. #endif
  16401.  
  16402. #ifdef __LITTLE_ENDIAN__
  16403. __ai uint16x8_t vmulq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  16404.   uint16x8_t __ret;
  16405.   __ret = __p0 * __p1;
  16406.   return __ret;
  16407. }
  16408. #else
  16409. __ai uint16x8_t vmulq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  16410.   uint16x8_t __ret;
  16411.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16412.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  16413.   __ret = __rev0 * __rev1;
  16414.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16415.   return __ret;
  16416. }
  16417. #endif
  16418.  
  16419. #ifdef __LITTLE_ENDIAN__
  16420. __ai int8x16_t vmulq_s8(int8x16_t __p0, int8x16_t __p1) {
  16421.   int8x16_t __ret;
  16422.   __ret = __p0 * __p1;
  16423.   return __ret;
  16424. }
  16425. #else
  16426. __ai int8x16_t vmulq_s8(int8x16_t __p0, int8x16_t __p1) {
  16427.   int8x16_t __ret;
  16428.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  16429.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  16430.   __ret = __rev0 * __rev1;
  16431.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  16432.   return __ret;
  16433. }
  16434. #endif
  16435.  
  16436. #ifdef __LITTLE_ENDIAN__
  16437. __ai float32x4_t vmulq_f32(float32x4_t __p0, float32x4_t __p1) {
  16438.   float32x4_t __ret;
  16439.   __ret = __p0 * __p1;
  16440.   return __ret;
  16441. }
  16442. #else
  16443. __ai float32x4_t vmulq_f32(float32x4_t __p0, float32x4_t __p1) {
  16444.   float32x4_t __ret;
  16445.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16446.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  16447.   __ret = __rev0 * __rev1;
  16448.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16449.   return __ret;
  16450. }
  16451. #endif
  16452.  
  16453. #ifdef __LITTLE_ENDIAN__
  16454. __ai int32x4_t vmulq_s32(int32x4_t __p0, int32x4_t __p1) {
  16455.   int32x4_t __ret;
  16456.   __ret = __p0 * __p1;
  16457.   return __ret;
  16458. }
  16459. #else
  16460. __ai int32x4_t vmulq_s32(int32x4_t __p0, int32x4_t __p1) {
  16461.   int32x4_t __ret;
  16462.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16463.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  16464.   __ret = __rev0 * __rev1;
  16465.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16466.   return __ret;
  16467. }
  16468. #endif
  16469.  
  16470. #ifdef __LITTLE_ENDIAN__
  16471. __ai int16x8_t vmulq_s16(int16x8_t __p0, int16x8_t __p1) {
  16472.   int16x8_t __ret;
  16473.   __ret = __p0 * __p1;
  16474.   return __ret;
  16475. }
  16476. #else
  16477. __ai int16x8_t vmulq_s16(int16x8_t __p0, int16x8_t __p1) {
  16478.   int16x8_t __ret;
  16479.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16480.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  16481.   __ret = __rev0 * __rev1;
  16482.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16483.   return __ret;
  16484. }
  16485. #endif
  16486.  
  16487. #ifdef __LITTLE_ENDIAN__
  16488. __ai uint8x8_t vmul_u8(uint8x8_t __p0, uint8x8_t __p1) {
  16489.   uint8x8_t __ret;
  16490.   __ret = __p0 * __p1;
  16491.   return __ret;
  16492. }
  16493. #else
  16494. __ai uint8x8_t vmul_u8(uint8x8_t __p0, uint8x8_t __p1) {
  16495.   uint8x8_t __ret;
  16496.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16497.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  16498.   __ret = __rev0 * __rev1;
  16499.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16500.   return __ret;
  16501. }
  16502. #endif
  16503.  
  16504. #ifdef __LITTLE_ENDIAN__
  16505. __ai uint32x2_t vmul_u32(uint32x2_t __p0, uint32x2_t __p1) {
  16506.   uint32x2_t __ret;
  16507.   __ret = __p0 * __p1;
  16508.   return __ret;
  16509. }
  16510. #else
  16511. __ai uint32x2_t vmul_u32(uint32x2_t __p0, uint32x2_t __p1) {
  16512.   uint32x2_t __ret;
  16513.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16514.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  16515.   __ret = __rev0 * __rev1;
  16516.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16517.   return __ret;
  16518. }
  16519. #endif
  16520.  
  16521. #ifdef __LITTLE_ENDIAN__
  16522. __ai uint16x4_t vmul_u16(uint16x4_t __p0, uint16x4_t __p1) {
  16523.   uint16x4_t __ret;
  16524.   __ret = __p0 * __p1;
  16525.   return __ret;
  16526. }
  16527. #else
  16528. __ai uint16x4_t vmul_u16(uint16x4_t __p0, uint16x4_t __p1) {
  16529.   uint16x4_t __ret;
  16530.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16531.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  16532.   __ret = __rev0 * __rev1;
  16533.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16534.   return __ret;
  16535. }
  16536. #endif
  16537.  
  16538. #ifdef __LITTLE_ENDIAN__
  16539. __ai int8x8_t vmul_s8(int8x8_t __p0, int8x8_t __p1) {
  16540.   int8x8_t __ret;
  16541.   __ret = __p0 * __p1;
  16542.   return __ret;
  16543. }
  16544. #else
  16545. __ai int8x8_t vmul_s8(int8x8_t __p0, int8x8_t __p1) {
  16546.   int8x8_t __ret;
  16547.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16548.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  16549.   __ret = __rev0 * __rev1;
  16550.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16551.   return __ret;
  16552. }
  16553. #endif
  16554.  
  16555. #ifdef __LITTLE_ENDIAN__
  16556. __ai float32x2_t vmul_f32(float32x2_t __p0, float32x2_t __p1) {
  16557.   float32x2_t __ret;
  16558.   __ret = __p0 * __p1;
  16559.   return __ret;
  16560. }
  16561. #else
  16562. __ai float32x2_t vmul_f32(float32x2_t __p0, float32x2_t __p1) {
  16563.   float32x2_t __ret;
  16564.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16565.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  16566.   __ret = __rev0 * __rev1;
  16567.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16568.   return __ret;
  16569. }
  16570. #endif
  16571.  
  16572. #ifdef __LITTLE_ENDIAN__
  16573. __ai int32x2_t vmul_s32(int32x2_t __p0, int32x2_t __p1) {
  16574.   int32x2_t __ret;
  16575.   __ret = __p0 * __p1;
  16576.   return __ret;
  16577. }
  16578. #else
  16579. __ai int32x2_t vmul_s32(int32x2_t __p0, int32x2_t __p1) {
  16580.   int32x2_t __ret;
  16581.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16582.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  16583.   __ret = __rev0 * __rev1;
  16584.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16585.   return __ret;
  16586. }
  16587. #endif
  16588.  
  16589. #ifdef __LITTLE_ENDIAN__
  16590. __ai int16x4_t vmul_s16(int16x4_t __p0, int16x4_t __p1) {
  16591.   int16x4_t __ret;
  16592.   __ret = __p0 * __p1;
  16593.   return __ret;
  16594. }
  16595. #else
  16596. __ai int16x4_t vmul_s16(int16x4_t __p0, int16x4_t __p1) {
  16597.   int16x4_t __ret;
  16598.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16599.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  16600.   __ret = __rev0 * __rev1;
  16601.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16602.   return __ret;
  16603. }
  16604. #endif
  16605.  
  16606. #ifdef __LITTLE_ENDIAN__
  16607. __ai poly8x8_t vmul_p8(poly8x8_t __p0, poly8x8_t __p1) {
  16608.   poly8x8_t __ret;
  16609.   __ret = (poly8x8_t) __builtin_neon_vmul_v((int8x8_t)__p0, (int8x8_t)__p1, 4);
  16610.   return __ret;
  16611. }
  16612. #else
  16613. __ai poly8x8_t vmul_p8(poly8x8_t __p0, poly8x8_t __p1) {
  16614.   poly8x8_t __ret;
  16615.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16616.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  16617.   __ret = (poly8x8_t) __builtin_neon_vmul_v((int8x8_t)__rev0, (int8x8_t)__rev1, 4);
  16618.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16619.   return __ret;
  16620. }
  16621. #endif
  16622.  
  16623. #ifdef __LITTLE_ENDIAN__
  16624. __ai poly8x16_t vmulq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  16625.   poly8x16_t __ret;
  16626.   __ret = (poly8x16_t) __builtin_neon_vmulq_v((int8x16_t)__p0, (int8x16_t)__p1, 36);
  16627.   return __ret;
  16628. }
  16629. #else
  16630. __ai poly8x16_t vmulq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  16631.   poly8x16_t __ret;
  16632.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  16633.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  16634.   __ret = (poly8x16_t) __builtin_neon_vmulq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 36);
  16635.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  16636.   return __ret;
  16637. }
  16638. #endif
  16639.  
  16640. #ifdef __LITTLE_ENDIAN__
  16641. #define vmulq_lane_u32(__p0_86, __p1_86, __p2_86) __extension__ ({ \
  16642.   uint32x4_t __ret_86; \
  16643.   uint32x4_t __s0_86 = __p0_86; \
  16644.   uint32x2_t __s1_86 = __p1_86; \
  16645.   __ret_86 = __s0_86 * splatq_lane_u32(__s1_86, __p2_86); \
  16646.   __ret_86; \
  16647. })
  16648. #else
  16649. #define vmulq_lane_u32(__p0_87, __p1_87, __p2_87) __extension__ ({ \
  16650.   uint32x4_t __ret_87; \
  16651.   uint32x4_t __s0_87 = __p0_87; \
  16652.   uint32x2_t __s1_87 = __p1_87; \
  16653.   uint32x4_t __rev0_87;  __rev0_87 = __builtin_shufflevector(__s0_87, __s0_87, 3, 2, 1, 0); \
  16654.   uint32x2_t __rev1_87;  __rev1_87 = __builtin_shufflevector(__s1_87, __s1_87, 1, 0); \
  16655.   __ret_87 = __rev0_87 * __noswap_splatq_lane_u32(__rev1_87, __p2_87); \
  16656.   __ret_87 = __builtin_shufflevector(__ret_87, __ret_87, 3, 2, 1, 0); \
  16657.   __ret_87; \
  16658. })
  16659. #endif
  16660.  
  16661. #ifdef __LITTLE_ENDIAN__
  16662. #define vmulq_lane_u16(__p0_88, __p1_88, __p2_88) __extension__ ({ \
  16663.   uint16x8_t __ret_88; \
  16664.   uint16x8_t __s0_88 = __p0_88; \
  16665.   uint16x4_t __s1_88 = __p1_88; \
  16666.   __ret_88 = __s0_88 * splatq_lane_u16(__s1_88, __p2_88); \
  16667.   __ret_88; \
  16668. })
  16669. #else
  16670. #define vmulq_lane_u16(__p0_89, __p1_89, __p2_89) __extension__ ({ \
  16671.   uint16x8_t __ret_89; \
  16672.   uint16x8_t __s0_89 = __p0_89; \
  16673.   uint16x4_t __s1_89 = __p1_89; \
  16674.   uint16x8_t __rev0_89;  __rev0_89 = __builtin_shufflevector(__s0_89, __s0_89, 7, 6, 5, 4, 3, 2, 1, 0); \
  16675.   uint16x4_t __rev1_89;  __rev1_89 = __builtin_shufflevector(__s1_89, __s1_89, 3, 2, 1, 0); \
  16676.   __ret_89 = __rev0_89 * __noswap_splatq_lane_u16(__rev1_89, __p2_89); \
  16677.   __ret_89 = __builtin_shufflevector(__ret_89, __ret_89, 7, 6, 5, 4, 3, 2, 1, 0); \
  16678.   __ret_89; \
  16679. })
  16680. #endif
  16681.  
  16682. #ifdef __LITTLE_ENDIAN__
  16683. #define vmulq_lane_f32(__p0_90, __p1_90, __p2_90) __extension__ ({ \
  16684.   float32x4_t __ret_90; \
  16685.   float32x4_t __s0_90 = __p0_90; \
  16686.   float32x2_t __s1_90 = __p1_90; \
  16687.   __ret_90 = __s0_90 * splatq_lane_f32(__s1_90, __p2_90); \
  16688.   __ret_90; \
  16689. })
  16690. #else
  16691. #define vmulq_lane_f32(__p0_91, __p1_91, __p2_91) __extension__ ({ \
  16692.   float32x4_t __ret_91; \
  16693.   float32x4_t __s0_91 = __p0_91; \
  16694.   float32x2_t __s1_91 = __p1_91; \
  16695.   float32x4_t __rev0_91;  __rev0_91 = __builtin_shufflevector(__s0_91, __s0_91, 3, 2, 1, 0); \
  16696.   float32x2_t __rev1_91;  __rev1_91 = __builtin_shufflevector(__s1_91, __s1_91, 1, 0); \
  16697.   __ret_91 = __rev0_91 * __noswap_splatq_lane_f32(__rev1_91, __p2_91); \
  16698.   __ret_91 = __builtin_shufflevector(__ret_91, __ret_91, 3, 2, 1, 0); \
  16699.   __ret_91; \
  16700. })
  16701. #endif
  16702.  
  16703. #ifdef __LITTLE_ENDIAN__
  16704. #define vmulq_lane_s32(__p0_92, __p1_92, __p2_92) __extension__ ({ \
  16705.   int32x4_t __ret_92; \
  16706.   int32x4_t __s0_92 = __p0_92; \
  16707.   int32x2_t __s1_92 = __p1_92; \
  16708.   __ret_92 = __s0_92 * splatq_lane_s32(__s1_92, __p2_92); \
  16709.   __ret_92; \
  16710. })
  16711. #else
  16712. #define vmulq_lane_s32(__p0_93, __p1_93, __p2_93) __extension__ ({ \
  16713.   int32x4_t __ret_93; \
  16714.   int32x4_t __s0_93 = __p0_93; \
  16715.   int32x2_t __s1_93 = __p1_93; \
  16716.   int32x4_t __rev0_93;  __rev0_93 = __builtin_shufflevector(__s0_93, __s0_93, 3, 2, 1, 0); \
  16717.   int32x2_t __rev1_93;  __rev1_93 = __builtin_shufflevector(__s1_93, __s1_93, 1, 0); \
  16718.   __ret_93 = __rev0_93 * __noswap_splatq_lane_s32(__rev1_93, __p2_93); \
  16719.   __ret_93 = __builtin_shufflevector(__ret_93, __ret_93, 3, 2, 1, 0); \
  16720.   __ret_93; \
  16721. })
  16722. #endif
  16723.  
  16724. #ifdef __LITTLE_ENDIAN__
  16725. #define vmulq_lane_s16(__p0_94, __p1_94, __p2_94) __extension__ ({ \
  16726.   int16x8_t __ret_94; \
  16727.   int16x8_t __s0_94 = __p0_94; \
  16728.   int16x4_t __s1_94 = __p1_94; \
  16729.   __ret_94 = __s0_94 * splatq_lane_s16(__s1_94, __p2_94); \
  16730.   __ret_94; \
  16731. })
  16732. #else
  16733. #define vmulq_lane_s16(__p0_95, __p1_95, __p2_95) __extension__ ({ \
  16734.   int16x8_t __ret_95; \
  16735.   int16x8_t __s0_95 = __p0_95; \
  16736.   int16x4_t __s1_95 = __p1_95; \
  16737.   int16x8_t __rev0_95;  __rev0_95 = __builtin_shufflevector(__s0_95, __s0_95, 7, 6, 5, 4, 3, 2, 1, 0); \
  16738.   int16x4_t __rev1_95;  __rev1_95 = __builtin_shufflevector(__s1_95, __s1_95, 3, 2, 1, 0); \
  16739.   __ret_95 = __rev0_95 * __noswap_splatq_lane_s16(__rev1_95, __p2_95); \
  16740.   __ret_95 = __builtin_shufflevector(__ret_95, __ret_95, 7, 6, 5, 4, 3, 2, 1, 0); \
  16741.   __ret_95; \
  16742. })
  16743. #endif
  16744.  
  16745. #ifdef __LITTLE_ENDIAN__
  16746. #define vmul_lane_u32(__p0_96, __p1_96, __p2_96) __extension__ ({ \
  16747.   uint32x2_t __ret_96; \
  16748.   uint32x2_t __s0_96 = __p0_96; \
  16749.   uint32x2_t __s1_96 = __p1_96; \
  16750.   __ret_96 = __s0_96 * splat_lane_u32(__s1_96, __p2_96); \
  16751.   __ret_96; \
  16752. })
  16753. #else
  16754. #define vmul_lane_u32(__p0_97, __p1_97, __p2_97) __extension__ ({ \
  16755.   uint32x2_t __ret_97; \
  16756.   uint32x2_t __s0_97 = __p0_97; \
  16757.   uint32x2_t __s1_97 = __p1_97; \
  16758.   uint32x2_t __rev0_97;  __rev0_97 = __builtin_shufflevector(__s0_97, __s0_97, 1, 0); \
  16759.   uint32x2_t __rev1_97;  __rev1_97 = __builtin_shufflevector(__s1_97, __s1_97, 1, 0); \
  16760.   __ret_97 = __rev0_97 * __noswap_splat_lane_u32(__rev1_97, __p2_97); \
  16761.   __ret_97 = __builtin_shufflevector(__ret_97, __ret_97, 1, 0); \
  16762.   __ret_97; \
  16763. })
  16764. #endif
  16765.  
  16766. #ifdef __LITTLE_ENDIAN__
  16767. #define vmul_lane_u16(__p0_98, __p1_98, __p2_98) __extension__ ({ \
  16768.   uint16x4_t __ret_98; \
  16769.   uint16x4_t __s0_98 = __p0_98; \
  16770.   uint16x4_t __s1_98 = __p1_98; \
  16771.   __ret_98 = __s0_98 * splat_lane_u16(__s1_98, __p2_98); \
  16772.   __ret_98; \
  16773. })
  16774. #else
  16775. #define vmul_lane_u16(__p0_99, __p1_99, __p2_99) __extension__ ({ \
  16776.   uint16x4_t __ret_99; \
  16777.   uint16x4_t __s0_99 = __p0_99; \
  16778.   uint16x4_t __s1_99 = __p1_99; \
  16779.   uint16x4_t __rev0_99;  __rev0_99 = __builtin_shufflevector(__s0_99, __s0_99, 3, 2, 1, 0); \
  16780.   uint16x4_t __rev1_99;  __rev1_99 = __builtin_shufflevector(__s1_99, __s1_99, 3, 2, 1, 0); \
  16781.   __ret_99 = __rev0_99 * __noswap_splat_lane_u16(__rev1_99, __p2_99); \
  16782.   __ret_99 = __builtin_shufflevector(__ret_99, __ret_99, 3, 2, 1, 0); \
  16783.   __ret_99; \
  16784. })
  16785. #endif
  16786.  
  16787. #ifdef __LITTLE_ENDIAN__
  16788. #define vmul_lane_f32(__p0_100, __p1_100, __p2_100) __extension__ ({ \
  16789.   float32x2_t __ret_100; \
  16790.   float32x2_t __s0_100 = __p0_100; \
  16791.   float32x2_t __s1_100 = __p1_100; \
  16792.   __ret_100 = __s0_100 * splat_lane_f32(__s1_100, __p2_100); \
  16793.   __ret_100; \
  16794. })
  16795. #else
  16796. #define vmul_lane_f32(__p0_101, __p1_101, __p2_101) __extension__ ({ \
  16797.   float32x2_t __ret_101; \
  16798.   float32x2_t __s0_101 = __p0_101; \
  16799.   float32x2_t __s1_101 = __p1_101; \
  16800.   float32x2_t __rev0_101;  __rev0_101 = __builtin_shufflevector(__s0_101, __s0_101, 1, 0); \
  16801.   float32x2_t __rev1_101;  __rev1_101 = __builtin_shufflevector(__s1_101, __s1_101, 1, 0); \
  16802.   __ret_101 = __rev0_101 * __noswap_splat_lane_f32(__rev1_101, __p2_101); \
  16803.   __ret_101 = __builtin_shufflevector(__ret_101, __ret_101, 1, 0); \
  16804.   __ret_101; \
  16805. })
  16806. #endif
  16807.  
  16808. #ifdef __LITTLE_ENDIAN__
  16809. #define vmul_lane_s32(__p0_102, __p1_102, __p2_102) __extension__ ({ \
  16810.   int32x2_t __ret_102; \
  16811.   int32x2_t __s0_102 = __p0_102; \
  16812.   int32x2_t __s1_102 = __p1_102; \
  16813.   __ret_102 = __s0_102 * splat_lane_s32(__s1_102, __p2_102); \
  16814.   __ret_102; \
  16815. })
  16816. #else
  16817. #define vmul_lane_s32(__p0_103, __p1_103, __p2_103) __extension__ ({ \
  16818.   int32x2_t __ret_103; \
  16819.   int32x2_t __s0_103 = __p0_103; \
  16820.   int32x2_t __s1_103 = __p1_103; \
  16821.   int32x2_t __rev0_103;  __rev0_103 = __builtin_shufflevector(__s0_103, __s0_103, 1, 0); \
  16822.   int32x2_t __rev1_103;  __rev1_103 = __builtin_shufflevector(__s1_103, __s1_103, 1, 0); \
  16823.   __ret_103 = __rev0_103 * __noswap_splat_lane_s32(__rev1_103, __p2_103); \
  16824.   __ret_103 = __builtin_shufflevector(__ret_103, __ret_103, 1, 0); \
  16825.   __ret_103; \
  16826. })
  16827. #endif
  16828.  
  16829. #ifdef __LITTLE_ENDIAN__
  16830. #define vmul_lane_s16(__p0_104, __p1_104, __p2_104) __extension__ ({ \
  16831.   int16x4_t __ret_104; \
  16832.   int16x4_t __s0_104 = __p0_104; \
  16833.   int16x4_t __s1_104 = __p1_104; \
  16834.   __ret_104 = __s0_104 * splat_lane_s16(__s1_104, __p2_104); \
  16835.   __ret_104; \
  16836. })
  16837. #else
  16838. #define vmul_lane_s16(__p0_105, __p1_105, __p2_105) __extension__ ({ \
  16839.   int16x4_t __ret_105; \
  16840.   int16x4_t __s0_105 = __p0_105; \
  16841.   int16x4_t __s1_105 = __p1_105; \
  16842.   int16x4_t __rev0_105;  __rev0_105 = __builtin_shufflevector(__s0_105, __s0_105, 3, 2, 1, 0); \
  16843.   int16x4_t __rev1_105;  __rev1_105 = __builtin_shufflevector(__s1_105, __s1_105, 3, 2, 1, 0); \
  16844.   __ret_105 = __rev0_105 * __noswap_splat_lane_s16(__rev1_105, __p2_105); \
  16845.   __ret_105 = __builtin_shufflevector(__ret_105, __ret_105, 3, 2, 1, 0); \
  16846.   __ret_105; \
  16847. })
  16848. #endif
  16849.  
  16850. #ifdef __LITTLE_ENDIAN__
  16851. __ai uint32x4_t vmulq_n_u32(uint32x4_t __p0, uint32_t __p1) {
  16852.   uint32x4_t __ret;
  16853.   __ret = __p0 * (uint32x4_t) {__p1, __p1, __p1, __p1};
  16854.   return __ret;
  16855. }
  16856. #else
  16857. __ai uint32x4_t vmulq_n_u32(uint32x4_t __p0, uint32_t __p1) {
  16858.   uint32x4_t __ret;
  16859.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16860.   __ret = __rev0 * (uint32x4_t) {__p1, __p1, __p1, __p1};
  16861.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16862.   return __ret;
  16863. }
  16864. #endif
  16865.  
  16866. #ifdef __LITTLE_ENDIAN__
  16867. __ai uint16x8_t vmulq_n_u16(uint16x8_t __p0, uint16_t __p1) {
  16868.   uint16x8_t __ret;
  16869.   __ret = __p0 * (uint16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
  16870.   return __ret;
  16871. }
  16872. #else
  16873. __ai uint16x8_t vmulq_n_u16(uint16x8_t __p0, uint16_t __p1) {
  16874.   uint16x8_t __ret;
  16875.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16876.   __ret = __rev0 * (uint16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
  16877.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16878.   return __ret;
  16879. }
  16880. #endif
  16881.  
  16882. #ifdef __LITTLE_ENDIAN__
  16883. __ai float32x4_t vmulq_n_f32(float32x4_t __p0, float32_t __p1) {
  16884.   float32x4_t __ret;
  16885.   __ret = __p0 * (float32x4_t) {__p1, __p1, __p1, __p1};
  16886.   return __ret;
  16887. }
  16888. #else
  16889. __ai float32x4_t vmulq_n_f32(float32x4_t __p0, float32_t __p1) {
  16890.   float32x4_t __ret;
  16891.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16892.   __ret = __rev0 * (float32x4_t) {__p1, __p1, __p1, __p1};
  16893.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16894.   return __ret;
  16895. }
  16896. #endif
  16897.  
  16898. #ifdef __LITTLE_ENDIAN__
  16899. __ai int32x4_t vmulq_n_s32(int32x4_t __p0, int32_t __p1) {
  16900.   int32x4_t __ret;
  16901.   __ret = __p0 * (int32x4_t) {__p1, __p1, __p1, __p1};
  16902.   return __ret;
  16903. }
  16904. #else
  16905. __ai int32x4_t vmulq_n_s32(int32x4_t __p0, int32_t __p1) {
  16906.   int32x4_t __ret;
  16907.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16908.   __ret = __rev0 * (int32x4_t) {__p1, __p1, __p1, __p1};
  16909.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16910.   return __ret;
  16911. }
  16912. #endif
  16913.  
  16914. #ifdef __LITTLE_ENDIAN__
  16915. __ai int16x8_t vmulq_n_s16(int16x8_t __p0, int16_t __p1) {
  16916.   int16x8_t __ret;
  16917.   __ret = __p0 * (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
  16918.   return __ret;
  16919. }
  16920. #else
  16921. __ai int16x8_t vmulq_n_s16(int16x8_t __p0, int16_t __p1) {
  16922.   int16x8_t __ret;
  16923.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  16924.   __ret = __rev0 * (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
  16925.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  16926.   return __ret;
  16927. }
  16928. #endif
  16929.  
  16930. #ifdef __LITTLE_ENDIAN__
  16931. __ai uint32x2_t vmul_n_u32(uint32x2_t __p0, uint32_t __p1) {
  16932.   uint32x2_t __ret;
  16933.   __ret = __p0 * (uint32x2_t) {__p1, __p1};
  16934.   return __ret;
  16935. }
  16936. #else
  16937. __ai uint32x2_t vmul_n_u32(uint32x2_t __p0, uint32_t __p1) {
  16938.   uint32x2_t __ret;
  16939.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16940.   __ret = __rev0 * (uint32x2_t) {__p1, __p1};
  16941.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16942.   return __ret;
  16943. }
  16944. #endif
  16945.  
  16946. #ifdef __LITTLE_ENDIAN__
  16947. __ai uint16x4_t vmul_n_u16(uint16x4_t __p0, uint16_t __p1) {
  16948.   uint16x4_t __ret;
  16949.   __ret = __p0 * (uint16x4_t) {__p1, __p1, __p1, __p1};
  16950.   return __ret;
  16951. }
  16952. #else
  16953. __ai uint16x4_t vmul_n_u16(uint16x4_t __p0, uint16_t __p1) {
  16954.   uint16x4_t __ret;
  16955.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  16956.   __ret = __rev0 * (uint16x4_t) {__p1, __p1, __p1, __p1};
  16957.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  16958.   return __ret;
  16959. }
  16960. #endif
  16961.  
  16962. #ifdef __LITTLE_ENDIAN__
  16963. __ai float32x2_t vmul_n_f32(float32x2_t __p0, float32_t __p1) {
  16964.   float32x2_t __ret;
  16965.   __ret = __p0 * (float32x2_t) {__p1, __p1};
  16966.   return __ret;
  16967. }
  16968. #else
  16969. __ai float32x2_t vmul_n_f32(float32x2_t __p0, float32_t __p1) {
  16970.   float32x2_t __ret;
  16971.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16972.   __ret = __rev0 * (float32x2_t) {__p1, __p1};
  16973.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16974.   return __ret;
  16975. }
  16976. #endif
  16977.  
  16978. #ifdef __LITTLE_ENDIAN__
  16979. __ai int32x2_t vmul_n_s32(int32x2_t __p0, int32_t __p1) {
  16980.   int32x2_t __ret;
  16981.   __ret = __p0 * (int32x2_t) {__p1, __p1};
  16982.   return __ret;
  16983. }
  16984. #else
  16985. __ai int32x2_t vmul_n_s32(int32x2_t __p0, int32_t __p1) {
  16986.   int32x2_t __ret;
  16987.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  16988.   __ret = __rev0 * (int32x2_t) {__p1, __p1};
  16989.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  16990.   return __ret;
  16991. }
  16992. #endif
  16993.  
  16994. #ifdef __LITTLE_ENDIAN__
  16995. __ai int16x4_t vmul_n_s16(int16x4_t __p0, int16_t __p1) {
  16996.   int16x4_t __ret;
  16997.   __ret = __p0 * (int16x4_t) {__p1, __p1, __p1, __p1};
  16998.   return __ret;
  16999. }
  17000. #else
  17001. __ai int16x4_t vmul_n_s16(int16x4_t __p0, int16_t __p1) {
  17002.   int16x4_t __ret;
  17003.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17004.   __ret = __rev0 * (int16x4_t) {__p1, __p1, __p1, __p1};
  17005.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17006.   return __ret;
  17007. }
  17008. #endif
  17009.  
  17010. #ifdef __LITTLE_ENDIAN__
  17011. __ai poly16x8_t vmull_p8(poly8x8_t __p0, poly8x8_t __p1) {
  17012.   poly16x8_t __ret;
  17013.   __ret = (poly16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 37);
  17014.   return __ret;
  17015. }
  17016. #else
  17017. __ai poly16x8_t vmull_p8(poly8x8_t __p0, poly8x8_t __p1) {
  17018.   poly16x8_t __ret;
  17019.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17020.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  17021.   __ret = (poly16x8_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 37);
  17022.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17023.   return __ret;
  17024. }
  17025. __ai poly16x8_t __noswap_vmull_p8(poly8x8_t __p0, poly8x8_t __p1) {
  17026.   poly16x8_t __ret;
  17027.   __ret = (poly16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 37);
  17028.   return __ret;
  17029. }
  17030. #endif
  17031.  
  17032. #ifdef __LITTLE_ENDIAN__
  17033. __ai uint16x8_t vmull_u8(uint8x8_t __p0, uint8x8_t __p1) {
  17034.   uint16x8_t __ret;
  17035.   __ret = (uint16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 49);
  17036.   return __ret;
  17037. }
  17038. #else
  17039. __ai uint16x8_t vmull_u8(uint8x8_t __p0, uint8x8_t __p1) {
  17040.   uint16x8_t __ret;
  17041.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17042.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  17043.   __ret = (uint16x8_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 49);
  17044.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17045.   return __ret;
  17046. }
  17047. __ai uint16x8_t __noswap_vmull_u8(uint8x8_t __p0, uint8x8_t __p1) {
  17048.   uint16x8_t __ret;
  17049.   __ret = (uint16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 49);
  17050.   return __ret;
  17051. }
  17052. #endif
  17053.  
  17054. #ifdef __LITTLE_ENDIAN__
  17055. __ai uint64x2_t vmull_u32(uint32x2_t __p0, uint32x2_t __p1) {
  17056.   uint64x2_t __ret;
  17057.   __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 51);
  17058.   return __ret;
  17059. }
  17060. #else
  17061. __ai uint64x2_t vmull_u32(uint32x2_t __p0, uint32x2_t __p1) {
  17062.   uint64x2_t __ret;
  17063.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17064.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  17065.   __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 51);
  17066.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17067.   return __ret;
  17068. }
  17069. __ai uint64x2_t __noswap_vmull_u32(uint32x2_t __p0, uint32x2_t __p1) {
  17070.   uint64x2_t __ret;
  17071.   __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 51);
  17072.   return __ret;
  17073. }
  17074. #endif
  17075.  
  17076. #ifdef __LITTLE_ENDIAN__
  17077. __ai uint32x4_t vmull_u16(uint16x4_t __p0, uint16x4_t __p1) {
  17078.   uint32x4_t __ret;
  17079.   __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 50);
  17080.   return __ret;
  17081. }
  17082. #else
  17083. __ai uint32x4_t vmull_u16(uint16x4_t __p0, uint16x4_t __p1) {
  17084.   uint32x4_t __ret;
  17085.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17086.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  17087.   __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 50);
  17088.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17089.   return __ret;
  17090. }
  17091. __ai uint32x4_t __noswap_vmull_u16(uint16x4_t __p0, uint16x4_t __p1) {
  17092.   uint32x4_t __ret;
  17093.   __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 50);
  17094.   return __ret;
  17095. }
  17096. #endif
  17097.  
  17098. #ifdef __LITTLE_ENDIAN__
  17099. __ai int16x8_t vmull_s8(int8x8_t __p0, int8x8_t __p1) {
  17100.   int16x8_t __ret;
  17101.   __ret = (int16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 33);
  17102.   return __ret;
  17103. }
  17104. #else
  17105. __ai int16x8_t vmull_s8(int8x8_t __p0, int8x8_t __p1) {
  17106.   int16x8_t __ret;
  17107.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17108.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  17109.   __ret = (int16x8_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 33);
  17110.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17111.   return __ret;
  17112. }
  17113. __ai int16x8_t __noswap_vmull_s8(int8x8_t __p0, int8x8_t __p1) {
  17114.   int16x8_t __ret;
  17115.   __ret = (int16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 33);
  17116.   return __ret;
  17117. }
  17118. #endif
  17119.  
  17120. #ifdef __LITTLE_ENDIAN__
  17121. __ai int64x2_t vmull_s32(int32x2_t __p0, int32x2_t __p1) {
  17122.   int64x2_t __ret;
  17123.   __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
  17124.   return __ret;
  17125. }
  17126. #else
  17127. __ai int64x2_t vmull_s32(int32x2_t __p0, int32x2_t __p1) {
  17128.   int64x2_t __ret;
  17129.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17130.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  17131.   __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 35);
  17132.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17133.   return __ret;
  17134. }
  17135. __ai int64x2_t __noswap_vmull_s32(int32x2_t __p0, int32x2_t __p1) {
  17136.   int64x2_t __ret;
  17137.   __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
  17138.   return __ret;
  17139. }
  17140. #endif
  17141.  
  17142. #ifdef __LITTLE_ENDIAN__
  17143. __ai int32x4_t vmull_s16(int16x4_t __p0, int16x4_t __p1) {
  17144.   int32x4_t __ret;
  17145.   __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
  17146.   return __ret;
  17147. }
  17148. #else
  17149. __ai int32x4_t vmull_s16(int16x4_t __p0, int16x4_t __p1) {
  17150.   int32x4_t __ret;
  17151.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17152.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  17153.   __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 34);
  17154.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17155.   return __ret;
  17156. }
  17157. __ai int32x4_t __noswap_vmull_s16(int16x4_t __p0, int16x4_t __p1) {
  17158.   int32x4_t __ret;
  17159.   __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
  17160.   return __ret;
  17161. }
  17162. #endif
  17163.  
  17164. #ifdef __LITTLE_ENDIAN__
  17165. #define vmull_lane_u32(__p0_106, __p1_106, __p2_106) __extension__ ({ \
  17166.   uint64x2_t __ret_106; \
  17167.   uint32x2_t __s0_106 = __p0_106; \
  17168.   uint32x2_t __s1_106 = __p1_106; \
  17169.   __ret_106 = vmull_u32(__s0_106, splat_lane_u32(__s1_106, __p2_106)); \
  17170.   __ret_106; \
  17171. })
  17172. #else
  17173. #define vmull_lane_u32(__p0_107, __p1_107, __p2_107) __extension__ ({ \
  17174.   uint64x2_t __ret_107; \
  17175.   uint32x2_t __s0_107 = __p0_107; \
  17176.   uint32x2_t __s1_107 = __p1_107; \
  17177.   uint32x2_t __rev0_107;  __rev0_107 = __builtin_shufflevector(__s0_107, __s0_107, 1, 0); \
  17178.   uint32x2_t __rev1_107;  __rev1_107 = __builtin_shufflevector(__s1_107, __s1_107, 1, 0); \
  17179.   __ret_107 = __noswap_vmull_u32(__rev0_107, __noswap_splat_lane_u32(__rev1_107, __p2_107)); \
  17180.   __ret_107 = __builtin_shufflevector(__ret_107, __ret_107, 1, 0); \
  17181.   __ret_107; \
  17182. })
  17183. #endif
  17184.  
  17185. #ifdef __LITTLE_ENDIAN__
  17186. #define vmull_lane_u16(__p0_108, __p1_108, __p2_108) __extension__ ({ \
  17187.   uint32x4_t __ret_108; \
  17188.   uint16x4_t __s0_108 = __p0_108; \
  17189.   uint16x4_t __s1_108 = __p1_108; \
  17190.   __ret_108 = vmull_u16(__s0_108, splat_lane_u16(__s1_108, __p2_108)); \
  17191.   __ret_108; \
  17192. })
  17193. #else
  17194. #define vmull_lane_u16(__p0_109, __p1_109, __p2_109) __extension__ ({ \
  17195.   uint32x4_t __ret_109; \
  17196.   uint16x4_t __s0_109 = __p0_109; \
  17197.   uint16x4_t __s1_109 = __p1_109; \
  17198.   uint16x4_t __rev0_109;  __rev0_109 = __builtin_shufflevector(__s0_109, __s0_109, 3, 2, 1, 0); \
  17199.   uint16x4_t __rev1_109;  __rev1_109 = __builtin_shufflevector(__s1_109, __s1_109, 3, 2, 1, 0); \
  17200.   __ret_109 = __noswap_vmull_u16(__rev0_109, __noswap_splat_lane_u16(__rev1_109, __p2_109)); \
  17201.   __ret_109 = __builtin_shufflevector(__ret_109, __ret_109, 3, 2, 1, 0); \
  17202.   __ret_109; \
  17203. })
  17204. #endif
  17205.  
  17206. #ifdef __LITTLE_ENDIAN__
  17207. #define vmull_lane_s32(__p0_110, __p1_110, __p2_110) __extension__ ({ \
  17208.   int64x2_t __ret_110; \
  17209.   int32x2_t __s0_110 = __p0_110; \
  17210.   int32x2_t __s1_110 = __p1_110; \
  17211.   __ret_110 = vmull_s32(__s0_110, splat_lane_s32(__s1_110, __p2_110)); \
  17212.   __ret_110; \
  17213. })
  17214. #else
  17215. #define vmull_lane_s32(__p0_111, __p1_111, __p2_111) __extension__ ({ \
  17216.   int64x2_t __ret_111; \
  17217.   int32x2_t __s0_111 = __p0_111; \
  17218.   int32x2_t __s1_111 = __p1_111; \
  17219.   int32x2_t __rev0_111;  __rev0_111 = __builtin_shufflevector(__s0_111, __s0_111, 1, 0); \
  17220.   int32x2_t __rev1_111;  __rev1_111 = __builtin_shufflevector(__s1_111, __s1_111, 1, 0); \
  17221.   __ret_111 = __noswap_vmull_s32(__rev0_111, __noswap_splat_lane_s32(__rev1_111, __p2_111)); \
  17222.   __ret_111 = __builtin_shufflevector(__ret_111, __ret_111, 1, 0); \
  17223.   __ret_111; \
  17224. })
  17225. #endif
  17226.  
  17227. #ifdef __LITTLE_ENDIAN__
  17228. #define vmull_lane_s16(__p0_112, __p1_112, __p2_112) __extension__ ({ \
  17229.   int32x4_t __ret_112; \
  17230.   int16x4_t __s0_112 = __p0_112; \
  17231.   int16x4_t __s1_112 = __p1_112; \
  17232.   __ret_112 = vmull_s16(__s0_112, splat_lane_s16(__s1_112, __p2_112)); \
  17233.   __ret_112; \
  17234. })
  17235. #else
  17236. #define vmull_lane_s16(__p0_113, __p1_113, __p2_113) __extension__ ({ \
  17237.   int32x4_t __ret_113; \
  17238.   int16x4_t __s0_113 = __p0_113; \
  17239.   int16x4_t __s1_113 = __p1_113; \
  17240.   int16x4_t __rev0_113;  __rev0_113 = __builtin_shufflevector(__s0_113, __s0_113, 3, 2, 1, 0); \
  17241.   int16x4_t __rev1_113;  __rev1_113 = __builtin_shufflevector(__s1_113, __s1_113, 3, 2, 1, 0); \
  17242.   __ret_113 = __noswap_vmull_s16(__rev0_113, __noswap_splat_lane_s16(__rev1_113, __p2_113)); \
  17243.   __ret_113 = __builtin_shufflevector(__ret_113, __ret_113, 3, 2, 1, 0); \
  17244.   __ret_113; \
  17245. })
  17246. #endif
  17247.  
  17248. #ifdef __LITTLE_ENDIAN__
  17249. __ai uint64x2_t vmull_n_u32(uint32x2_t __p0, uint32_t __p1) {
  17250.   uint64x2_t __ret;
  17251.   __ret = vmull_u32(__p0, (uint32x2_t) {__p1, __p1});
  17252.   return __ret;
  17253. }
  17254. #else
  17255. __ai uint64x2_t vmull_n_u32(uint32x2_t __p0, uint32_t __p1) {
  17256.   uint64x2_t __ret;
  17257.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17258.   __ret = __noswap_vmull_u32(__rev0, (uint32x2_t) {__p1, __p1});
  17259.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17260.   return __ret;
  17261. }
  17262. __ai uint64x2_t __noswap_vmull_n_u32(uint32x2_t __p0, uint32_t __p1) {
  17263.   uint64x2_t __ret;
  17264.   __ret = __noswap_vmull_u32(__p0, (uint32x2_t) {__p1, __p1});
  17265.   return __ret;
  17266. }
  17267. #endif
  17268.  
  17269. #ifdef __LITTLE_ENDIAN__
  17270. __ai uint32x4_t vmull_n_u16(uint16x4_t __p0, uint16_t __p1) {
  17271.   uint32x4_t __ret;
  17272.   __ret = vmull_u16(__p0, (uint16x4_t) {__p1, __p1, __p1, __p1});
  17273.   return __ret;
  17274. }
  17275. #else
  17276. __ai uint32x4_t vmull_n_u16(uint16x4_t __p0, uint16_t __p1) {
  17277.   uint32x4_t __ret;
  17278.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17279.   __ret = __noswap_vmull_u16(__rev0, (uint16x4_t) {__p1, __p1, __p1, __p1});
  17280.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17281.   return __ret;
  17282. }
  17283. __ai uint32x4_t __noswap_vmull_n_u16(uint16x4_t __p0, uint16_t __p1) {
  17284.   uint32x4_t __ret;
  17285.   __ret = __noswap_vmull_u16(__p0, (uint16x4_t) {__p1, __p1, __p1, __p1});
  17286.   return __ret;
  17287. }
  17288. #endif
  17289.  
  17290. #ifdef __LITTLE_ENDIAN__
  17291. __ai int64x2_t vmull_n_s32(int32x2_t __p0, int32_t __p1) {
  17292.   int64x2_t __ret;
  17293.   __ret = vmull_s32(__p0, (int32x2_t) {__p1, __p1});
  17294.   return __ret;
  17295. }
  17296. #else
  17297. __ai int64x2_t vmull_n_s32(int32x2_t __p0, int32_t __p1) {
  17298.   int64x2_t __ret;
  17299.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17300.   __ret = __noswap_vmull_s32(__rev0, (int32x2_t) {__p1, __p1});
  17301.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17302.   return __ret;
  17303. }
  17304. __ai int64x2_t __noswap_vmull_n_s32(int32x2_t __p0, int32_t __p1) {
  17305.   int64x2_t __ret;
  17306.   __ret = __noswap_vmull_s32(__p0, (int32x2_t) {__p1, __p1});
  17307.   return __ret;
  17308. }
  17309. #endif
  17310.  
  17311. #ifdef __LITTLE_ENDIAN__
  17312. __ai int32x4_t vmull_n_s16(int16x4_t __p0, int16_t __p1) {
  17313.   int32x4_t __ret;
  17314.   __ret = vmull_s16(__p0, (int16x4_t) {__p1, __p1, __p1, __p1});
  17315.   return __ret;
  17316. }
  17317. #else
  17318. __ai int32x4_t vmull_n_s16(int16x4_t __p0, int16_t __p1) {
  17319.   int32x4_t __ret;
  17320.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17321.   __ret = __noswap_vmull_s16(__rev0, (int16x4_t) {__p1, __p1, __p1, __p1});
  17322.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17323.   return __ret;
  17324. }
  17325. __ai int32x4_t __noswap_vmull_n_s16(int16x4_t __p0, int16_t __p1) {
  17326.   int32x4_t __ret;
  17327.   __ret = __noswap_vmull_s16(__p0, (int16x4_t) {__p1, __p1, __p1, __p1});
  17328.   return __ret;
  17329. }
  17330. #endif
  17331.  
  17332. #ifdef __LITTLE_ENDIAN__
  17333. __ai poly8x8_t vmvn_p8(poly8x8_t __p0) {
  17334.   poly8x8_t __ret;
  17335.   __ret = ~__p0;
  17336.   return __ret;
  17337. }
  17338. #else
  17339. __ai poly8x8_t vmvn_p8(poly8x8_t __p0) {
  17340.   poly8x8_t __ret;
  17341.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17342.   __ret = ~__rev0;
  17343.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17344.   return __ret;
  17345. }
  17346. #endif
  17347.  
  17348. #ifdef __LITTLE_ENDIAN__
  17349. __ai poly8x16_t vmvnq_p8(poly8x16_t __p0) {
  17350.   poly8x16_t __ret;
  17351.   __ret = ~__p0;
  17352.   return __ret;
  17353. }
  17354. #else
  17355. __ai poly8x16_t vmvnq_p8(poly8x16_t __p0) {
  17356.   poly8x16_t __ret;
  17357.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17358.   __ret = ~__rev0;
  17359.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17360.   return __ret;
  17361. }
  17362. #endif
  17363.  
  17364. #ifdef __LITTLE_ENDIAN__
  17365. __ai uint8x16_t vmvnq_u8(uint8x16_t __p0) {
  17366.   uint8x16_t __ret;
  17367.   __ret = ~__p0;
  17368.   return __ret;
  17369. }
  17370. #else
  17371. __ai uint8x16_t vmvnq_u8(uint8x16_t __p0) {
  17372.   uint8x16_t __ret;
  17373.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17374.   __ret = ~__rev0;
  17375.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17376.   return __ret;
  17377. }
  17378. #endif
  17379.  
  17380. #ifdef __LITTLE_ENDIAN__
  17381. __ai uint32x4_t vmvnq_u32(uint32x4_t __p0) {
  17382.   uint32x4_t __ret;
  17383.   __ret = ~__p0;
  17384.   return __ret;
  17385. }
  17386. #else
  17387. __ai uint32x4_t vmvnq_u32(uint32x4_t __p0) {
  17388.   uint32x4_t __ret;
  17389.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17390.   __ret = ~__rev0;
  17391.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17392.   return __ret;
  17393. }
  17394. #endif
  17395.  
  17396. #ifdef __LITTLE_ENDIAN__
  17397. __ai uint16x8_t vmvnq_u16(uint16x8_t __p0) {
  17398.   uint16x8_t __ret;
  17399.   __ret = ~__p0;
  17400.   return __ret;
  17401. }
  17402. #else
  17403. __ai uint16x8_t vmvnq_u16(uint16x8_t __p0) {
  17404.   uint16x8_t __ret;
  17405.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17406.   __ret = ~__rev0;
  17407.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17408.   return __ret;
  17409. }
  17410. #endif
  17411.  
  17412. #ifdef __LITTLE_ENDIAN__
  17413. __ai int8x16_t vmvnq_s8(int8x16_t __p0) {
  17414.   int8x16_t __ret;
  17415.   __ret = ~__p0;
  17416.   return __ret;
  17417. }
  17418. #else
  17419. __ai int8x16_t vmvnq_s8(int8x16_t __p0) {
  17420.   int8x16_t __ret;
  17421.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17422.   __ret = ~__rev0;
  17423.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17424.   return __ret;
  17425. }
  17426. #endif
  17427.  
  17428. #ifdef __LITTLE_ENDIAN__
  17429. __ai int32x4_t vmvnq_s32(int32x4_t __p0) {
  17430.   int32x4_t __ret;
  17431.   __ret = ~__p0;
  17432.   return __ret;
  17433. }
  17434. #else
  17435. __ai int32x4_t vmvnq_s32(int32x4_t __p0) {
  17436.   int32x4_t __ret;
  17437.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17438.   __ret = ~__rev0;
  17439.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17440.   return __ret;
  17441. }
  17442. #endif
  17443.  
  17444. #ifdef __LITTLE_ENDIAN__
  17445. __ai int16x8_t vmvnq_s16(int16x8_t __p0) {
  17446.   int16x8_t __ret;
  17447.   __ret = ~__p0;
  17448.   return __ret;
  17449. }
  17450. #else
  17451. __ai int16x8_t vmvnq_s16(int16x8_t __p0) {
  17452.   int16x8_t __ret;
  17453.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17454.   __ret = ~__rev0;
  17455.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17456.   return __ret;
  17457. }
  17458. #endif
  17459.  
  17460. #ifdef __LITTLE_ENDIAN__
  17461. __ai uint8x8_t vmvn_u8(uint8x8_t __p0) {
  17462.   uint8x8_t __ret;
  17463.   __ret = ~__p0;
  17464.   return __ret;
  17465. }
  17466. #else
  17467. __ai uint8x8_t vmvn_u8(uint8x8_t __p0) {
  17468.   uint8x8_t __ret;
  17469.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17470.   __ret = ~__rev0;
  17471.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17472.   return __ret;
  17473. }
  17474. #endif
  17475.  
  17476. #ifdef __LITTLE_ENDIAN__
  17477. __ai uint32x2_t vmvn_u32(uint32x2_t __p0) {
  17478.   uint32x2_t __ret;
  17479.   __ret = ~__p0;
  17480.   return __ret;
  17481. }
  17482. #else
  17483. __ai uint32x2_t vmvn_u32(uint32x2_t __p0) {
  17484.   uint32x2_t __ret;
  17485.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17486.   __ret = ~__rev0;
  17487.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17488.   return __ret;
  17489. }
  17490. #endif
  17491.  
  17492. #ifdef __LITTLE_ENDIAN__
  17493. __ai uint16x4_t vmvn_u16(uint16x4_t __p0) {
  17494.   uint16x4_t __ret;
  17495.   __ret = ~__p0;
  17496.   return __ret;
  17497. }
  17498. #else
  17499. __ai uint16x4_t vmvn_u16(uint16x4_t __p0) {
  17500.   uint16x4_t __ret;
  17501.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17502.   __ret = ~__rev0;
  17503.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17504.   return __ret;
  17505. }
  17506. #endif
  17507.  
  17508. #ifdef __LITTLE_ENDIAN__
  17509. __ai int8x8_t vmvn_s8(int8x8_t __p0) {
  17510.   int8x8_t __ret;
  17511.   __ret = ~__p0;
  17512.   return __ret;
  17513. }
  17514. #else
  17515. __ai int8x8_t vmvn_s8(int8x8_t __p0) {
  17516.   int8x8_t __ret;
  17517.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17518.   __ret = ~__rev0;
  17519.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17520.   return __ret;
  17521. }
  17522. #endif
  17523.  
  17524. #ifdef __LITTLE_ENDIAN__
  17525. __ai int32x2_t vmvn_s32(int32x2_t __p0) {
  17526.   int32x2_t __ret;
  17527.   __ret = ~__p0;
  17528.   return __ret;
  17529. }
  17530. #else
  17531. __ai int32x2_t vmvn_s32(int32x2_t __p0) {
  17532.   int32x2_t __ret;
  17533.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17534.   __ret = ~__rev0;
  17535.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17536.   return __ret;
  17537. }
  17538. #endif
  17539.  
  17540. #ifdef __LITTLE_ENDIAN__
  17541. __ai int16x4_t vmvn_s16(int16x4_t __p0) {
  17542.   int16x4_t __ret;
  17543.   __ret = ~__p0;
  17544.   return __ret;
  17545. }
  17546. #else
  17547. __ai int16x4_t vmvn_s16(int16x4_t __p0) {
  17548.   int16x4_t __ret;
  17549.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17550.   __ret = ~__rev0;
  17551.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17552.   return __ret;
  17553. }
  17554. #endif
  17555.  
  17556. #ifdef __LITTLE_ENDIAN__
  17557. __ai int8x16_t vnegq_s8(int8x16_t __p0) {
  17558.   int8x16_t __ret;
  17559.   __ret = -__p0;
  17560.   return __ret;
  17561. }
  17562. #else
  17563. __ai int8x16_t vnegq_s8(int8x16_t __p0) {
  17564.   int8x16_t __ret;
  17565.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17566.   __ret = -__rev0;
  17567.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17568.   return __ret;
  17569. }
  17570. #endif
  17571.  
  17572. #ifdef __LITTLE_ENDIAN__
  17573. __ai float32x4_t vnegq_f32(float32x4_t __p0) {
  17574.   float32x4_t __ret;
  17575.   __ret = -__p0;
  17576.   return __ret;
  17577. }
  17578. #else
  17579. __ai float32x4_t vnegq_f32(float32x4_t __p0) {
  17580.   float32x4_t __ret;
  17581.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17582.   __ret = -__rev0;
  17583.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17584.   return __ret;
  17585. }
  17586. #endif
  17587.  
  17588. #ifdef __LITTLE_ENDIAN__
  17589. __ai int32x4_t vnegq_s32(int32x4_t __p0) {
  17590.   int32x4_t __ret;
  17591.   __ret = -__p0;
  17592.   return __ret;
  17593. }
  17594. #else
  17595. __ai int32x4_t vnegq_s32(int32x4_t __p0) {
  17596.   int32x4_t __ret;
  17597.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17598.   __ret = -__rev0;
  17599.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17600.   return __ret;
  17601. }
  17602. #endif
  17603.  
  17604. #ifdef __LITTLE_ENDIAN__
  17605. __ai int16x8_t vnegq_s16(int16x8_t __p0) {
  17606.   int16x8_t __ret;
  17607.   __ret = -__p0;
  17608.   return __ret;
  17609. }
  17610. #else
  17611. __ai int16x8_t vnegq_s16(int16x8_t __p0) {
  17612.   int16x8_t __ret;
  17613.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17614.   __ret = -__rev0;
  17615.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17616.   return __ret;
  17617. }
  17618. #endif
  17619.  
  17620. #ifdef __LITTLE_ENDIAN__
  17621. __ai int8x8_t vneg_s8(int8x8_t __p0) {
  17622.   int8x8_t __ret;
  17623.   __ret = -__p0;
  17624.   return __ret;
  17625. }
  17626. #else
  17627. __ai int8x8_t vneg_s8(int8x8_t __p0) {
  17628.   int8x8_t __ret;
  17629.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17630.   __ret = -__rev0;
  17631.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17632.   return __ret;
  17633. }
  17634. #endif
  17635.  
  17636. #ifdef __LITTLE_ENDIAN__
  17637. __ai float32x2_t vneg_f32(float32x2_t __p0) {
  17638.   float32x2_t __ret;
  17639.   __ret = -__p0;
  17640.   return __ret;
  17641. }
  17642. #else
  17643. __ai float32x2_t vneg_f32(float32x2_t __p0) {
  17644.   float32x2_t __ret;
  17645.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17646.   __ret = -__rev0;
  17647.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17648.   return __ret;
  17649. }
  17650. #endif
  17651.  
  17652. #ifdef __LITTLE_ENDIAN__
  17653. __ai int32x2_t vneg_s32(int32x2_t __p0) {
  17654.   int32x2_t __ret;
  17655.   __ret = -__p0;
  17656.   return __ret;
  17657. }
  17658. #else
  17659. __ai int32x2_t vneg_s32(int32x2_t __p0) {
  17660.   int32x2_t __ret;
  17661.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17662.   __ret = -__rev0;
  17663.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17664.   return __ret;
  17665. }
  17666. #endif
  17667.  
  17668. #ifdef __LITTLE_ENDIAN__
  17669. __ai int16x4_t vneg_s16(int16x4_t __p0) {
  17670.   int16x4_t __ret;
  17671.   __ret = -__p0;
  17672.   return __ret;
  17673. }
  17674. #else
  17675. __ai int16x4_t vneg_s16(int16x4_t __p0) {
  17676.   int16x4_t __ret;
  17677.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17678.   __ret = -__rev0;
  17679.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17680.   return __ret;
  17681. }
  17682. #endif
  17683.  
  17684. #ifdef __LITTLE_ENDIAN__
  17685. __ai uint8x16_t vornq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  17686.   uint8x16_t __ret;
  17687.   __ret = __p0 | ~__p1;
  17688.   return __ret;
  17689. }
  17690. #else
  17691. __ai uint8x16_t vornq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  17692.   uint8x16_t __ret;
  17693.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17694.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17695.   __ret = __rev0 | ~__rev1;
  17696.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17697.   return __ret;
  17698. }
  17699. #endif
  17700.  
  17701. #ifdef __LITTLE_ENDIAN__
  17702. __ai uint32x4_t vornq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  17703.   uint32x4_t __ret;
  17704.   __ret = __p0 | ~__p1;
  17705.   return __ret;
  17706. }
  17707. #else
  17708. __ai uint32x4_t vornq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  17709.   uint32x4_t __ret;
  17710.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17711.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  17712.   __ret = __rev0 | ~__rev1;
  17713.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17714.   return __ret;
  17715. }
  17716. #endif
  17717.  
  17718. #ifdef __LITTLE_ENDIAN__
  17719. __ai uint64x2_t vornq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  17720.   uint64x2_t __ret;
  17721.   __ret = __p0 | ~__p1;
  17722.   return __ret;
  17723. }
  17724. #else
  17725. __ai uint64x2_t vornq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  17726.   uint64x2_t __ret;
  17727.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17728.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  17729.   __ret = __rev0 | ~__rev1;
  17730.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17731.   return __ret;
  17732. }
  17733. #endif
  17734.  
  17735. #ifdef __LITTLE_ENDIAN__
  17736. __ai uint16x8_t vornq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  17737.   uint16x8_t __ret;
  17738.   __ret = __p0 | ~__p1;
  17739.   return __ret;
  17740. }
  17741. #else
  17742. __ai uint16x8_t vornq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  17743.   uint16x8_t __ret;
  17744.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17745.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  17746.   __ret = __rev0 | ~__rev1;
  17747.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17748.   return __ret;
  17749. }
  17750. #endif
  17751.  
  17752. #ifdef __LITTLE_ENDIAN__
  17753. __ai int8x16_t vornq_s8(int8x16_t __p0, int8x16_t __p1) {
  17754.   int8x16_t __ret;
  17755.   __ret = __p0 | ~__p1;
  17756.   return __ret;
  17757. }
  17758. #else
  17759. __ai int8x16_t vornq_s8(int8x16_t __p0, int8x16_t __p1) {
  17760.   int8x16_t __ret;
  17761.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17762.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17763.   __ret = __rev0 | ~__rev1;
  17764.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17765.   return __ret;
  17766. }
  17767. #endif
  17768.  
  17769. #ifdef __LITTLE_ENDIAN__
  17770. __ai int32x4_t vornq_s32(int32x4_t __p0, int32x4_t __p1) {
  17771.   int32x4_t __ret;
  17772.   __ret = __p0 | ~__p1;
  17773.   return __ret;
  17774. }
  17775. #else
  17776. __ai int32x4_t vornq_s32(int32x4_t __p0, int32x4_t __p1) {
  17777.   int32x4_t __ret;
  17778.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17779.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  17780.   __ret = __rev0 | ~__rev1;
  17781.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17782.   return __ret;
  17783. }
  17784. #endif
  17785.  
  17786. #ifdef __LITTLE_ENDIAN__
  17787. __ai int64x2_t vornq_s64(int64x2_t __p0, int64x2_t __p1) {
  17788.   int64x2_t __ret;
  17789.   __ret = __p0 | ~__p1;
  17790.   return __ret;
  17791. }
  17792. #else
  17793. __ai int64x2_t vornq_s64(int64x2_t __p0, int64x2_t __p1) {
  17794.   int64x2_t __ret;
  17795.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17796.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  17797.   __ret = __rev0 | ~__rev1;
  17798.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17799.   return __ret;
  17800. }
  17801. #endif
  17802.  
  17803. #ifdef __LITTLE_ENDIAN__
  17804. __ai int16x8_t vornq_s16(int16x8_t __p0, int16x8_t __p1) {
  17805.   int16x8_t __ret;
  17806.   __ret = __p0 | ~__p1;
  17807.   return __ret;
  17808. }
  17809. #else
  17810. __ai int16x8_t vornq_s16(int16x8_t __p0, int16x8_t __p1) {
  17811.   int16x8_t __ret;
  17812.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17813.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  17814.   __ret = __rev0 | ~__rev1;
  17815.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17816.   return __ret;
  17817. }
  17818. #endif
  17819.  
  17820. #ifdef __LITTLE_ENDIAN__
  17821. __ai uint8x8_t vorn_u8(uint8x8_t __p0, uint8x8_t __p1) {
  17822.   uint8x8_t __ret;
  17823.   __ret = __p0 | ~__p1;
  17824.   return __ret;
  17825. }
  17826. #else
  17827. __ai uint8x8_t vorn_u8(uint8x8_t __p0, uint8x8_t __p1) {
  17828.   uint8x8_t __ret;
  17829.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17830.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  17831.   __ret = __rev0 | ~__rev1;
  17832.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17833.   return __ret;
  17834. }
  17835. #endif
  17836.  
  17837. #ifdef __LITTLE_ENDIAN__
  17838. __ai uint32x2_t vorn_u32(uint32x2_t __p0, uint32x2_t __p1) {
  17839.   uint32x2_t __ret;
  17840.   __ret = __p0 | ~__p1;
  17841.   return __ret;
  17842. }
  17843. #else
  17844. __ai uint32x2_t vorn_u32(uint32x2_t __p0, uint32x2_t __p1) {
  17845.   uint32x2_t __ret;
  17846.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17847.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  17848.   __ret = __rev0 | ~__rev1;
  17849.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17850.   return __ret;
  17851. }
  17852. #endif
  17853.  
  17854. __ai uint64x1_t vorn_u64(uint64x1_t __p0, uint64x1_t __p1) {
  17855.   uint64x1_t __ret;
  17856.   __ret = __p0 | ~__p1;
  17857.   return __ret;
  17858. }
  17859. #ifdef __LITTLE_ENDIAN__
  17860. __ai uint16x4_t vorn_u16(uint16x4_t __p0, uint16x4_t __p1) {
  17861.   uint16x4_t __ret;
  17862.   __ret = __p0 | ~__p1;
  17863.   return __ret;
  17864. }
  17865. #else
  17866. __ai uint16x4_t vorn_u16(uint16x4_t __p0, uint16x4_t __p1) {
  17867.   uint16x4_t __ret;
  17868.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17869.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  17870.   __ret = __rev0 | ~__rev1;
  17871.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17872.   return __ret;
  17873. }
  17874. #endif
  17875.  
  17876. #ifdef __LITTLE_ENDIAN__
  17877. __ai int8x8_t vorn_s8(int8x8_t __p0, int8x8_t __p1) {
  17878.   int8x8_t __ret;
  17879.   __ret = __p0 | ~__p1;
  17880.   return __ret;
  17881. }
  17882. #else
  17883. __ai int8x8_t vorn_s8(int8x8_t __p0, int8x8_t __p1) {
  17884.   int8x8_t __ret;
  17885.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17886.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  17887.   __ret = __rev0 | ~__rev1;
  17888.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17889.   return __ret;
  17890. }
  17891. #endif
  17892.  
  17893. #ifdef __LITTLE_ENDIAN__
  17894. __ai int32x2_t vorn_s32(int32x2_t __p0, int32x2_t __p1) {
  17895.   int32x2_t __ret;
  17896.   __ret = __p0 | ~__p1;
  17897.   return __ret;
  17898. }
  17899. #else
  17900. __ai int32x2_t vorn_s32(int32x2_t __p0, int32x2_t __p1) {
  17901.   int32x2_t __ret;
  17902.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17903.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  17904.   __ret = __rev0 | ~__rev1;
  17905.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17906.   return __ret;
  17907. }
  17908. #endif
  17909.  
  17910. __ai int64x1_t vorn_s64(int64x1_t __p0, int64x1_t __p1) {
  17911.   int64x1_t __ret;
  17912.   __ret = __p0 | ~__p1;
  17913.   return __ret;
  17914. }
  17915. #ifdef __LITTLE_ENDIAN__
  17916. __ai int16x4_t vorn_s16(int16x4_t __p0, int16x4_t __p1) {
  17917.   int16x4_t __ret;
  17918.   __ret = __p0 | ~__p1;
  17919.   return __ret;
  17920. }
  17921. #else
  17922. __ai int16x4_t vorn_s16(int16x4_t __p0, int16x4_t __p1) {
  17923.   int16x4_t __ret;
  17924.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17925.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  17926.   __ret = __rev0 | ~__rev1;
  17927.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17928.   return __ret;
  17929. }
  17930. #endif
  17931.  
  17932. #ifdef __LITTLE_ENDIAN__
  17933. __ai uint8x16_t vorrq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  17934.   uint8x16_t __ret;
  17935.   __ret = __p0 | __p1;
  17936.   return __ret;
  17937. }
  17938. #else
  17939. __ai uint8x16_t vorrq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  17940.   uint8x16_t __ret;
  17941.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17942.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17943.   __ret = __rev0 | __rev1;
  17944.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  17945.   return __ret;
  17946. }
  17947. #endif
  17948.  
  17949. #ifdef __LITTLE_ENDIAN__
  17950. __ai uint32x4_t vorrq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  17951.   uint32x4_t __ret;
  17952.   __ret = __p0 | __p1;
  17953.   return __ret;
  17954. }
  17955. #else
  17956. __ai uint32x4_t vorrq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  17957.   uint32x4_t __ret;
  17958.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  17959.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  17960.   __ret = __rev0 | __rev1;
  17961.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  17962.   return __ret;
  17963. }
  17964. #endif
  17965.  
  17966. #ifdef __LITTLE_ENDIAN__
  17967. __ai uint64x2_t vorrq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  17968.   uint64x2_t __ret;
  17969.   __ret = __p0 | __p1;
  17970.   return __ret;
  17971. }
  17972. #else
  17973. __ai uint64x2_t vorrq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  17974.   uint64x2_t __ret;
  17975.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  17976.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  17977.   __ret = __rev0 | __rev1;
  17978.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  17979.   return __ret;
  17980. }
  17981. #endif
  17982.  
  17983. #ifdef __LITTLE_ENDIAN__
  17984. __ai uint16x8_t vorrq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  17985.   uint16x8_t __ret;
  17986.   __ret = __p0 | __p1;
  17987.   return __ret;
  17988. }
  17989. #else
  17990. __ai uint16x8_t vorrq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  17991.   uint16x8_t __ret;
  17992.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  17993.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  17994.   __ret = __rev0 | __rev1;
  17995.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  17996.   return __ret;
  17997. }
  17998. #endif
  17999.  
  18000. #ifdef __LITTLE_ENDIAN__
  18001. __ai int8x16_t vorrq_s8(int8x16_t __p0, int8x16_t __p1) {
  18002.   int8x16_t __ret;
  18003.   __ret = __p0 | __p1;
  18004.   return __ret;
  18005. }
  18006. #else
  18007. __ai int8x16_t vorrq_s8(int8x16_t __p0, int8x16_t __p1) {
  18008.   int8x16_t __ret;
  18009.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  18010.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  18011.   __ret = __rev0 | __rev1;
  18012.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  18013.   return __ret;
  18014. }
  18015. #endif
  18016.  
  18017. #ifdef __LITTLE_ENDIAN__
  18018. __ai int32x4_t vorrq_s32(int32x4_t __p0, int32x4_t __p1) {
  18019.   int32x4_t __ret;
  18020.   __ret = __p0 | __p1;
  18021.   return __ret;
  18022. }
  18023. #else
  18024. __ai int32x4_t vorrq_s32(int32x4_t __p0, int32x4_t __p1) {
  18025.   int32x4_t __ret;
  18026.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18027.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18028.   __ret = __rev0 | __rev1;
  18029.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18030.   return __ret;
  18031. }
  18032. #endif
  18033.  
  18034. #ifdef __LITTLE_ENDIAN__
  18035. __ai int64x2_t vorrq_s64(int64x2_t __p0, int64x2_t __p1) {
  18036.   int64x2_t __ret;
  18037.   __ret = __p0 | __p1;
  18038.   return __ret;
  18039. }
  18040. #else
  18041. __ai int64x2_t vorrq_s64(int64x2_t __p0, int64x2_t __p1) {
  18042.   int64x2_t __ret;
  18043.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18044.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18045.   __ret = __rev0 | __rev1;
  18046.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18047.   return __ret;
  18048. }
  18049. #endif
  18050.  
  18051. #ifdef __LITTLE_ENDIAN__
  18052. __ai int16x8_t vorrq_s16(int16x8_t __p0, int16x8_t __p1) {
  18053.   int16x8_t __ret;
  18054.   __ret = __p0 | __p1;
  18055.   return __ret;
  18056. }
  18057. #else
  18058. __ai int16x8_t vorrq_s16(int16x8_t __p0, int16x8_t __p1) {
  18059.   int16x8_t __ret;
  18060.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18061.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18062.   __ret = __rev0 | __rev1;
  18063.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18064.   return __ret;
  18065. }
  18066. #endif
  18067.  
  18068. #ifdef __LITTLE_ENDIAN__
  18069. __ai uint8x8_t vorr_u8(uint8x8_t __p0, uint8x8_t __p1) {
  18070.   uint8x8_t __ret;
  18071.   __ret = __p0 | __p1;
  18072.   return __ret;
  18073. }
  18074. #else
  18075. __ai uint8x8_t vorr_u8(uint8x8_t __p0, uint8x8_t __p1) {
  18076.   uint8x8_t __ret;
  18077.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18078.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18079.   __ret = __rev0 | __rev1;
  18080.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18081.   return __ret;
  18082. }
  18083. #endif
  18084.  
  18085. #ifdef __LITTLE_ENDIAN__
  18086. __ai uint32x2_t vorr_u32(uint32x2_t __p0, uint32x2_t __p1) {
  18087.   uint32x2_t __ret;
  18088.   __ret = __p0 | __p1;
  18089.   return __ret;
  18090. }
  18091. #else
  18092. __ai uint32x2_t vorr_u32(uint32x2_t __p0, uint32x2_t __p1) {
  18093.   uint32x2_t __ret;
  18094.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18095.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18096.   __ret = __rev0 | __rev1;
  18097.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18098.   return __ret;
  18099. }
  18100. #endif
  18101.  
  18102. __ai uint64x1_t vorr_u64(uint64x1_t __p0, uint64x1_t __p1) {
  18103.   uint64x1_t __ret;
  18104.   __ret = __p0 | __p1;
  18105.   return __ret;
  18106. }
  18107. #ifdef __LITTLE_ENDIAN__
  18108. __ai uint16x4_t vorr_u16(uint16x4_t __p0, uint16x4_t __p1) {
  18109.   uint16x4_t __ret;
  18110.   __ret = __p0 | __p1;
  18111.   return __ret;
  18112. }
  18113. #else
  18114. __ai uint16x4_t vorr_u16(uint16x4_t __p0, uint16x4_t __p1) {
  18115.   uint16x4_t __ret;
  18116.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18117.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18118.   __ret = __rev0 | __rev1;
  18119.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18120.   return __ret;
  18121. }
  18122. #endif
  18123.  
  18124. #ifdef __LITTLE_ENDIAN__
  18125. __ai int8x8_t vorr_s8(int8x8_t __p0, int8x8_t __p1) {
  18126.   int8x8_t __ret;
  18127.   __ret = __p0 | __p1;
  18128.   return __ret;
  18129. }
  18130. #else
  18131. __ai int8x8_t vorr_s8(int8x8_t __p0, int8x8_t __p1) {
  18132.   int8x8_t __ret;
  18133.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18134.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18135.   __ret = __rev0 | __rev1;
  18136.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18137.   return __ret;
  18138. }
  18139. #endif
  18140.  
  18141. #ifdef __LITTLE_ENDIAN__
  18142. __ai int32x2_t vorr_s32(int32x2_t __p0, int32x2_t __p1) {
  18143.   int32x2_t __ret;
  18144.   __ret = __p0 | __p1;
  18145.   return __ret;
  18146. }
  18147. #else
  18148. __ai int32x2_t vorr_s32(int32x2_t __p0, int32x2_t __p1) {
  18149.   int32x2_t __ret;
  18150.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18151.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18152.   __ret = __rev0 | __rev1;
  18153.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18154.   return __ret;
  18155. }
  18156. #endif
  18157.  
  18158. __ai int64x1_t vorr_s64(int64x1_t __p0, int64x1_t __p1) {
  18159.   int64x1_t __ret;
  18160.   __ret = __p0 | __p1;
  18161.   return __ret;
  18162. }
  18163. #ifdef __LITTLE_ENDIAN__
  18164. __ai int16x4_t vorr_s16(int16x4_t __p0, int16x4_t __p1) {
  18165.   int16x4_t __ret;
  18166.   __ret = __p0 | __p1;
  18167.   return __ret;
  18168. }
  18169. #else
  18170. __ai int16x4_t vorr_s16(int16x4_t __p0, int16x4_t __p1) {
  18171.   int16x4_t __ret;
  18172.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18173.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18174.   __ret = __rev0 | __rev1;
  18175.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18176.   return __ret;
  18177. }
  18178. #endif
  18179.  
  18180. #ifdef __LITTLE_ENDIAN__
  18181. __ai uint16x8_t vpadalq_u8(uint16x8_t __p0, uint8x16_t __p1) {
  18182.   uint16x8_t __ret;
  18183.   __ret = (uint16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  18184.   return __ret;
  18185. }
  18186. #else
  18187. __ai uint16x8_t vpadalq_u8(uint16x8_t __p0, uint8x16_t __p1) {
  18188.   uint16x8_t __ret;
  18189.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18190.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  18191.   __ret = (uint16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  18192.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18193.   return __ret;
  18194. }
  18195. #endif
  18196.  
  18197. #ifdef __LITTLE_ENDIAN__
  18198. __ai uint64x2_t vpadalq_u32(uint64x2_t __p0, uint32x4_t __p1) {
  18199.   uint64x2_t __ret;
  18200.   __ret = (uint64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  18201.   return __ret;
  18202. }
  18203. #else
  18204. __ai uint64x2_t vpadalq_u32(uint64x2_t __p0, uint32x4_t __p1) {
  18205.   uint64x2_t __ret;
  18206.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18207.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18208.   __ret = (uint64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  18209.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18210.   return __ret;
  18211. }
  18212. #endif
  18213.  
  18214. #ifdef __LITTLE_ENDIAN__
  18215. __ai uint32x4_t vpadalq_u16(uint32x4_t __p0, uint16x8_t __p1) {
  18216.   uint32x4_t __ret;
  18217.   __ret = (uint32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  18218.   return __ret;
  18219. }
  18220. #else
  18221. __ai uint32x4_t vpadalq_u16(uint32x4_t __p0, uint16x8_t __p1) {
  18222.   uint32x4_t __ret;
  18223.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18224.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18225.   __ret = (uint32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  18226.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18227.   return __ret;
  18228. }
  18229. #endif
  18230.  
  18231. #ifdef __LITTLE_ENDIAN__
  18232. __ai int16x8_t vpadalq_s8(int16x8_t __p0, int8x16_t __p1) {
  18233.   int16x8_t __ret;
  18234.   __ret = (int16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  18235.   return __ret;
  18236. }
  18237. #else
  18238. __ai int16x8_t vpadalq_s8(int16x8_t __p0, int8x16_t __p1) {
  18239.   int16x8_t __ret;
  18240.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18241.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  18242.   __ret = (int16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  18243.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18244.   return __ret;
  18245. }
  18246. #endif
  18247.  
  18248. #ifdef __LITTLE_ENDIAN__
  18249. __ai int64x2_t vpadalq_s32(int64x2_t __p0, int32x4_t __p1) {
  18250.   int64x2_t __ret;
  18251.   __ret = (int64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
  18252.   return __ret;
  18253. }
  18254. #else
  18255. __ai int64x2_t vpadalq_s32(int64x2_t __p0, int32x4_t __p1) {
  18256.   int64x2_t __ret;
  18257.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18258.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18259.   __ret = (int64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
  18260.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18261.   return __ret;
  18262. }
  18263. #endif
  18264.  
  18265. #ifdef __LITTLE_ENDIAN__
  18266. __ai int32x4_t vpadalq_s16(int32x4_t __p0, int16x8_t __p1) {
  18267.   int32x4_t __ret;
  18268.   __ret = (int32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  18269.   return __ret;
  18270. }
  18271. #else
  18272. __ai int32x4_t vpadalq_s16(int32x4_t __p0, int16x8_t __p1) {
  18273.   int32x4_t __ret;
  18274.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18275.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18276.   __ret = (int32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  18277.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18278.   return __ret;
  18279. }
  18280. #endif
  18281.  
  18282. #ifdef __LITTLE_ENDIAN__
  18283. __ai uint16x4_t vpadal_u8(uint16x4_t __p0, uint8x8_t __p1) {
  18284.   uint16x4_t __ret;
  18285.   __ret = (uint16x4_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  18286.   return __ret;
  18287. }
  18288. #else
  18289. __ai uint16x4_t vpadal_u8(uint16x4_t __p0, uint8x8_t __p1) {
  18290.   uint16x4_t __ret;
  18291.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18292.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18293.   __ret = (uint16x4_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  18294.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18295.   return __ret;
  18296. }
  18297. #endif
  18298.  
  18299. #ifdef __LITTLE_ENDIAN__
  18300. __ai uint64x1_t vpadal_u32(uint64x1_t __p0, uint32x2_t __p1) {
  18301.   uint64x1_t __ret;
  18302.   __ret = (uint64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  18303.   return __ret;
  18304. }
  18305. #else
  18306. __ai uint64x1_t vpadal_u32(uint64x1_t __p0, uint32x2_t __p1) {
  18307.   uint64x1_t __ret;
  18308.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18309.   __ret = (uint64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__rev1, 19);
  18310.   return __ret;
  18311. }
  18312. #endif
  18313.  
  18314. #ifdef __LITTLE_ENDIAN__
  18315. __ai uint32x2_t vpadal_u16(uint32x2_t __p0, uint16x4_t __p1) {
  18316.   uint32x2_t __ret;
  18317.   __ret = (uint32x2_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  18318.   return __ret;
  18319. }
  18320. #else
  18321. __ai uint32x2_t vpadal_u16(uint32x2_t __p0, uint16x4_t __p1) {
  18322.   uint32x2_t __ret;
  18323.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18324.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18325.   __ret = (uint32x2_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  18326.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18327.   return __ret;
  18328. }
  18329. #endif
  18330.  
  18331. #ifdef __LITTLE_ENDIAN__
  18332. __ai int16x4_t vpadal_s8(int16x4_t __p0, int8x8_t __p1) {
  18333.   int16x4_t __ret;
  18334.   __ret = (int16x4_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  18335.   return __ret;
  18336. }
  18337. #else
  18338. __ai int16x4_t vpadal_s8(int16x4_t __p0, int8x8_t __p1) {
  18339.   int16x4_t __ret;
  18340.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18341.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18342.   __ret = (int16x4_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  18343.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18344.   return __ret;
  18345. }
  18346. #endif
  18347.  
  18348. #ifdef __LITTLE_ENDIAN__
  18349. __ai int64x1_t vpadal_s32(int64x1_t __p0, int32x2_t __p1) {
  18350.   int64x1_t __ret;
  18351.   __ret = (int64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
  18352.   return __ret;
  18353. }
  18354. #else
  18355. __ai int64x1_t vpadal_s32(int64x1_t __p0, int32x2_t __p1) {
  18356.   int64x1_t __ret;
  18357.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18358.   __ret = (int64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__rev1, 3);
  18359.   return __ret;
  18360. }
  18361. #endif
  18362.  
  18363. #ifdef __LITTLE_ENDIAN__
  18364. __ai int32x2_t vpadal_s16(int32x2_t __p0, int16x4_t __p1) {
  18365.   int32x2_t __ret;
  18366.   __ret = (int32x2_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  18367.   return __ret;
  18368. }
  18369. #else
  18370. __ai int32x2_t vpadal_s16(int32x2_t __p0, int16x4_t __p1) {
  18371.   int32x2_t __ret;
  18372.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18373.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18374.   __ret = (int32x2_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  18375.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18376.   return __ret;
  18377. }
  18378. #endif
  18379.  
  18380. #ifdef __LITTLE_ENDIAN__
  18381. __ai uint8x8_t vpadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  18382.   uint8x8_t __ret;
  18383.   __ret = (uint8x8_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  18384.   return __ret;
  18385. }
  18386. #else
  18387. __ai uint8x8_t vpadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  18388.   uint8x8_t __ret;
  18389.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18390.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18391.   __ret = (uint8x8_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  18392.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18393.   return __ret;
  18394. }
  18395. #endif
  18396.  
  18397. #ifdef __LITTLE_ENDIAN__
  18398. __ai uint32x2_t vpadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  18399.   uint32x2_t __ret;
  18400.   __ret = (uint32x2_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  18401.   return __ret;
  18402. }
  18403. #else
  18404. __ai uint32x2_t vpadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  18405.   uint32x2_t __ret;
  18406.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18407.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18408.   __ret = (uint32x2_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  18409.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18410.   return __ret;
  18411. }
  18412. #endif
  18413.  
  18414. #ifdef __LITTLE_ENDIAN__
  18415. __ai uint16x4_t vpadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  18416.   uint16x4_t __ret;
  18417.   __ret = (uint16x4_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  18418.   return __ret;
  18419. }
  18420. #else
  18421. __ai uint16x4_t vpadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  18422.   uint16x4_t __ret;
  18423.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18424.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18425.   __ret = (uint16x4_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  18426.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18427.   return __ret;
  18428. }
  18429. #endif
  18430.  
  18431. #ifdef __LITTLE_ENDIAN__
  18432. __ai int8x8_t vpadd_s8(int8x8_t __p0, int8x8_t __p1) {
  18433.   int8x8_t __ret;
  18434.   __ret = (int8x8_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  18435.   return __ret;
  18436. }
  18437. #else
  18438. __ai int8x8_t vpadd_s8(int8x8_t __p0, int8x8_t __p1) {
  18439.   int8x8_t __ret;
  18440.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18441.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18442.   __ret = (int8x8_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  18443.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18444.   return __ret;
  18445. }
  18446. #endif
  18447.  
  18448. #ifdef __LITTLE_ENDIAN__
  18449. __ai float32x2_t vpadd_f32(float32x2_t __p0, float32x2_t __p1) {
  18450.   float32x2_t __ret;
  18451.   __ret = (float32x2_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  18452.   return __ret;
  18453. }
  18454. #else
  18455. __ai float32x2_t vpadd_f32(float32x2_t __p0, float32x2_t __p1) {
  18456.   float32x2_t __ret;
  18457.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18458.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18459.   __ret = (float32x2_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  18460.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18461.   return __ret;
  18462. }
  18463. #endif
  18464.  
  18465. #ifdef __LITTLE_ENDIAN__
  18466. __ai int32x2_t vpadd_s32(int32x2_t __p0, int32x2_t __p1) {
  18467.   int32x2_t __ret;
  18468.   __ret = (int32x2_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  18469.   return __ret;
  18470. }
  18471. #else
  18472. __ai int32x2_t vpadd_s32(int32x2_t __p0, int32x2_t __p1) {
  18473.   int32x2_t __ret;
  18474.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18475.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18476.   __ret = (int32x2_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  18477.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18478.   return __ret;
  18479. }
  18480. #endif
  18481.  
  18482. #ifdef __LITTLE_ENDIAN__
  18483. __ai int16x4_t vpadd_s16(int16x4_t __p0, int16x4_t __p1) {
  18484.   int16x4_t __ret;
  18485.   __ret = (int16x4_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  18486.   return __ret;
  18487. }
  18488. #else
  18489. __ai int16x4_t vpadd_s16(int16x4_t __p0, int16x4_t __p1) {
  18490.   int16x4_t __ret;
  18491.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18492.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18493.   __ret = (int16x4_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  18494.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18495.   return __ret;
  18496. }
  18497. #endif
  18498.  
  18499. #ifdef __LITTLE_ENDIAN__
  18500. __ai uint16x8_t vpaddlq_u8(uint8x16_t __p0) {
  18501.   uint16x8_t __ret;
  18502.   __ret = (uint16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 49);
  18503.   return __ret;
  18504. }
  18505. #else
  18506. __ai uint16x8_t vpaddlq_u8(uint8x16_t __p0) {
  18507.   uint16x8_t __ret;
  18508.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  18509.   __ret = (uint16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 49);
  18510.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18511.   return __ret;
  18512. }
  18513. #endif
  18514.  
  18515. #ifdef __LITTLE_ENDIAN__
  18516. __ai uint64x2_t vpaddlq_u32(uint32x4_t __p0) {
  18517.   uint64x2_t __ret;
  18518.   __ret = (uint64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 51);
  18519.   return __ret;
  18520. }
  18521. #else
  18522. __ai uint64x2_t vpaddlq_u32(uint32x4_t __p0) {
  18523.   uint64x2_t __ret;
  18524.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18525.   __ret = (uint64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 51);
  18526.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18527.   return __ret;
  18528. }
  18529. #endif
  18530.  
  18531. #ifdef __LITTLE_ENDIAN__
  18532. __ai uint32x4_t vpaddlq_u16(uint16x8_t __p0) {
  18533.   uint32x4_t __ret;
  18534.   __ret = (uint32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 50);
  18535.   return __ret;
  18536. }
  18537. #else
  18538. __ai uint32x4_t vpaddlq_u16(uint16x8_t __p0) {
  18539.   uint32x4_t __ret;
  18540.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18541.   __ret = (uint32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 50);
  18542.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18543.   return __ret;
  18544. }
  18545. #endif
  18546.  
  18547. #ifdef __LITTLE_ENDIAN__
  18548. __ai int16x8_t vpaddlq_s8(int8x16_t __p0) {
  18549.   int16x8_t __ret;
  18550.   __ret = (int16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 33);
  18551.   return __ret;
  18552. }
  18553. #else
  18554. __ai int16x8_t vpaddlq_s8(int8x16_t __p0) {
  18555.   int16x8_t __ret;
  18556.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  18557.   __ret = (int16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 33);
  18558.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18559.   return __ret;
  18560. }
  18561. #endif
  18562.  
  18563. #ifdef __LITTLE_ENDIAN__
  18564. __ai int64x2_t vpaddlq_s32(int32x4_t __p0) {
  18565.   int64x2_t __ret;
  18566.   __ret = (int64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 35);
  18567.   return __ret;
  18568. }
  18569. #else
  18570. __ai int64x2_t vpaddlq_s32(int32x4_t __p0) {
  18571.   int64x2_t __ret;
  18572.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18573.   __ret = (int64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 35);
  18574.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18575.   return __ret;
  18576. }
  18577. #endif
  18578.  
  18579. #ifdef __LITTLE_ENDIAN__
  18580. __ai int32x4_t vpaddlq_s16(int16x8_t __p0) {
  18581.   int32x4_t __ret;
  18582.   __ret = (int32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 34);
  18583.   return __ret;
  18584. }
  18585. #else
  18586. __ai int32x4_t vpaddlq_s16(int16x8_t __p0) {
  18587.   int32x4_t __ret;
  18588.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18589.   __ret = (int32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 34);
  18590.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18591.   return __ret;
  18592. }
  18593. #endif
  18594.  
  18595. #ifdef __LITTLE_ENDIAN__
  18596. __ai uint16x4_t vpaddl_u8(uint8x8_t __p0) {
  18597.   uint16x4_t __ret;
  18598.   __ret = (uint16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 17);
  18599.   return __ret;
  18600. }
  18601. #else
  18602. __ai uint16x4_t vpaddl_u8(uint8x8_t __p0) {
  18603.   uint16x4_t __ret;
  18604.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18605.   __ret = (uint16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 17);
  18606.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18607.   return __ret;
  18608. }
  18609. #endif
  18610.  
  18611. #ifdef __LITTLE_ENDIAN__
  18612. __ai uint64x1_t vpaddl_u32(uint32x2_t __p0) {
  18613.   uint64x1_t __ret;
  18614.   __ret = (uint64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 19);
  18615.   return __ret;
  18616. }
  18617. #else
  18618. __ai uint64x1_t vpaddl_u32(uint32x2_t __p0) {
  18619.   uint64x1_t __ret;
  18620.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18621.   __ret = (uint64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 19);
  18622.   return __ret;
  18623. }
  18624. #endif
  18625.  
  18626. #ifdef __LITTLE_ENDIAN__
  18627. __ai uint32x2_t vpaddl_u16(uint16x4_t __p0) {
  18628.   uint32x2_t __ret;
  18629.   __ret = (uint32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 18);
  18630.   return __ret;
  18631. }
  18632. #else
  18633. __ai uint32x2_t vpaddl_u16(uint16x4_t __p0) {
  18634.   uint32x2_t __ret;
  18635.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18636.   __ret = (uint32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 18);
  18637.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18638.   return __ret;
  18639. }
  18640. #endif
  18641.  
  18642. #ifdef __LITTLE_ENDIAN__
  18643. __ai int16x4_t vpaddl_s8(int8x8_t __p0) {
  18644.   int16x4_t __ret;
  18645.   __ret = (int16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 1);
  18646.   return __ret;
  18647. }
  18648. #else
  18649. __ai int16x4_t vpaddl_s8(int8x8_t __p0) {
  18650.   int16x4_t __ret;
  18651.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18652.   __ret = (int16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 1);
  18653.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18654.   return __ret;
  18655. }
  18656. #endif
  18657.  
  18658. #ifdef __LITTLE_ENDIAN__
  18659. __ai int64x1_t vpaddl_s32(int32x2_t __p0) {
  18660.   int64x1_t __ret;
  18661.   __ret = (int64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 3);
  18662.   return __ret;
  18663. }
  18664. #else
  18665. __ai int64x1_t vpaddl_s32(int32x2_t __p0) {
  18666.   int64x1_t __ret;
  18667.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18668.   __ret = (int64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 3);
  18669.   return __ret;
  18670. }
  18671. #endif
  18672.  
  18673. #ifdef __LITTLE_ENDIAN__
  18674. __ai int32x2_t vpaddl_s16(int16x4_t __p0) {
  18675.   int32x2_t __ret;
  18676.   __ret = (int32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 2);
  18677.   return __ret;
  18678. }
  18679. #else
  18680. __ai int32x2_t vpaddl_s16(int16x4_t __p0) {
  18681.   int32x2_t __ret;
  18682.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18683.   __ret = (int32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 2);
  18684.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18685.   return __ret;
  18686. }
  18687. #endif
  18688.  
  18689. #ifdef __LITTLE_ENDIAN__
  18690. __ai uint8x8_t vpmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
  18691.   uint8x8_t __ret;
  18692.   __ret = (uint8x8_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  18693.   return __ret;
  18694. }
  18695. #else
  18696. __ai uint8x8_t vpmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
  18697.   uint8x8_t __ret;
  18698.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18699.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18700.   __ret = (uint8x8_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  18701.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18702.   return __ret;
  18703. }
  18704. #endif
  18705.  
  18706. #ifdef __LITTLE_ENDIAN__
  18707. __ai uint32x2_t vpmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
  18708.   uint32x2_t __ret;
  18709.   __ret = (uint32x2_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  18710.   return __ret;
  18711. }
  18712. #else
  18713. __ai uint32x2_t vpmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
  18714.   uint32x2_t __ret;
  18715.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18716.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18717.   __ret = (uint32x2_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  18718.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18719.   return __ret;
  18720. }
  18721. #endif
  18722.  
  18723. #ifdef __LITTLE_ENDIAN__
  18724. __ai uint16x4_t vpmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
  18725.   uint16x4_t __ret;
  18726.   __ret = (uint16x4_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  18727.   return __ret;
  18728. }
  18729. #else
  18730. __ai uint16x4_t vpmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
  18731.   uint16x4_t __ret;
  18732.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18733.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18734.   __ret = (uint16x4_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  18735.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18736.   return __ret;
  18737. }
  18738. #endif
  18739.  
  18740. #ifdef __LITTLE_ENDIAN__
  18741. __ai int8x8_t vpmax_s8(int8x8_t __p0, int8x8_t __p1) {
  18742.   int8x8_t __ret;
  18743.   __ret = (int8x8_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  18744.   return __ret;
  18745. }
  18746. #else
  18747. __ai int8x8_t vpmax_s8(int8x8_t __p0, int8x8_t __p1) {
  18748.   int8x8_t __ret;
  18749.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18750.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18751.   __ret = (int8x8_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  18752.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18753.   return __ret;
  18754. }
  18755. #endif
  18756.  
  18757. #ifdef __LITTLE_ENDIAN__
  18758. __ai float32x2_t vpmax_f32(float32x2_t __p0, float32x2_t __p1) {
  18759.   float32x2_t __ret;
  18760.   __ret = (float32x2_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  18761.   return __ret;
  18762. }
  18763. #else
  18764. __ai float32x2_t vpmax_f32(float32x2_t __p0, float32x2_t __p1) {
  18765.   float32x2_t __ret;
  18766.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18767.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18768.   __ret = (float32x2_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  18769.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18770.   return __ret;
  18771. }
  18772. #endif
  18773.  
  18774. #ifdef __LITTLE_ENDIAN__
  18775. __ai int32x2_t vpmax_s32(int32x2_t __p0, int32x2_t __p1) {
  18776.   int32x2_t __ret;
  18777.   __ret = (int32x2_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  18778.   return __ret;
  18779. }
  18780. #else
  18781. __ai int32x2_t vpmax_s32(int32x2_t __p0, int32x2_t __p1) {
  18782.   int32x2_t __ret;
  18783.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18784.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18785.   __ret = (int32x2_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  18786.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18787.   return __ret;
  18788. }
  18789. #endif
  18790.  
  18791. #ifdef __LITTLE_ENDIAN__
  18792. __ai int16x4_t vpmax_s16(int16x4_t __p0, int16x4_t __p1) {
  18793.   int16x4_t __ret;
  18794.   __ret = (int16x4_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  18795.   return __ret;
  18796. }
  18797. #else
  18798. __ai int16x4_t vpmax_s16(int16x4_t __p0, int16x4_t __p1) {
  18799.   int16x4_t __ret;
  18800.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18801.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18802.   __ret = (int16x4_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  18803.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18804.   return __ret;
  18805. }
  18806. #endif
  18807.  
  18808. #ifdef __LITTLE_ENDIAN__
  18809. __ai uint8x8_t vpmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
  18810.   uint8x8_t __ret;
  18811.   __ret = (uint8x8_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  18812.   return __ret;
  18813. }
  18814. #else
  18815. __ai uint8x8_t vpmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
  18816.   uint8x8_t __ret;
  18817.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18818.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18819.   __ret = (uint8x8_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  18820.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18821.   return __ret;
  18822. }
  18823. #endif
  18824.  
  18825. #ifdef __LITTLE_ENDIAN__
  18826. __ai uint32x2_t vpmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
  18827.   uint32x2_t __ret;
  18828.   __ret = (uint32x2_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  18829.   return __ret;
  18830. }
  18831. #else
  18832. __ai uint32x2_t vpmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
  18833.   uint32x2_t __ret;
  18834.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18835.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18836.   __ret = (uint32x2_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  18837.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18838.   return __ret;
  18839. }
  18840. #endif
  18841.  
  18842. #ifdef __LITTLE_ENDIAN__
  18843. __ai uint16x4_t vpmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
  18844.   uint16x4_t __ret;
  18845.   __ret = (uint16x4_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  18846.   return __ret;
  18847. }
  18848. #else
  18849. __ai uint16x4_t vpmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
  18850.   uint16x4_t __ret;
  18851.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18852.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18853.   __ret = (uint16x4_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  18854.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18855.   return __ret;
  18856. }
  18857. #endif
  18858.  
  18859. #ifdef __LITTLE_ENDIAN__
  18860. __ai int8x8_t vpmin_s8(int8x8_t __p0, int8x8_t __p1) {
  18861.   int8x8_t __ret;
  18862.   __ret = (int8x8_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  18863.   return __ret;
  18864. }
  18865. #else
  18866. __ai int8x8_t vpmin_s8(int8x8_t __p0, int8x8_t __p1) {
  18867.   int8x8_t __ret;
  18868.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18869.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  18870.   __ret = (int8x8_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  18871.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18872.   return __ret;
  18873. }
  18874. #endif
  18875.  
  18876. #ifdef __LITTLE_ENDIAN__
  18877. __ai float32x2_t vpmin_f32(float32x2_t __p0, float32x2_t __p1) {
  18878.   float32x2_t __ret;
  18879.   __ret = (float32x2_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  18880.   return __ret;
  18881. }
  18882. #else
  18883. __ai float32x2_t vpmin_f32(float32x2_t __p0, float32x2_t __p1) {
  18884.   float32x2_t __ret;
  18885.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18886.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18887.   __ret = (float32x2_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  18888.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18889.   return __ret;
  18890. }
  18891. #endif
  18892.  
  18893. #ifdef __LITTLE_ENDIAN__
  18894. __ai int32x2_t vpmin_s32(int32x2_t __p0, int32x2_t __p1) {
  18895.   int32x2_t __ret;
  18896.   __ret = (int32x2_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  18897.   return __ret;
  18898. }
  18899. #else
  18900. __ai int32x2_t vpmin_s32(int32x2_t __p0, int32x2_t __p1) {
  18901.   int32x2_t __ret;
  18902.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  18903.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  18904.   __ret = (int32x2_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  18905.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  18906.   return __ret;
  18907. }
  18908. #endif
  18909.  
  18910. #ifdef __LITTLE_ENDIAN__
  18911. __ai int16x4_t vpmin_s16(int16x4_t __p0, int16x4_t __p1) {
  18912.   int16x4_t __ret;
  18913.   __ret = (int16x4_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  18914.   return __ret;
  18915. }
  18916. #else
  18917. __ai int16x4_t vpmin_s16(int16x4_t __p0, int16x4_t __p1) {
  18918.   int16x4_t __ret;
  18919.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18920.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  18921.   __ret = (int16x4_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  18922.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18923.   return __ret;
  18924. }
  18925. #endif
  18926.  
  18927. #ifdef __LITTLE_ENDIAN__
  18928. __ai int8x16_t vqabsq_s8(int8x16_t __p0) {
  18929.   int8x16_t __ret;
  18930.   __ret = (int8x16_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 32);
  18931.   return __ret;
  18932. }
  18933. #else
  18934. __ai int8x16_t vqabsq_s8(int8x16_t __p0) {
  18935.   int8x16_t __ret;
  18936.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  18937.   __ret = (int8x16_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 32);
  18938.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  18939.   return __ret;
  18940. }
  18941. #endif
  18942.  
  18943. #ifdef __LITTLE_ENDIAN__
  18944. __ai int32x4_t vqabsq_s32(int32x4_t __p0) {
  18945.   int32x4_t __ret;
  18946.   __ret = (int32x4_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 34);
  18947.   return __ret;
  18948. }
  18949. #else
  18950. __ai int32x4_t vqabsq_s32(int32x4_t __p0) {
  18951.   int32x4_t __ret;
  18952.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  18953.   __ret = (int32x4_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 34);
  18954.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  18955.   return __ret;
  18956. }
  18957. #endif
  18958.  
  18959. #ifdef __LITTLE_ENDIAN__
  18960. __ai int16x8_t vqabsq_s16(int16x8_t __p0) {
  18961.   int16x8_t __ret;
  18962.   __ret = (int16x8_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 33);
  18963.   return __ret;
  18964. }
  18965. #else
  18966. __ai int16x8_t vqabsq_s16(int16x8_t __p0) {
  18967.   int16x8_t __ret;
  18968.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18969.   __ret = (int16x8_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 33);
  18970.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18971.   return __ret;
  18972. }
  18973. #endif
  18974.  
  18975. #ifdef __LITTLE_ENDIAN__
  18976. __ai int8x8_t vqabs_s8(int8x8_t __p0) {
  18977.   int8x8_t __ret;
  18978.   __ret = (int8x8_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 0);
  18979.   return __ret;
  18980. }
  18981. #else
  18982. __ai int8x8_t vqabs_s8(int8x8_t __p0) {
  18983.   int8x8_t __ret;
  18984.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  18985.   __ret = (int8x8_t) __builtin_neon_vqabs_v((int8x8_t)__rev0, 0);
  18986.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  18987.   return __ret;
  18988. }
  18989. #endif
  18990.  
  18991. #ifdef __LITTLE_ENDIAN__
  18992. __ai int32x2_t vqabs_s32(int32x2_t __p0) {
  18993.   int32x2_t __ret;
  18994.   __ret = (int32x2_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 2);
  18995.   return __ret;
  18996. }
  18997. #else
  18998. __ai int32x2_t vqabs_s32(int32x2_t __p0) {
  18999.   int32x2_t __ret;
  19000.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19001.   __ret = (int32x2_t) __builtin_neon_vqabs_v((int8x8_t)__rev0, 2);
  19002.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19003.   return __ret;
  19004. }
  19005. #endif
  19006.  
  19007. #ifdef __LITTLE_ENDIAN__
  19008. __ai int16x4_t vqabs_s16(int16x4_t __p0) {
  19009.   int16x4_t __ret;
  19010.   __ret = (int16x4_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 1);
  19011.   return __ret;
  19012. }
  19013. #else
  19014. __ai int16x4_t vqabs_s16(int16x4_t __p0) {
  19015.   int16x4_t __ret;
  19016.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19017.   __ret = (int16x4_t) __builtin_neon_vqabs_v((int8x8_t)__rev0, 1);
  19018.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19019.   return __ret;
  19020. }
  19021. #endif
  19022.  
  19023. #ifdef __LITTLE_ENDIAN__
  19024. __ai uint8x16_t vqaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  19025.   uint8x16_t __ret;
  19026.   __ret = (uint8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  19027.   return __ret;
  19028. }
  19029. #else
  19030. __ai uint8x16_t vqaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  19031.   uint8x16_t __ret;
  19032.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  19033.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  19034.   __ret = (uint8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  19035.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  19036.   return __ret;
  19037. }
  19038. #endif
  19039.  
  19040. #ifdef __LITTLE_ENDIAN__
  19041. __ai uint32x4_t vqaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  19042.   uint32x4_t __ret;
  19043.   __ret = (uint32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  19044.   return __ret;
  19045. }
  19046. #else
  19047. __ai uint32x4_t vqaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  19048.   uint32x4_t __ret;
  19049.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19050.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19051.   __ret = (uint32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  19052.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19053.   return __ret;
  19054. }
  19055. #endif
  19056.  
  19057. #ifdef __LITTLE_ENDIAN__
  19058. __ai uint64x2_t vqaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  19059.   uint64x2_t __ret;
  19060.   __ret = (uint64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  19061.   return __ret;
  19062. }
  19063. #else
  19064. __ai uint64x2_t vqaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  19065.   uint64x2_t __ret;
  19066.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19067.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19068.   __ret = (uint64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  19069.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19070.   return __ret;
  19071. }
  19072. #endif
  19073.  
  19074. #ifdef __LITTLE_ENDIAN__
  19075. __ai uint16x8_t vqaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  19076.   uint16x8_t __ret;
  19077.   __ret = (uint16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  19078.   return __ret;
  19079. }
  19080. #else
  19081. __ai uint16x8_t vqaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  19082.   uint16x8_t __ret;
  19083.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  19084.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  19085.   __ret = (uint16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  19086.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  19087.   return __ret;
  19088. }
  19089. #endif
  19090.  
  19091. #ifdef __LITTLE_ENDIAN__
  19092. __ai int8x16_t vqaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  19093.   int8x16_t __ret;
  19094.   __ret = (int8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  19095.   return __ret;
  19096. }
  19097. #else
  19098. __ai int8x16_t vqaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  19099.   int8x16_t __ret;
  19100.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  19101.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  19102.   __ret = (int8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  19103.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  19104.   return __ret;
  19105. }
  19106. #endif
  19107.  
  19108. #ifdef __LITTLE_ENDIAN__
  19109. __ai int32x4_t vqaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  19110.   int32x4_t __ret;
  19111.   __ret = (int32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  19112.   return __ret;
  19113. }
  19114. #else
  19115. __ai int32x4_t vqaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  19116.   int32x4_t __ret;
  19117.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19118.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19119.   __ret = (int32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  19120.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19121.   return __ret;
  19122. }
  19123. #endif
  19124.  
  19125. #ifdef __LITTLE_ENDIAN__
  19126. __ai int64x2_t vqaddq_s64(int64x2_t __p0, int64x2_t __p1) {
  19127.   int64x2_t __ret;
  19128.   __ret = (int64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
  19129.   return __ret;
  19130. }
  19131. #else
  19132. __ai int64x2_t vqaddq_s64(int64x2_t __p0, int64x2_t __p1) {
  19133.   int64x2_t __ret;
  19134.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19135.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19136.   __ret = (int64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
  19137.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19138.   return __ret;
  19139. }
  19140. #endif
  19141.  
  19142. #ifdef __LITTLE_ENDIAN__
  19143. __ai int16x8_t vqaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  19144.   int16x8_t __ret;
  19145.   __ret = (int16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  19146.   return __ret;
  19147. }
  19148. #else
  19149. __ai int16x8_t vqaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  19150.   int16x8_t __ret;
  19151.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  19152.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  19153.   __ret = (int16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  19154.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  19155.   return __ret;
  19156. }
  19157. #endif
  19158.  
  19159. #ifdef __LITTLE_ENDIAN__
  19160. __ai uint8x8_t vqadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  19161.   uint8x8_t __ret;
  19162.   __ret = (uint8x8_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  19163.   return __ret;
  19164. }
  19165. #else
  19166. __ai uint8x8_t vqadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  19167.   uint8x8_t __ret;
  19168.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  19169.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  19170.   __ret = (uint8x8_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  19171.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  19172.   return __ret;
  19173. }
  19174. #endif
  19175.  
  19176. #ifdef __LITTLE_ENDIAN__
  19177. __ai uint32x2_t vqadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  19178.   uint32x2_t __ret;
  19179.   __ret = (uint32x2_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  19180.   return __ret;
  19181. }
  19182. #else
  19183. __ai uint32x2_t vqadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  19184.   uint32x2_t __ret;
  19185.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19186.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19187.   __ret = (uint32x2_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  19188.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19189.   return __ret;
  19190. }
  19191. #endif
  19192.  
  19193. __ai uint64x1_t vqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
  19194.   uint64x1_t __ret;
  19195.   __ret = (uint64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  19196.   return __ret;
  19197. }
  19198. #ifdef __LITTLE_ENDIAN__
  19199. __ai uint16x4_t vqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  19200.   uint16x4_t __ret;
  19201.   __ret = (uint16x4_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  19202.   return __ret;
  19203. }
  19204. #else
  19205. __ai uint16x4_t vqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  19206.   uint16x4_t __ret;
  19207.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19208.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19209.   __ret = (uint16x4_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  19210.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19211.   return __ret;
  19212. }
  19213. #endif
  19214.  
  19215. #ifdef __LITTLE_ENDIAN__
  19216. __ai int8x8_t vqadd_s8(int8x8_t __p0, int8x8_t __p1) {
  19217.   int8x8_t __ret;
  19218.   __ret = (int8x8_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  19219.   return __ret;
  19220. }
  19221. #else
  19222. __ai int8x8_t vqadd_s8(int8x8_t __p0, int8x8_t __p1) {
  19223.   int8x8_t __ret;
  19224.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  19225.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  19226.   __ret = (int8x8_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  19227.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  19228.   return __ret;
  19229. }
  19230. #endif
  19231.  
  19232. #ifdef __LITTLE_ENDIAN__
  19233. __ai int32x2_t vqadd_s32(int32x2_t __p0, int32x2_t __p1) {
  19234.   int32x2_t __ret;
  19235.   __ret = (int32x2_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  19236.   return __ret;
  19237. }
  19238. #else
  19239. __ai int32x2_t vqadd_s32(int32x2_t __p0, int32x2_t __p1) {
  19240.   int32x2_t __ret;
  19241.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19242.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19243.   __ret = (int32x2_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  19244.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19245.   return __ret;
  19246. }
  19247. #endif
  19248.  
  19249. __ai int64x1_t vqadd_s64(int64x1_t __p0, int64x1_t __p1) {
  19250.   int64x1_t __ret;
  19251.   __ret = (int64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
  19252.   return __ret;
  19253. }
  19254. #ifdef __LITTLE_ENDIAN__
  19255. __ai int16x4_t vqadd_s16(int16x4_t __p0, int16x4_t __p1) {
  19256.   int16x4_t __ret;
  19257.   __ret = (int16x4_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  19258.   return __ret;
  19259. }
  19260. #else
  19261. __ai int16x4_t vqadd_s16(int16x4_t __p0, int16x4_t __p1) {
  19262.   int16x4_t __ret;
  19263.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19264.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19265.   __ret = (int16x4_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  19266.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19267.   return __ret;
  19268. }
  19269. #endif
  19270.  
  19271. #ifdef __LITTLE_ENDIAN__
  19272. __ai int64x2_t vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  19273.   int64x2_t __ret;
  19274.   __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
  19275.   return __ret;
  19276. }
  19277. #else
  19278. __ai int64x2_t vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  19279.   int64x2_t __ret;
  19280.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19281.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19282.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  19283.   __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 35);
  19284.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19285.   return __ret;
  19286. }
  19287. __ai int64x2_t __noswap_vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  19288.   int64x2_t __ret;
  19289.   __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
  19290.   return __ret;
  19291. }
  19292. #endif
  19293.  
  19294. #ifdef __LITTLE_ENDIAN__
  19295. __ai int32x4_t vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  19296.   int32x4_t __ret;
  19297.   __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
  19298.   return __ret;
  19299. }
  19300. #else
  19301. __ai int32x4_t vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  19302.   int32x4_t __ret;
  19303.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19304.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19305.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  19306.   __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 34);
  19307.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19308.   return __ret;
  19309. }
  19310. __ai int32x4_t __noswap_vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  19311.   int32x4_t __ret;
  19312.   __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
  19313.   return __ret;
  19314. }
  19315. #endif
  19316.  
  19317. #ifdef __LITTLE_ENDIAN__
  19318. #define vqdmlal_lane_s32(__p0_114, __p1_114, __p2_114, __p3_114) __extension__ ({ \
  19319.   int64x2_t __ret_114; \
  19320.   int64x2_t __s0_114 = __p0_114; \
  19321.   int32x2_t __s1_114 = __p1_114; \
  19322.   int32x2_t __s2_114 = __p2_114; \
  19323.   __ret_114 = vqdmlal_s32(__s0_114, __s1_114, splat_lane_s32(__s2_114, __p3_114)); \
  19324.   __ret_114; \
  19325. })
  19326. #else
  19327. #define vqdmlal_lane_s32(__p0_115, __p1_115, __p2_115, __p3_115) __extension__ ({ \
  19328.   int64x2_t __ret_115; \
  19329.   int64x2_t __s0_115 = __p0_115; \
  19330.   int32x2_t __s1_115 = __p1_115; \
  19331.   int32x2_t __s2_115 = __p2_115; \
  19332.   int64x2_t __rev0_115;  __rev0_115 = __builtin_shufflevector(__s0_115, __s0_115, 1, 0); \
  19333.   int32x2_t __rev1_115;  __rev1_115 = __builtin_shufflevector(__s1_115, __s1_115, 1, 0); \
  19334.   int32x2_t __rev2_115;  __rev2_115 = __builtin_shufflevector(__s2_115, __s2_115, 1, 0); \
  19335.   __ret_115 = __noswap_vqdmlal_s32(__rev0_115, __rev1_115, __noswap_splat_lane_s32(__rev2_115, __p3_115)); \
  19336.   __ret_115 = __builtin_shufflevector(__ret_115, __ret_115, 1, 0); \
  19337.   __ret_115; \
  19338. })
  19339. #endif
  19340.  
  19341. #ifdef __LITTLE_ENDIAN__
  19342. #define vqdmlal_lane_s16(__p0_116, __p1_116, __p2_116, __p3_116) __extension__ ({ \
  19343.   int32x4_t __ret_116; \
  19344.   int32x4_t __s0_116 = __p0_116; \
  19345.   int16x4_t __s1_116 = __p1_116; \
  19346.   int16x4_t __s2_116 = __p2_116; \
  19347.   __ret_116 = vqdmlal_s16(__s0_116, __s1_116, splat_lane_s16(__s2_116, __p3_116)); \
  19348.   __ret_116; \
  19349. })
  19350. #else
  19351. #define vqdmlal_lane_s16(__p0_117, __p1_117, __p2_117, __p3_117) __extension__ ({ \
  19352.   int32x4_t __ret_117; \
  19353.   int32x4_t __s0_117 = __p0_117; \
  19354.   int16x4_t __s1_117 = __p1_117; \
  19355.   int16x4_t __s2_117 = __p2_117; \
  19356.   int32x4_t __rev0_117;  __rev0_117 = __builtin_shufflevector(__s0_117, __s0_117, 3, 2, 1, 0); \
  19357.   int16x4_t __rev1_117;  __rev1_117 = __builtin_shufflevector(__s1_117, __s1_117, 3, 2, 1, 0); \
  19358.   int16x4_t __rev2_117;  __rev2_117 = __builtin_shufflevector(__s2_117, __s2_117, 3, 2, 1, 0); \
  19359.   __ret_117 = __noswap_vqdmlal_s16(__rev0_117, __rev1_117, __noswap_splat_lane_s16(__rev2_117, __p3_117)); \
  19360.   __ret_117 = __builtin_shufflevector(__ret_117, __ret_117, 3, 2, 1, 0); \
  19361.   __ret_117; \
  19362. })
  19363. #endif
  19364.  
  19365. #ifdef __LITTLE_ENDIAN__
  19366. __ai int64x2_t vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  19367.   int64x2_t __ret;
  19368.   __ret = vqdmlal_s32(__p0, __p1, (int32x2_t) {__p2, __p2});
  19369.   return __ret;
  19370. }
  19371. #else
  19372. __ai int64x2_t vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  19373.   int64x2_t __ret;
  19374.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19375.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19376.   __ret = __noswap_vqdmlal_s32(__rev0, __rev1, (int32x2_t) {__p2, __p2});
  19377.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19378.   return __ret;
  19379. }
  19380. __ai int64x2_t __noswap_vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  19381.   int64x2_t __ret;
  19382.   __ret = __noswap_vqdmlal_s32(__p0, __p1, (int32x2_t) {__p2, __p2});
  19383.   return __ret;
  19384. }
  19385. #endif
  19386.  
  19387. #ifdef __LITTLE_ENDIAN__
  19388. __ai int32x4_t vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  19389.   int32x4_t __ret;
  19390.   __ret = vqdmlal_s16(__p0, __p1, (int16x4_t) {__p2, __p2, __p2, __p2});
  19391.   return __ret;
  19392. }
  19393. #else
  19394. __ai int32x4_t vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  19395.   int32x4_t __ret;
  19396.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19397.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19398.   __ret = __noswap_vqdmlal_s16(__rev0, __rev1, (int16x4_t) {__p2, __p2, __p2, __p2});
  19399.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19400.   return __ret;
  19401. }
  19402. __ai int32x4_t __noswap_vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  19403.   int32x4_t __ret;
  19404.   __ret = __noswap_vqdmlal_s16(__p0, __p1, (int16x4_t) {__p2, __p2, __p2, __p2});
  19405.   return __ret;
  19406. }
  19407. #endif
  19408.  
  19409. #ifdef __LITTLE_ENDIAN__
  19410. __ai int64x2_t vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  19411.   int64x2_t __ret;
  19412.   __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
  19413.   return __ret;
  19414. }
  19415. #else
  19416. __ai int64x2_t vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  19417.   int64x2_t __ret;
  19418.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19419.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19420.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  19421.   __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 35);
  19422.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19423.   return __ret;
  19424. }
  19425. __ai int64x2_t __noswap_vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  19426.   int64x2_t __ret;
  19427.   __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
  19428.   return __ret;
  19429. }
  19430. #endif
  19431.  
  19432. #ifdef __LITTLE_ENDIAN__
  19433. __ai int32x4_t vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  19434.   int32x4_t __ret;
  19435.   __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
  19436.   return __ret;
  19437. }
  19438. #else
  19439. __ai int32x4_t vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  19440.   int32x4_t __ret;
  19441.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19442.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19443.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  19444.   __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 34);
  19445.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19446.   return __ret;
  19447. }
  19448. __ai int32x4_t __noswap_vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  19449.   int32x4_t __ret;
  19450.   __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
  19451.   return __ret;
  19452. }
  19453. #endif
  19454.  
  19455. #ifdef __LITTLE_ENDIAN__
  19456. #define vqdmlsl_lane_s32(__p0_118, __p1_118, __p2_118, __p3_118) __extension__ ({ \
  19457.   int64x2_t __ret_118; \
  19458.   int64x2_t __s0_118 = __p0_118; \
  19459.   int32x2_t __s1_118 = __p1_118; \
  19460.   int32x2_t __s2_118 = __p2_118; \
  19461.   __ret_118 = vqdmlsl_s32(__s0_118, __s1_118, splat_lane_s32(__s2_118, __p3_118)); \
  19462.   __ret_118; \
  19463. })
  19464. #else
  19465. #define vqdmlsl_lane_s32(__p0_119, __p1_119, __p2_119, __p3_119) __extension__ ({ \
  19466.   int64x2_t __ret_119; \
  19467.   int64x2_t __s0_119 = __p0_119; \
  19468.   int32x2_t __s1_119 = __p1_119; \
  19469.   int32x2_t __s2_119 = __p2_119; \
  19470.   int64x2_t __rev0_119;  __rev0_119 = __builtin_shufflevector(__s0_119, __s0_119, 1, 0); \
  19471.   int32x2_t __rev1_119;  __rev1_119 = __builtin_shufflevector(__s1_119, __s1_119, 1, 0); \
  19472.   int32x2_t __rev2_119;  __rev2_119 = __builtin_shufflevector(__s2_119, __s2_119, 1, 0); \
  19473.   __ret_119 = __noswap_vqdmlsl_s32(__rev0_119, __rev1_119, __noswap_splat_lane_s32(__rev2_119, __p3_119)); \
  19474.   __ret_119 = __builtin_shufflevector(__ret_119, __ret_119, 1, 0); \
  19475.   __ret_119; \
  19476. })
  19477. #endif
  19478.  
  19479. #ifdef __LITTLE_ENDIAN__
  19480. #define vqdmlsl_lane_s16(__p0_120, __p1_120, __p2_120, __p3_120) __extension__ ({ \
  19481.   int32x4_t __ret_120; \
  19482.   int32x4_t __s0_120 = __p0_120; \
  19483.   int16x4_t __s1_120 = __p1_120; \
  19484.   int16x4_t __s2_120 = __p2_120; \
  19485.   __ret_120 = vqdmlsl_s16(__s0_120, __s1_120, splat_lane_s16(__s2_120, __p3_120)); \
  19486.   __ret_120; \
  19487. })
  19488. #else
  19489. #define vqdmlsl_lane_s16(__p0_121, __p1_121, __p2_121, __p3_121) __extension__ ({ \
  19490.   int32x4_t __ret_121; \
  19491.   int32x4_t __s0_121 = __p0_121; \
  19492.   int16x4_t __s1_121 = __p1_121; \
  19493.   int16x4_t __s2_121 = __p2_121; \
  19494.   int32x4_t __rev0_121;  __rev0_121 = __builtin_shufflevector(__s0_121, __s0_121, 3, 2, 1, 0); \
  19495.   int16x4_t __rev1_121;  __rev1_121 = __builtin_shufflevector(__s1_121, __s1_121, 3, 2, 1, 0); \
  19496.   int16x4_t __rev2_121;  __rev2_121 = __builtin_shufflevector(__s2_121, __s2_121, 3, 2, 1, 0); \
  19497.   __ret_121 = __noswap_vqdmlsl_s16(__rev0_121, __rev1_121, __noswap_splat_lane_s16(__rev2_121, __p3_121)); \
  19498.   __ret_121 = __builtin_shufflevector(__ret_121, __ret_121, 3, 2, 1, 0); \
  19499.   __ret_121; \
  19500. })
  19501. #endif
  19502.  
  19503. #ifdef __LITTLE_ENDIAN__
  19504. __ai int64x2_t vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  19505.   int64x2_t __ret;
  19506.   __ret = vqdmlsl_s32(__p0, __p1, (int32x2_t) {__p2, __p2});
  19507.   return __ret;
  19508. }
  19509. #else
  19510. __ai int64x2_t vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  19511.   int64x2_t __ret;
  19512.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19513.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19514.   __ret = __noswap_vqdmlsl_s32(__rev0, __rev1, (int32x2_t) {__p2, __p2});
  19515.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19516.   return __ret;
  19517. }
  19518. __ai int64x2_t __noswap_vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  19519.   int64x2_t __ret;
  19520.   __ret = __noswap_vqdmlsl_s32(__p0, __p1, (int32x2_t) {__p2, __p2});
  19521.   return __ret;
  19522. }
  19523. #endif
  19524.  
  19525. #ifdef __LITTLE_ENDIAN__
  19526. __ai int32x4_t vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  19527.   int32x4_t __ret;
  19528.   __ret = vqdmlsl_s16(__p0, __p1, (int16x4_t) {__p2, __p2, __p2, __p2});
  19529.   return __ret;
  19530. }
  19531. #else
  19532. __ai int32x4_t vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  19533.   int32x4_t __ret;
  19534.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19535.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19536.   __ret = __noswap_vqdmlsl_s16(__rev0, __rev1, (int16x4_t) {__p2, __p2, __p2, __p2});
  19537.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19538.   return __ret;
  19539. }
  19540. __ai int32x4_t __noswap_vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  19541.   int32x4_t __ret;
  19542.   __ret = __noswap_vqdmlsl_s16(__p0, __p1, (int16x4_t) {__p2, __p2, __p2, __p2});
  19543.   return __ret;
  19544. }
  19545. #endif
  19546.  
  19547. #ifdef __LITTLE_ENDIAN__
  19548. __ai int32x4_t vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
  19549.   int32x4_t __ret;
  19550.   __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  19551.   return __ret;
  19552. }
  19553. #else
  19554. __ai int32x4_t vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
  19555.   int32x4_t __ret;
  19556.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19557.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19558.   __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  19559.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19560.   return __ret;
  19561. }
  19562. __ai int32x4_t __noswap_vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
  19563.   int32x4_t __ret;
  19564.   __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  19565.   return __ret;
  19566. }
  19567. #endif
  19568.  
  19569. #ifdef __LITTLE_ENDIAN__
  19570. __ai int16x8_t vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
  19571.   int16x8_t __ret;
  19572.   __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  19573.   return __ret;
  19574. }
  19575. #else
  19576. __ai int16x8_t vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
  19577.   int16x8_t __ret;
  19578.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  19579.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  19580.   __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  19581.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  19582.   return __ret;
  19583. }
  19584. __ai int16x8_t __noswap_vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
  19585.   int16x8_t __ret;
  19586.   __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  19587.   return __ret;
  19588. }
  19589. #endif
  19590.  
  19591. #ifdef __LITTLE_ENDIAN__
  19592. __ai int32x2_t vqdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
  19593.   int32x2_t __ret;
  19594.   __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  19595.   return __ret;
  19596. }
  19597. #else
  19598. __ai int32x2_t vqdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
  19599.   int32x2_t __ret;
  19600.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19601.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19602.   __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  19603.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19604.   return __ret;
  19605. }
  19606. __ai int32x2_t __noswap_vqdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
  19607.   int32x2_t __ret;
  19608.   __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  19609.   return __ret;
  19610. }
  19611. #endif
  19612.  
  19613. #ifdef __LITTLE_ENDIAN__
  19614. __ai int16x4_t vqdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
  19615.   int16x4_t __ret;
  19616.   __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  19617.   return __ret;
  19618. }
  19619. #else
  19620. __ai int16x4_t vqdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
  19621.   int16x4_t __ret;
  19622.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19623.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19624.   __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  19625.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19626.   return __ret;
  19627. }
  19628. __ai int16x4_t __noswap_vqdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
  19629.   int16x4_t __ret;
  19630.   __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  19631.   return __ret;
  19632. }
  19633. #endif
  19634.  
  19635. #ifdef __LITTLE_ENDIAN__
  19636. __ai int32x4_t vqdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
  19637.   int32x4_t __ret;
  19638.   __ret = vqdmulhq_s32(__p0, (int32x4_t) {__p1, __p1, __p1, __p1});
  19639.   return __ret;
  19640. }
  19641. #else
  19642. __ai int32x4_t vqdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
  19643.   int32x4_t __ret;
  19644.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19645.   __ret = __noswap_vqdmulhq_s32(__rev0, (int32x4_t) {__p1, __p1, __p1, __p1});
  19646.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19647.   return __ret;
  19648. }
  19649. #endif
  19650.  
  19651. #ifdef __LITTLE_ENDIAN__
  19652. __ai int16x8_t vqdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
  19653.   int16x8_t __ret;
  19654.   __ret = vqdmulhq_s16(__p0, (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1});
  19655.   return __ret;
  19656. }
  19657. #else
  19658. __ai int16x8_t vqdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
  19659.   int16x8_t __ret;
  19660.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  19661.   __ret = __noswap_vqdmulhq_s16(__rev0, (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1});
  19662.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  19663.   return __ret;
  19664. }
  19665. #endif
  19666.  
  19667. #ifdef __LITTLE_ENDIAN__
  19668. __ai int32x2_t vqdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
  19669.   int32x2_t __ret;
  19670.   __ret = vqdmulh_s32(__p0, (int32x2_t) {__p1, __p1});
  19671.   return __ret;
  19672. }
  19673. #else
  19674. __ai int32x2_t vqdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
  19675.   int32x2_t __ret;
  19676.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19677.   __ret = __noswap_vqdmulh_s32(__rev0, (int32x2_t) {__p1, __p1});
  19678.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19679.   return __ret;
  19680. }
  19681. #endif
  19682.  
  19683. #ifdef __LITTLE_ENDIAN__
  19684. __ai int16x4_t vqdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
  19685.   int16x4_t __ret;
  19686.   __ret = vqdmulh_s16(__p0, (int16x4_t) {__p1, __p1, __p1, __p1});
  19687.   return __ret;
  19688. }
  19689. #else
  19690. __ai int16x4_t vqdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
  19691.   int16x4_t __ret;
  19692.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19693.   __ret = __noswap_vqdmulh_s16(__rev0, (int16x4_t) {__p1, __p1, __p1, __p1});
  19694.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19695.   return __ret;
  19696. }
  19697. #endif
  19698.  
  19699. #ifdef __LITTLE_ENDIAN__
  19700. __ai int64x2_t vqdmull_s32(int32x2_t __p0, int32x2_t __p1) {
  19701.   int64x2_t __ret;
  19702.   __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
  19703.   return __ret;
  19704. }
  19705. #else
  19706. __ai int64x2_t vqdmull_s32(int32x2_t __p0, int32x2_t __p1) {
  19707.   int64x2_t __ret;
  19708.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19709.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  19710.   __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 35);
  19711.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19712.   return __ret;
  19713. }
  19714. __ai int64x2_t __noswap_vqdmull_s32(int32x2_t __p0, int32x2_t __p1) {
  19715.   int64x2_t __ret;
  19716.   __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
  19717.   return __ret;
  19718. }
  19719. #endif
  19720.  
  19721. #ifdef __LITTLE_ENDIAN__
  19722. __ai int32x4_t vqdmull_s16(int16x4_t __p0, int16x4_t __p1) {
  19723.   int32x4_t __ret;
  19724.   __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
  19725.   return __ret;
  19726. }
  19727. #else
  19728. __ai int32x4_t vqdmull_s16(int16x4_t __p0, int16x4_t __p1) {
  19729.   int32x4_t __ret;
  19730.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19731.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  19732.   __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 34);
  19733.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19734.   return __ret;
  19735. }
  19736. __ai int32x4_t __noswap_vqdmull_s16(int16x4_t __p0, int16x4_t __p1) {
  19737.   int32x4_t __ret;
  19738.   __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
  19739.   return __ret;
  19740. }
  19741. #endif
  19742.  
  19743. #ifdef __LITTLE_ENDIAN__
  19744. #define vqdmull_lane_s32(__p0_122, __p1_122, __p2_122) __extension__ ({ \
  19745.   int64x2_t __ret_122; \
  19746.   int32x2_t __s0_122 = __p0_122; \
  19747.   int32x2_t __s1_122 = __p1_122; \
  19748.   __ret_122 = vqdmull_s32(__s0_122, splat_lane_s32(__s1_122, __p2_122)); \
  19749.   __ret_122; \
  19750. })
  19751. #else
  19752. #define vqdmull_lane_s32(__p0_123, __p1_123, __p2_123) __extension__ ({ \
  19753.   int64x2_t __ret_123; \
  19754.   int32x2_t __s0_123 = __p0_123; \
  19755.   int32x2_t __s1_123 = __p1_123; \
  19756.   int32x2_t __rev0_123;  __rev0_123 = __builtin_shufflevector(__s0_123, __s0_123, 1, 0); \
  19757.   int32x2_t __rev1_123;  __rev1_123 = __builtin_shufflevector(__s1_123, __s1_123, 1, 0); \
  19758.   __ret_123 = __noswap_vqdmull_s32(__rev0_123, __noswap_splat_lane_s32(__rev1_123, __p2_123)); \
  19759.   __ret_123 = __builtin_shufflevector(__ret_123, __ret_123, 1, 0); \
  19760.   __ret_123; \
  19761. })
  19762. #endif
  19763.  
  19764. #ifdef __LITTLE_ENDIAN__
  19765. #define vqdmull_lane_s16(__p0_124, __p1_124, __p2_124) __extension__ ({ \
  19766.   int32x4_t __ret_124; \
  19767.   int16x4_t __s0_124 = __p0_124; \
  19768.   int16x4_t __s1_124 = __p1_124; \
  19769.   __ret_124 = vqdmull_s16(__s0_124, splat_lane_s16(__s1_124, __p2_124)); \
  19770.   __ret_124; \
  19771. })
  19772. #else
  19773. #define vqdmull_lane_s16(__p0_125, __p1_125, __p2_125) __extension__ ({ \
  19774.   int32x4_t __ret_125; \
  19775.   int16x4_t __s0_125 = __p0_125; \
  19776.   int16x4_t __s1_125 = __p1_125; \
  19777.   int16x4_t __rev0_125;  __rev0_125 = __builtin_shufflevector(__s0_125, __s0_125, 3, 2, 1, 0); \
  19778.   int16x4_t __rev1_125;  __rev1_125 = __builtin_shufflevector(__s1_125, __s1_125, 3, 2, 1, 0); \
  19779.   __ret_125 = __noswap_vqdmull_s16(__rev0_125, __noswap_splat_lane_s16(__rev1_125, __p2_125)); \
  19780.   __ret_125 = __builtin_shufflevector(__ret_125, __ret_125, 3, 2, 1, 0); \
  19781.   __ret_125; \
  19782. })
  19783. #endif
  19784.  
  19785. #ifdef __LITTLE_ENDIAN__
  19786. __ai int64x2_t vqdmull_n_s32(int32x2_t __p0, int32_t __p1) {
  19787.   int64x2_t __ret;
  19788.   __ret = vqdmull_s32(__p0, (int32x2_t) {__p1, __p1});
  19789.   return __ret;
  19790. }
  19791. #else
  19792. __ai int64x2_t vqdmull_n_s32(int32x2_t __p0, int32_t __p1) {
  19793.   int64x2_t __ret;
  19794.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19795.   __ret = __noswap_vqdmull_s32(__rev0, (int32x2_t) {__p1, __p1});
  19796.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19797.   return __ret;
  19798. }
  19799. __ai int64x2_t __noswap_vqdmull_n_s32(int32x2_t __p0, int32_t __p1) {
  19800.   int64x2_t __ret;
  19801.   __ret = __noswap_vqdmull_s32(__p0, (int32x2_t) {__p1, __p1});
  19802.   return __ret;
  19803. }
  19804. #endif
  19805.  
  19806. #ifdef __LITTLE_ENDIAN__
  19807. __ai int32x4_t vqdmull_n_s16(int16x4_t __p0, int16_t __p1) {
  19808.   int32x4_t __ret;
  19809.   __ret = vqdmull_s16(__p0, (int16x4_t) {__p1, __p1, __p1, __p1});
  19810.   return __ret;
  19811. }
  19812. #else
  19813. __ai int32x4_t vqdmull_n_s16(int16x4_t __p0, int16_t __p1) {
  19814.   int32x4_t __ret;
  19815.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19816.   __ret = __noswap_vqdmull_s16(__rev0, (int16x4_t) {__p1, __p1, __p1, __p1});
  19817.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19818.   return __ret;
  19819. }
  19820. __ai int32x4_t __noswap_vqdmull_n_s16(int16x4_t __p0, int16_t __p1) {
  19821.   int32x4_t __ret;
  19822.   __ret = __noswap_vqdmull_s16(__p0, (int16x4_t) {__p1, __p1, __p1, __p1});
  19823.   return __ret;
  19824. }
  19825. #endif
  19826.  
  19827. #ifdef __LITTLE_ENDIAN__
  19828. __ai uint16x4_t vqmovn_u32(uint32x4_t __p0) {
  19829.   uint16x4_t __ret;
  19830.   __ret = (uint16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 17);
  19831.   return __ret;
  19832. }
  19833. #else
  19834. __ai uint16x4_t vqmovn_u32(uint32x4_t __p0) {
  19835.   uint16x4_t __ret;
  19836.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19837.   __ret = (uint16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 17);
  19838.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19839.   return __ret;
  19840. }
  19841. __ai uint16x4_t __noswap_vqmovn_u32(uint32x4_t __p0) {
  19842.   uint16x4_t __ret;
  19843.   __ret = (uint16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 17);
  19844.   return __ret;
  19845. }
  19846. #endif
  19847.  
  19848. #ifdef __LITTLE_ENDIAN__
  19849. __ai uint32x2_t vqmovn_u64(uint64x2_t __p0) {
  19850.   uint32x2_t __ret;
  19851.   __ret = (uint32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 18);
  19852.   return __ret;
  19853. }
  19854. #else
  19855. __ai uint32x2_t vqmovn_u64(uint64x2_t __p0) {
  19856.   uint32x2_t __ret;
  19857.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19858.   __ret = (uint32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 18);
  19859.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19860.   return __ret;
  19861. }
  19862. __ai uint32x2_t __noswap_vqmovn_u64(uint64x2_t __p0) {
  19863.   uint32x2_t __ret;
  19864.   __ret = (uint32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 18);
  19865.   return __ret;
  19866. }
  19867. #endif
  19868.  
  19869. #ifdef __LITTLE_ENDIAN__
  19870. __ai uint8x8_t vqmovn_u16(uint16x8_t __p0) {
  19871.   uint8x8_t __ret;
  19872.   __ret = (uint8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 16);
  19873.   return __ret;
  19874. }
  19875. #else
  19876. __ai uint8x8_t vqmovn_u16(uint16x8_t __p0) {
  19877.   uint8x8_t __ret;
  19878.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  19879.   __ret = (uint8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 16);
  19880.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  19881.   return __ret;
  19882. }
  19883. __ai uint8x8_t __noswap_vqmovn_u16(uint16x8_t __p0) {
  19884.   uint8x8_t __ret;
  19885.   __ret = (uint8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 16);
  19886.   return __ret;
  19887. }
  19888. #endif
  19889.  
  19890. #ifdef __LITTLE_ENDIAN__
  19891. __ai int16x4_t vqmovn_s32(int32x4_t __p0) {
  19892.   int16x4_t __ret;
  19893.   __ret = (int16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 1);
  19894.   return __ret;
  19895. }
  19896. #else
  19897. __ai int16x4_t vqmovn_s32(int32x4_t __p0) {
  19898.   int16x4_t __ret;
  19899.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19900.   __ret = (int16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 1);
  19901.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19902.   return __ret;
  19903. }
  19904. __ai int16x4_t __noswap_vqmovn_s32(int32x4_t __p0) {
  19905.   int16x4_t __ret;
  19906.   __ret = (int16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 1);
  19907.   return __ret;
  19908. }
  19909. #endif
  19910.  
  19911. #ifdef __LITTLE_ENDIAN__
  19912. __ai int32x2_t vqmovn_s64(int64x2_t __p0) {
  19913.   int32x2_t __ret;
  19914.   __ret = (int32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 2);
  19915.   return __ret;
  19916. }
  19917. #else
  19918. __ai int32x2_t vqmovn_s64(int64x2_t __p0) {
  19919.   int32x2_t __ret;
  19920.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19921.   __ret = (int32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 2);
  19922.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19923.   return __ret;
  19924. }
  19925. __ai int32x2_t __noswap_vqmovn_s64(int64x2_t __p0) {
  19926.   int32x2_t __ret;
  19927.   __ret = (int32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 2);
  19928.   return __ret;
  19929. }
  19930. #endif
  19931.  
  19932. #ifdef __LITTLE_ENDIAN__
  19933. __ai int8x8_t vqmovn_s16(int16x8_t __p0) {
  19934.   int8x8_t __ret;
  19935.   __ret = (int8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 0);
  19936.   return __ret;
  19937. }
  19938. #else
  19939. __ai int8x8_t vqmovn_s16(int16x8_t __p0) {
  19940.   int8x8_t __ret;
  19941.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  19942.   __ret = (int8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 0);
  19943.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  19944.   return __ret;
  19945. }
  19946. __ai int8x8_t __noswap_vqmovn_s16(int16x8_t __p0) {
  19947.   int8x8_t __ret;
  19948.   __ret = (int8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 0);
  19949.   return __ret;
  19950. }
  19951. #endif
  19952.  
  19953. #ifdef __LITTLE_ENDIAN__
  19954. __ai uint16x4_t vqmovun_s32(int32x4_t __p0) {
  19955.   uint16x4_t __ret;
  19956.   __ret = (uint16x4_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 17);
  19957.   return __ret;
  19958. }
  19959. #else
  19960. __ai uint16x4_t vqmovun_s32(int32x4_t __p0) {
  19961.   uint16x4_t __ret;
  19962.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  19963.   __ret = (uint16x4_t) __builtin_neon_vqmovun_v((int8x16_t)__rev0, 17);
  19964.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  19965.   return __ret;
  19966. }
  19967. __ai uint16x4_t __noswap_vqmovun_s32(int32x4_t __p0) {
  19968.   uint16x4_t __ret;
  19969.   __ret = (uint16x4_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 17);
  19970.   return __ret;
  19971. }
  19972. #endif
  19973.  
  19974. #ifdef __LITTLE_ENDIAN__
  19975. __ai uint32x2_t vqmovun_s64(int64x2_t __p0) {
  19976.   uint32x2_t __ret;
  19977.   __ret = (uint32x2_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 18);
  19978.   return __ret;
  19979. }
  19980. #else
  19981. __ai uint32x2_t vqmovun_s64(int64x2_t __p0) {
  19982.   uint32x2_t __ret;
  19983.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  19984.   __ret = (uint32x2_t) __builtin_neon_vqmovun_v((int8x16_t)__rev0, 18);
  19985.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  19986.   return __ret;
  19987. }
  19988. __ai uint32x2_t __noswap_vqmovun_s64(int64x2_t __p0) {
  19989.   uint32x2_t __ret;
  19990.   __ret = (uint32x2_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 18);
  19991.   return __ret;
  19992. }
  19993. #endif
  19994.  
  19995. #ifdef __LITTLE_ENDIAN__
  19996. __ai uint8x8_t vqmovun_s16(int16x8_t __p0) {
  19997.   uint8x8_t __ret;
  19998.   __ret = (uint8x8_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 16);
  19999.   return __ret;
  20000. }
  20001. #else
  20002. __ai uint8x8_t vqmovun_s16(int16x8_t __p0) {
  20003.   uint8x8_t __ret;
  20004.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20005.   __ret = (uint8x8_t) __builtin_neon_vqmovun_v((int8x16_t)__rev0, 16);
  20006.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20007.   return __ret;
  20008. }
  20009. __ai uint8x8_t __noswap_vqmovun_s16(int16x8_t __p0) {
  20010.   uint8x8_t __ret;
  20011.   __ret = (uint8x8_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 16);
  20012.   return __ret;
  20013. }
  20014. #endif
  20015.  
  20016. #ifdef __LITTLE_ENDIAN__
  20017. __ai int8x16_t vqnegq_s8(int8x16_t __p0) {
  20018.   int8x16_t __ret;
  20019.   __ret = (int8x16_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 32);
  20020.   return __ret;
  20021. }
  20022. #else
  20023. __ai int8x16_t vqnegq_s8(int8x16_t __p0) {
  20024.   int8x16_t __ret;
  20025.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20026.   __ret = (int8x16_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 32);
  20027.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20028.   return __ret;
  20029. }
  20030. #endif
  20031.  
  20032. #ifdef __LITTLE_ENDIAN__
  20033. __ai int32x4_t vqnegq_s32(int32x4_t __p0) {
  20034.   int32x4_t __ret;
  20035.   __ret = (int32x4_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 34);
  20036.   return __ret;
  20037. }
  20038. #else
  20039. __ai int32x4_t vqnegq_s32(int32x4_t __p0) {
  20040.   int32x4_t __ret;
  20041.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20042.   __ret = (int32x4_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 34);
  20043.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20044.   return __ret;
  20045. }
  20046. #endif
  20047.  
  20048. #ifdef __LITTLE_ENDIAN__
  20049. __ai int16x8_t vqnegq_s16(int16x8_t __p0) {
  20050.   int16x8_t __ret;
  20051.   __ret = (int16x8_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 33);
  20052.   return __ret;
  20053. }
  20054. #else
  20055. __ai int16x8_t vqnegq_s16(int16x8_t __p0) {
  20056.   int16x8_t __ret;
  20057.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20058.   __ret = (int16x8_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 33);
  20059.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20060.   return __ret;
  20061. }
  20062. #endif
  20063.  
  20064. #ifdef __LITTLE_ENDIAN__
  20065. __ai int8x8_t vqneg_s8(int8x8_t __p0) {
  20066.   int8x8_t __ret;
  20067.   __ret = (int8x8_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 0);
  20068.   return __ret;
  20069. }
  20070. #else
  20071. __ai int8x8_t vqneg_s8(int8x8_t __p0) {
  20072.   int8x8_t __ret;
  20073.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20074.   __ret = (int8x8_t) __builtin_neon_vqneg_v((int8x8_t)__rev0, 0);
  20075.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20076.   return __ret;
  20077. }
  20078. #endif
  20079.  
  20080. #ifdef __LITTLE_ENDIAN__
  20081. __ai int32x2_t vqneg_s32(int32x2_t __p0) {
  20082.   int32x2_t __ret;
  20083.   __ret = (int32x2_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 2);
  20084.   return __ret;
  20085. }
  20086. #else
  20087. __ai int32x2_t vqneg_s32(int32x2_t __p0) {
  20088.   int32x2_t __ret;
  20089.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20090.   __ret = (int32x2_t) __builtin_neon_vqneg_v((int8x8_t)__rev0, 2);
  20091.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20092.   return __ret;
  20093. }
  20094. #endif
  20095.  
  20096. #ifdef __LITTLE_ENDIAN__
  20097. __ai int16x4_t vqneg_s16(int16x4_t __p0) {
  20098.   int16x4_t __ret;
  20099.   __ret = (int16x4_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 1);
  20100.   return __ret;
  20101. }
  20102. #else
  20103. __ai int16x4_t vqneg_s16(int16x4_t __p0) {
  20104.   int16x4_t __ret;
  20105.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20106.   __ret = (int16x4_t) __builtin_neon_vqneg_v((int8x8_t)__rev0, 1);
  20107.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20108.   return __ret;
  20109. }
  20110. #endif
  20111.  
  20112. #ifdef __LITTLE_ENDIAN__
  20113. __ai int32x4_t vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
  20114.   int32x4_t __ret;
  20115.   __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  20116.   return __ret;
  20117. }
  20118. #else
  20119. __ai int32x4_t vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
  20120.   int32x4_t __ret;
  20121.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20122.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20123.   __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  20124.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20125.   return __ret;
  20126. }
  20127. __ai int32x4_t __noswap_vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
  20128.   int32x4_t __ret;
  20129.   __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  20130.   return __ret;
  20131. }
  20132. #endif
  20133.  
  20134. #ifdef __LITTLE_ENDIAN__
  20135. __ai int16x8_t vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
  20136.   int16x8_t __ret;
  20137.   __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  20138.   return __ret;
  20139. }
  20140. #else
  20141. __ai int16x8_t vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
  20142.   int16x8_t __ret;
  20143.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20144.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  20145.   __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  20146.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20147.   return __ret;
  20148. }
  20149. __ai int16x8_t __noswap_vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
  20150.   int16x8_t __ret;
  20151.   __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  20152.   return __ret;
  20153. }
  20154. #endif
  20155.  
  20156. #ifdef __LITTLE_ENDIAN__
  20157. __ai int32x2_t vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
  20158.   int32x2_t __ret;
  20159.   __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  20160.   return __ret;
  20161. }
  20162. #else
  20163. __ai int32x2_t vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
  20164.   int32x2_t __ret;
  20165.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20166.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  20167.   __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  20168.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20169.   return __ret;
  20170. }
  20171. __ai int32x2_t __noswap_vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
  20172.   int32x2_t __ret;
  20173.   __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  20174.   return __ret;
  20175. }
  20176. #endif
  20177.  
  20178. #ifdef __LITTLE_ENDIAN__
  20179. __ai int16x4_t vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
  20180.   int16x4_t __ret;
  20181.   __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  20182.   return __ret;
  20183. }
  20184. #else
  20185. __ai int16x4_t vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
  20186.   int16x4_t __ret;
  20187.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20188.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20189.   __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  20190.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20191.   return __ret;
  20192. }
  20193. __ai int16x4_t __noswap_vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
  20194.   int16x4_t __ret;
  20195.   __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  20196.   return __ret;
  20197. }
  20198. #endif
  20199.  
  20200. #ifdef __LITTLE_ENDIAN__
  20201. __ai int32x4_t vqrdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
  20202.   int32x4_t __ret;
  20203.   __ret = vqrdmulhq_s32(__p0, (int32x4_t) {__p1, __p1, __p1, __p1});
  20204.   return __ret;
  20205. }
  20206. #else
  20207. __ai int32x4_t vqrdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
  20208.   int32x4_t __ret;
  20209.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20210.   __ret = __noswap_vqrdmulhq_s32(__rev0, (int32x4_t) {__p1, __p1, __p1, __p1});
  20211.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20212.   return __ret;
  20213. }
  20214. #endif
  20215.  
  20216. #ifdef __LITTLE_ENDIAN__
  20217. __ai int16x8_t vqrdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
  20218.   int16x8_t __ret;
  20219.   __ret = vqrdmulhq_s16(__p0, (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1});
  20220.   return __ret;
  20221. }
  20222. #else
  20223. __ai int16x8_t vqrdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
  20224.   int16x8_t __ret;
  20225.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20226.   __ret = __noswap_vqrdmulhq_s16(__rev0, (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1});
  20227.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20228.   return __ret;
  20229. }
  20230. #endif
  20231.  
  20232. #ifdef __LITTLE_ENDIAN__
  20233. __ai int32x2_t vqrdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
  20234.   int32x2_t __ret;
  20235.   __ret = vqrdmulh_s32(__p0, (int32x2_t) {__p1, __p1});
  20236.   return __ret;
  20237. }
  20238. #else
  20239. __ai int32x2_t vqrdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
  20240.   int32x2_t __ret;
  20241.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20242.   __ret = __noswap_vqrdmulh_s32(__rev0, (int32x2_t) {__p1, __p1});
  20243.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20244.   return __ret;
  20245. }
  20246. #endif
  20247.  
  20248. #ifdef __LITTLE_ENDIAN__
  20249. __ai int16x4_t vqrdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
  20250.   int16x4_t __ret;
  20251.   __ret = vqrdmulh_s16(__p0, (int16x4_t) {__p1, __p1, __p1, __p1});
  20252.   return __ret;
  20253. }
  20254. #else
  20255. __ai int16x4_t vqrdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
  20256.   int16x4_t __ret;
  20257.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20258.   __ret = __noswap_vqrdmulh_s16(__rev0, (int16x4_t) {__p1, __p1, __p1, __p1});
  20259.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20260.   return __ret;
  20261. }
  20262. #endif
  20263.  
  20264. #ifdef __LITTLE_ENDIAN__
  20265. __ai uint8x16_t vqrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
  20266.   uint8x16_t __ret;
  20267.   __ret = (uint8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  20268.   return __ret;
  20269. }
  20270. #else
  20271. __ai uint8x16_t vqrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
  20272.   uint8x16_t __ret;
  20273.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20274.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20275.   __ret = (uint8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  20276.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20277.   return __ret;
  20278. }
  20279. #endif
  20280.  
  20281. #ifdef __LITTLE_ENDIAN__
  20282. __ai uint32x4_t vqrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
  20283.   uint32x4_t __ret;
  20284.   __ret = (uint32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  20285.   return __ret;
  20286. }
  20287. #else
  20288. __ai uint32x4_t vqrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
  20289.   uint32x4_t __ret;
  20290.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20291.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20292.   __ret = (uint32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  20293.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20294.   return __ret;
  20295. }
  20296. #endif
  20297.  
  20298. #ifdef __LITTLE_ENDIAN__
  20299. __ai uint64x2_t vqrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
  20300.   uint64x2_t __ret;
  20301.   __ret = (uint64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  20302.   return __ret;
  20303. }
  20304. #else
  20305. __ai uint64x2_t vqrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
  20306.   uint64x2_t __ret;
  20307.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20308.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  20309.   __ret = (uint64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  20310.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20311.   return __ret;
  20312. }
  20313. #endif
  20314.  
  20315. #ifdef __LITTLE_ENDIAN__
  20316. __ai uint16x8_t vqrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
  20317.   uint16x8_t __ret;
  20318.   __ret = (uint16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  20319.   return __ret;
  20320. }
  20321. #else
  20322. __ai uint16x8_t vqrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
  20323.   uint16x8_t __ret;
  20324.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20325.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  20326.   __ret = (uint16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  20327.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20328.   return __ret;
  20329. }
  20330. #endif
  20331.  
  20332. #ifdef __LITTLE_ENDIAN__
  20333. __ai int8x16_t vqrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
  20334.   int8x16_t __ret;
  20335.   __ret = (int8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  20336.   return __ret;
  20337. }
  20338. #else
  20339. __ai int8x16_t vqrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
  20340.   int8x16_t __ret;
  20341.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20342.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20343.   __ret = (int8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  20344.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20345.   return __ret;
  20346. }
  20347. #endif
  20348.  
  20349. #ifdef __LITTLE_ENDIAN__
  20350. __ai int32x4_t vqrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
  20351.   int32x4_t __ret;
  20352.   __ret = (int32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  20353.   return __ret;
  20354. }
  20355. #else
  20356. __ai int32x4_t vqrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
  20357.   int32x4_t __ret;
  20358.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20359.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20360.   __ret = (int32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  20361.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20362.   return __ret;
  20363. }
  20364. #endif
  20365.  
  20366. #ifdef __LITTLE_ENDIAN__
  20367. __ai int64x2_t vqrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
  20368.   int64x2_t __ret;
  20369.   __ret = (int64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
  20370.   return __ret;
  20371. }
  20372. #else
  20373. __ai int64x2_t vqrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
  20374.   int64x2_t __ret;
  20375.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20376.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  20377.   __ret = (int64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
  20378.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20379.   return __ret;
  20380. }
  20381. #endif
  20382.  
  20383. #ifdef __LITTLE_ENDIAN__
  20384. __ai int16x8_t vqrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
  20385.   int16x8_t __ret;
  20386.   __ret = (int16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  20387.   return __ret;
  20388. }
  20389. #else
  20390. __ai int16x8_t vqrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
  20391.   int16x8_t __ret;
  20392.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20393.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  20394.   __ret = (int16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  20395.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20396.   return __ret;
  20397. }
  20398. #endif
  20399.  
  20400. #ifdef __LITTLE_ENDIAN__
  20401. __ai uint8x8_t vqrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
  20402.   uint8x8_t __ret;
  20403.   __ret = (uint8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  20404.   return __ret;
  20405. }
  20406. #else
  20407. __ai uint8x8_t vqrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
  20408.   uint8x8_t __ret;
  20409.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20410.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  20411.   __ret = (uint8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  20412.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20413.   return __ret;
  20414. }
  20415. #endif
  20416.  
  20417. #ifdef __LITTLE_ENDIAN__
  20418. __ai uint32x2_t vqrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
  20419.   uint32x2_t __ret;
  20420.   __ret = (uint32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  20421.   return __ret;
  20422. }
  20423. #else
  20424. __ai uint32x2_t vqrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
  20425.   uint32x2_t __ret;
  20426.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20427.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  20428.   __ret = (uint32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  20429.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20430.   return __ret;
  20431. }
  20432. #endif
  20433.  
  20434. __ai uint64x1_t vqrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
  20435.   uint64x1_t __ret;
  20436.   __ret = (uint64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  20437.   return __ret;
  20438. }
  20439. #ifdef __LITTLE_ENDIAN__
  20440. __ai uint16x4_t vqrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
  20441.   uint16x4_t __ret;
  20442.   __ret = (uint16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  20443.   return __ret;
  20444. }
  20445. #else
  20446. __ai uint16x4_t vqrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
  20447.   uint16x4_t __ret;
  20448.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20449.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20450.   __ret = (uint16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  20451.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20452.   return __ret;
  20453. }
  20454. #endif
  20455.  
  20456. #ifdef __LITTLE_ENDIAN__
  20457. __ai int8x8_t vqrshl_s8(int8x8_t __p0, int8x8_t __p1) {
  20458.   int8x8_t __ret;
  20459.   __ret = (int8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  20460.   return __ret;
  20461. }
  20462. #else
  20463. __ai int8x8_t vqrshl_s8(int8x8_t __p0, int8x8_t __p1) {
  20464.   int8x8_t __ret;
  20465.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20466.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  20467.   __ret = (int8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  20468.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20469.   return __ret;
  20470. }
  20471. #endif
  20472.  
  20473. #ifdef __LITTLE_ENDIAN__
  20474. __ai int32x2_t vqrshl_s32(int32x2_t __p0, int32x2_t __p1) {
  20475.   int32x2_t __ret;
  20476.   __ret = (int32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  20477.   return __ret;
  20478. }
  20479. #else
  20480. __ai int32x2_t vqrshl_s32(int32x2_t __p0, int32x2_t __p1) {
  20481.   int32x2_t __ret;
  20482.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20483.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  20484.   __ret = (int32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  20485.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20486.   return __ret;
  20487. }
  20488. #endif
  20489.  
  20490. __ai int64x1_t vqrshl_s64(int64x1_t __p0, int64x1_t __p1) {
  20491.   int64x1_t __ret;
  20492.   __ret = (int64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
  20493.   return __ret;
  20494. }
  20495. #ifdef __LITTLE_ENDIAN__
  20496. __ai int16x4_t vqrshl_s16(int16x4_t __p0, int16x4_t __p1) {
  20497.   int16x4_t __ret;
  20498.   __ret = (int16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  20499.   return __ret;
  20500. }
  20501. #else
  20502. __ai int16x4_t vqrshl_s16(int16x4_t __p0, int16x4_t __p1) {
  20503.   int16x4_t __ret;
  20504.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20505.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20506.   __ret = (int16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  20507.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20508.   return __ret;
  20509. }
  20510. #endif
  20511.  
  20512. #ifdef __LITTLE_ENDIAN__
  20513. #define vqrshrn_n_u32(__p0, __p1) __extension__ ({ \
  20514.   uint16x4_t __ret; \
  20515.   uint32x4_t __s0 = __p0; \
  20516.   __ret = (uint16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 17); \
  20517.   __ret; \
  20518. })
  20519. #else
  20520. #define vqrshrn_n_u32(__p0, __p1) __extension__ ({ \
  20521.   uint16x4_t __ret; \
  20522.   uint32x4_t __s0 = __p0; \
  20523.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  20524.   __ret = (uint16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 17); \
  20525.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  20526.   __ret; \
  20527. })
  20528. #define __noswap_vqrshrn_n_u32(__p0, __p1) __extension__ ({ \
  20529.   uint16x4_t __ret; \
  20530.   uint32x4_t __s0 = __p0; \
  20531.   __ret = (uint16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 17); \
  20532.   __ret; \
  20533. })
  20534. #endif
  20535.  
  20536. #ifdef __LITTLE_ENDIAN__
  20537. #define vqrshrn_n_u64(__p0, __p1) __extension__ ({ \
  20538.   uint32x2_t __ret; \
  20539.   uint64x2_t __s0 = __p0; \
  20540.   __ret = (uint32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 18); \
  20541.   __ret; \
  20542. })
  20543. #else
  20544. #define vqrshrn_n_u64(__p0, __p1) __extension__ ({ \
  20545.   uint32x2_t __ret; \
  20546.   uint64x2_t __s0 = __p0; \
  20547.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  20548.   __ret = (uint32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 18); \
  20549.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  20550.   __ret; \
  20551. })
  20552. #define __noswap_vqrshrn_n_u64(__p0, __p1) __extension__ ({ \
  20553.   uint32x2_t __ret; \
  20554.   uint64x2_t __s0 = __p0; \
  20555.   __ret = (uint32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 18); \
  20556.   __ret; \
  20557. })
  20558. #endif
  20559.  
  20560. #ifdef __LITTLE_ENDIAN__
  20561. #define vqrshrn_n_u16(__p0, __p1) __extension__ ({ \
  20562.   uint8x8_t __ret; \
  20563.   uint16x8_t __s0 = __p0; \
  20564.   __ret = (uint8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 16); \
  20565.   __ret; \
  20566. })
  20567. #else
  20568. #define vqrshrn_n_u16(__p0, __p1) __extension__ ({ \
  20569.   uint8x8_t __ret; \
  20570.   uint16x8_t __s0 = __p0; \
  20571.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  20572.   __ret = (uint8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 16); \
  20573.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  20574.   __ret; \
  20575. })
  20576. #define __noswap_vqrshrn_n_u16(__p0, __p1) __extension__ ({ \
  20577.   uint8x8_t __ret; \
  20578.   uint16x8_t __s0 = __p0; \
  20579.   __ret = (uint8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 16); \
  20580.   __ret; \
  20581. })
  20582. #endif
  20583.  
  20584. #ifdef __LITTLE_ENDIAN__
  20585. #define vqrshrn_n_s32(__p0, __p1) __extension__ ({ \
  20586.   int16x4_t __ret; \
  20587.   int32x4_t __s0 = __p0; \
  20588.   __ret = (int16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 1); \
  20589.   __ret; \
  20590. })
  20591. #else
  20592. #define vqrshrn_n_s32(__p0, __p1) __extension__ ({ \
  20593.   int16x4_t __ret; \
  20594.   int32x4_t __s0 = __p0; \
  20595.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  20596.   __ret = (int16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 1); \
  20597.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  20598.   __ret; \
  20599. })
  20600. #define __noswap_vqrshrn_n_s32(__p0, __p1) __extension__ ({ \
  20601.   int16x4_t __ret; \
  20602.   int32x4_t __s0 = __p0; \
  20603.   __ret = (int16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 1); \
  20604.   __ret; \
  20605. })
  20606. #endif
  20607.  
  20608. #ifdef __LITTLE_ENDIAN__
  20609. #define vqrshrn_n_s64(__p0, __p1) __extension__ ({ \
  20610.   int32x2_t __ret; \
  20611.   int64x2_t __s0 = __p0; \
  20612.   __ret = (int32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 2); \
  20613.   __ret; \
  20614. })
  20615. #else
  20616. #define vqrshrn_n_s64(__p0, __p1) __extension__ ({ \
  20617.   int32x2_t __ret; \
  20618.   int64x2_t __s0 = __p0; \
  20619.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  20620.   __ret = (int32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 2); \
  20621.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  20622.   __ret; \
  20623. })
  20624. #define __noswap_vqrshrn_n_s64(__p0, __p1) __extension__ ({ \
  20625.   int32x2_t __ret; \
  20626.   int64x2_t __s0 = __p0; \
  20627.   __ret = (int32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 2); \
  20628.   __ret; \
  20629. })
  20630. #endif
  20631.  
  20632. #ifdef __LITTLE_ENDIAN__
  20633. #define vqrshrn_n_s16(__p0, __p1) __extension__ ({ \
  20634.   int8x8_t __ret; \
  20635.   int16x8_t __s0 = __p0; \
  20636.   __ret = (int8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 0); \
  20637.   __ret; \
  20638. })
  20639. #else
  20640. #define vqrshrn_n_s16(__p0, __p1) __extension__ ({ \
  20641.   int8x8_t __ret; \
  20642.   int16x8_t __s0 = __p0; \
  20643.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  20644.   __ret = (int8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 0); \
  20645.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  20646.   __ret; \
  20647. })
  20648. #define __noswap_vqrshrn_n_s16(__p0, __p1) __extension__ ({ \
  20649.   int8x8_t __ret; \
  20650.   int16x8_t __s0 = __p0; \
  20651.   __ret = (int8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 0); \
  20652.   __ret; \
  20653. })
  20654. #endif
  20655.  
  20656. #ifdef __LITTLE_ENDIAN__
  20657. #define vqrshrun_n_s32(__p0, __p1) __extension__ ({ \
  20658.   uint16x4_t __ret; \
  20659.   int32x4_t __s0 = __p0; \
  20660.   __ret = (uint16x4_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 17); \
  20661.   __ret; \
  20662. })
  20663. #else
  20664. #define vqrshrun_n_s32(__p0, __p1) __extension__ ({ \
  20665.   uint16x4_t __ret; \
  20666.   int32x4_t __s0 = __p0; \
  20667.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  20668.   __ret = (uint16x4_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__rev0, __p1, 17); \
  20669.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  20670.   __ret; \
  20671. })
  20672. #define __noswap_vqrshrun_n_s32(__p0, __p1) __extension__ ({ \
  20673.   uint16x4_t __ret; \
  20674.   int32x4_t __s0 = __p0; \
  20675.   __ret = (uint16x4_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 17); \
  20676.   __ret; \
  20677. })
  20678. #endif
  20679.  
  20680. #ifdef __LITTLE_ENDIAN__
  20681. #define vqrshrun_n_s64(__p0, __p1) __extension__ ({ \
  20682.   uint32x2_t __ret; \
  20683.   int64x2_t __s0 = __p0; \
  20684.   __ret = (uint32x2_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 18); \
  20685.   __ret; \
  20686. })
  20687. #else
  20688. #define vqrshrun_n_s64(__p0, __p1) __extension__ ({ \
  20689.   uint32x2_t __ret; \
  20690.   int64x2_t __s0 = __p0; \
  20691.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  20692.   __ret = (uint32x2_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__rev0, __p1, 18); \
  20693.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  20694.   __ret; \
  20695. })
  20696. #define __noswap_vqrshrun_n_s64(__p0, __p1) __extension__ ({ \
  20697.   uint32x2_t __ret; \
  20698.   int64x2_t __s0 = __p0; \
  20699.   __ret = (uint32x2_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 18); \
  20700.   __ret; \
  20701. })
  20702. #endif
  20703.  
  20704. #ifdef __LITTLE_ENDIAN__
  20705. #define vqrshrun_n_s16(__p0, __p1) __extension__ ({ \
  20706.   uint8x8_t __ret; \
  20707.   int16x8_t __s0 = __p0; \
  20708.   __ret = (uint8x8_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 16); \
  20709.   __ret; \
  20710. })
  20711. #else
  20712. #define vqrshrun_n_s16(__p0, __p1) __extension__ ({ \
  20713.   uint8x8_t __ret; \
  20714.   int16x8_t __s0 = __p0; \
  20715.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  20716.   __ret = (uint8x8_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__rev0, __p1, 16); \
  20717.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  20718.   __ret; \
  20719. })
  20720. #define __noswap_vqrshrun_n_s16(__p0, __p1) __extension__ ({ \
  20721.   uint8x8_t __ret; \
  20722.   int16x8_t __s0 = __p0; \
  20723.   __ret = (uint8x8_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 16); \
  20724.   __ret; \
  20725. })
  20726. #endif
  20727.  
  20728. #ifdef __LITTLE_ENDIAN__
  20729. __ai uint8x16_t vqshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
  20730.   uint8x16_t __ret;
  20731.   __ret = (uint8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  20732.   return __ret;
  20733. }
  20734. #else
  20735. __ai uint8x16_t vqshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
  20736.   uint8x16_t __ret;
  20737.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20738.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20739.   __ret = (uint8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  20740.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20741.   return __ret;
  20742. }
  20743. #endif
  20744.  
  20745. #ifdef __LITTLE_ENDIAN__
  20746. __ai uint32x4_t vqshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
  20747.   uint32x4_t __ret;
  20748.   __ret = (uint32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  20749.   return __ret;
  20750. }
  20751. #else
  20752. __ai uint32x4_t vqshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
  20753.   uint32x4_t __ret;
  20754.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20755.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20756.   __ret = (uint32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  20757.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20758.   return __ret;
  20759. }
  20760. #endif
  20761.  
  20762. #ifdef __LITTLE_ENDIAN__
  20763. __ai uint64x2_t vqshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
  20764.   uint64x2_t __ret;
  20765.   __ret = (uint64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  20766.   return __ret;
  20767. }
  20768. #else
  20769. __ai uint64x2_t vqshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
  20770.   uint64x2_t __ret;
  20771.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20772.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  20773.   __ret = (uint64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  20774.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20775.   return __ret;
  20776. }
  20777. #endif
  20778.  
  20779. #ifdef __LITTLE_ENDIAN__
  20780. __ai uint16x8_t vqshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
  20781.   uint16x8_t __ret;
  20782.   __ret = (uint16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  20783.   return __ret;
  20784. }
  20785. #else
  20786. __ai uint16x8_t vqshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
  20787.   uint16x8_t __ret;
  20788.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20789.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  20790.   __ret = (uint16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  20791.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20792.   return __ret;
  20793. }
  20794. #endif
  20795.  
  20796. #ifdef __LITTLE_ENDIAN__
  20797. __ai int8x16_t vqshlq_s8(int8x16_t __p0, int8x16_t __p1) {
  20798.   int8x16_t __ret;
  20799.   __ret = (int8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  20800.   return __ret;
  20801. }
  20802. #else
  20803. __ai int8x16_t vqshlq_s8(int8x16_t __p0, int8x16_t __p1) {
  20804.   int8x16_t __ret;
  20805.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20806.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20807.   __ret = (int8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  20808.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  20809.   return __ret;
  20810. }
  20811. #endif
  20812.  
  20813. #ifdef __LITTLE_ENDIAN__
  20814. __ai int32x4_t vqshlq_s32(int32x4_t __p0, int32x4_t __p1) {
  20815.   int32x4_t __ret;
  20816.   __ret = (int32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  20817.   return __ret;
  20818. }
  20819. #else
  20820. __ai int32x4_t vqshlq_s32(int32x4_t __p0, int32x4_t __p1) {
  20821.   int32x4_t __ret;
  20822.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20823.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20824.   __ret = (int32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  20825.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20826.   return __ret;
  20827. }
  20828. #endif
  20829.  
  20830. #ifdef __LITTLE_ENDIAN__
  20831. __ai int64x2_t vqshlq_s64(int64x2_t __p0, int64x2_t __p1) {
  20832.   int64x2_t __ret;
  20833.   __ret = (int64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
  20834.   return __ret;
  20835. }
  20836. #else
  20837. __ai int64x2_t vqshlq_s64(int64x2_t __p0, int64x2_t __p1) {
  20838.   int64x2_t __ret;
  20839.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20840.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  20841.   __ret = (int64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
  20842.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20843.   return __ret;
  20844. }
  20845. #endif
  20846.  
  20847. #ifdef __LITTLE_ENDIAN__
  20848. __ai int16x8_t vqshlq_s16(int16x8_t __p0, int16x8_t __p1) {
  20849.   int16x8_t __ret;
  20850.   __ret = (int16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  20851.   return __ret;
  20852. }
  20853. #else
  20854. __ai int16x8_t vqshlq_s16(int16x8_t __p0, int16x8_t __p1) {
  20855.   int16x8_t __ret;
  20856.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20857.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  20858.   __ret = (int16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  20859.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20860.   return __ret;
  20861. }
  20862. #endif
  20863.  
  20864. #ifdef __LITTLE_ENDIAN__
  20865. __ai uint8x8_t vqshl_u8(uint8x8_t __p0, int8x8_t __p1) {
  20866.   uint8x8_t __ret;
  20867.   __ret = (uint8x8_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  20868.   return __ret;
  20869. }
  20870. #else
  20871. __ai uint8x8_t vqshl_u8(uint8x8_t __p0, int8x8_t __p1) {
  20872.   uint8x8_t __ret;
  20873.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20874.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  20875.   __ret = (uint8x8_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  20876.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20877.   return __ret;
  20878. }
  20879. #endif
  20880.  
  20881. #ifdef __LITTLE_ENDIAN__
  20882. __ai uint32x2_t vqshl_u32(uint32x2_t __p0, int32x2_t __p1) {
  20883.   uint32x2_t __ret;
  20884.   __ret = (uint32x2_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  20885.   return __ret;
  20886. }
  20887. #else
  20888. __ai uint32x2_t vqshl_u32(uint32x2_t __p0, int32x2_t __p1) {
  20889.   uint32x2_t __ret;
  20890.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20891.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  20892.   __ret = (uint32x2_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  20893.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20894.   return __ret;
  20895. }
  20896. #endif
  20897.  
  20898. __ai uint64x1_t vqshl_u64(uint64x1_t __p0, int64x1_t __p1) {
  20899.   uint64x1_t __ret;
  20900.   __ret = (uint64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  20901.   return __ret;
  20902. }
  20903. #ifdef __LITTLE_ENDIAN__
  20904. __ai uint16x4_t vqshl_u16(uint16x4_t __p0, int16x4_t __p1) {
  20905.   uint16x4_t __ret;
  20906.   __ret = (uint16x4_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  20907.   return __ret;
  20908. }
  20909. #else
  20910. __ai uint16x4_t vqshl_u16(uint16x4_t __p0, int16x4_t __p1) {
  20911.   uint16x4_t __ret;
  20912.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20913.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20914.   __ret = (uint16x4_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  20915.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20916.   return __ret;
  20917. }
  20918. #endif
  20919.  
  20920. #ifdef __LITTLE_ENDIAN__
  20921. __ai int8x8_t vqshl_s8(int8x8_t __p0, int8x8_t __p1) {
  20922.   int8x8_t __ret;
  20923.   __ret = (int8x8_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  20924.   return __ret;
  20925. }
  20926. #else
  20927. __ai int8x8_t vqshl_s8(int8x8_t __p0, int8x8_t __p1) {
  20928.   int8x8_t __ret;
  20929.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  20930.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  20931.   __ret = (int8x8_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  20932.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  20933.   return __ret;
  20934. }
  20935. #endif
  20936.  
  20937. #ifdef __LITTLE_ENDIAN__
  20938. __ai int32x2_t vqshl_s32(int32x2_t __p0, int32x2_t __p1) {
  20939.   int32x2_t __ret;
  20940.   __ret = (int32x2_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  20941.   return __ret;
  20942. }
  20943. #else
  20944. __ai int32x2_t vqshl_s32(int32x2_t __p0, int32x2_t __p1) {
  20945.   int32x2_t __ret;
  20946.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  20947.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  20948.   __ret = (int32x2_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  20949.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  20950.   return __ret;
  20951. }
  20952. #endif
  20953.  
  20954. __ai int64x1_t vqshl_s64(int64x1_t __p0, int64x1_t __p1) {
  20955.   int64x1_t __ret;
  20956.   __ret = (int64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
  20957.   return __ret;
  20958. }
  20959. #ifdef __LITTLE_ENDIAN__
  20960. __ai int16x4_t vqshl_s16(int16x4_t __p0, int16x4_t __p1) {
  20961.   int16x4_t __ret;
  20962.   __ret = (int16x4_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  20963.   return __ret;
  20964. }
  20965. #else
  20966. __ai int16x4_t vqshl_s16(int16x4_t __p0, int16x4_t __p1) {
  20967.   int16x4_t __ret;
  20968.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  20969.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  20970.   __ret = (int16x4_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  20971.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  20972.   return __ret;
  20973. }
  20974. #endif
  20975.  
  20976. #ifdef __LITTLE_ENDIAN__
  20977. #define vqshlq_n_u8(__p0, __p1) __extension__ ({ \
  20978.   uint8x16_t __ret; \
  20979.   uint8x16_t __s0 = __p0; \
  20980.   __ret = (uint8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 48); \
  20981.   __ret; \
  20982. })
  20983. #else
  20984. #define vqshlq_n_u8(__p0, __p1) __extension__ ({ \
  20985.   uint8x16_t __ret; \
  20986.   uint8x16_t __s0 = __p0; \
  20987.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  20988.   __ret = (uint8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 48); \
  20989.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  20990.   __ret; \
  20991. })
  20992. #endif
  20993.  
  20994. #ifdef __LITTLE_ENDIAN__
  20995. #define vqshlq_n_u32(__p0, __p1) __extension__ ({ \
  20996.   uint32x4_t __ret; \
  20997.   uint32x4_t __s0 = __p0; \
  20998.   __ret = (uint32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 50); \
  20999.   __ret; \
  21000. })
  21001. #else
  21002. #define vqshlq_n_u32(__p0, __p1) __extension__ ({ \
  21003.   uint32x4_t __ret; \
  21004.   uint32x4_t __s0 = __p0; \
  21005.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  21006.   __ret = (uint32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 50); \
  21007.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  21008.   __ret; \
  21009. })
  21010. #endif
  21011.  
  21012. #ifdef __LITTLE_ENDIAN__
  21013. #define vqshlq_n_u64(__p0, __p1) __extension__ ({ \
  21014.   uint64x2_t __ret; \
  21015.   uint64x2_t __s0 = __p0; \
  21016.   __ret = (uint64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 51); \
  21017.   __ret; \
  21018. })
  21019. #else
  21020. #define vqshlq_n_u64(__p0, __p1) __extension__ ({ \
  21021.   uint64x2_t __ret; \
  21022.   uint64x2_t __s0 = __p0; \
  21023.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  21024.   __ret = (uint64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 51); \
  21025.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  21026.   __ret; \
  21027. })
  21028. #endif
  21029.  
  21030. #ifdef __LITTLE_ENDIAN__
  21031. #define vqshlq_n_u16(__p0, __p1) __extension__ ({ \
  21032.   uint16x8_t __ret; \
  21033.   uint16x8_t __s0 = __p0; \
  21034.   __ret = (uint16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 49); \
  21035.   __ret; \
  21036. })
  21037. #else
  21038. #define vqshlq_n_u16(__p0, __p1) __extension__ ({ \
  21039.   uint16x8_t __ret; \
  21040.   uint16x8_t __s0 = __p0; \
  21041.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  21042.   __ret = (uint16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 49); \
  21043.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  21044.   __ret; \
  21045. })
  21046. #endif
  21047.  
  21048. #ifdef __LITTLE_ENDIAN__
  21049. #define vqshlq_n_s8(__p0, __p1) __extension__ ({ \
  21050.   int8x16_t __ret; \
  21051.   int8x16_t __s0 = __p0; \
  21052.   __ret = (int8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 32); \
  21053.   __ret; \
  21054. })
  21055. #else
  21056. #define vqshlq_n_s8(__p0, __p1) __extension__ ({ \
  21057.   int8x16_t __ret; \
  21058.   int8x16_t __s0 = __p0; \
  21059.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  21060.   __ret = (int8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 32); \
  21061.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  21062.   __ret; \
  21063. })
  21064. #endif
  21065.  
  21066. #ifdef __LITTLE_ENDIAN__
  21067. #define vqshlq_n_s32(__p0, __p1) __extension__ ({ \
  21068.   int32x4_t __ret; \
  21069.   int32x4_t __s0 = __p0; \
  21070.   __ret = (int32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 34); \
  21071.   __ret; \
  21072. })
  21073. #else
  21074. #define vqshlq_n_s32(__p0, __p1) __extension__ ({ \
  21075.   int32x4_t __ret; \
  21076.   int32x4_t __s0 = __p0; \
  21077.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  21078.   __ret = (int32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 34); \
  21079.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  21080.   __ret; \
  21081. })
  21082. #endif
  21083.  
  21084. #ifdef __LITTLE_ENDIAN__
  21085. #define vqshlq_n_s64(__p0, __p1) __extension__ ({ \
  21086.   int64x2_t __ret; \
  21087.   int64x2_t __s0 = __p0; \
  21088.   __ret = (int64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 35); \
  21089.   __ret; \
  21090. })
  21091. #else
  21092. #define vqshlq_n_s64(__p0, __p1) __extension__ ({ \
  21093.   int64x2_t __ret; \
  21094.   int64x2_t __s0 = __p0; \
  21095.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  21096.   __ret = (int64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 35); \
  21097.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  21098.   __ret; \
  21099. })
  21100. #endif
  21101.  
  21102. #ifdef __LITTLE_ENDIAN__
  21103. #define vqshlq_n_s16(__p0, __p1) __extension__ ({ \
  21104.   int16x8_t __ret; \
  21105.   int16x8_t __s0 = __p0; \
  21106.   __ret = (int16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 33); \
  21107.   __ret; \
  21108. })
  21109. #else
  21110. #define vqshlq_n_s16(__p0, __p1) __extension__ ({ \
  21111.   int16x8_t __ret; \
  21112.   int16x8_t __s0 = __p0; \
  21113.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  21114.   __ret = (int16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 33); \
  21115.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  21116.   __ret; \
  21117. })
  21118. #endif
  21119.  
  21120. #ifdef __LITTLE_ENDIAN__
  21121. #define vqshl_n_u8(__p0, __p1) __extension__ ({ \
  21122.   uint8x8_t __ret; \
  21123.   uint8x8_t __s0 = __p0; \
  21124.   __ret = (uint8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 16); \
  21125.   __ret; \
  21126. })
  21127. #else
  21128. #define vqshl_n_u8(__p0, __p1) __extension__ ({ \
  21129.   uint8x8_t __ret; \
  21130.   uint8x8_t __s0 = __p0; \
  21131.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  21132.   __ret = (uint8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 16); \
  21133.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  21134.   __ret; \
  21135. })
  21136. #endif
  21137.  
  21138. #ifdef __LITTLE_ENDIAN__
  21139. #define vqshl_n_u32(__p0, __p1) __extension__ ({ \
  21140.   uint32x2_t __ret; \
  21141.   uint32x2_t __s0 = __p0; \
  21142.   __ret = (uint32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 18); \
  21143.   __ret; \
  21144. })
  21145. #else
  21146. #define vqshl_n_u32(__p0, __p1) __extension__ ({ \
  21147.   uint32x2_t __ret; \
  21148.   uint32x2_t __s0 = __p0; \
  21149.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  21150.   __ret = (uint32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 18); \
  21151.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  21152.   __ret; \
  21153. })
  21154. #endif
  21155.  
  21156. #define vqshl_n_u64(__p0, __p1) __extension__ ({ \
  21157.   uint64x1_t __ret; \
  21158.   uint64x1_t __s0 = __p0; \
  21159.   __ret = (uint64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 19); \
  21160.   __ret; \
  21161. })
  21162. #ifdef __LITTLE_ENDIAN__
  21163. #define vqshl_n_u16(__p0, __p1) __extension__ ({ \
  21164.   uint16x4_t __ret; \
  21165.   uint16x4_t __s0 = __p0; \
  21166.   __ret = (uint16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 17); \
  21167.   __ret; \
  21168. })
  21169. #else
  21170. #define vqshl_n_u16(__p0, __p1) __extension__ ({ \
  21171.   uint16x4_t __ret; \
  21172.   uint16x4_t __s0 = __p0; \
  21173.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  21174.   __ret = (uint16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 17); \
  21175.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  21176.   __ret; \
  21177. })
  21178. #endif
  21179.  
  21180. #ifdef __LITTLE_ENDIAN__
  21181. #define vqshl_n_s8(__p0, __p1) __extension__ ({ \
  21182.   int8x8_t __ret; \
  21183.   int8x8_t __s0 = __p0; \
  21184.   __ret = (int8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 0); \
  21185.   __ret; \
  21186. })
  21187. #else
  21188. #define vqshl_n_s8(__p0, __p1) __extension__ ({ \
  21189.   int8x8_t __ret; \
  21190.   int8x8_t __s0 = __p0; \
  21191.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  21192.   __ret = (int8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 0); \
  21193.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  21194.   __ret; \
  21195. })
  21196. #endif
  21197.  
  21198. #ifdef __LITTLE_ENDIAN__
  21199. #define vqshl_n_s32(__p0, __p1) __extension__ ({ \
  21200.   int32x2_t __ret; \
  21201.   int32x2_t __s0 = __p0; \
  21202.   __ret = (int32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 2); \
  21203.   __ret; \
  21204. })
  21205. #else
  21206. #define vqshl_n_s32(__p0, __p1) __extension__ ({ \
  21207.   int32x2_t __ret; \
  21208.   int32x2_t __s0 = __p0; \
  21209.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  21210.   __ret = (int32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 2); \
  21211.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  21212.   __ret; \
  21213. })
  21214. #endif
  21215.  
  21216. #define vqshl_n_s64(__p0, __p1) __extension__ ({ \
  21217.   int64x1_t __ret; \
  21218.   int64x1_t __s0 = __p0; \
  21219.   __ret = (int64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 3); \
  21220.   __ret; \
  21221. })
  21222. #ifdef __LITTLE_ENDIAN__
  21223. #define vqshl_n_s16(__p0, __p1) __extension__ ({ \
  21224.   int16x4_t __ret; \
  21225.   int16x4_t __s0 = __p0; \
  21226.   __ret = (int16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 1); \
  21227.   __ret; \
  21228. })
  21229. #else
  21230. #define vqshl_n_s16(__p0, __p1) __extension__ ({ \
  21231.   int16x4_t __ret; \
  21232.   int16x4_t __s0 = __p0; \
  21233.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  21234.   __ret = (int16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 1); \
  21235.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  21236.   __ret; \
  21237. })
  21238. #endif
  21239.  
  21240. #ifdef __LITTLE_ENDIAN__
  21241. #define vqshluq_n_s8(__p0, __p1) __extension__ ({ \
  21242.   uint8x16_t __ret; \
  21243.   int8x16_t __s0 = __p0; \
  21244.   __ret = (uint8x16_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 48); \
  21245.   __ret; \
  21246. })
  21247. #else
  21248. #define vqshluq_n_s8(__p0, __p1) __extension__ ({ \
  21249.   uint8x16_t __ret; \
  21250.   int8x16_t __s0 = __p0; \
  21251.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  21252.   __ret = (uint8x16_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 48); \
  21253.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  21254.   __ret; \
  21255. })
  21256. #endif
  21257.  
  21258. #ifdef __LITTLE_ENDIAN__
  21259. #define vqshluq_n_s32(__p0, __p1) __extension__ ({ \
  21260.   uint32x4_t __ret; \
  21261.   int32x4_t __s0 = __p0; \
  21262.   __ret = (uint32x4_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 50); \
  21263.   __ret; \
  21264. })
  21265. #else
  21266. #define vqshluq_n_s32(__p0, __p1) __extension__ ({ \
  21267.   uint32x4_t __ret; \
  21268.   int32x4_t __s0 = __p0; \
  21269.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  21270.   __ret = (uint32x4_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 50); \
  21271.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  21272.   __ret; \
  21273. })
  21274. #endif
  21275.  
  21276. #ifdef __LITTLE_ENDIAN__
  21277. #define vqshluq_n_s64(__p0, __p1) __extension__ ({ \
  21278.   uint64x2_t __ret; \
  21279.   int64x2_t __s0 = __p0; \
  21280.   __ret = (uint64x2_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 51); \
  21281.   __ret; \
  21282. })
  21283. #else
  21284. #define vqshluq_n_s64(__p0, __p1) __extension__ ({ \
  21285.   uint64x2_t __ret; \
  21286.   int64x2_t __s0 = __p0; \
  21287.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  21288.   __ret = (uint64x2_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 51); \
  21289.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  21290.   __ret; \
  21291. })
  21292. #endif
  21293.  
  21294. #ifdef __LITTLE_ENDIAN__
  21295. #define vqshluq_n_s16(__p0, __p1) __extension__ ({ \
  21296.   uint16x8_t __ret; \
  21297.   int16x8_t __s0 = __p0; \
  21298.   __ret = (uint16x8_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 49); \
  21299.   __ret; \
  21300. })
  21301. #else
  21302. #define vqshluq_n_s16(__p0, __p1) __extension__ ({ \
  21303.   uint16x8_t __ret; \
  21304.   int16x8_t __s0 = __p0; \
  21305.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  21306.   __ret = (uint16x8_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 49); \
  21307.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  21308.   __ret; \
  21309. })
  21310. #endif
  21311.  
  21312. #ifdef __LITTLE_ENDIAN__
  21313. #define vqshlu_n_s8(__p0, __p1) __extension__ ({ \
  21314.   uint8x8_t __ret; \
  21315.   int8x8_t __s0 = __p0; \
  21316.   __ret = (uint8x8_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 16); \
  21317.   __ret; \
  21318. })
  21319. #else
  21320. #define vqshlu_n_s8(__p0, __p1) __extension__ ({ \
  21321.   uint8x8_t __ret; \
  21322.   int8x8_t __s0 = __p0; \
  21323.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  21324.   __ret = (uint8x8_t) __builtin_neon_vqshlu_n_v((int8x8_t)__rev0, __p1, 16); \
  21325.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  21326.   __ret; \
  21327. })
  21328. #endif
  21329.  
  21330. #ifdef __LITTLE_ENDIAN__
  21331. #define vqshlu_n_s32(__p0, __p1) __extension__ ({ \
  21332.   uint32x2_t __ret; \
  21333.   int32x2_t __s0 = __p0; \
  21334.   __ret = (uint32x2_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 18); \
  21335.   __ret; \
  21336. })
  21337. #else
  21338. #define vqshlu_n_s32(__p0, __p1) __extension__ ({ \
  21339.   uint32x2_t __ret; \
  21340.   int32x2_t __s0 = __p0; \
  21341.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  21342.   __ret = (uint32x2_t) __builtin_neon_vqshlu_n_v((int8x8_t)__rev0, __p1, 18); \
  21343.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  21344.   __ret; \
  21345. })
  21346. #endif
  21347.  
  21348. #define vqshlu_n_s64(__p0, __p1) __extension__ ({ \
  21349.   uint64x1_t __ret; \
  21350.   int64x1_t __s0 = __p0; \
  21351.   __ret = (uint64x1_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 19); \
  21352.   __ret; \
  21353. })
  21354. #ifdef __LITTLE_ENDIAN__
  21355. #define vqshlu_n_s16(__p0, __p1) __extension__ ({ \
  21356.   uint16x4_t __ret; \
  21357.   int16x4_t __s0 = __p0; \
  21358.   __ret = (uint16x4_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 17); \
  21359.   __ret; \
  21360. })
  21361. #else
  21362. #define vqshlu_n_s16(__p0, __p1) __extension__ ({ \
  21363.   uint16x4_t __ret; \
  21364.   int16x4_t __s0 = __p0; \
  21365.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  21366.   __ret = (uint16x4_t) __builtin_neon_vqshlu_n_v((int8x8_t)__rev0, __p1, 17); \
  21367.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  21368.   __ret; \
  21369. })
  21370. #endif
  21371.  
  21372. #ifdef __LITTLE_ENDIAN__
  21373. #define vqshrn_n_u32(__p0, __p1) __extension__ ({ \
  21374.   uint16x4_t __ret; \
  21375.   uint32x4_t __s0 = __p0; \
  21376.   __ret = (uint16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 17); \
  21377.   __ret; \
  21378. })
  21379. #else
  21380. #define vqshrn_n_u32(__p0, __p1) __extension__ ({ \
  21381.   uint16x4_t __ret; \
  21382.   uint32x4_t __s0 = __p0; \
  21383.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  21384.   __ret = (uint16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 17); \
  21385.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  21386.   __ret; \
  21387. })
  21388. #define __noswap_vqshrn_n_u32(__p0, __p1) __extension__ ({ \
  21389.   uint16x4_t __ret; \
  21390.   uint32x4_t __s0 = __p0; \
  21391.   __ret = (uint16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 17); \
  21392.   __ret; \
  21393. })
  21394. #endif
  21395.  
  21396. #ifdef __LITTLE_ENDIAN__
  21397. #define vqshrn_n_u64(__p0, __p1) __extension__ ({ \
  21398.   uint32x2_t __ret; \
  21399.   uint64x2_t __s0 = __p0; \
  21400.   __ret = (uint32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 18); \
  21401.   __ret; \
  21402. })
  21403. #else
  21404. #define vqshrn_n_u64(__p0, __p1) __extension__ ({ \
  21405.   uint32x2_t __ret; \
  21406.   uint64x2_t __s0 = __p0; \
  21407.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  21408.   __ret = (uint32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 18); \
  21409.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  21410.   __ret; \
  21411. })
  21412. #define __noswap_vqshrn_n_u64(__p0, __p1) __extension__ ({ \
  21413.   uint32x2_t __ret; \
  21414.   uint64x2_t __s0 = __p0; \
  21415.   __ret = (uint32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 18); \
  21416.   __ret; \
  21417. })
  21418. #endif
  21419.  
  21420. #ifdef __LITTLE_ENDIAN__
  21421. #define vqshrn_n_u16(__p0, __p1) __extension__ ({ \
  21422.   uint8x8_t __ret; \
  21423.   uint16x8_t __s0 = __p0; \
  21424.   __ret = (uint8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 16); \
  21425.   __ret; \
  21426. })
  21427. #else
  21428. #define vqshrn_n_u16(__p0, __p1) __extension__ ({ \
  21429.   uint8x8_t __ret; \
  21430.   uint16x8_t __s0 = __p0; \
  21431.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  21432.   __ret = (uint8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 16); \
  21433.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  21434.   __ret; \
  21435. })
  21436. #define __noswap_vqshrn_n_u16(__p0, __p1) __extension__ ({ \
  21437.   uint8x8_t __ret; \
  21438.   uint16x8_t __s0 = __p0; \
  21439.   __ret = (uint8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 16); \
  21440.   __ret; \
  21441. })
  21442. #endif
  21443.  
  21444. #ifdef __LITTLE_ENDIAN__
  21445. #define vqshrn_n_s32(__p0, __p1) __extension__ ({ \
  21446.   int16x4_t __ret; \
  21447.   int32x4_t __s0 = __p0; \
  21448.   __ret = (int16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 1); \
  21449.   __ret; \
  21450. })
  21451. #else
  21452. #define vqshrn_n_s32(__p0, __p1) __extension__ ({ \
  21453.   int16x4_t __ret; \
  21454.   int32x4_t __s0 = __p0; \
  21455.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  21456.   __ret = (int16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 1); \
  21457.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  21458.   __ret; \
  21459. })
  21460. #define __noswap_vqshrn_n_s32(__p0, __p1) __extension__ ({ \
  21461.   int16x4_t __ret; \
  21462.   int32x4_t __s0 = __p0; \
  21463.   __ret = (int16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 1); \
  21464.   __ret; \
  21465. })
  21466. #endif
  21467.  
  21468. #ifdef __LITTLE_ENDIAN__
  21469. #define vqshrn_n_s64(__p0, __p1) __extension__ ({ \
  21470.   int32x2_t __ret; \
  21471.   int64x2_t __s0 = __p0; \
  21472.   __ret = (int32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 2); \
  21473.   __ret; \
  21474. })
  21475. #else
  21476. #define vqshrn_n_s64(__p0, __p1) __extension__ ({ \
  21477.   int32x2_t __ret; \
  21478.   int64x2_t __s0 = __p0; \
  21479.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  21480.   __ret = (int32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 2); \
  21481.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  21482.   __ret; \
  21483. })
  21484. #define __noswap_vqshrn_n_s64(__p0, __p1) __extension__ ({ \
  21485.   int32x2_t __ret; \
  21486.   int64x2_t __s0 = __p0; \
  21487.   __ret = (int32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 2); \
  21488.   __ret; \
  21489. })
  21490. #endif
  21491.  
  21492. #ifdef __LITTLE_ENDIAN__
  21493. #define vqshrn_n_s16(__p0, __p1) __extension__ ({ \
  21494.   int8x8_t __ret; \
  21495.   int16x8_t __s0 = __p0; \
  21496.   __ret = (int8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 0); \
  21497.   __ret; \
  21498. })
  21499. #else
  21500. #define vqshrn_n_s16(__p0, __p1) __extension__ ({ \
  21501.   int8x8_t __ret; \
  21502.   int16x8_t __s0 = __p0; \
  21503.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  21504.   __ret = (int8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 0); \
  21505.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  21506.   __ret; \
  21507. })
  21508. #define __noswap_vqshrn_n_s16(__p0, __p1) __extension__ ({ \
  21509.   int8x8_t __ret; \
  21510.   int16x8_t __s0 = __p0; \
  21511.   __ret = (int8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 0); \
  21512.   __ret; \
  21513. })
  21514. #endif
  21515.  
  21516. #ifdef __LITTLE_ENDIAN__
  21517. #define vqshrun_n_s32(__p0, __p1) __extension__ ({ \
  21518.   uint16x4_t __ret; \
  21519.   int32x4_t __s0 = __p0; \
  21520.   __ret = (uint16x4_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 17); \
  21521.   __ret; \
  21522. })
  21523. #else
  21524. #define vqshrun_n_s32(__p0, __p1) __extension__ ({ \
  21525.   uint16x4_t __ret; \
  21526.   int32x4_t __s0 = __p0; \
  21527.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  21528.   __ret = (uint16x4_t) __builtin_neon_vqshrun_n_v((int8x16_t)__rev0, __p1, 17); \
  21529.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  21530.   __ret; \
  21531. })
  21532. #define __noswap_vqshrun_n_s32(__p0, __p1) __extension__ ({ \
  21533.   uint16x4_t __ret; \
  21534.   int32x4_t __s0 = __p0; \
  21535.   __ret = (uint16x4_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 17); \
  21536.   __ret; \
  21537. })
  21538. #endif
  21539.  
  21540. #ifdef __LITTLE_ENDIAN__
  21541. #define vqshrun_n_s64(__p0, __p1) __extension__ ({ \
  21542.   uint32x2_t __ret; \
  21543.   int64x2_t __s0 = __p0; \
  21544.   __ret = (uint32x2_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 18); \
  21545.   __ret; \
  21546. })
  21547. #else
  21548. #define vqshrun_n_s64(__p0, __p1) __extension__ ({ \
  21549.   uint32x2_t __ret; \
  21550.   int64x2_t __s0 = __p0; \
  21551.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  21552.   __ret = (uint32x2_t) __builtin_neon_vqshrun_n_v((int8x16_t)__rev0, __p1, 18); \
  21553.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  21554.   __ret; \
  21555. })
  21556. #define __noswap_vqshrun_n_s64(__p0, __p1) __extension__ ({ \
  21557.   uint32x2_t __ret; \
  21558.   int64x2_t __s0 = __p0; \
  21559.   __ret = (uint32x2_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 18); \
  21560.   __ret; \
  21561. })
  21562. #endif
  21563.  
  21564. #ifdef __LITTLE_ENDIAN__
  21565. #define vqshrun_n_s16(__p0, __p1) __extension__ ({ \
  21566.   uint8x8_t __ret; \
  21567.   int16x8_t __s0 = __p0; \
  21568.   __ret = (uint8x8_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 16); \
  21569.   __ret; \
  21570. })
  21571. #else
  21572. #define vqshrun_n_s16(__p0, __p1) __extension__ ({ \
  21573.   uint8x8_t __ret; \
  21574.   int16x8_t __s0 = __p0; \
  21575.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  21576.   __ret = (uint8x8_t) __builtin_neon_vqshrun_n_v((int8x16_t)__rev0, __p1, 16); \
  21577.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  21578.   __ret; \
  21579. })
  21580. #define __noswap_vqshrun_n_s16(__p0, __p1) __extension__ ({ \
  21581.   uint8x8_t __ret; \
  21582.   int16x8_t __s0 = __p0; \
  21583.   __ret = (uint8x8_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 16); \
  21584.   __ret; \
  21585. })
  21586. #endif
  21587.  
  21588. #ifdef __LITTLE_ENDIAN__
  21589. __ai uint8x16_t vqsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  21590.   uint8x16_t __ret;
  21591.   __ret = (uint8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  21592.   return __ret;
  21593. }
  21594. #else
  21595. __ai uint8x16_t vqsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  21596.   uint8x16_t __ret;
  21597.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  21598.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  21599.   __ret = (uint8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  21600.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  21601.   return __ret;
  21602. }
  21603. #endif
  21604.  
  21605. #ifdef __LITTLE_ENDIAN__
  21606. __ai uint32x4_t vqsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  21607.   uint32x4_t __ret;
  21608.   __ret = (uint32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  21609.   return __ret;
  21610. }
  21611. #else
  21612. __ai uint32x4_t vqsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  21613.   uint32x4_t __ret;
  21614.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  21615.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  21616.   __ret = (uint32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  21617.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  21618.   return __ret;
  21619. }
  21620. #endif
  21621.  
  21622. #ifdef __LITTLE_ENDIAN__
  21623. __ai uint64x2_t vqsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  21624.   uint64x2_t __ret;
  21625.   __ret = (uint64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  21626.   return __ret;
  21627. }
  21628. #else
  21629. __ai uint64x2_t vqsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  21630.   uint64x2_t __ret;
  21631.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  21632.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  21633.   __ret = (uint64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  21634.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  21635.   return __ret;
  21636. }
  21637. #endif
  21638.  
  21639. #ifdef __LITTLE_ENDIAN__
  21640. __ai uint16x8_t vqsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  21641.   uint16x8_t __ret;
  21642.   __ret = (uint16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  21643.   return __ret;
  21644. }
  21645. #else
  21646. __ai uint16x8_t vqsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  21647.   uint16x8_t __ret;
  21648.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  21649.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  21650.   __ret = (uint16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  21651.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  21652.   return __ret;
  21653. }
  21654. #endif
  21655.  
  21656. #ifdef __LITTLE_ENDIAN__
  21657. __ai int8x16_t vqsubq_s8(int8x16_t __p0, int8x16_t __p1) {
  21658.   int8x16_t __ret;
  21659.   __ret = (int8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  21660.   return __ret;
  21661. }
  21662. #else
  21663. __ai int8x16_t vqsubq_s8(int8x16_t __p0, int8x16_t __p1) {
  21664.   int8x16_t __ret;
  21665.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  21666.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  21667.   __ret = (int8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  21668.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  21669.   return __ret;
  21670. }
  21671. #endif
  21672.  
  21673. #ifdef __LITTLE_ENDIAN__
  21674. __ai int32x4_t vqsubq_s32(int32x4_t __p0, int32x4_t __p1) {
  21675.   int32x4_t __ret;
  21676.   __ret = (int32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  21677.   return __ret;
  21678. }
  21679. #else
  21680. __ai int32x4_t vqsubq_s32(int32x4_t __p0, int32x4_t __p1) {
  21681.   int32x4_t __ret;
  21682.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  21683.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  21684.   __ret = (int32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  21685.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  21686.   return __ret;
  21687. }
  21688. #endif
  21689.  
  21690. #ifdef __LITTLE_ENDIAN__
  21691. __ai int64x2_t vqsubq_s64(int64x2_t __p0, int64x2_t __p1) {
  21692.   int64x2_t __ret;
  21693.   __ret = (int64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
  21694.   return __ret;
  21695. }
  21696. #else
  21697. __ai int64x2_t vqsubq_s64(int64x2_t __p0, int64x2_t __p1) {
  21698.   int64x2_t __ret;
  21699.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  21700.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  21701.   __ret = (int64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
  21702.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  21703.   return __ret;
  21704. }
  21705. #endif
  21706.  
  21707. #ifdef __LITTLE_ENDIAN__
  21708. __ai int16x8_t vqsubq_s16(int16x8_t __p0, int16x8_t __p1) {
  21709.   int16x8_t __ret;
  21710.   __ret = (int16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  21711.   return __ret;
  21712. }
  21713. #else
  21714. __ai int16x8_t vqsubq_s16(int16x8_t __p0, int16x8_t __p1) {
  21715.   int16x8_t __ret;
  21716.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  21717.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  21718.   __ret = (int16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  21719.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  21720.   return __ret;
  21721. }
  21722. #endif
  21723.  
  21724. #ifdef __LITTLE_ENDIAN__
  21725. __ai uint8x8_t vqsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
  21726.   uint8x8_t __ret;
  21727.   __ret = (uint8x8_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  21728.   return __ret;
  21729. }
  21730. #else
  21731. __ai uint8x8_t vqsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
  21732.   uint8x8_t __ret;
  21733.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  21734.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  21735.   __ret = (uint8x8_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  21736.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  21737.   return __ret;
  21738. }
  21739. #endif
  21740.  
  21741. #ifdef __LITTLE_ENDIAN__
  21742. __ai uint32x2_t vqsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
  21743.   uint32x2_t __ret;
  21744.   __ret = (uint32x2_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  21745.   return __ret;
  21746. }
  21747. #else
  21748. __ai uint32x2_t vqsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
  21749.   uint32x2_t __ret;
  21750.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  21751.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  21752.   __ret = (uint32x2_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  21753.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  21754.   return __ret;
  21755. }
  21756. #endif
  21757.  
  21758. __ai uint64x1_t vqsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
  21759.   uint64x1_t __ret;
  21760.   __ret = (uint64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  21761.   return __ret;
  21762. }
  21763. #ifdef __LITTLE_ENDIAN__
  21764. __ai uint16x4_t vqsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
  21765.   uint16x4_t __ret;
  21766.   __ret = (uint16x4_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  21767.   return __ret;
  21768. }
  21769. #else
  21770. __ai uint16x4_t vqsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
  21771.   uint16x4_t __ret;
  21772.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  21773.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  21774.   __ret = (uint16x4_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  21775.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  21776.   return __ret;
  21777. }
  21778. #endif
  21779.  
  21780. #ifdef __LITTLE_ENDIAN__
  21781. __ai int8x8_t vqsub_s8(int8x8_t __p0, int8x8_t __p1) {
  21782.   int8x8_t __ret;
  21783.   __ret = (int8x8_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  21784.   return __ret;
  21785. }
  21786. #else
  21787. __ai int8x8_t vqsub_s8(int8x8_t __p0, int8x8_t __p1) {
  21788.   int8x8_t __ret;
  21789.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  21790.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  21791.   __ret = (int8x8_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  21792.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  21793.   return __ret;
  21794. }
  21795. #endif
  21796.  
  21797. #ifdef __LITTLE_ENDIAN__
  21798. __ai int32x2_t vqsub_s32(int32x2_t __p0, int32x2_t __p1) {
  21799.   int32x2_t __ret;
  21800.   __ret = (int32x2_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  21801.   return __ret;
  21802. }
  21803. #else
  21804. __ai int32x2_t vqsub_s32(int32x2_t __p0, int32x2_t __p1) {
  21805.   int32x2_t __ret;
  21806.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  21807.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  21808.   __ret = (int32x2_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  21809.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  21810.   return __ret;
  21811. }
  21812. #endif
  21813.  
  21814. __ai int64x1_t vqsub_s64(int64x1_t __p0, int64x1_t __p1) {
  21815.   int64x1_t __ret;
  21816.   __ret = (int64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
  21817.   return __ret;
  21818. }
  21819. #ifdef __LITTLE_ENDIAN__
  21820. __ai int16x4_t vqsub_s16(int16x4_t __p0, int16x4_t __p1) {
  21821.   int16x4_t __ret;
  21822.   __ret = (int16x4_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  21823.   return __ret;
  21824. }
  21825. #else
  21826. __ai int16x4_t vqsub_s16(int16x4_t __p0, int16x4_t __p1) {
  21827.   int16x4_t __ret;
  21828.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  21829.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  21830.   __ret = (int16x4_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  21831.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  21832.   return __ret;
  21833. }
  21834. #endif
  21835.  
  21836. #ifdef __LITTLE_ENDIAN__
  21837. __ai uint16x4_t vraddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  21838.   uint16x4_t __ret;
  21839.   __ret = (uint16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  21840.   return __ret;
  21841. }
  21842. #else
  21843. __ai uint16x4_t vraddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  21844.   uint16x4_t __ret;
  21845.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  21846.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  21847.   __ret = (uint16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
  21848.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  21849.   return __ret;
  21850. }
  21851. __ai uint16x4_t __noswap_vraddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  21852.   uint16x4_t __ret;
  21853.   __ret = (uint16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  21854.   return __ret;
  21855. }
  21856. #endif
  21857.  
  21858. #ifdef __LITTLE_ENDIAN__
  21859. __ai uint32x2_t vraddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  21860.   uint32x2_t __ret;
  21861.   __ret = (uint32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  21862.   return __ret;
  21863. }
  21864. #else
  21865. __ai uint32x2_t vraddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  21866.   uint32x2_t __ret;
  21867.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  21868.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  21869.   __ret = (uint32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
  21870.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  21871.   return __ret;
  21872. }
  21873. __ai uint32x2_t __noswap_vraddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  21874.   uint32x2_t __ret;
  21875.   __ret = (uint32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  21876.   return __ret;
  21877. }
  21878. #endif
  21879.  
  21880. #ifdef __LITTLE_ENDIAN__
  21881. __ai uint8x8_t vraddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  21882.   uint8x8_t __ret;
  21883.   __ret = (uint8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  21884.   return __ret;
  21885. }
  21886. #else
  21887. __ai uint8x8_t vraddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  21888.   uint8x8_t __ret;
  21889.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  21890.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  21891.   __ret = (uint8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
  21892.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  21893.   return __ret;
  21894. }
  21895. __ai uint8x8_t __noswap_vraddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  21896.   uint8x8_t __ret;
  21897.   __ret = (uint8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  21898.   return __ret;
  21899. }
  21900. #endif
  21901.  
  21902. #ifdef __LITTLE_ENDIAN__
  21903. __ai int16x4_t vraddhn_s32(int32x4_t __p0, int32x4_t __p1) {
  21904.   int16x4_t __ret;
  21905.   __ret = (int16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  21906.   return __ret;
  21907. }
  21908. #else
  21909. __ai int16x4_t vraddhn_s32(int32x4_t __p0, int32x4_t __p1) {
  21910.   int16x4_t __ret;
  21911.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  21912.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  21913.   __ret = (int16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
  21914.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  21915.   return __ret;
  21916. }
  21917. __ai int16x4_t __noswap_vraddhn_s32(int32x4_t __p0, int32x4_t __p1) {
  21918.   int16x4_t __ret;
  21919.   __ret = (int16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  21920.   return __ret;
  21921. }
  21922. #endif
  21923.  
  21924. #ifdef __LITTLE_ENDIAN__
  21925. __ai int32x2_t vraddhn_s64(int64x2_t __p0, int64x2_t __p1) {
  21926.   int32x2_t __ret;
  21927.   __ret = (int32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  21928.   return __ret;
  21929. }
  21930. #else
  21931. __ai int32x2_t vraddhn_s64(int64x2_t __p0, int64x2_t __p1) {
  21932.   int32x2_t __ret;
  21933.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  21934.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  21935.   __ret = (int32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
  21936.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  21937.   return __ret;
  21938. }
  21939. __ai int32x2_t __noswap_vraddhn_s64(int64x2_t __p0, int64x2_t __p1) {
  21940.   int32x2_t __ret;
  21941.   __ret = (int32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  21942.   return __ret;
  21943. }
  21944. #endif
  21945.  
  21946. #ifdef __LITTLE_ENDIAN__
  21947. __ai int8x8_t vraddhn_s16(int16x8_t __p0, int16x8_t __p1) {
  21948.   int8x8_t __ret;
  21949.   __ret = (int8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  21950.   return __ret;
  21951. }
  21952. #else
  21953. __ai int8x8_t vraddhn_s16(int16x8_t __p0, int16x8_t __p1) {
  21954.   int8x8_t __ret;
  21955.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  21956.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  21957.   __ret = (int8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
  21958.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  21959.   return __ret;
  21960. }
  21961. __ai int8x8_t __noswap_vraddhn_s16(int16x8_t __p0, int16x8_t __p1) {
  21962.   int8x8_t __ret;
  21963.   __ret = (int8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  21964.   return __ret;
  21965. }
  21966. #endif
  21967.  
  21968. #ifdef __LITTLE_ENDIAN__
  21969. __ai uint32x4_t vrecpeq_u32(uint32x4_t __p0) {
  21970.   uint32x4_t __ret;
  21971.   __ret = (uint32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 50);
  21972.   return __ret;
  21973. }
  21974. #else
  21975. __ai uint32x4_t vrecpeq_u32(uint32x4_t __p0) {
  21976.   uint32x4_t __ret;
  21977.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  21978.   __ret = (uint32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 50);
  21979.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  21980.   return __ret;
  21981. }
  21982. #endif
  21983.  
  21984. #ifdef __LITTLE_ENDIAN__
  21985. __ai float32x4_t vrecpeq_f32(float32x4_t __p0) {
  21986.   float32x4_t __ret;
  21987.   __ret = (float32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 41);
  21988.   return __ret;
  21989. }
  21990. #else
  21991. __ai float32x4_t vrecpeq_f32(float32x4_t __p0) {
  21992.   float32x4_t __ret;
  21993.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  21994.   __ret = (float32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 41);
  21995.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  21996.   return __ret;
  21997. }
  21998. #endif
  21999.  
  22000. #ifdef __LITTLE_ENDIAN__
  22001. __ai uint32x2_t vrecpe_u32(uint32x2_t __p0) {
  22002.   uint32x2_t __ret;
  22003.   __ret = (uint32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 18);
  22004.   return __ret;
  22005. }
  22006. #else
  22007. __ai uint32x2_t vrecpe_u32(uint32x2_t __p0) {
  22008.   uint32x2_t __ret;
  22009.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22010.   __ret = (uint32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__rev0, 18);
  22011.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22012.   return __ret;
  22013. }
  22014. #endif
  22015.  
  22016. #ifdef __LITTLE_ENDIAN__
  22017. __ai float32x2_t vrecpe_f32(float32x2_t __p0) {
  22018.   float32x2_t __ret;
  22019.   __ret = (float32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 9);
  22020.   return __ret;
  22021. }
  22022. #else
  22023. __ai float32x2_t vrecpe_f32(float32x2_t __p0) {
  22024.   float32x2_t __ret;
  22025.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22026.   __ret = (float32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__rev0, 9);
  22027.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22028.   return __ret;
  22029. }
  22030. #endif
  22031.  
  22032. #ifdef __LITTLE_ENDIAN__
  22033. __ai float32x4_t vrecpsq_f32(float32x4_t __p0, float32x4_t __p1) {
  22034.   float32x4_t __ret;
  22035.   __ret = (float32x4_t) __builtin_neon_vrecpsq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  22036.   return __ret;
  22037. }
  22038. #else
  22039. __ai float32x4_t vrecpsq_f32(float32x4_t __p0, float32x4_t __p1) {
  22040.   float32x4_t __ret;
  22041.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22042.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  22043.   __ret = (float32x4_t) __builtin_neon_vrecpsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  22044.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22045.   return __ret;
  22046. }
  22047. #endif
  22048.  
  22049. #ifdef __LITTLE_ENDIAN__
  22050. __ai float32x2_t vrecps_f32(float32x2_t __p0, float32x2_t __p1) {
  22051.   float32x2_t __ret;
  22052.   __ret = (float32x2_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  22053.   return __ret;
  22054. }
  22055. #else
  22056. __ai float32x2_t vrecps_f32(float32x2_t __p0, float32x2_t __p1) {
  22057.   float32x2_t __ret;
  22058.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22059.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  22060.   __ret = (float32x2_t) __builtin_neon_vrecps_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  22061.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22062.   return __ret;
  22063. }
  22064. #endif
  22065.  
  22066. #ifdef __LITTLE_ENDIAN__
  22067. __ai poly8x8_t vrev16_p8(poly8x8_t __p0) {
  22068.   poly8x8_t __ret;
  22069.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
  22070.   return __ret;
  22071. }
  22072. #else
  22073. __ai poly8x8_t vrev16_p8(poly8x8_t __p0) {
  22074.   poly8x8_t __ret;
  22075.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22076.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
  22077.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22078.   return __ret;
  22079. }
  22080. #endif
  22081.  
  22082. #ifdef __LITTLE_ENDIAN__
  22083. __ai poly8x16_t vrev16q_p8(poly8x16_t __p0) {
  22084.   poly8x16_t __ret;
  22085.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
  22086.   return __ret;
  22087. }
  22088. #else
  22089. __ai poly8x16_t vrev16q_p8(poly8x16_t __p0) {
  22090.   poly8x16_t __ret;
  22091.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22092.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
  22093.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22094.   return __ret;
  22095. }
  22096. #endif
  22097.  
  22098. #ifdef __LITTLE_ENDIAN__
  22099. __ai uint8x16_t vrev16q_u8(uint8x16_t __p0) {
  22100.   uint8x16_t __ret;
  22101.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
  22102.   return __ret;
  22103. }
  22104. #else
  22105. __ai uint8x16_t vrev16q_u8(uint8x16_t __p0) {
  22106.   uint8x16_t __ret;
  22107.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22108.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
  22109.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22110.   return __ret;
  22111. }
  22112. #endif
  22113.  
  22114. #ifdef __LITTLE_ENDIAN__
  22115. __ai int8x16_t vrev16q_s8(int8x16_t __p0) {
  22116.   int8x16_t __ret;
  22117.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
  22118.   return __ret;
  22119. }
  22120. #else
  22121. __ai int8x16_t vrev16q_s8(int8x16_t __p0) {
  22122.   int8x16_t __ret;
  22123.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22124.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
  22125.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22126.   return __ret;
  22127. }
  22128. #endif
  22129.  
  22130. #ifdef __LITTLE_ENDIAN__
  22131. __ai uint8x8_t vrev16_u8(uint8x8_t __p0) {
  22132.   uint8x8_t __ret;
  22133.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
  22134.   return __ret;
  22135. }
  22136. #else
  22137. __ai uint8x8_t vrev16_u8(uint8x8_t __p0) {
  22138.   uint8x8_t __ret;
  22139.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22140.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
  22141.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22142.   return __ret;
  22143. }
  22144. #endif
  22145.  
  22146. #ifdef __LITTLE_ENDIAN__
  22147. __ai int8x8_t vrev16_s8(int8x8_t __p0) {
  22148.   int8x8_t __ret;
  22149.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
  22150.   return __ret;
  22151. }
  22152. #else
  22153. __ai int8x8_t vrev16_s8(int8x8_t __p0) {
  22154.   int8x8_t __ret;
  22155.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22156.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
  22157.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22158.   return __ret;
  22159. }
  22160. #endif
  22161.  
  22162. #ifdef __LITTLE_ENDIAN__
  22163. __ai poly8x8_t vrev32_p8(poly8x8_t __p0) {
  22164.   poly8x8_t __ret;
  22165.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
  22166.   return __ret;
  22167. }
  22168. #else
  22169. __ai poly8x8_t vrev32_p8(poly8x8_t __p0) {
  22170.   poly8x8_t __ret;
  22171.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22172.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
  22173.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22174.   return __ret;
  22175. }
  22176. #endif
  22177.  
  22178. #ifdef __LITTLE_ENDIAN__
  22179. __ai poly16x4_t vrev32_p16(poly16x4_t __p0) {
  22180.   poly16x4_t __ret;
  22181.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
  22182.   return __ret;
  22183. }
  22184. #else
  22185. __ai poly16x4_t vrev32_p16(poly16x4_t __p0) {
  22186.   poly16x4_t __ret;
  22187.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22188.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
  22189.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22190.   return __ret;
  22191. }
  22192. #endif
  22193.  
  22194. #ifdef __LITTLE_ENDIAN__
  22195. __ai poly8x16_t vrev32q_p8(poly8x16_t __p0) {
  22196.   poly8x16_t __ret;
  22197.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
  22198.   return __ret;
  22199. }
  22200. #else
  22201. __ai poly8x16_t vrev32q_p8(poly8x16_t __p0) {
  22202.   poly8x16_t __ret;
  22203.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22204.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
  22205.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22206.   return __ret;
  22207. }
  22208. #endif
  22209.  
  22210. #ifdef __LITTLE_ENDIAN__
  22211. __ai poly16x8_t vrev32q_p16(poly16x8_t __p0) {
  22212.   poly16x8_t __ret;
  22213.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
  22214.   return __ret;
  22215. }
  22216. #else
  22217. __ai poly16x8_t vrev32q_p16(poly16x8_t __p0) {
  22218.   poly16x8_t __ret;
  22219.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22220.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
  22221.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22222.   return __ret;
  22223. }
  22224. #endif
  22225.  
  22226. #ifdef __LITTLE_ENDIAN__
  22227. __ai uint8x16_t vrev32q_u8(uint8x16_t __p0) {
  22228.   uint8x16_t __ret;
  22229.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
  22230.   return __ret;
  22231. }
  22232. #else
  22233. __ai uint8x16_t vrev32q_u8(uint8x16_t __p0) {
  22234.   uint8x16_t __ret;
  22235.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22236.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
  22237.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22238.   return __ret;
  22239. }
  22240. #endif
  22241.  
  22242. #ifdef __LITTLE_ENDIAN__
  22243. __ai uint16x8_t vrev32q_u16(uint16x8_t __p0) {
  22244.   uint16x8_t __ret;
  22245.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
  22246.   return __ret;
  22247. }
  22248. #else
  22249. __ai uint16x8_t vrev32q_u16(uint16x8_t __p0) {
  22250.   uint16x8_t __ret;
  22251.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22252.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
  22253.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22254.   return __ret;
  22255. }
  22256. #endif
  22257.  
  22258. #ifdef __LITTLE_ENDIAN__
  22259. __ai int8x16_t vrev32q_s8(int8x16_t __p0) {
  22260.   int8x16_t __ret;
  22261.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
  22262.   return __ret;
  22263. }
  22264. #else
  22265. __ai int8x16_t vrev32q_s8(int8x16_t __p0) {
  22266.   int8x16_t __ret;
  22267.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22268.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
  22269.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22270.   return __ret;
  22271. }
  22272. #endif
  22273.  
  22274. #ifdef __LITTLE_ENDIAN__
  22275. __ai int16x8_t vrev32q_s16(int16x8_t __p0) {
  22276.   int16x8_t __ret;
  22277.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
  22278.   return __ret;
  22279. }
  22280. #else
  22281. __ai int16x8_t vrev32q_s16(int16x8_t __p0) {
  22282.   int16x8_t __ret;
  22283.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22284.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
  22285.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22286.   return __ret;
  22287. }
  22288. #endif
  22289.  
  22290. #ifdef __LITTLE_ENDIAN__
  22291. __ai uint8x8_t vrev32_u8(uint8x8_t __p0) {
  22292.   uint8x8_t __ret;
  22293.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
  22294.   return __ret;
  22295. }
  22296. #else
  22297. __ai uint8x8_t vrev32_u8(uint8x8_t __p0) {
  22298.   uint8x8_t __ret;
  22299.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22300.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
  22301.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22302.   return __ret;
  22303. }
  22304. #endif
  22305.  
  22306. #ifdef __LITTLE_ENDIAN__
  22307. __ai uint16x4_t vrev32_u16(uint16x4_t __p0) {
  22308.   uint16x4_t __ret;
  22309.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
  22310.   return __ret;
  22311. }
  22312. #else
  22313. __ai uint16x4_t vrev32_u16(uint16x4_t __p0) {
  22314.   uint16x4_t __ret;
  22315.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22316.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
  22317.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22318.   return __ret;
  22319. }
  22320. #endif
  22321.  
  22322. #ifdef __LITTLE_ENDIAN__
  22323. __ai int8x8_t vrev32_s8(int8x8_t __p0) {
  22324.   int8x8_t __ret;
  22325.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
  22326.   return __ret;
  22327. }
  22328. #else
  22329. __ai int8x8_t vrev32_s8(int8x8_t __p0) {
  22330.   int8x8_t __ret;
  22331.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22332.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
  22333.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22334.   return __ret;
  22335. }
  22336. #endif
  22337.  
  22338. #ifdef __LITTLE_ENDIAN__
  22339. __ai int16x4_t vrev32_s16(int16x4_t __p0) {
  22340.   int16x4_t __ret;
  22341.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
  22342.   return __ret;
  22343. }
  22344. #else
  22345. __ai int16x4_t vrev32_s16(int16x4_t __p0) {
  22346.   int16x4_t __ret;
  22347.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22348.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
  22349.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22350.   return __ret;
  22351. }
  22352. #endif
  22353.  
  22354. #ifdef __LITTLE_ENDIAN__
  22355. __ai poly8x8_t vrev64_p8(poly8x8_t __p0) {
  22356.   poly8x8_t __ret;
  22357.   __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22358.   return __ret;
  22359. }
  22360. #else
  22361. __ai poly8x8_t vrev64_p8(poly8x8_t __p0) {
  22362.   poly8x8_t __ret;
  22363.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22364.   __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0);
  22365.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22366.   return __ret;
  22367. }
  22368. #endif
  22369.  
  22370. #ifdef __LITTLE_ENDIAN__
  22371. __ai poly16x4_t vrev64_p16(poly16x4_t __p0) {
  22372.   poly16x4_t __ret;
  22373.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22374.   return __ret;
  22375. }
  22376. #else
  22377. __ai poly16x4_t vrev64_p16(poly16x4_t __p0) {
  22378.   poly16x4_t __ret;
  22379.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22380.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
  22381.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22382.   return __ret;
  22383. }
  22384. #endif
  22385.  
  22386. #ifdef __LITTLE_ENDIAN__
  22387. __ai poly8x16_t vrev64q_p8(poly8x16_t __p0) {
  22388.   poly8x16_t __ret;
  22389.   __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
  22390.   return __ret;
  22391. }
  22392. #else
  22393. __ai poly8x16_t vrev64q_p8(poly8x16_t __p0) {
  22394.   poly8x16_t __ret;
  22395.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22396.   __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
  22397.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22398.   return __ret;
  22399. }
  22400. #endif
  22401.  
  22402. #ifdef __LITTLE_ENDIAN__
  22403. __ai poly16x8_t vrev64q_p16(poly16x8_t __p0) {
  22404.   poly16x8_t __ret;
  22405.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
  22406.   return __ret;
  22407. }
  22408. #else
  22409. __ai poly16x8_t vrev64q_p16(poly16x8_t __p0) {
  22410.   poly16x8_t __ret;
  22411.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22412.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
  22413.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22414.   return __ret;
  22415. }
  22416. #endif
  22417.  
  22418. #ifdef __LITTLE_ENDIAN__
  22419. __ai uint8x16_t vrev64q_u8(uint8x16_t __p0) {
  22420.   uint8x16_t __ret;
  22421.   __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
  22422.   return __ret;
  22423. }
  22424. #else
  22425. __ai uint8x16_t vrev64q_u8(uint8x16_t __p0) {
  22426.   uint8x16_t __ret;
  22427.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22428.   __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
  22429.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22430.   return __ret;
  22431. }
  22432. #endif
  22433.  
  22434. #ifdef __LITTLE_ENDIAN__
  22435. __ai uint32x4_t vrev64q_u32(uint32x4_t __p0) {
  22436.   uint32x4_t __ret;
  22437.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
  22438.   return __ret;
  22439. }
  22440. #else
  22441. __ai uint32x4_t vrev64q_u32(uint32x4_t __p0) {
  22442.   uint32x4_t __ret;
  22443.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22444.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
  22445.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22446.   return __ret;
  22447. }
  22448. #endif
  22449.  
  22450. #ifdef __LITTLE_ENDIAN__
  22451. __ai uint16x8_t vrev64q_u16(uint16x8_t __p0) {
  22452.   uint16x8_t __ret;
  22453.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
  22454.   return __ret;
  22455. }
  22456. #else
  22457. __ai uint16x8_t vrev64q_u16(uint16x8_t __p0) {
  22458.   uint16x8_t __ret;
  22459.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22460.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
  22461.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22462.   return __ret;
  22463. }
  22464. #endif
  22465.  
  22466. #ifdef __LITTLE_ENDIAN__
  22467. __ai int8x16_t vrev64q_s8(int8x16_t __p0) {
  22468.   int8x16_t __ret;
  22469.   __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
  22470.   return __ret;
  22471. }
  22472. #else
  22473. __ai int8x16_t vrev64q_s8(int8x16_t __p0) {
  22474.   int8x16_t __ret;
  22475.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22476.   __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
  22477.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22478.   return __ret;
  22479. }
  22480. #endif
  22481.  
  22482. #ifdef __LITTLE_ENDIAN__
  22483. __ai float32x4_t vrev64q_f32(float32x4_t __p0) {
  22484.   float32x4_t __ret;
  22485.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
  22486.   return __ret;
  22487. }
  22488. #else
  22489. __ai float32x4_t vrev64q_f32(float32x4_t __p0) {
  22490.   float32x4_t __ret;
  22491.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22492.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
  22493.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22494.   return __ret;
  22495. }
  22496. #endif
  22497.  
  22498. #ifdef __LITTLE_ENDIAN__
  22499. __ai int32x4_t vrev64q_s32(int32x4_t __p0) {
  22500.   int32x4_t __ret;
  22501.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
  22502.   return __ret;
  22503. }
  22504. #else
  22505. __ai int32x4_t vrev64q_s32(int32x4_t __p0) {
  22506.   int32x4_t __ret;
  22507.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22508.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
  22509.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22510.   return __ret;
  22511. }
  22512. #endif
  22513.  
  22514. #ifdef __LITTLE_ENDIAN__
  22515. __ai int16x8_t vrev64q_s16(int16x8_t __p0) {
  22516.   int16x8_t __ret;
  22517.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
  22518.   return __ret;
  22519. }
  22520. #else
  22521. __ai int16x8_t vrev64q_s16(int16x8_t __p0) {
  22522.   int16x8_t __ret;
  22523.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22524.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
  22525.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22526.   return __ret;
  22527. }
  22528. #endif
  22529.  
  22530. #ifdef __LITTLE_ENDIAN__
  22531. __ai uint8x8_t vrev64_u8(uint8x8_t __p0) {
  22532.   uint8x8_t __ret;
  22533.   __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22534.   return __ret;
  22535. }
  22536. #else
  22537. __ai uint8x8_t vrev64_u8(uint8x8_t __p0) {
  22538.   uint8x8_t __ret;
  22539.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22540.   __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0);
  22541.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22542.   return __ret;
  22543. }
  22544. #endif
  22545.  
  22546. #ifdef __LITTLE_ENDIAN__
  22547. __ai uint32x2_t vrev64_u32(uint32x2_t __p0) {
  22548.   uint32x2_t __ret;
  22549.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0);
  22550.   return __ret;
  22551. }
  22552. #else
  22553. __ai uint32x2_t vrev64_u32(uint32x2_t __p0) {
  22554.   uint32x2_t __ret;
  22555.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22556.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0);
  22557.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22558.   return __ret;
  22559. }
  22560. #endif
  22561.  
  22562. #ifdef __LITTLE_ENDIAN__
  22563. __ai uint16x4_t vrev64_u16(uint16x4_t __p0) {
  22564.   uint16x4_t __ret;
  22565.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22566.   return __ret;
  22567. }
  22568. #else
  22569. __ai uint16x4_t vrev64_u16(uint16x4_t __p0) {
  22570.   uint16x4_t __ret;
  22571.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22572.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
  22573.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22574.   return __ret;
  22575. }
  22576. #endif
  22577.  
  22578. #ifdef __LITTLE_ENDIAN__
  22579. __ai int8x8_t vrev64_s8(int8x8_t __p0) {
  22580.   int8x8_t __ret;
  22581.   __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22582.   return __ret;
  22583. }
  22584. #else
  22585. __ai int8x8_t vrev64_s8(int8x8_t __p0) {
  22586.   int8x8_t __ret;
  22587.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22588.   __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0);
  22589.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22590.   return __ret;
  22591. }
  22592. #endif
  22593.  
  22594. #ifdef __LITTLE_ENDIAN__
  22595. __ai float32x2_t vrev64_f32(float32x2_t __p0) {
  22596.   float32x2_t __ret;
  22597.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0);
  22598.   return __ret;
  22599. }
  22600. #else
  22601. __ai float32x2_t vrev64_f32(float32x2_t __p0) {
  22602.   float32x2_t __ret;
  22603.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22604.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0);
  22605.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22606.   return __ret;
  22607. }
  22608. #endif
  22609.  
  22610. #ifdef __LITTLE_ENDIAN__
  22611. __ai int32x2_t vrev64_s32(int32x2_t __p0) {
  22612.   int32x2_t __ret;
  22613.   __ret = __builtin_shufflevector(__p0, __p0, 1, 0);
  22614.   return __ret;
  22615. }
  22616. #else
  22617. __ai int32x2_t vrev64_s32(int32x2_t __p0) {
  22618.   int32x2_t __ret;
  22619.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22620.   __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0);
  22621.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22622.   return __ret;
  22623. }
  22624. #endif
  22625.  
  22626. #ifdef __LITTLE_ENDIAN__
  22627. __ai int16x4_t vrev64_s16(int16x4_t __p0) {
  22628.   int16x4_t __ret;
  22629.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22630.   return __ret;
  22631. }
  22632. #else
  22633. __ai int16x4_t vrev64_s16(int16x4_t __p0) {
  22634.   int16x4_t __ret;
  22635.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22636.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
  22637.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22638.   return __ret;
  22639. }
  22640. #endif
  22641.  
  22642. #ifdef __LITTLE_ENDIAN__
  22643. __ai uint8x16_t vrhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  22644.   uint8x16_t __ret;
  22645.   __ret = (uint8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  22646.   return __ret;
  22647. }
  22648. #else
  22649. __ai uint8x16_t vrhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  22650.   uint8x16_t __ret;
  22651.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22652.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22653.   __ret = (uint8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  22654.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22655.   return __ret;
  22656. }
  22657. #endif
  22658.  
  22659. #ifdef __LITTLE_ENDIAN__
  22660. __ai uint32x4_t vrhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  22661.   uint32x4_t __ret;
  22662.   __ret = (uint32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  22663.   return __ret;
  22664. }
  22665. #else
  22666. __ai uint32x4_t vrhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  22667.   uint32x4_t __ret;
  22668.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22669.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  22670.   __ret = (uint32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  22671.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22672.   return __ret;
  22673. }
  22674. #endif
  22675.  
  22676. #ifdef __LITTLE_ENDIAN__
  22677. __ai uint16x8_t vrhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  22678.   uint16x8_t __ret;
  22679.   __ret = (uint16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  22680.   return __ret;
  22681. }
  22682. #else
  22683. __ai uint16x8_t vrhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  22684.   uint16x8_t __ret;
  22685.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22686.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  22687.   __ret = (uint16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  22688.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22689.   return __ret;
  22690. }
  22691. #endif
  22692.  
  22693. #ifdef __LITTLE_ENDIAN__
  22694. __ai int8x16_t vrhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  22695.   int8x16_t __ret;
  22696.   __ret = (int8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  22697.   return __ret;
  22698. }
  22699. #else
  22700. __ai int8x16_t vrhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  22701.   int8x16_t __ret;
  22702.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22703.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22704.   __ret = (int8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  22705.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22706.   return __ret;
  22707. }
  22708. #endif
  22709.  
  22710. #ifdef __LITTLE_ENDIAN__
  22711. __ai int32x4_t vrhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  22712.   int32x4_t __ret;
  22713.   __ret = (int32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  22714.   return __ret;
  22715. }
  22716. #else
  22717. __ai int32x4_t vrhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  22718.   int32x4_t __ret;
  22719.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22720.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  22721.   __ret = (int32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  22722.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22723.   return __ret;
  22724. }
  22725. #endif
  22726.  
  22727. #ifdef __LITTLE_ENDIAN__
  22728. __ai int16x8_t vrhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  22729.   int16x8_t __ret;
  22730.   __ret = (int16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  22731.   return __ret;
  22732. }
  22733. #else
  22734. __ai int16x8_t vrhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  22735.   int16x8_t __ret;
  22736.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22737.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  22738.   __ret = (int16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  22739.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22740.   return __ret;
  22741. }
  22742. #endif
  22743.  
  22744. #ifdef __LITTLE_ENDIAN__
  22745. __ai uint8x8_t vrhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  22746.   uint8x8_t __ret;
  22747.   __ret = (uint8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  22748.   return __ret;
  22749. }
  22750. #else
  22751. __ai uint8x8_t vrhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
  22752.   uint8x8_t __ret;
  22753.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22754.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  22755.   __ret = (uint8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  22756.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22757.   return __ret;
  22758. }
  22759. #endif
  22760.  
  22761. #ifdef __LITTLE_ENDIAN__
  22762. __ai uint32x2_t vrhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  22763.   uint32x2_t __ret;
  22764.   __ret = (uint32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  22765.   return __ret;
  22766. }
  22767. #else
  22768. __ai uint32x2_t vrhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
  22769.   uint32x2_t __ret;
  22770.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22771.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  22772.   __ret = (uint32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  22773.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22774.   return __ret;
  22775. }
  22776. #endif
  22777.  
  22778. #ifdef __LITTLE_ENDIAN__
  22779. __ai uint16x4_t vrhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  22780.   uint16x4_t __ret;
  22781.   __ret = (uint16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  22782.   return __ret;
  22783. }
  22784. #else
  22785. __ai uint16x4_t vrhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
  22786.   uint16x4_t __ret;
  22787.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22788.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  22789.   __ret = (uint16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  22790.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22791.   return __ret;
  22792. }
  22793. #endif
  22794.  
  22795. #ifdef __LITTLE_ENDIAN__
  22796. __ai int8x8_t vrhadd_s8(int8x8_t __p0, int8x8_t __p1) {
  22797.   int8x8_t __ret;
  22798.   __ret = (int8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  22799.   return __ret;
  22800. }
  22801. #else
  22802. __ai int8x8_t vrhadd_s8(int8x8_t __p0, int8x8_t __p1) {
  22803.   int8x8_t __ret;
  22804.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22805.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  22806.   __ret = (int8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  22807.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22808.   return __ret;
  22809. }
  22810. #endif
  22811.  
  22812. #ifdef __LITTLE_ENDIAN__
  22813. __ai int32x2_t vrhadd_s32(int32x2_t __p0, int32x2_t __p1) {
  22814.   int32x2_t __ret;
  22815.   __ret = (int32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  22816.   return __ret;
  22817. }
  22818. #else
  22819. __ai int32x2_t vrhadd_s32(int32x2_t __p0, int32x2_t __p1) {
  22820.   int32x2_t __ret;
  22821.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22822.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  22823.   __ret = (int32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  22824.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22825.   return __ret;
  22826. }
  22827. #endif
  22828.  
  22829. #ifdef __LITTLE_ENDIAN__
  22830. __ai int16x4_t vrhadd_s16(int16x4_t __p0, int16x4_t __p1) {
  22831.   int16x4_t __ret;
  22832.   __ret = (int16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  22833.   return __ret;
  22834. }
  22835. #else
  22836. __ai int16x4_t vrhadd_s16(int16x4_t __p0, int16x4_t __p1) {
  22837.   int16x4_t __ret;
  22838.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22839.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  22840.   __ret = (int16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  22841.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22842.   return __ret;
  22843. }
  22844. #endif
  22845.  
  22846. #ifdef __LITTLE_ENDIAN__
  22847. __ai uint8x16_t vrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
  22848.   uint8x16_t __ret;
  22849.   __ret = (uint8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  22850.   return __ret;
  22851. }
  22852. #else
  22853. __ai uint8x16_t vrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
  22854.   uint8x16_t __ret;
  22855.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22856.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22857.   __ret = (uint8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  22858.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22859.   return __ret;
  22860. }
  22861. #endif
  22862.  
  22863. #ifdef __LITTLE_ENDIAN__
  22864. __ai uint32x4_t vrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
  22865.   uint32x4_t __ret;
  22866.   __ret = (uint32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  22867.   return __ret;
  22868. }
  22869. #else
  22870. __ai uint32x4_t vrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
  22871.   uint32x4_t __ret;
  22872.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22873.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  22874.   __ret = (uint32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  22875.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22876.   return __ret;
  22877. }
  22878. #endif
  22879.  
  22880. #ifdef __LITTLE_ENDIAN__
  22881. __ai uint64x2_t vrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
  22882.   uint64x2_t __ret;
  22883.   __ret = (uint64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  22884.   return __ret;
  22885. }
  22886. #else
  22887. __ai uint64x2_t vrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
  22888.   uint64x2_t __ret;
  22889.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22890.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  22891.   __ret = (uint64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  22892.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22893.   return __ret;
  22894. }
  22895. #endif
  22896.  
  22897. #ifdef __LITTLE_ENDIAN__
  22898. __ai uint16x8_t vrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
  22899.   uint16x8_t __ret;
  22900.   __ret = (uint16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  22901.   return __ret;
  22902. }
  22903. #else
  22904. __ai uint16x8_t vrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
  22905.   uint16x8_t __ret;
  22906.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22907.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  22908.   __ret = (uint16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  22909.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22910.   return __ret;
  22911. }
  22912. #endif
  22913.  
  22914. #ifdef __LITTLE_ENDIAN__
  22915. __ai int8x16_t vrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
  22916.   int8x16_t __ret;
  22917.   __ret = (int8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  22918.   return __ret;
  22919. }
  22920. #else
  22921. __ai int8x16_t vrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
  22922.   int8x16_t __ret;
  22923.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22924.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22925.   __ret = (int8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  22926.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  22927.   return __ret;
  22928. }
  22929. #endif
  22930.  
  22931. #ifdef __LITTLE_ENDIAN__
  22932. __ai int32x4_t vrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
  22933.   int32x4_t __ret;
  22934.   __ret = (int32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  22935.   return __ret;
  22936. }
  22937. #else
  22938. __ai int32x4_t vrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
  22939.   int32x4_t __ret;
  22940.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  22941.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  22942.   __ret = (int32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  22943.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  22944.   return __ret;
  22945. }
  22946. #endif
  22947.  
  22948. #ifdef __LITTLE_ENDIAN__
  22949. __ai int64x2_t vrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
  22950.   int64x2_t __ret;
  22951.   __ret = (int64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
  22952.   return __ret;
  22953. }
  22954. #else
  22955. __ai int64x2_t vrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
  22956.   int64x2_t __ret;
  22957.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  22958.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  22959.   __ret = (int64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
  22960.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  22961.   return __ret;
  22962. }
  22963. #endif
  22964.  
  22965. #ifdef __LITTLE_ENDIAN__
  22966. __ai int16x8_t vrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
  22967.   int16x8_t __ret;
  22968.   __ret = (int16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  22969.   return __ret;
  22970. }
  22971. #else
  22972. __ai int16x8_t vrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
  22973.   int16x8_t __ret;
  22974.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22975.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  22976.   __ret = (int16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  22977.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22978.   return __ret;
  22979. }
  22980. #endif
  22981.  
  22982. #ifdef __LITTLE_ENDIAN__
  22983. __ai uint8x8_t vrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
  22984.   uint8x8_t __ret;
  22985.   __ret = (uint8x8_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  22986.   return __ret;
  22987. }
  22988. #else
  22989. __ai uint8x8_t vrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
  22990.   uint8x8_t __ret;
  22991.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  22992.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  22993.   __ret = (uint8x8_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  22994.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  22995.   return __ret;
  22996. }
  22997. #endif
  22998.  
  22999. #ifdef __LITTLE_ENDIAN__
  23000. __ai uint32x2_t vrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
  23001.   uint32x2_t __ret;
  23002.   __ret = (uint32x2_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  23003.   return __ret;
  23004. }
  23005. #else
  23006. __ai uint32x2_t vrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
  23007.   uint32x2_t __ret;
  23008.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  23009.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  23010.   __ret = (uint32x2_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  23011.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  23012.   return __ret;
  23013. }
  23014. #endif
  23015.  
  23016. __ai uint64x1_t vrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
  23017.   uint64x1_t __ret;
  23018.   __ret = (uint64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  23019.   return __ret;
  23020. }
  23021. #ifdef __LITTLE_ENDIAN__
  23022. __ai uint16x4_t vrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
  23023.   uint16x4_t __ret;
  23024.   __ret = (uint16x4_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  23025.   return __ret;
  23026. }
  23027. #else
  23028. __ai uint16x4_t vrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
  23029.   uint16x4_t __ret;
  23030.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  23031.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  23032.   __ret = (uint16x4_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  23033.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  23034.   return __ret;
  23035. }
  23036. #endif
  23037.  
  23038. #ifdef __LITTLE_ENDIAN__
  23039. __ai int8x8_t vrshl_s8(int8x8_t __p0, int8x8_t __p1) {
  23040.   int8x8_t __ret;
  23041.   __ret = (int8x8_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  23042.   return __ret;
  23043. }
  23044. #else
  23045. __ai int8x8_t vrshl_s8(int8x8_t __p0, int8x8_t __p1) {
  23046.   int8x8_t __ret;
  23047.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  23048.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  23049.   __ret = (int8x8_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  23050.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  23051.   return __ret;
  23052. }
  23053. #endif
  23054.  
  23055. #ifdef __LITTLE_ENDIAN__
  23056. __ai int32x2_t vrshl_s32(int32x2_t __p0, int32x2_t __p1) {
  23057.   int32x2_t __ret;
  23058.   __ret = (int32x2_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  23059.   return __ret;
  23060. }
  23061. #else
  23062. __ai int32x2_t vrshl_s32(int32x2_t __p0, int32x2_t __p1) {
  23063.   int32x2_t __ret;
  23064.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  23065.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  23066.   __ret = (int32x2_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  23067.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  23068.   return __ret;
  23069. }
  23070. #endif
  23071.  
  23072. __ai int64x1_t vrshl_s64(int64x1_t __p0, int64x1_t __p1) {
  23073.   int64x1_t __ret;
  23074.   __ret = (int64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
  23075.   return __ret;
  23076. }
  23077. #ifdef __LITTLE_ENDIAN__
  23078. __ai int16x4_t vrshl_s16(int16x4_t __p0, int16x4_t __p1) {
  23079.   int16x4_t __ret;
  23080.   __ret = (int16x4_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  23081.   return __ret;
  23082. }
  23083. #else
  23084. __ai int16x4_t vrshl_s16(int16x4_t __p0, int16x4_t __p1) {
  23085.   int16x4_t __ret;
  23086.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  23087.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  23088.   __ret = (int16x4_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  23089.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  23090.   return __ret;
  23091. }
  23092. #endif
  23093.  
  23094. #ifdef __LITTLE_ENDIAN__
  23095. #define vrshrq_n_u8(__p0, __p1) __extension__ ({ \
  23096.   uint8x16_t __ret; \
  23097.   uint8x16_t __s0 = __p0; \
  23098.   __ret = (uint8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 48); \
  23099.   __ret; \
  23100. })
  23101. #else
  23102. #define vrshrq_n_u8(__p0, __p1) __extension__ ({ \
  23103.   uint8x16_t __ret; \
  23104.   uint8x16_t __s0 = __p0; \
  23105.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23106.   __ret = (uint8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 48); \
  23107.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23108.   __ret; \
  23109. })
  23110. #endif
  23111.  
  23112. #ifdef __LITTLE_ENDIAN__
  23113. #define vrshrq_n_u32(__p0, __p1) __extension__ ({ \
  23114.   uint32x4_t __ret; \
  23115.   uint32x4_t __s0 = __p0; \
  23116.   __ret = (uint32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 50); \
  23117.   __ret; \
  23118. })
  23119. #else
  23120. #define vrshrq_n_u32(__p0, __p1) __extension__ ({ \
  23121.   uint32x4_t __ret; \
  23122.   uint32x4_t __s0 = __p0; \
  23123.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23124.   __ret = (uint32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 50); \
  23125.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23126.   __ret; \
  23127. })
  23128. #endif
  23129.  
  23130. #ifdef __LITTLE_ENDIAN__
  23131. #define vrshrq_n_u64(__p0, __p1) __extension__ ({ \
  23132.   uint64x2_t __ret; \
  23133.   uint64x2_t __s0 = __p0; \
  23134.   __ret = (uint64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 51); \
  23135.   __ret; \
  23136. })
  23137. #else
  23138. #define vrshrq_n_u64(__p0, __p1) __extension__ ({ \
  23139.   uint64x2_t __ret; \
  23140.   uint64x2_t __s0 = __p0; \
  23141.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23142.   __ret = (uint64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 51); \
  23143.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23144.   __ret; \
  23145. })
  23146. #endif
  23147.  
  23148. #ifdef __LITTLE_ENDIAN__
  23149. #define vrshrq_n_u16(__p0, __p1) __extension__ ({ \
  23150.   uint16x8_t __ret; \
  23151.   uint16x8_t __s0 = __p0; \
  23152.   __ret = (uint16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 49); \
  23153.   __ret; \
  23154. })
  23155. #else
  23156. #define vrshrq_n_u16(__p0, __p1) __extension__ ({ \
  23157.   uint16x8_t __ret; \
  23158.   uint16x8_t __s0 = __p0; \
  23159.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23160.   __ret = (uint16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 49); \
  23161.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23162.   __ret; \
  23163. })
  23164. #endif
  23165.  
  23166. #ifdef __LITTLE_ENDIAN__
  23167. #define vrshrq_n_s8(__p0, __p1) __extension__ ({ \
  23168.   int8x16_t __ret; \
  23169.   int8x16_t __s0 = __p0; \
  23170.   __ret = (int8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 32); \
  23171.   __ret; \
  23172. })
  23173. #else
  23174. #define vrshrq_n_s8(__p0, __p1) __extension__ ({ \
  23175.   int8x16_t __ret; \
  23176.   int8x16_t __s0 = __p0; \
  23177.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23178.   __ret = (int8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 32); \
  23179.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23180.   __ret; \
  23181. })
  23182. #endif
  23183.  
  23184. #ifdef __LITTLE_ENDIAN__
  23185. #define vrshrq_n_s32(__p0, __p1) __extension__ ({ \
  23186.   int32x4_t __ret; \
  23187.   int32x4_t __s0 = __p0; \
  23188.   __ret = (int32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 34); \
  23189.   __ret; \
  23190. })
  23191. #else
  23192. #define vrshrq_n_s32(__p0, __p1) __extension__ ({ \
  23193.   int32x4_t __ret; \
  23194.   int32x4_t __s0 = __p0; \
  23195.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23196.   __ret = (int32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 34); \
  23197.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23198.   __ret; \
  23199. })
  23200. #endif
  23201.  
  23202. #ifdef __LITTLE_ENDIAN__
  23203. #define vrshrq_n_s64(__p0, __p1) __extension__ ({ \
  23204.   int64x2_t __ret; \
  23205.   int64x2_t __s0 = __p0; \
  23206.   __ret = (int64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 35); \
  23207.   __ret; \
  23208. })
  23209. #else
  23210. #define vrshrq_n_s64(__p0, __p1) __extension__ ({ \
  23211.   int64x2_t __ret; \
  23212.   int64x2_t __s0 = __p0; \
  23213.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23214.   __ret = (int64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 35); \
  23215.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23216.   __ret; \
  23217. })
  23218. #endif
  23219.  
  23220. #ifdef __LITTLE_ENDIAN__
  23221. #define vrshrq_n_s16(__p0, __p1) __extension__ ({ \
  23222.   int16x8_t __ret; \
  23223.   int16x8_t __s0 = __p0; \
  23224.   __ret = (int16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 33); \
  23225.   __ret; \
  23226. })
  23227. #else
  23228. #define vrshrq_n_s16(__p0, __p1) __extension__ ({ \
  23229.   int16x8_t __ret; \
  23230.   int16x8_t __s0 = __p0; \
  23231.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23232.   __ret = (int16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 33); \
  23233.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23234.   __ret; \
  23235. })
  23236. #endif
  23237.  
  23238. #ifdef __LITTLE_ENDIAN__
  23239. #define vrshr_n_u8(__p0, __p1) __extension__ ({ \
  23240.   uint8x8_t __ret; \
  23241.   uint8x8_t __s0 = __p0; \
  23242.   __ret = (uint8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 16); \
  23243.   __ret; \
  23244. })
  23245. #else
  23246. #define vrshr_n_u8(__p0, __p1) __extension__ ({ \
  23247.   uint8x8_t __ret; \
  23248.   uint8x8_t __s0 = __p0; \
  23249.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23250.   __ret = (uint8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 16); \
  23251.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23252.   __ret; \
  23253. })
  23254. #endif
  23255.  
  23256. #ifdef __LITTLE_ENDIAN__
  23257. #define vrshr_n_u32(__p0, __p1) __extension__ ({ \
  23258.   uint32x2_t __ret; \
  23259.   uint32x2_t __s0 = __p0; \
  23260.   __ret = (uint32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 18); \
  23261.   __ret; \
  23262. })
  23263. #else
  23264. #define vrshr_n_u32(__p0, __p1) __extension__ ({ \
  23265.   uint32x2_t __ret; \
  23266.   uint32x2_t __s0 = __p0; \
  23267.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23268.   __ret = (uint32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 18); \
  23269.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23270.   __ret; \
  23271. })
  23272. #endif
  23273.  
  23274. #define vrshr_n_u64(__p0, __p1) __extension__ ({ \
  23275.   uint64x1_t __ret; \
  23276.   uint64x1_t __s0 = __p0; \
  23277.   __ret = (uint64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 19); \
  23278.   __ret; \
  23279. })
  23280. #ifdef __LITTLE_ENDIAN__
  23281. #define vrshr_n_u16(__p0, __p1) __extension__ ({ \
  23282.   uint16x4_t __ret; \
  23283.   uint16x4_t __s0 = __p0; \
  23284.   __ret = (uint16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 17); \
  23285.   __ret; \
  23286. })
  23287. #else
  23288. #define vrshr_n_u16(__p0, __p1) __extension__ ({ \
  23289.   uint16x4_t __ret; \
  23290.   uint16x4_t __s0 = __p0; \
  23291.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23292.   __ret = (uint16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 17); \
  23293.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23294.   __ret; \
  23295. })
  23296. #endif
  23297.  
  23298. #ifdef __LITTLE_ENDIAN__
  23299. #define vrshr_n_s8(__p0, __p1) __extension__ ({ \
  23300.   int8x8_t __ret; \
  23301.   int8x8_t __s0 = __p0; \
  23302.   __ret = (int8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 0); \
  23303.   __ret; \
  23304. })
  23305. #else
  23306. #define vrshr_n_s8(__p0, __p1) __extension__ ({ \
  23307.   int8x8_t __ret; \
  23308.   int8x8_t __s0 = __p0; \
  23309.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23310.   __ret = (int8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 0); \
  23311.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23312.   __ret; \
  23313. })
  23314. #endif
  23315.  
  23316. #ifdef __LITTLE_ENDIAN__
  23317. #define vrshr_n_s32(__p0, __p1) __extension__ ({ \
  23318.   int32x2_t __ret; \
  23319.   int32x2_t __s0 = __p0; \
  23320.   __ret = (int32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 2); \
  23321.   __ret; \
  23322. })
  23323. #else
  23324. #define vrshr_n_s32(__p0, __p1) __extension__ ({ \
  23325.   int32x2_t __ret; \
  23326.   int32x2_t __s0 = __p0; \
  23327.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23328.   __ret = (int32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 2); \
  23329.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23330.   __ret; \
  23331. })
  23332. #endif
  23333.  
  23334. #define vrshr_n_s64(__p0, __p1) __extension__ ({ \
  23335.   int64x1_t __ret; \
  23336.   int64x1_t __s0 = __p0; \
  23337.   __ret = (int64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 3); \
  23338.   __ret; \
  23339. })
  23340. #ifdef __LITTLE_ENDIAN__
  23341. #define vrshr_n_s16(__p0, __p1) __extension__ ({ \
  23342.   int16x4_t __ret; \
  23343.   int16x4_t __s0 = __p0; \
  23344.   __ret = (int16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 1); \
  23345.   __ret; \
  23346. })
  23347. #else
  23348. #define vrshr_n_s16(__p0, __p1) __extension__ ({ \
  23349.   int16x4_t __ret; \
  23350.   int16x4_t __s0 = __p0; \
  23351.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23352.   __ret = (int16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 1); \
  23353.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23354.   __ret; \
  23355. })
  23356. #endif
  23357.  
  23358. #ifdef __LITTLE_ENDIAN__
  23359. #define vrshrn_n_u32(__p0, __p1) __extension__ ({ \
  23360.   uint16x4_t __ret; \
  23361.   uint32x4_t __s0 = __p0; \
  23362.   __ret = (uint16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 17); \
  23363.   __ret; \
  23364. })
  23365. #else
  23366. #define vrshrn_n_u32(__p0, __p1) __extension__ ({ \
  23367.   uint16x4_t __ret; \
  23368.   uint32x4_t __s0 = __p0; \
  23369.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23370.   __ret = (uint16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 17); \
  23371.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23372.   __ret; \
  23373. })
  23374. #define __noswap_vrshrn_n_u32(__p0, __p1) __extension__ ({ \
  23375.   uint16x4_t __ret; \
  23376.   uint32x4_t __s0 = __p0; \
  23377.   __ret = (uint16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 17); \
  23378.   __ret; \
  23379. })
  23380. #endif
  23381.  
  23382. #ifdef __LITTLE_ENDIAN__
  23383. #define vrshrn_n_u64(__p0, __p1) __extension__ ({ \
  23384.   uint32x2_t __ret; \
  23385.   uint64x2_t __s0 = __p0; \
  23386.   __ret = (uint32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 18); \
  23387.   __ret; \
  23388. })
  23389. #else
  23390. #define vrshrn_n_u64(__p0, __p1) __extension__ ({ \
  23391.   uint32x2_t __ret; \
  23392.   uint64x2_t __s0 = __p0; \
  23393.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23394.   __ret = (uint32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 18); \
  23395.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23396.   __ret; \
  23397. })
  23398. #define __noswap_vrshrn_n_u64(__p0, __p1) __extension__ ({ \
  23399.   uint32x2_t __ret; \
  23400.   uint64x2_t __s0 = __p0; \
  23401.   __ret = (uint32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 18); \
  23402.   __ret; \
  23403. })
  23404. #endif
  23405.  
  23406. #ifdef __LITTLE_ENDIAN__
  23407. #define vrshrn_n_u16(__p0, __p1) __extension__ ({ \
  23408.   uint8x8_t __ret; \
  23409.   uint16x8_t __s0 = __p0; \
  23410.   __ret = (uint8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 16); \
  23411.   __ret; \
  23412. })
  23413. #else
  23414. #define vrshrn_n_u16(__p0, __p1) __extension__ ({ \
  23415.   uint8x8_t __ret; \
  23416.   uint16x8_t __s0 = __p0; \
  23417.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23418.   __ret = (uint8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 16); \
  23419.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23420.   __ret; \
  23421. })
  23422. #define __noswap_vrshrn_n_u16(__p0, __p1) __extension__ ({ \
  23423.   uint8x8_t __ret; \
  23424.   uint16x8_t __s0 = __p0; \
  23425.   __ret = (uint8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 16); \
  23426.   __ret; \
  23427. })
  23428. #endif
  23429.  
  23430. #ifdef __LITTLE_ENDIAN__
  23431. #define vrshrn_n_s32(__p0, __p1) __extension__ ({ \
  23432.   int16x4_t __ret; \
  23433.   int32x4_t __s0 = __p0; \
  23434.   __ret = (int16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 1); \
  23435.   __ret; \
  23436. })
  23437. #else
  23438. #define vrshrn_n_s32(__p0, __p1) __extension__ ({ \
  23439.   int16x4_t __ret; \
  23440.   int32x4_t __s0 = __p0; \
  23441.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23442.   __ret = (int16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 1); \
  23443.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23444.   __ret; \
  23445. })
  23446. #define __noswap_vrshrn_n_s32(__p0, __p1) __extension__ ({ \
  23447.   int16x4_t __ret; \
  23448.   int32x4_t __s0 = __p0; \
  23449.   __ret = (int16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 1); \
  23450.   __ret; \
  23451. })
  23452. #endif
  23453.  
  23454. #ifdef __LITTLE_ENDIAN__
  23455. #define vrshrn_n_s64(__p0, __p1) __extension__ ({ \
  23456.   int32x2_t __ret; \
  23457.   int64x2_t __s0 = __p0; \
  23458.   __ret = (int32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 2); \
  23459.   __ret; \
  23460. })
  23461. #else
  23462. #define vrshrn_n_s64(__p0, __p1) __extension__ ({ \
  23463.   int32x2_t __ret; \
  23464.   int64x2_t __s0 = __p0; \
  23465.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23466.   __ret = (int32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 2); \
  23467.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23468.   __ret; \
  23469. })
  23470. #define __noswap_vrshrn_n_s64(__p0, __p1) __extension__ ({ \
  23471.   int32x2_t __ret; \
  23472.   int64x2_t __s0 = __p0; \
  23473.   __ret = (int32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 2); \
  23474.   __ret; \
  23475. })
  23476. #endif
  23477.  
  23478. #ifdef __LITTLE_ENDIAN__
  23479. #define vrshrn_n_s16(__p0, __p1) __extension__ ({ \
  23480.   int8x8_t __ret; \
  23481.   int16x8_t __s0 = __p0; \
  23482.   __ret = (int8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 0); \
  23483.   __ret; \
  23484. })
  23485. #else
  23486. #define vrshrn_n_s16(__p0, __p1) __extension__ ({ \
  23487.   int8x8_t __ret; \
  23488.   int16x8_t __s0 = __p0; \
  23489.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23490.   __ret = (int8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 0); \
  23491.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23492.   __ret; \
  23493. })
  23494. #define __noswap_vrshrn_n_s16(__p0, __p1) __extension__ ({ \
  23495.   int8x8_t __ret; \
  23496.   int16x8_t __s0 = __p0; \
  23497.   __ret = (int8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 0); \
  23498.   __ret; \
  23499. })
  23500. #endif
  23501.  
  23502. #ifdef __LITTLE_ENDIAN__
  23503. __ai uint32x4_t vrsqrteq_u32(uint32x4_t __p0) {
  23504.   uint32x4_t __ret;
  23505.   __ret = (uint32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 50);
  23506.   return __ret;
  23507. }
  23508. #else
  23509. __ai uint32x4_t vrsqrteq_u32(uint32x4_t __p0) {
  23510.   uint32x4_t __ret;
  23511.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  23512.   __ret = (uint32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 50);
  23513.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  23514.   return __ret;
  23515. }
  23516. #endif
  23517.  
  23518. #ifdef __LITTLE_ENDIAN__
  23519. __ai float32x4_t vrsqrteq_f32(float32x4_t __p0) {
  23520.   float32x4_t __ret;
  23521.   __ret = (float32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 41);
  23522.   return __ret;
  23523. }
  23524. #else
  23525. __ai float32x4_t vrsqrteq_f32(float32x4_t __p0) {
  23526.   float32x4_t __ret;
  23527.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  23528.   __ret = (float32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 41);
  23529.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  23530.   return __ret;
  23531. }
  23532. #endif
  23533.  
  23534. #ifdef __LITTLE_ENDIAN__
  23535. __ai uint32x2_t vrsqrte_u32(uint32x2_t __p0) {
  23536.   uint32x2_t __ret;
  23537.   __ret = (uint32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 18);
  23538.   return __ret;
  23539. }
  23540. #else
  23541. __ai uint32x2_t vrsqrte_u32(uint32x2_t __p0) {
  23542.   uint32x2_t __ret;
  23543.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  23544.   __ret = (uint32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__rev0, 18);
  23545.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  23546.   return __ret;
  23547. }
  23548. #endif
  23549.  
  23550. #ifdef __LITTLE_ENDIAN__
  23551. __ai float32x2_t vrsqrte_f32(float32x2_t __p0) {
  23552.   float32x2_t __ret;
  23553.   __ret = (float32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 9);
  23554.   return __ret;
  23555. }
  23556. #else
  23557. __ai float32x2_t vrsqrte_f32(float32x2_t __p0) {
  23558.   float32x2_t __ret;
  23559.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  23560.   __ret = (float32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__rev0, 9);
  23561.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  23562.   return __ret;
  23563. }
  23564. #endif
  23565.  
  23566. #ifdef __LITTLE_ENDIAN__
  23567. __ai float32x4_t vrsqrtsq_f32(float32x4_t __p0, float32x4_t __p1) {
  23568.   float32x4_t __ret;
  23569.   __ret = (float32x4_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  23570.   return __ret;
  23571. }
  23572. #else
  23573. __ai float32x4_t vrsqrtsq_f32(float32x4_t __p0, float32x4_t __p1) {
  23574.   float32x4_t __ret;
  23575.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  23576.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  23577.   __ret = (float32x4_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  23578.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  23579.   return __ret;
  23580. }
  23581. #endif
  23582.  
  23583. #ifdef __LITTLE_ENDIAN__
  23584. __ai float32x2_t vrsqrts_f32(float32x2_t __p0, float32x2_t __p1) {
  23585.   float32x2_t __ret;
  23586.   __ret = (float32x2_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  23587.   return __ret;
  23588. }
  23589. #else
  23590. __ai float32x2_t vrsqrts_f32(float32x2_t __p0, float32x2_t __p1) {
  23591.   float32x2_t __ret;
  23592.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  23593.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  23594.   __ret = (float32x2_t) __builtin_neon_vrsqrts_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  23595.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  23596.   return __ret;
  23597. }
  23598. #endif
  23599.  
  23600. #ifdef __LITTLE_ENDIAN__
  23601. #define vrsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
  23602.   uint8x16_t __ret; \
  23603.   uint8x16_t __s0 = __p0; \
  23604.   uint8x16_t __s1 = __p1; \
  23605.   __ret = (uint8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
  23606.   __ret; \
  23607. })
  23608. #else
  23609. #define vrsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
  23610.   uint8x16_t __ret; \
  23611.   uint8x16_t __s0 = __p0; \
  23612.   uint8x16_t __s1 = __p1; \
  23613.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23614.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23615.   __ret = (uint8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
  23616.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23617.   __ret; \
  23618. })
  23619. #endif
  23620.  
  23621. #ifdef __LITTLE_ENDIAN__
  23622. #define vrsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
  23623.   uint32x4_t __ret; \
  23624.   uint32x4_t __s0 = __p0; \
  23625.   uint32x4_t __s1 = __p1; \
  23626.   __ret = (uint32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
  23627.   __ret; \
  23628. })
  23629. #else
  23630. #define vrsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
  23631.   uint32x4_t __ret; \
  23632.   uint32x4_t __s0 = __p0; \
  23633.   uint32x4_t __s1 = __p1; \
  23634.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23635.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  23636.   __ret = (uint32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
  23637.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23638.   __ret; \
  23639. })
  23640. #endif
  23641.  
  23642. #ifdef __LITTLE_ENDIAN__
  23643. #define vrsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
  23644.   uint64x2_t __ret; \
  23645.   uint64x2_t __s0 = __p0; \
  23646.   uint64x2_t __s1 = __p1; \
  23647.   __ret = (uint64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
  23648.   __ret; \
  23649. })
  23650. #else
  23651. #define vrsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
  23652.   uint64x2_t __ret; \
  23653.   uint64x2_t __s0 = __p0; \
  23654.   uint64x2_t __s1 = __p1; \
  23655.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23656.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  23657.   __ret = (uint64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
  23658.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23659.   __ret; \
  23660. })
  23661. #endif
  23662.  
  23663. #ifdef __LITTLE_ENDIAN__
  23664. #define vrsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
  23665.   uint16x8_t __ret; \
  23666.   uint16x8_t __s0 = __p0; \
  23667.   uint16x8_t __s1 = __p1; \
  23668.   __ret = (uint16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
  23669.   __ret; \
  23670. })
  23671. #else
  23672. #define vrsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
  23673.   uint16x8_t __ret; \
  23674.   uint16x8_t __s0 = __p0; \
  23675.   uint16x8_t __s1 = __p1; \
  23676.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23677.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  23678.   __ret = (uint16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
  23679.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23680.   __ret; \
  23681. })
  23682. #endif
  23683.  
  23684. #ifdef __LITTLE_ENDIAN__
  23685. #define vrsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
  23686.   int8x16_t __ret; \
  23687.   int8x16_t __s0 = __p0; \
  23688.   int8x16_t __s1 = __p1; \
  23689.   __ret = (int8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
  23690.   __ret; \
  23691. })
  23692. #else
  23693. #define vrsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
  23694.   int8x16_t __ret; \
  23695.   int8x16_t __s0 = __p0; \
  23696.   int8x16_t __s1 = __p1; \
  23697.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23698.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23699.   __ret = (int8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
  23700.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  23701.   __ret; \
  23702. })
  23703. #endif
  23704.  
  23705. #ifdef __LITTLE_ENDIAN__
  23706. #define vrsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
  23707.   int32x4_t __ret; \
  23708.   int32x4_t __s0 = __p0; \
  23709.   int32x4_t __s1 = __p1; \
  23710.   __ret = (int32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
  23711.   __ret; \
  23712. })
  23713. #else
  23714. #define vrsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
  23715.   int32x4_t __ret; \
  23716.   int32x4_t __s0 = __p0; \
  23717.   int32x4_t __s1 = __p1; \
  23718.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23719.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  23720.   __ret = (int32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
  23721.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23722.   __ret; \
  23723. })
  23724. #endif
  23725.  
  23726. #ifdef __LITTLE_ENDIAN__
  23727. #define vrsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
  23728.   int64x2_t __ret; \
  23729.   int64x2_t __s0 = __p0; \
  23730.   int64x2_t __s1 = __p1; \
  23731.   __ret = (int64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
  23732.   __ret; \
  23733. })
  23734. #else
  23735. #define vrsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
  23736.   int64x2_t __ret; \
  23737.   int64x2_t __s0 = __p0; \
  23738.   int64x2_t __s1 = __p1; \
  23739.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23740.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  23741.   __ret = (int64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
  23742.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23743.   __ret; \
  23744. })
  23745. #endif
  23746.  
  23747. #ifdef __LITTLE_ENDIAN__
  23748. #define vrsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
  23749.   int16x8_t __ret; \
  23750.   int16x8_t __s0 = __p0; \
  23751.   int16x8_t __s1 = __p1; \
  23752.   __ret = (int16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
  23753.   __ret; \
  23754. })
  23755. #else
  23756. #define vrsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
  23757.   int16x8_t __ret; \
  23758.   int16x8_t __s0 = __p0; \
  23759.   int16x8_t __s1 = __p1; \
  23760.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23761.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  23762.   __ret = (int16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
  23763.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23764.   __ret; \
  23765. })
  23766. #endif
  23767.  
  23768. #ifdef __LITTLE_ENDIAN__
  23769. #define vrsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
  23770.   uint8x8_t __ret; \
  23771.   uint8x8_t __s0 = __p0; \
  23772.   uint8x8_t __s1 = __p1; \
  23773.   __ret = (uint8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
  23774.   __ret; \
  23775. })
  23776. #else
  23777. #define vrsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
  23778.   uint8x8_t __ret; \
  23779.   uint8x8_t __s0 = __p0; \
  23780.   uint8x8_t __s1 = __p1; \
  23781.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23782.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  23783.   __ret = (uint8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
  23784.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23785.   __ret; \
  23786. })
  23787. #endif
  23788.  
  23789. #ifdef __LITTLE_ENDIAN__
  23790. #define vrsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
  23791.   uint32x2_t __ret; \
  23792.   uint32x2_t __s0 = __p0; \
  23793.   uint32x2_t __s1 = __p1; \
  23794.   __ret = (uint32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
  23795.   __ret; \
  23796. })
  23797. #else
  23798. #define vrsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
  23799.   uint32x2_t __ret; \
  23800.   uint32x2_t __s0 = __p0; \
  23801.   uint32x2_t __s1 = __p1; \
  23802.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23803.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  23804.   __ret = (uint32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
  23805.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23806.   __ret; \
  23807. })
  23808. #endif
  23809.  
  23810. #define vrsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
  23811.   uint64x1_t __ret; \
  23812.   uint64x1_t __s0 = __p0; \
  23813.   uint64x1_t __s1 = __p1; \
  23814.   __ret = (uint64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
  23815.   __ret; \
  23816. })
  23817. #ifdef __LITTLE_ENDIAN__
  23818. #define vrsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
  23819.   uint16x4_t __ret; \
  23820.   uint16x4_t __s0 = __p0; \
  23821.   uint16x4_t __s1 = __p1; \
  23822.   __ret = (uint16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
  23823.   __ret; \
  23824. })
  23825. #else
  23826. #define vrsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
  23827.   uint16x4_t __ret; \
  23828.   uint16x4_t __s0 = __p0; \
  23829.   uint16x4_t __s1 = __p1; \
  23830.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23831.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  23832.   __ret = (uint16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
  23833.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23834.   __ret; \
  23835. })
  23836. #endif
  23837.  
  23838. #ifdef __LITTLE_ENDIAN__
  23839. #define vrsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
  23840.   int8x8_t __ret; \
  23841.   int8x8_t __s0 = __p0; \
  23842.   int8x8_t __s1 = __p1; \
  23843.   __ret = (int8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
  23844.   __ret; \
  23845. })
  23846. #else
  23847. #define vrsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
  23848.   int8x8_t __ret; \
  23849.   int8x8_t __s0 = __p0; \
  23850.   int8x8_t __s1 = __p1; \
  23851.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  23852.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  23853.   __ret = (int8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
  23854.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  23855.   __ret; \
  23856. })
  23857. #endif
  23858.  
  23859. #ifdef __LITTLE_ENDIAN__
  23860. #define vrsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
  23861.   int32x2_t __ret; \
  23862.   int32x2_t __s0 = __p0; \
  23863.   int32x2_t __s1 = __p1; \
  23864.   __ret = (int32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
  23865.   __ret; \
  23866. })
  23867. #else
  23868. #define vrsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
  23869.   int32x2_t __ret; \
  23870.   int32x2_t __s0 = __p0; \
  23871.   int32x2_t __s1 = __p1; \
  23872.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  23873.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  23874.   __ret = (int32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  23875.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  23876.   __ret; \
  23877. })
  23878. #endif
  23879.  
  23880. #define vrsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
  23881.   int64x1_t __ret; \
  23882.   int64x1_t __s0 = __p0; \
  23883.   int64x1_t __s1 = __p1; \
  23884.   __ret = (int64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
  23885.   __ret; \
  23886. })
  23887. #ifdef __LITTLE_ENDIAN__
  23888. #define vrsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
  23889.   int16x4_t __ret; \
  23890.   int16x4_t __s0 = __p0; \
  23891.   int16x4_t __s1 = __p1; \
  23892.   __ret = (int16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
  23893.   __ret; \
  23894. })
  23895. #else
  23896. #define vrsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
  23897.   int16x4_t __ret; \
  23898.   int16x4_t __s0 = __p0; \
  23899.   int16x4_t __s1 = __p1; \
  23900.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  23901.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  23902.   __ret = (int16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  23903.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  23904.   __ret; \
  23905. })
  23906. #endif
  23907.  
  23908. #ifdef __LITTLE_ENDIAN__
  23909. __ai uint16x4_t vrsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  23910.   uint16x4_t __ret;
  23911.   __ret = (uint16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  23912.   return __ret;
  23913. }
  23914. #else
  23915. __ai uint16x4_t vrsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  23916.   uint16x4_t __ret;
  23917.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  23918.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  23919.   __ret = (uint16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
  23920.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  23921.   return __ret;
  23922. }
  23923. __ai uint16x4_t __noswap_vrsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  23924.   uint16x4_t __ret;
  23925.   __ret = (uint16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  23926.   return __ret;
  23927. }
  23928. #endif
  23929.  
  23930. #ifdef __LITTLE_ENDIAN__
  23931. __ai uint32x2_t vrsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  23932.   uint32x2_t __ret;
  23933.   __ret = (uint32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  23934.   return __ret;
  23935. }
  23936. #else
  23937. __ai uint32x2_t vrsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  23938.   uint32x2_t __ret;
  23939.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  23940.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  23941.   __ret = (uint32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
  23942.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  23943.   return __ret;
  23944. }
  23945. __ai uint32x2_t __noswap_vrsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  23946.   uint32x2_t __ret;
  23947.   __ret = (uint32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  23948.   return __ret;
  23949. }
  23950. #endif
  23951.  
  23952. #ifdef __LITTLE_ENDIAN__
  23953. __ai uint8x8_t vrsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  23954.   uint8x8_t __ret;
  23955.   __ret = (uint8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  23956.   return __ret;
  23957. }
  23958. #else
  23959. __ai uint8x8_t vrsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  23960.   uint8x8_t __ret;
  23961.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  23962.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  23963.   __ret = (uint8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
  23964.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  23965.   return __ret;
  23966. }
  23967. __ai uint8x8_t __noswap_vrsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  23968.   uint8x8_t __ret;
  23969.   __ret = (uint8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  23970.   return __ret;
  23971. }
  23972. #endif
  23973.  
  23974. #ifdef __LITTLE_ENDIAN__
  23975. __ai int16x4_t vrsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
  23976.   int16x4_t __ret;
  23977.   __ret = (int16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  23978.   return __ret;
  23979. }
  23980. #else
  23981. __ai int16x4_t vrsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
  23982.   int16x4_t __ret;
  23983.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  23984.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  23985.   __ret = (int16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
  23986.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  23987.   return __ret;
  23988. }
  23989. __ai int16x4_t __noswap_vrsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
  23990.   int16x4_t __ret;
  23991.   __ret = (int16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  23992.   return __ret;
  23993. }
  23994. #endif
  23995.  
  23996. #ifdef __LITTLE_ENDIAN__
  23997. __ai int32x2_t vrsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
  23998.   int32x2_t __ret;
  23999.   __ret = (int32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  24000.   return __ret;
  24001. }
  24002. #else
  24003. __ai int32x2_t vrsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
  24004.   int32x2_t __ret;
  24005.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  24006.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  24007.   __ret = (int32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
  24008.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  24009.   return __ret;
  24010. }
  24011. __ai int32x2_t __noswap_vrsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
  24012.   int32x2_t __ret;
  24013.   __ret = (int32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  24014.   return __ret;
  24015. }
  24016. #endif
  24017.  
  24018. #ifdef __LITTLE_ENDIAN__
  24019. __ai int8x8_t vrsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
  24020.   int8x8_t __ret;
  24021.   __ret = (int8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  24022.   return __ret;
  24023. }
  24024. #else
  24025. __ai int8x8_t vrsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
  24026.   int8x8_t __ret;
  24027.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  24028.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  24029.   __ret = (int8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
  24030.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  24031.   return __ret;
  24032. }
  24033. __ai int8x8_t __noswap_vrsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
  24034.   int8x8_t __ret;
  24035.   __ret = (int8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  24036.   return __ret;
  24037. }
  24038. #endif
  24039.  
  24040. #ifdef __LITTLE_ENDIAN__
  24041. #define vset_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  24042.   poly8x8_t __ret; \
  24043.   poly8_t __s0 = __p0; \
  24044.   poly8x8_t __s1 = __p1; \
  24045.   __ret = (poly8x8_t) __builtin_neon_vset_lane_i8(__s0, (poly8x8_t)__s1, __p2); \
  24046.   __ret; \
  24047. })
  24048. #else
  24049. #define vset_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  24050.   poly8x8_t __ret; \
  24051.   poly8_t __s0 = __p0; \
  24052.   poly8x8_t __s1 = __p1; \
  24053.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  24054.   __ret = (poly8x8_t) __builtin_neon_vset_lane_i8(__s0, (poly8x8_t)__rev1, __p2); \
  24055.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  24056.   __ret; \
  24057. })
  24058. #define __noswap_vset_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  24059.   poly8x8_t __ret; \
  24060.   poly8_t __s0 = __p0; \
  24061.   poly8x8_t __s1 = __p1; \
  24062.   __ret = (poly8x8_t) __builtin_neon_vset_lane_i8(__s0, (poly8x8_t)__s1, __p2); \
  24063.   __ret; \
  24064. })
  24065. #endif
  24066.  
  24067. #ifdef __LITTLE_ENDIAN__
  24068. #define vset_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  24069.   poly16x4_t __ret; \
  24070.   poly16_t __s0 = __p0; \
  24071.   poly16x4_t __s1 = __p1; \
  24072.   __ret = (poly16x4_t) __builtin_neon_vset_lane_i16(__s0, (poly16x4_t)__s1, __p2); \
  24073.   __ret; \
  24074. })
  24075. #else
  24076. #define vset_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  24077.   poly16x4_t __ret; \
  24078.   poly16_t __s0 = __p0; \
  24079.   poly16x4_t __s1 = __p1; \
  24080.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  24081.   __ret = (poly16x4_t) __builtin_neon_vset_lane_i16(__s0, (poly16x4_t)__rev1, __p2); \
  24082.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  24083.   __ret; \
  24084. })
  24085. #define __noswap_vset_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  24086.   poly16x4_t __ret; \
  24087.   poly16_t __s0 = __p0; \
  24088.   poly16x4_t __s1 = __p1; \
  24089.   __ret = (poly16x4_t) __builtin_neon_vset_lane_i16(__s0, (poly16x4_t)__s1, __p2); \
  24090.   __ret; \
  24091. })
  24092. #endif
  24093.  
  24094. #ifdef __LITTLE_ENDIAN__
  24095. #define vsetq_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  24096.   poly8x16_t __ret; \
  24097.   poly8_t __s0 = __p0; \
  24098.   poly8x16_t __s1 = __p1; \
  24099.   __ret = (poly8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (poly8x16_t)__s1, __p2); \
  24100.   __ret; \
  24101. })
  24102. #else
  24103. #define vsetq_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  24104.   poly8x16_t __ret; \
  24105.   poly8_t __s0 = __p0; \
  24106.   poly8x16_t __s1 = __p1; \
  24107.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24108.   __ret = (poly8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (poly8x16_t)__rev1, __p2); \
  24109.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24110.   __ret; \
  24111. })
  24112. #define __noswap_vsetq_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  24113.   poly8x16_t __ret; \
  24114.   poly8_t __s0 = __p0; \
  24115.   poly8x16_t __s1 = __p1; \
  24116.   __ret = (poly8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (poly8x16_t)__s1, __p2); \
  24117.   __ret; \
  24118. })
  24119. #endif
  24120.  
  24121. #ifdef __LITTLE_ENDIAN__
  24122. #define vsetq_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  24123.   poly16x8_t __ret; \
  24124.   poly16_t __s0 = __p0; \
  24125.   poly16x8_t __s1 = __p1; \
  24126.   __ret = (poly16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (poly16x8_t)__s1, __p2); \
  24127.   __ret; \
  24128. })
  24129. #else
  24130. #define vsetq_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  24131.   poly16x8_t __ret; \
  24132.   poly16_t __s0 = __p0; \
  24133.   poly16x8_t __s1 = __p1; \
  24134.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  24135.   __ret = (poly16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (poly16x8_t)__rev1, __p2); \
  24136.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  24137.   __ret; \
  24138. })
  24139. #define __noswap_vsetq_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  24140.   poly16x8_t __ret; \
  24141.   poly16_t __s0 = __p0; \
  24142.   poly16x8_t __s1 = __p1; \
  24143.   __ret = (poly16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (poly16x8_t)__s1, __p2); \
  24144.   __ret; \
  24145. })
  24146. #endif
  24147.  
  24148. #ifdef __LITTLE_ENDIAN__
  24149. #define vsetq_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  24150.   uint8x16_t __ret; \
  24151.   uint8_t __s0 = __p0; \
  24152.   uint8x16_t __s1 = __p1; \
  24153.   __ret = (uint8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
  24154.   __ret; \
  24155. })
  24156. #else
  24157. #define vsetq_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  24158.   uint8x16_t __ret; \
  24159.   uint8_t __s0 = __p0; \
  24160.   uint8x16_t __s1 = __p1; \
  24161.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24162.   __ret = (uint8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__rev1, __p2); \
  24163.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24164.   __ret; \
  24165. })
  24166. #define __noswap_vsetq_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  24167.   uint8x16_t __ret; \
  24168.   uint8_t __s0 = __p0; \
  24169.   uint8x16_t __s1 = __p1; \
  24170.   __ret = (uint8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
  24171.   __ret; \
  24172. })
  24173. #endif
  24174.  
  24175. #ifdef __LITTLE_ENDIAN__
  24176. #define vsetq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  24177.   uint32x4_t __ret; \
  24178.   uint32_t __s0 = __p0; \
  24179.   uint32x4_t __s1 = __p1; \
  24180.   __ret = (uint32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int32x4_t)__s1, __p2); \
  24181.   __ret; \
  24182. })
  24183. #else
  24184. #define vsetq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  24185.   uint32x4_t __ret; \
  24186.   uint32_t __s0 = __p0; \
  24187.   uint32x4_t __s1 = __p1; \
  24188.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  24189.   __ret = (uint32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int32x4_t)__rev1, __p2); \
  24190.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  24191.   __ret; \
  24192. })
  24193. #define __noswap_vsetq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  24194.   uint32x4_t __ret; \
  24195.   uint32_t __s0 = __p0; \
  24196.   uint32x4_t __s1 = __p1; \
  24197.   __ret = (uint32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int32x4_t)__s1, __p2); \
  24198.   __ret; \
  24199. })
  24200. #endif
  24201.  
  24202. #ifdef __LITTLE_ENDIAN__
  24203. #define vsetq_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  24204.   uint64x2_t __ret; \
  24205.   uint64_t __s0 = __p0; \
  24206.   uint64x2_t __s1 = __p1; \
  24207.   __ret = (uint64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int64x2_t)__s1, __p2); \
  24208.   __ret; \
  24209. })
  24210. #else
  24211. #define vsetq_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  24212.   uint64x2_t __ret; \
  24213.   uint64_t __s0 = __p0; \
  24214.   uint64x2_t __s1 = __p1; \
  24215.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  24216.   __ret = (uint64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int64x2_t)__rev1, __p2); \
  24217.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  24218.   __ret; \
  24219. })
  24220. #define __noswap_vsetq_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  24221.   uint64x2_t __ret; \
  24222.   uint64_t __s0 = __p0; \
  24223.   uint64x2_t __s1 = __p1; \
  24224.   __ret = (uint64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int64x2_t)__s1, __p2); \
  24225.   __ret; \
  24226. })
  24227. #endif
  24228.  
  24229. #ifdef __LITTLE_ENDIAN__
  24230. #define vsetq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  24231.   uint16x8_t __ret; \
  24232.   uint16_t __s0 = __p0; \
  24233.   uint16x8_t __s1 = __p1; \
  24234.   __ret = (uint16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int16x8_t)__s1, __p2); \
  24235.   __ret; \
  24236. })
  24237. #else
  24238. #define vsetq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  24239.   uint16x8_t __ret; \
  24240.   uint16_t __s0 = __p0; \
  24241.   uint16x8_t __s1 = __p1; \
  24242.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  24243.   __ret = (uint16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int16x8_t)__rev1, __p2); \
  24244.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  24245.   __ret; \
  24246. })
  24247. #define __noswap_vsetq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  24248.   uint16x8_t __ret; \
  24249.   uint16_t __s0 = __p0; \
  24250.   uint16x8_t __s1 = __p1; \
  24251.   __ret = (uint16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int16x8_t)__s1, __p2); \
  24252.   __ret; \
  24253. })
  24254. #endif
  24255.  
  24256. #ifdef __LITTLE_ENDIAN__
  24257. #define vsetq_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  24258.   int8x16_t __ret; \
  24259.   int8_t __s0 = __p0; \
  24260.   int8x16_t __s1 = __p1; \
  24261.   __ret = (int8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
  24262.   __ret; \
  24263. })
  24264. #else
  24265. #define vsetq_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  24266.   int8x16_t __ret; \
  24267.   int8_t __s0 = __p0; \
  24268.   int8x16_t __s1 = __p1; \
  24269.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24270.   __ret = (int8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__rev1, __p2); \
  24271.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24272.   __ret; \
  24273. })
  24274. #define __noswap_vsetq_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  24275.   int8x16_t __ret; \
  24276.   int8_t __s0 = __p0; \
  24277.   int8x16_t __s1 = __p1; \
  24278.   __ret = (int8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
  24279.   __ret; \
  24280. })
  24281. #endif
  24282.  
  24283. #ifdef __LITTLE_ENDIAN__
  24284. #define vsetq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  24285.   float32x4_t __ret; \
  24286.   float32_t __s0 = __p0; \
  24287.   float32x4_t __s1 = __p1; \
  24288.   __ret = (float32x4_t) __builtin_neon_vsetq_lane_f32(__s0, (float32x4_t)__s1, __p2); \
  24289.   __ret; \
  24290. })
  24291. #else
  24292. #define vsetq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  24293.   float32x4_t __ret; \
  24294.   float32_t __s0 = __p0; \
  24295.   float32x4_t __s1 = __p1; \
  24296.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  24297.   __ret = (float32x4_t) __builtin_neon_vsetq_lane_f32(__s0, (float32x4_t)__rev1, __p2); \
  24298.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  24299.   __ret; \
  24300. })
  24301. #define __noswap_vsetq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  24302.   float32x4_t __ret; \
  24303.   float32_t __s0 = __p0; \
  24304.   float32x4_t __s1 = __p1; \
  24305.   __ret = (float32x4_t) __builtin_neon_vsetq_lane_f32(__s0, (float32x4_t)__s1, __p2); \
  24306.   __ret; \
  24307. })
  24308. #endif
  24309.  
  24310. #ifdef __LITTLE_ENDIAN__
  24311. #define vsetq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  24312.   int32x4_t __ret; \
  24313.   int32_t __s0 = __p0; \
  24314.   int32x4_t __s1 = __p1; \
  24315.   __ret = (int32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int32x4_t)__s1, __p2); \
  24316.   __ret; \
  24317. })
  24318. #else
  24319. #define vsetq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  24320.   int32x4_t __ret; \
  24321.   int32_t __s0 = __p0; \
  24322.   int32x4_t __s1 = __p1; \
  24323.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  24324.   __ret = (int32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int32x4_t)__rev1, __p2); \
  24325.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  24326.   __ret; \
  24327. })
  24328. #define __noswap_vsetq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  24329.   int32x4_t __ret; \
  24330.   int32_t __s0 = __p0; \
  24331.   int32x4_t __s1 = __p1; \
  24332.   __ret = (int32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int32x4_t)__s1, __p2); \
  24333.   __ret; \
  24334. })
  24335. #endif
  24336.  
  24337. #ifdef __LITTLE_ENDIAN__
  24338. #define vsetq_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  24339.   int64x2_t __ret; \
  24340.   int64_t __s0 = __p0; \
  24341.   int64x2_t __s1 = __p1; \
  24342.   __ret = (int64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int64x2_t)__s1, __p2); \
  24343.   __ret; \
  24344. })
  24345. #else
  24346. #define vsetq_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  24347.   int64x2_t __ret; \
  24348.   int64_t __s0 = __p0; \
  24349.   int64x2_t __s1 = __p1; \
  24350.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  24351.   __ret = (int64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int64x2_t)__rev1, __p2); \
  24352.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  24353.   __ret; \
  24354. })
  24355. #define __noswap_vsetq_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  24356.   int64x2_t __ret; \
  24357.   int64_t __s0 = __p0; \
  24358.   int64x2_t __s1 = __p1; \
  24359.   __ret = (int64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int64x2_t)__s1, __p2); \
  24360.   __ret; \
  24361. })
  24362. #endif
  24363.  
  24364. #ifdef __LITTLE_ENDIAN__
  24365. #define vsetq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  24366.   int16x8_t __ret; \
  24367.   int16_t __s0 = __p0; \
  24368.   int16x8_t __s1 = __p1; \
  24369.   __ret = (int16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int16x8_t)__s1, __p2); \
  24370.   __ret; \
  24371. })
  24372. #else
  24373. #define vsetq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  24374.   int16x8_t __ret; \
  24375.   int16_t __s0 = __p0; \
  24376.   int16x8_t __s1 = __p1; \
  24377.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  24378.   __ret = (int16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int16x8_t)__rev1, __p2); \
  24379.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  24380.   __ret; \
  24381. })
  24382. #define __noswap_vsetq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  24383.   int16x8_t __ret; \
  24384.   int16_t __s0 = __p0; \
  24385.   int16x8_t __s1 = __p1; \
  24386.   __ret = (int16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int16x8_t)__s1, __p2); \
  24387.   __ret; \
  24388. })
  24389. #endif
  24390.  
  24391. #ifdef __LITTLE_ENDIAN__
  24392. #define vset_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  24393.   uint8x8_t __ret; \
  24394.   uint8_t __s0 = __p0; \
  24395.   uint8x8_t __s1 = __p1; \
  24396.   __ret = (uint8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
  24397.   __ret; \
  24398. })
  24399. #else
  24400. #define vset_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  24401.   uint8x8_t __ret; \
  24402.   uint8_t __s0 = __p0; \
  24403.   uint8x8_t __s1 = __p1; \
  24404.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  24405.   __ret = (uint8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__rev1, __p2); \
  24406.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  24407.   __ret; \
  24408. })
  24409. #define __noswap_vset_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  24410.   uint8x8_t __ret; \
  24411.   uint8_t __s0 = __p0; \
  24412.   uint8x8_t __s1 = __p1; \
  24413.   __ret = (uint8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
  24414.   __ret; \
  24415. })
  24416. #endif
  24417.  
  24418. #ifdef __LITTLE_ENDIAN__
  24419. #define vset_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  24420.   uint32x2_t __ret; \
  24421.   uint32_t __s0 = __p0; \
  24422.   uint32x2_t __s1 = __p1; \
  24423.   __ret = (uint32x2_t) __builtin_neon_vset_lane_i32(__s0, (int32x2_t)__s1, __p2); \
  24424.   __ret; \
  24425. })
  24426. #else
  24427. #define vset_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  24428.   uint32x2_t __ret; \
  24429.   uint32_t __s0 = __p0; \
  24430.   uint32x2_t __s1 = __p1; \
  24431.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  24432.   __ret = (uint32x2_t) __builtin_neon_vset_lane_i32(__s0, (int32x2_t)__rev1, __p2); \
  24433.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  24434.   __ret; \
  24435. })
  24436. #define __noswap_vset_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  24437.   uint32x2_t __ret; \
  24438.   uint32_t __s0 = __p0; \
  24439.   uint32x2_t __s1 = __p1; \
  24440.   __ret = (uint32x2_t) __builtin_neon_vset_lane_i32(__s0, (int32x2_t)__s1, __p2); \
  24441.   __ret; \
  24442. })
  24443. #endif
  24444.  
  24445. #define vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  24446.   uint64x1_t __ret; \
  24447.   uint64_t __s0 = __p0; \
  24448.   uint64x1_t __s1 = __p1; \
  24449.   __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int64x1_t)__s1, __p2); \
  24450.   __ret; \
  24451. })
  24452. #ifdef __LITTLE_ENDIAN__
  24453. #define vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  24454.   uint16x4_t __ret; \
  24455.   uint16_t __s0 = __p0; \
  24456.   uint16x4_t __s1 = __p1; \
  24457.   __ret = (uint16x4_t) __builtin_neon_vset_lane_i16(__s0, (int16x4_t)__s1, __p2); \
  24458.   __ret; \
  24459. })
  24460. #else
  24461. #define vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  24462.   uint16x4_t __ret; \
  24463.   uint16_t __s0 = __p0; \
  24464.   uint16x4_t __s1 = __p1; \
  24465.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  24466.   __ret = (uint16x4_t) __builtin_neon_vset_lane_i16(__s0, (int16x4_t)__rev1, __p2); \
  24467.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  24468.   __ret; \
  24469. })
  24470. #define __noswap_vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  24471.   uint16x4_t __ret; \
  24472.   uint16_t __s0 = __p0; \
  24473.   uint16x4_t __s1 = __p1; \
  24474.   __ret = (uint16x4_t) __builtin_neon_vset_lane_i16(__s0, (int16x4_t)__s1, __p2); \
  24475.   __ret; \
  24476. })
  24477. #endif
  24478.  
  24479. #ifdef __LITTLE_ENDIAN__
  24480. #define vset_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  24481.   int8x8_t __ret; \
  24482.   int8_t __s0 = __p0; \
  24483.   int8x8_t __s1 = __p1; \
  24484.   __ret = (int8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
  24485.   __ret; \
  24486. })
  24487. #else
  24488. #define vset_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  24489.   int8x8_t __ret; \
  24490.   int8_t __s0 = __p0; \
  24491.   int8x8_t __s1 = __p1; \
  24492.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  24493.   __ret = (int8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__rev1, __p2); \
  24494.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  24495.   __ret; \
  24496. })
  24497. #define __noswap_vset_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  24498.   int8x8_t __ret; \
  24499.   int8_t __s0 = __p0; \
  24500.   int8x8_t __s1 = __p1; \
  24501.   __ret = (int8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
  24502.   __ret; \
  24503. })
  24504. #endif
  24505.  
  24506. #ifdef __LITTLE_ENDIAN__
  24507. #define vset_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  24508.   float32x2_t __ret; \
  24509.   float32_t __s0 = __p0; \
  24510.   float32x2_t __s1 = __p1; \
  24511.   __ret = (float32x2_t) __builtin_neon_vset_lane_f32(__s0, (float32x2_t)__s1, __p2); \
  24512.   __ret; \
  24513. })
  24514. #else
  24515. #define vset_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  24516.   float32x2_t __ret; \
  24517.   float32_t __s0 = __p0; \
  24518.   float32x2_t __s1 = __p1; \
  24519.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  24520.   __ret = (float32x2_t) __builtin_neon_vset_lane_f32(__s0, (float32x2_t)__rev1, __p2); \
  24521.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  24522.   __ret; \
  24523. })
  24524. #define __noswap_vset_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  24525.   float32x2_t __ret; \
  24526.   float32_t __s0 = __p0; \
  24527.   float32x2_t __s1 = __p1; \
  24528.   __ret = (float32x2_t) __builtin_neon_vset_lane_f32(__s0, (float32x2_t)__s1, __p2); \
  24529.   __ret; \
  24530. })
  24531. #endif
  24532.  
  24533. #ifdef __LITTLE_ENDIAN__
  24534. #define vset_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  24535.   int32x2_t __ret; \
  24536.   int32_t __s0 = __p0; \
  24537.   int32x2_t __s1 = __p1; \
  24538.   __ret = (int32x2_t) __builtin_neon_vset_lane_i32(__s0, (int32x2_t)__s1, __p2); \
  24539.   __ret; \
  24540. })
  24541. #else
  24542. #define vset_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  24543.   int32x2_t __ret; \
  24544.   int32_t __s0 = __p0; \
  24545.   int32x2_t __s1 = __p1; \
  24546.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  24547.   __ret = (int32x2_t) __builtin_neon_vset_lane_i32(__s0, (int32x2_t)__rev1, __p2); \
  24548.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  24549.   __ret; \
  24550. })
  24551. #define __noswap_vset_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  24552.   int32x2_t __ret; \
  24553.   int32_t __s0 = __p0; \
  24554.   int32x2_t __s1 = __p1; \
  24555.   __ret = (int32x2_t) __builtin_neon_vset_lane_i32(__s0, (int32x2_t)__s1, __p2); \
  24556.   __ret; \
  24557. })
  24558. #endif
  24559.  
  24560. #define vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  24561.   int64x1_t __ret; \
  24562.   int64_t __s0 = __p0; \
  24563.   int64x1_t __s1 = __p1; \
  24564.   __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int64x1_t)__s1, __p2); \
  24565.   __ret; \
  24566. })
  24567. #ifdef __LITTLE_ENDIAN__
  24568. #define vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  24569.   int16x4_t __ret; \
  24570.   int16_t __s0 = __p0; \
  24571.   int16x4_t __s1 = __p1; \
  24572.   __ret = (int16x4_t) __builtin_neon_vset_lane_i16(__s0, (int16x4_t)__s1, __p2); \
  24573.   __ret; \
  24574. })
  24575. #else
  24576. #define vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  24577.   int16x4_t __ret; \
  24578.   int16_t __s0 = __p0; \
  24579.   int16x4_t __s1 = __p1; \
  24580.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  24581.   __ret = (int16x4_t) __builtin_neon_vset_lane_i16(__s0, (int16x4_t)__rev1, __p2); \
  24582.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  24583.   __ret; \
  24584. })
  24585. #define __noswap_vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  24586.   int16x4_t __ret; \
  24587.   int16_t __s0 = __p0; \
  24588.   int16x4_t __s1 = __p1; \
  24589.   __ret = (int16x4_t) __builtin_neon_vset_lane_i16(__s0, (int16x4_t)__s1, __p2); \
  24590.   __ret; \
  24591. })
  24592. #endif
  24593.  
  24594. #ifdef __LITTLE_ENDIAN__
  24595. __ai uint8x16_t vshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
  24596.   uint8x16_t __ret;
  24597.   __ret = (uint8x16_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  24598.   return __ret;
  24599. }
  24600. #else
  24601. __ai uint8x16_t vshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
  24602.   uint8x16_t __ret;
  24603.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  24604.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  24605.   __ret = (uint8x16_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  24606.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  24607.   return __ret;
  24608. }
  24609. #endif
  24610.  
  24611. #ifdef __LITTLE_ENDIAN__
  24612. __ai uint32x4_t vshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
  24613.   uint32x4_t __ret;
  24614.   __ret = (uint32x4_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  24615.   return __ret;
  24616. }
  24617. #else
  24618. __ai uint32x4_t vshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
  24619.   uint32x4_t __ret;
  24620.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  24621.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  24622.   __ret = (uint32x4_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  24623.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  24624.   return __ret;
  24625. }
  24626. #endif
  24627.  
  24628. #ifdef __LITTLE_ENDIAN__
  24629. __ai uint64x2_t vshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
  24630.   uint64x2_t __ret;
  24631.   __ret = (uint64x2_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  24632.   return __ret;
  24633. }
  24634. #else
  24635. __ai uint64x2_t vshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
  24636.   uint64x2_t __ret;
  24637.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  24638.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  24639.   __ret = (uint64x2_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  24640.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  24641.   return __ret;
  24642. }
  24643. #endif
  24644.  
  24645. #ifdef __LITTLE_ENDIAN__
  24646. __ai uint16x8_t vshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
  24647.   uint16x8_t __ret;
  24648.   __ret = (uint16x8_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  24649.   return __ret;
  24650. }
  24651. #else
  24652. __ai uint16x8_t vshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
  24653.   uint16x8_t __ret;
  24654.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  24655.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  24656.   __ret = (uint16x8_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  24657.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  24658.   return __ret;
  24659. }
  24660. #endif
  24661.  
  24662. #ifdef __LITTLE_ENDIAN__
  24663. __ai int8x16_t vshlq_s8(int8x16_t __p0, int8x16_t __p1) {
  24664.   int8x16_t __ret;
  24665.   __ret = (int8x16_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  24666.   return __ret;
  24667. }
  24668. #else
  24669. __ai int8x16_t vshlq_s8(int8x16_t __p0, int8x16_t __p1) {
  24670.   int8x16_t __ret;
  24671.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  24672.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  24673.   __ret = (int8x16_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  24674.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  24675.   return __ret;
  24676. }
  24677. #endif
  24678.  
  24679. #ifdef __LITTLE_ENDIAN__
  24680. __ai int32x4_t vshlq_s32(int32x4_t __p0, int32x4_t __p1) {
  24681.   int32x4_t __ret;
  24682.   __ret = (int32x4_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  24683.   return __ret;
  24684. }
  24685. #else
  24686. __ai int32x4_t vshlq_s32(int32x4_t __p0, int32x4_t __p1) {
  24687.   int32x4_t __ret;
  24688.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  24689.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  24690.   __ret = (int32x4_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  24691.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  24692.   return __ret;
  24693. }
  24694. #endif
  24695.  
  24696. #ifdef __LITTLE_ENDIAN__
  24697. __ai int64x2_t vshlq_s64(int64x2_t __p0, int64x2_t __p1) {
  24698.   int64x2_t __ret;
  24699.   __ret = (int64x2_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
  24700.   return __ret;
  24701. }
  24702. #else
  24703. __ai int64x2_t vshlq_s64(int64x2_t __p0, int64x2_t __p1) {
  24704.   int64x2_t __ret;
  24705.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  24706.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  24707.   __ret = (int64x2_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
  24708.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  24709.   return __ret;
  24710. }
  24711. #endif
  24712.  
  24713. #ifdef __LITTLE_ENDIAN__
  24714. __ai int16x8_t vshlq_s16(int16x8_t __p0, int16x8_t __p1) {
  24715.   int16x8_t __ret;
  24716.   __ret = (int16x8_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  24717.   return __ret;
  24718. }
  24719. #else
  24720. __ai int16x8_t vshlq_s16(int16x8_t __p0, int16x8_t __p1) {
  24721.   int16x8_t __ret;
  24722.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  24723.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  24724.   __ret = (int16x8_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  24725.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  24726.   return __ret;
  24727. }
  24728. #endif
  24729.  
  24730. #ifdef __LITTLE_ENDIAN__
  24731. __ai uint8x8_t vshl_u8(uint8x8_t __p0, int8x8_t __p1) {
  24732.   uint8x8_t __ret;
  24733.   __ret = (uint8x8_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  24734.   return __ret;
  24735. }
  24736. #else
  24737. __ai uint8x8_t vshl_u8(uint8x8_t __p0, int8x8_t __p1) {
  24738.   uint8x8_t __ret;
  24739.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  24740.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  24741.   __ret = (uint8x8_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  24742.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  24743.   return __ret;
  24744. }
  24745. #endif
  24746.  
  24747. #ifdef __LITTLE_ENDIAN__
  24748. __ai uint32x2_t vshl_u32(uint32x2_t __p0, int32x2_t __p1) {
  24749.   uint32x2_t __ret;
  24750.   __ret = (uint32x2_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  24751.   return __ret;
  24752. }
  24753. #else
  24754. __ai uint32x2_t vshl_u32(uint32x2_t __p0, int32x2_t __p1) {
  24755.   uint32x2_t __ret;
  24756.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  24757.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  24758.   __ret = (uint32x2_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  24759.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  24760.   return __ret;
  24761. }
  24762. #endif
  24763.  
  24764. __ai uint64x1_t vshl_u64(uint64x1_t __p0, int64x1_t __p1) {
  24765.   uint64x1_t __ret;
  24766.   __ret = (uint64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  24767.   return __ret;
  24768. }
  24769. #ifdef __LITTLE_ENDIAN__
  24770. __ai uint16x4_t vshl_u16(uint16x4_t __p0, int16x4_t __p1) {
  24771.   uint16x4_t __ret;
  24772.   __ret = (uint16x4_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  24773.   return __ret;
  24774. }
  24775. #else
  24776. __ai uint16x4_t vshl_u16(uint16x4_t __p0, int16x4_t __p1) {
  24777.   uint16x4_t __ret;
  24778.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  24779.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  24780.   __ret = (uint16x4_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  24781.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  24782.   return __ret;
  24783. }
  24784. #endif
  24785.  
  24786. #ifdef __LITTLE_ENDIAN__
  24787. __ai int8x8_t vshl_s8(int8x8_t __p0, int8x8_t __p1) {
  24788.   int8x8_t __ret;
  24789.   __ret = (int8x8_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  24790.   return __ret;
  24791. }
  24792. #else
  24793. __ai int8x8_t vshl_s8(int8x8_t __p0, int8x8_t __p1) {
  24794.   int8x8_t __ret;
  24795.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  24796.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  24797.   __ret = (int8x8_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  24798.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  24799.   return __ret;
  24800. }
  24801. #endif
  24802.  
  24803. #ifdef __LITTLE_ENDIAN__
  24804. __ai int32x2_t vshl_s32(int32x2_t __p0, int32x2_t __p1) {
  24805.   int32x2_t __ret;
  24806.   __ret = (int32x2_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  24807.   return __ret;
  24808. }
  24809. #else
  24810. __ai int32x2_t vshl_s32(int32x2_t __p0, int32x2_t __p1) {
  24811.   int32x2_t __ret;
  24812.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  24813.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  24814.   __ret = (int32x2_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  24815.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  24816.   return __ret;
  24817. }
  24818. #endif
  24819.  
  24820. __ai int64x1_t vshl_s64(int64x1_t __p0, int64x1_t __p1) {
  24821.   int64x1_t __ret;
  24822.   __ret = (int64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
  24823.   return __ret;
  24824. }
  24825. #ifdef __LITTLE_ENDIAN__
  24826. __ai int16x4_t vshl_s16(int16x4_t __p0, int16x4_t __p1) {
  24827.   int16x4_t __ret;
  24828.   __ret = (int16x4_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  24829.   return __ret;
  24830. }
  24831. #else
  24832. __ai int16x4_t vshl_s16(int16x4_t __p0, int16x4_t __p1) {
  24833.   int16x4_t __ret;
  24834.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  24835.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  24836.   __ret = (int16x4_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  24837.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  24838.   return __ret;
  24839. }
  24840. #endif
  24841.  
  24842. #ifdef __LITTLE_ENDIAN__
  24843. #define vshlq_n_u8(__p0, __p1) __extension__ ({ \
  24844.   uint8x16_t __ret; \
  24845.   uint8x16_t __s0 = __p0; \
  24846.   __ret = (uint8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 48); \
  24847.   __ret; \
  24848. })
  24849. #else
  24850. #define vshlq_n_u8(__p0, __p1) __extension__ ({ \
  24851.   uint8x16_t __ret; \
  24852.   uint8x16_t __s0 = __p0; \
  24853.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24854.   __ret = (uint8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 48); \
  24855.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24856.   __ret; \
  24857. })
  24858. #endif
  24859.  
  24860. #ifdef __LITTLE_ENDIAN__
  24861. #define vshlq_n_u32(__p0, __p1) __extension__ ({ \
  24862.   uint32x4_t __ret; \
  24863.   uint32x4_t __s0 = __p0; \
  24864.   __ret = (uint32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 50); \
  24865.   __ret; \
  24866. })
  24867. #else
  24868. #define vshlq_n_u32(__p0, __p1) __extension__ ({ \
  24869.   uint32x4_t __ret; \
  24870.   uint32x4_t __s0 = __p0; \
  24871.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  24872.   __ret = (uint32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 50); \
  24873.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  24874.   __ret; \
  24875. })
  24876. #endif
  24877.  
  24878. #ifdef __LITTLE_ENDIAN__
  24879. #define vshlq_n_u64(__p0, __p1) __extension__ ({ \
  24880.   uint64x2_t __ret; \
  24881.   uint64x2_t __s0 = __p0; \
  24882.   __ret = (uint64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 51); \
  24883.   __ret; \
  24884. })
  24885. #else
  24886. #define vshlq_n_u64(__p0, __p1) __extension__ ({ \
  24887.   uint64x2_t __ret; \
  24888.   uint64x2_t __s0 = __p0; \
  24889.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  24890.   __ret = (uint64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 51); \
  24891.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  24892.   __ret; \
  24893. })
  24894. #endif
  24895.  
  24896. #ifdef __LITTLE_ENDIAN__
  24897. #define vshlq_n_u16(__p0, __p1) __extension__ ({ \
  24898.   uint16x8_t __ret; \
  24899.   uint16x8_t __s0 = __p0; \
  24900.   __ret = (uint16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 49); \
  24901.   __ret; \
  24902. })
  24903. #else
  24904. #define vshlq_n_u16(__p0, __p1) __extension__ ({ \
  24905.   uint16x8_t __ret; \
  24906.   uint16x8_t __s0 = __p0; \
  24907.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  24908.   __ret = (uint16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 49); \
  24909.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  24910.   __ret; \
  24911. })
  24912. #endif
  24913.  
  24914. #ifdef __LITTLE_ENDIAN__
  24915. #define vshlq_n_s8(__p0, __p1) __extension__ ({ \
  24916.   int8x16_t __ret; \
  24917.   int8x16_t __s0 = __p0; \
  24918.   __ret = (int8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 32); \
  24919.   __ret; \
  24920. })
  24921. #else
  24922. #define vshlq_n_s8(__p0, __p1) __extension__ ({ \
  24923.   int8x16_t __ret; \
  24924.   int8x16_t __s0 = __p0; \
  24925.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24926.   __ret = (int8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 32); \
  24927.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  24928.   __ret; \
  24929. })
  24930. #endif
  24931.  
  24932. #ifdef __LITTLE_ENDIAN__
  24933. #define vshlq_n_s32(__p0, __p1) __extension__ ({ \
  24934.   int32x4_t __ret; \
  24935.   int32x4_t __s0 = __p0; \
  24936.   __ret = (int32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 34); \
  24937.   __ret; \
  24938. })
  24939. #else
  24940. #define vshlq_n_s32(__p0, __p1) __extension__ ({ \
  24941.   int32x4_t __ret; \
  24942.   int32x4_t __s0 = __p0; \
  24943.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  24944.   __ret = (int32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 34); \
  24945.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  24946.   __ret; \
  24947. })
  24948. #endif
  24949.  
  24950. #ifdef __LITTLE_ENDIAN__
  24951. #define vshlq_n_s64(__p0, __p1) __extension__ ({ \
  24952.   int64x2_t __ret; \
  24953.   int64x2_t __s0 = __p0; \
  24954.   __ret = (int64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 35); \
  24955.   __ret; \
  24956. })
  24957. #else
  24958. #define vshlq_n_s64(__p0, __p1) __extension__ ({ \
  24959.   int64x2_t __ret; \
  24960.   int64x2_t __s0 = __p0; \
  24961.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  24962.   __ret = (int64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 35); \
  24963.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  24964.   __ret; \
  24965. })
  24966. #endif
  24967.  
  24968. #ifdef __LITTLE_ENDIAN__
  24969. #define vshlq_n_s16(__p0, __p1) __extension__ ({ \
  24970.   int16x8_t __ret; \
  24971.   int16x8_t __s0 = __p0; \
  24972.   __ret = (int16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 33); \
  24973.   __ret; \
  24974. })
  24975. #else
  24976. #define vshlq_n_s16(__p0, __p1) __extension__ ({ \
  24977.   int16x8_t __ret; \
  24978.   int16x8_t __s0 = __p0; \
  24979.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  24980.   __ret = (int16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 33); \
  24981.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  24982.   __ret; \
  24983. })
  24984. #endif
  24985.  
  24986. #ifdef __LITTLE_ENDIAN__
  24987. #define vshl_n_u8(__p0, __p1) __extension__ ({ \
  24988.   uint8x8_t __ret; \
  24989.   uint8x8_t __s0 = __p0; \
  24990.   __ret = (uint8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 16); \
  24991.   __ret; \
  24992. })
  24993. #else
  24994. #define vshl_n_u8(__p0, __p1) __extension__ ({ \
  24995.   uint8x8_t __ret; \
  24996.   uint8x8_t __s0 = __p0; \
  24997.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  24998.   __ret = (uint8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 16); \
  24999.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25000.   __ret; \
  25001. })
  25002. #endif
  25003.  
  25004. #ifdef __LITTLE_ENDIAN__
  25005. #define vshl_n_u32(__p0, __p1) __extension__ ({ \
  25006.   uint32x2_t __ret; \
  25007.   uint32x2_t __s0 = __p0; \
  25008.   __ret = (uint32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 18); \
  25009.   __ret; \
  25010. })
  25011. #else
  25012. #define vshl_n_u32(__p0, __p1) __extension__ ({ \
  25013.   uint32x2_t __ret; \
  25014.   uint32x2_t __s0 = __p0; \
  25015.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25016.   __ret = (uint32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 18); \
  25017.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25018.   __ret; \
  25019. })
  25020. #endif
  25021.  
  25022. #define vshl_n_u64(__p0, __p1) __extension__ ({ \
  25023.   uint64x1_t __ret; \
  25024.   uint64x1_t __s0 = __p0; \
  25025.   __ret = (uint64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 19); \
  25026.   __ret; \
  25027. })
  25028. #ifdef __LITTLE_ENDIAN__
  25029. #define vshl_n_u16(__p0, __p1) __extension__ ({ \
  25030.   uint16x4_t __ret; \
  25031.   uint16x4_t __s0 = __p0; \
  25032.   __ret = (uint16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 17); \
  25033.   __ret; \
  25034. })
  25035. #else
  25036. #define vshl_n_u16(__p0, __p1) __extension__ ({ \
  25037.   uint16x4_t __ret; \
  25038.   uint16x4_t __s0 = __p0; \
  25039.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25040.   __ret = (uint16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 17); \
  25041.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25042.   __ret; \
  25043. })
  25044. #endif
  25045.  
  25046. #ifdef __LITTLE_ENDIAN__
  25047. #define vshl_n_s8(__p0, __p1) __extension__ ({ \
  25048.   int8x8_t __ret; \
  25049.   int8x8_t __s0 = __p0; \
  25050.   __ret = (int8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 0); \
  25051.   __ret; \
  25052. })
  25053. #else
  25054. #define vshl_n_s8(__p0, __p1) __extension__ ({ \
  25055.   int8x8_t __ret; \
  25056.   int8x8_t __s0 = __p0; \
  25057.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25058.   __ret = (int8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 0); \
  25059.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25060.   __ret; \
  25061. })
  25062. #endif
  25063.  
  25064. #ifdef __LITTLE_ENDIAN__
  25065. #define vshl_n_s32(__p0, __p1) __extension__ ({ \
  25066.   int32x2_t __ret; \
  25067.   int32x2_t __s0 = __p0; \
  25068.   __ret = (int32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 2); \
  25069.   __ret; \
  25070. })
  25071. #else
  25072. #define vshl_n_s32(__p0, __p1) __extension__ ({ \
  25073.   int32x2_t __ret; \
  25074.   int32x2_t __s0 = __p0; \
  25075.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25076.   __ret = (int32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 2); \
  25077.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25078.   __ret; \
  25079. })
  25080. #endif
  25081.  
  25082. #define vshl_n_s64(__p0, __p1) __extension__ ({ \
  25083.   int64x1_t __ret; \
  25084.   int64x1_t __s0 = __p0; \
  25085.   __ret = (int64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 3); \
  25086.   __ret; \
  25087. })
  25088. #ifdef __LITTLE_ENDIAN__
  25089. #define vshl_n_s16(__p0, __p1) __extension__ ({ \
  25090.   int16x4_t __ret; \
  25091.   int16x4_t __s0 = __p0; \
  25092.   __ret = (int16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 1); \
  25093.   __ret; \
  25094. })
  25095. #else
  25096. #define vshl_n_s16(__p0, __p1) __extension__ ({ \
  25097.   int16x4_t __ret; \
  25098.   int16x4_t __s0 = __p0; \
  25099.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25100.   __ret = (int16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 1); \
  25101.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25102.   __ret; \
  25103. })
  25104. #endif
  25105.  
  25106. #ifdef __LITTLE_ENDIAN__
  25107. #define vshll_n_u8(__p0, __p1) __extension__ ({ \
  25108.   uint16x8_t __ret; \
  25109.   uint8x8_t __s0 = __p0; \
  25110.   __ret = (uint16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 49); \
  25111.   __ret; \
  25112. })
  25113. #else
  25114. #define vshll_n_u8(__p0, __p1) __extension__ ({ \
  25115.   uint16x8_t __ret; \
  25116.   uint8x8_t __s0 = __p0; \
  25117.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25118.   __ret = (uint16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 49); \
  25119.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25120.   __ret; \
  25121. })
  25122. #define __noswap_vshll_n_u8(__p0, __p1) __extension__ ({ \
  25123.   uint16x8_t __ret; \
  25124.   uint8x8_t __s0 = __p0; \
  25125.   __ret = (uint16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 49); \
  25126.   __ret; \
  25127. })
  25128. #endif
  25129.  
  25130. #ifdef __LITTLE_ENDIAN__
  25131. #define vshll_n_u32(__p0, __p1) __extension__ ({ \
  25132.   uint64x2_t __ret; \
  25133.   uint32x2_t __s0 = __p0; \
  25134.   __ret = (uint64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 51); \
  25135.   __ret; \
  25136. })
  25137. #else
  25138. #define vshll_n_u32(__p0, __p1) __extension__ ({ \
  25139.   uint64x2_t __ret; \
  25140.   uint32x2_t __s0 = __p0; \
  25141.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25142.   __ret = (uint64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 51); \
  25143.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25144.   __ret; \
  25145. })
  25146. #define __noswap_vshll_n_u32(__p0, __p1) __extension__ ({ \
  25147.   uint64x2_t __ret; \
  25148.   uint32x2_t __s0 = __p0; \
  25149.   __ret = (uint64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 51); \
  25150.   __ret; \
  25151. })
  25152. #endif
  25153.  
  25154. #ifdef __LITTLE_ENDIAN__
  25155. #define vshll_n_u16(__p0, __p1) __extension__ ({ \
  25156.   uint32x4_t __ret; \
  25157.   uint16x4_t __s0 = __p0; \
  25158.   __ret = (uint32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 50); \
  25159.   __ret; \
  25160. })
  25161. #else
  25162. #define vshll_n_u16(__p0, __p1) __extension__ ({ \
  25163.   uint32x4_t __ret; \
  25164.   uint16x4_t __s0 = __p0; \
  25165.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25166.   __ret = (uint32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 50); \
  25167.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25168.   __ret; \
  25169. })
  25170. #define __noswap_vshll_n_u16(__p0, __p1) __extension__ ({ \
  25171.   uint32x4_t __ret; \
  25172.   uint16x4_t __s0 = __p0; \
  25173.   __ret = (uint32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 50); \
  25174.   __ret; \
  25175. })
  25176. #endif
  25177.  
  25178. #ifdef __LITTLE_ENDIAN__
  25179. #define vshll_n_s8(__p0, __p1) __extension__ ({ \
  25180.   int16x8_t __ret; \
  25181.   int8x8_t __s0 = __p0; \
  25182.   __ret = (int16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 33); \
  25183.   __ret; \
  25184. })
  25185. #else
  25186. #define vshll_n_s8(__p0, __p1) __extension__ ({ \
  25187.   int16x8_t __ret; \
  25188.   int8x8_t __s0 = __p0; \
  25189.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25190.   __ret = (int16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 33); \
  25191.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25192.   __ret; \
  25193. })
  25194. #define __noswap_vshll_n_s8(__p0, __p1) __extension__ ({ \
  25195.   int16x8_t __ret; \
  25196.   int8x8_t __s0 = __p0; \
  25197.   __ret = (int16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 33); \
  25198.   __ret; \
  25199. })
  25200. #endif
  25201.  
  25202. #ifdef __LITTLE_ENDIAN__
  25203. #define vshll_n_s32(__p0, __p1) __extension__ ({ \
  25204.   int64x2_t __ret; \
  25205.   int32x2_t __s0 = __p0; \
  25206.   __ret = (int64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 35); \
  25207.   __ret; \
  25208. })
  25209. #else
  25210. #define vshll_n_s32(__p0, __p1) __extension__ ({ \
  25211.   int64x2_t __ret; \
  25212.   int32x2_t __s0 = __p0; \
  25213.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25214.   __ret = (int64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 35); \
  25215.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25216.   __ret; \
  25217. })
  25218. #define __noswap_vshll_n_s32(__p0, __p1) __extension__ ({ \
  25219.   int64x2_t __ret; \
  25220.   int32x2_t __s0 = __p0; \
  25221.   __ret = (int64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 35); \
  25222.   __ret; \
  25223. })
  25224. #endif
  25225.  
  25226. #ifdef __LITTLE_ENDIAN__
  25227. #define vshll_n_s16(__p0, __p1) __extension__ ({ \
  25228.   int32x4_t __ret; \
  25229.   int16x4_t __s0 = __p0; \
  25230.   __ret = (int32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 34); \
  25231.   __ret; \
  25232. })
  25233. #else
  25234. #define vshll_n_s16(__p0, __p1) __extension__ ({ \
  25235.   int32x4_t __ret; \
  25236.   int16x4_t __s0 = __p0; \
  25237.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25238.   __ret = (int32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 34); \
  25239.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25240.   __ret; \
  25241. })
  25242. #define __noswap_vshll_n_s16(__p0, __p1) __extension__ ({ \
  25243.   int32x4_t __ret; \
  25244.   int16x4_t __s0 = __p0; \
  25245.   __ret = (int32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 34); \
  25246.   __ret; \
  25247. })
  25248. #endif
  25249.  
  25250. #ifdef __LITTLE_ENDIAN__
  25251. #define vshrq_n_u8(__p0, __p1) __extension__ ({ \
  25252.   uint8x16_t __ret; \
  25253.   uint8x16_t __s0 = __p0; \
  25254.   __ret = (uint8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 48); \
  25255.   __ret; \
  25256. })
  25257. #else
  25258. #define vshrq_n_u8(__p0, __p1) __extension__ ({ \
  25259.   uint8x16_t __ret; \
  25260.   uint8x16_t __s0 = __p0; \
  25261.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25262.   __ret = (uint8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 48); \
  25263.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25264.   __ret; \
  25265. })
  25266. #endif
  25267.  
  25268. #ifdef __LITTLE_ENDIAN__
  25269. #define vshrq_n_u32(__p0, __p1) __extension__ ({ \
  25270.   uint32x4_t __ret; \
  25271.   uint32x4_t __s0 = __p0; \
  25272.   __ret = (uint32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 50); \
  25273.   __ret; \
  25274. })
  25275. #else
  25276. #define vshrq_n_u32(__p0, __p1) __extension__ ({ \
  25277.   uint32x4_t __ret; \
  25278.   uint32x4_t __s0 = __p0; \
  25279.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25280.   __ret = (uint32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 50); \
  25281.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25282.   __ret; \
  25283. })
  25284. #endif
  25285.  
  25286. #ifdef __LITTLE_ENDIAN__
  25287. #define vshrq_n_u64(__p0, __p1) __extension__ ({ \
  25288.   uint64x2_t __ret; \
  25289.   uint64x2_t __s0 = __p0; \
  25290.   __ret = (uint64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 51); \
  25291.   __ret; \
  25292. })
  25293. #else
  25294. #define vshrq_n_u64(__p0, __p1) __extension__ ({ \
  25295.   uint64x2_t __ret; \
  25296.   uint64x2_t __s0 = __p0; \
  25297.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25298.   __ret = (uint64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 51); \
  25299.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25300.   __ret; \
  25301. })
  25302. #endif
  25303.  
  25304. #ifdef __LITTLE_ENDIAN__
  25305. #define vshrq_n_u16(__p0, __p1) __extension__ ({ \
  25306.   uint16x8_t __ret; \
  25307.   uint16x8_t __s0 = __p0; \
  25308.   __ret = (uint16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 49); \
  25309.   __ret; \
  25310. })
  25311. #else
  25312. #define vshrq_n_u16(__p0, __p1) __extension__ ({ \
  25313.   uint16x8_t __ret; \
  25314.   uint16x8_t __s0 = __p0; \
  25315.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25316.   __ret = (uint16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 49); \
  25317.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25318.   __ret; \
  25319. })
  25320. #endif
  25321.  
  25322. #ifdef __LITTLE_ENDIAN__
  25323. #define vshrq_n_s8(__p0, __p1) __extension__ ({ \
  25324.   int8x16_t __ret; \
  25325.   int8x16_t __s0 = __p0; \
  25326.   __ret = (int8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 32); \
  25327.   __ret; \
  25328. })
  25329. #else
  25330. #define vshrq_n_s8(__p0, __p1) __extension__ ({ \
  25331.   int8x16_t __ret; \
  25332.   int8x16_t __s0 = __p0; \
  25333.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25334.   __ret = (int8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 32); \
  25335.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25336.   __ret; \
  25337. })
  25338. #endif
  25339.  
  25340. #ifdef __LITTLE_ENDIAN__
  25341. #define vshrq_n_s32(__p0, __p1) __extension__ ({ \
  25342.   int32x4_t __ret; \
  25343.   int32x4_t __s0 = __p0; \
  25344.   __ret = (int32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 34); \
  25345.   __ret; \
  25346. })
  25347. #else
  25348. #define vshrq_n_s32(__p0, __p1) __extension__ ({ \
  25349.   int32x4_t __ret; \
  25350.   int32x4_t __s0 = __p0; \
  25351.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25352.   __ret = (int32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 34); \
  25353.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25354.   __ret; \
  25355. })
  25356. #endif
  25357.  
  25358. #ifdef __LITTLE_ENDIAN__
  25359. #define vshrq_n_s64(__p0, __p1) __extension__ ({ \
  25360.   int64x2_t __ret; \
  25361.   int64x2_t __s0 = __p0; \
  25362.   __ret = (int64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 35); \
  25363.   __ret; \
  25364. })
  25365. #else
  25366. #define vshrq_n_s64(__p0, __p1) __extension__ ({ \
  25367.   int64x2_t __ret; \
  25368.   int64x2_t __s0 = __p0; \
  25369.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25370.   __ret = (int64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 35); \
  25371.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25372.   __ret; \
  25373. })
  25374. #endif
  25375.  
  25376. #ifdef __LITTLE_ENDIAN__
  25377. #define vshrq_n_s16(__p0, __p1) __extension__ ({ \
  25378.   int16x8_t __ret; \
  25379.   int16x8_t __s0 = __p0; \
  25380.   __ret = (int16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 33); \
  25381.   __ret; \
  25382. })
  25383. #else
  25384. #define vshrq_n_s16(__p0, __p1) __extension__ ({ \
  25385.   int16x8_t __ret; \
  25386.   int16x8_t __s0 = __p0; \
  25387.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25388.   __ret = (int16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 33); \
  25389.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25390.   __ret; \
  25391. })
  25392. #endif
  25393.  
  25394. #ifdef __LITTLE_ENDIAN__
  25395. #define vshr_n_u8(__p0, __p1) __extension__ ({ \
  25396.   uint8x8_t __ret; \
  25397.   uint8x8_t __s0 = __p0; \
  25398.   __ret = (uint8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 16); \
  25399.   __ret; \
  25400. })
  25401. #else
  25402. #define vshr_n_u8(__p0, __p1) __extension__ ({ \
  25403.   uint8x8_t __ret; \
  25404.   uint8x8_t __s0 = __p0; \
  25405.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25406.   __ret = (uint8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 16); \
  25407.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25408.   __ret; \
  25409. })
  25410. #endif
  25411.  
  25412. #ifdef __LITTLE_ENDIAN__
  25413. #define vshr_n_u32(__p0, __p1) __extension__ ({ \
  25414.   uint32x2_t __ret; \
  25415.   uint32x2_t __s0 = __p0; \
  25416.   __ret = (uint32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 18); \
  25417.   __ret; \
  25418. })
  25419. #else
  25420. #define vshr_n_u32(__p0, __p1) __extension__ ({ \
  25421.   uint32x2_t __ret; \
  25422.   uint32x2_t __s0 = __p0; \
  25423.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25424.   __ret = (uint32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 18); \
  25425.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25426.   __ret; \
  25427. })
  25428. #endif
  25429.  
  25430. #define vshr_n_u64(__p0, __p1) __extension__ ({ \
  25431.   uint64x1_t __ret; \
  25432.   uint64x1_t __s0 = __p0; \
  25433.   __ret = (uint64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 19); \
  25434.   __ret; \
  25435. })
  25436. #ifdef __LITTLE_ENDIAN__
  25437. #define vshr_n_u16(__p0, __p1) __extension__ ({ \
  25438.   uint16x4_t __ret; \
  25439.   uint16x4_t __s0 = __p0; \
  25440.   __ret = (uint16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 17); \
  25441.   __ret; \
  25442. })
  25443. #else
  25444. #define vshr_n_u16(__p0, __p1) __extension__ ({ \
  25445.   uint16x4_t __ret; \
  25446.   uint16x4_t __s0 = __p0; \
  25447.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25448.   __ret = (uint16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 17); \
  25449.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25450.   __ret; \
  25451. })
  25452. #endif
  25453.  
  25454. #ifdef __LITTLE_ENDIAN__
  25455. #define vshr_n_s8(__p0, __p1) __extension__ ({ \
  25456.   int8x8_t __ret; \
  25457.   int8x8_t __s0 = __p0; \
  25458.   __ret = (int8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 0); \
  25459.   __ret; \
  25460. })
  25461. #else
  25462. #define vshr_n_s8(__p0, __p1) __extension__ ({ \
  25463.   int8x8_t __ret; \
  25464.   int8x8_t __s0 = __p0; \
  25465.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25466.   __ret = (int8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 0); \
  25467.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25468.   __ret; \
  25469. })
  25470. #endif
  25471.  
  25472. #ifdef __LITTLE_ENDIAN__
  25473. #define vshr_n_s32(__p0, __p1) __extension__ ({ \
  25474.   int32x2_t __ret; \
  25475.   int32x2_t __s0 = __p0; \
  25476.   __ret = (int32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 2); \
  25477.   __ret; \
  25478. })
  25479. #else
  25480. #define vshr_n_s32(__p0, __p1) __extension__ ({ \
  25481.   int32x2_t __ret; \
  25482.   int32x2_t __s0 = __p0; \
  25483.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25484.   __ret = (int32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 2); \
  25485.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25486.   __ret; \
  25487. })
  25488. #endif
  25489.  
  25490. #define vshr_n_s64(__p0, __p1) __extension__ ({ \
  25491.   int64x1_t __ret; \
  25492.   int64x1_t __s0 = __p0; \
  25493.   __ret = (int64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 3); \
  25494.   __ret; \
  25495. })
  25496. #ifdef __LITTLE_ENDIAN__
  25497. #define vshr_n_s16(__p0, __p1) __extension__ ({ \
  25498.   int16x4_t __ret; \
  25499.   int16x4_t __s0 = __p0; \
  25500.   __ret = (int16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 1); \
  25501.   __ret; \
  25502. })
  25503. #else
  25504. #define vshr_n_s16(__p0, __p1) __extension__ ({ \
  25505.   int16x4_t __ret; \
  25506.   int16x4_t __s0 = __p0; \
  25507.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25508.   __ret = (int16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 1); \
  25509.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25510.   __ret; \
  25511. })
  25512. #endif
  25513.  
  25514. #ifdef __LITTLE_ENDIAN__
  25515. #define vshrn_n_u32(__p0, __p1) __extension__ ({ \
  25516.   uint16x4_t __ret; \
  25517.   uint32x4_t __s0 = __p0; \
  25518.   __ret = (uint16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 17); \
  25519.   __ret; \
  25520. })
  25521. #else
  25522. #define vshrn_n_u32(__p0, __p1) __extension__ ({ \
  25523.   uint16x4_t __ret; \
  25524.   uint32x4_t __s0 = __p0; \
  25525.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25526.   __ret = (uint16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 17); \
  25527.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25528.   __ret; \
  25529. })
  25530. #define __noswap_vshrn_n_u32(__p0, __p1) __extension__ ({ \
  25531.   uint16x4_t __ret; \
  25532.   uint32x4_t __s0 = __p0; \
  25533.   __ret = (uint16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 17); \
  25534.   __ret; \
  25535. })
  25536. #endif
  25537.  
  25538. #ifdef __LITTLE_ENDIAN__
  25539. #define vshrn_n_u64(__p0, __p1) __extension__ ({ \
  25540.   uint32x2_t __ret; \
  25541.   uint64x2_t __s0 = __p0; \
  25542.   __ret = (uint32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 18); \
  25543.   __ret; \
  25544. })
  25545. #else
  25546. #define vshrn_n_u64(__p0, __p1) __extension__ ({ \
  25547.   uint32x2_t __ret; \
  25548.   uint64x2_t __s0 = __p0; \
  25549.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25550.   __ret = (uint32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 18); \
  25551.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25552.   __ret; \
  25553. })
  25554. #define __noswap_vshrn_n_u64(__p0, __p1) __extension__ ({ \
  25555.   uint32x2_t __ret; \
  25556.   uint64x2_t __s0 = __p0; \
  25557.   __ret = (uint32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 18); \
  25558.   __ret; \
  25559. })
  25560. #endif
  25561.  
  25562. #ifdef __LITTLE_ENDIAN__
  25563. #define vshrn_n_u16(__p0, __p1) __extension__ ({ \
  25564.   uint8x8_t __ret; \
  25565.   uint16x8_t __s0 = __p0; \
  25566.   __ret = (uint8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 16); \
  25567.   __ret; \
  25568. })
  25569. #else
  25570. #define vshrn_n_u16(__p0, __p1) __extension__ ({ \
  25571.   uint8x8_t __ret; \
  25572.   uint16x8_t __s0 = __p0; \
  25573.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25574.   __ret = (uint8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 16); \
  25575.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25576.   __ret; \
  25577. })
  25578. #define __noswap_vshrn_n_u16(__p0, __p1) __extension__ ({ \
  25579.   uint8x8_t __ret; \
  25580.   uint16x8_t __s0 = __p0; \
  25581.   __ret = (uint8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 16); \
  25582.   __ret; \
  25583. })
  25584. #endif
  25585.  
  25586. #ifdef __LITTLE_ENDIAN__
  25587. #define vshrn_n_s32(__p0, __p1) __extension__ ({ \
  25588.   int16x4_t __ret; \
  25589.   int32x4_t __s0 = __p0; \
  25590.   __ret = (int16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 1); \
  25591.   __ret; \
  25592. })
  25593. #else
  25594. #define vshrn_n_s32(__p0, __p1) __extension__ ({ \
  25595.   int16x4_t __ret; \
  25596.   int32x4_t __s0 = __p0; \
  25597.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25598.   __ret = (int16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 1); \
  25599.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25600.   __ret; \
  25601. })
  25602. #define __noswap_vshrn_n_s32(__p0, __p1) __extension__ ({ \
  25603.   int16x4_t __ret; \
  25604.   int32x4_t __s0 = __p0; \
  25605.   __ret = (int16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 1); \
  25606.   __ret; \
  25607. })
  25608. #endif
  25609.  
  25610. #ifdef __LITTLE_ENDIAN__
  25611. #define vshrn_n_s64(__p0, __p1) __extension__ ({ \
  25612.   int32x2_t __ret; \
  25613.   int64x2_t __s0 = __p0; \
  25614.   __ret = (int32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 2); \
  25615.   __ret; \
  25616. })
  25617. #else
  25618. #define vshrn_n_s64(__p0, __p1) __extension__ ({ \
  25619.   int32x2_t __ret; \
  25620.   int64x2_t __s0 = __p0; \
  25621.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25622.   __ret = (int32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 2); \
  25623.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25624.   __ret; \
  25625. })
  25626. #define __noswap_vshrn_n_s64(__p0, __p1) __extension__ ({ \
  25627.   int32x2_t __ret; \
  25628.   int64x2_t __s0 = __p0; \
  25629.   __ret = (int32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 2); \
  25630.   __ret; \
  25631. })
  25632. #endif
  25633.  
  25634. #ifdef __LITTLE_ENDIAN__
  25635. #define vshrn_n_s16(__p0, __p1) __extension__ ({ \
  25636.   int8x8_t __ret; \
  25637.   int16x8_t __s0 = __p0; \
  25638.   __ret = (int8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 0); \
  25639.   __ret; \
  25640. })
  25641. #else
  25642. #define vshrn_n_s16(__p0, __p1) __extension__ ({ \
  25643.   int8x8_t __ret; \
  25644.   int16x8_t __s0 = __p0; \
  25645.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25646.   __ret = (int8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 0); \
  25647.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25648.   __ret; \
  25649. })
  25650. #define __noswap_vshrn_n_s16(__p0, __p1) __extension__ ({ \
  25651.   int8x8_t __ret; \
  25652.   int16x8_t __s0 = __p0; \
  25653.   __ret = (int8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 0); \
  25654.   __ret; \
  25655. })
  25656. #endif
  25657.  
  25658. #ifdef __LITTLE_ENDIAN__
  25659. #define vsli_n_p8(__p0, __p1, __p2) __extension__ ({ \
  25660.   poly8x8_t __ret; \
  25661.   poly8x8_t __s0 = __p0; \
  25662.   poly8x8_t __s1 = __p1; \
  25663.   __ret = (poly8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
  25664.   __ret; \
  25665. })
  25666. #else
  25667. #define vsli_n_p8(__p0, __p1, __p2) __extension__ ({ \
  25668.   poly8x8_t __ret; \
  25669.   poly8x8_t __s0 = __p0; \
  25670.   poly8x8_t __s1 = __p1; \
  25671.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25672.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  25673.   __ret = (poly8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
  25674.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25675.   __ret; \
  25676. })
  25677. #endif
  25678.  
  25679. #ifdef __LITTLE_ENDIAN__
  25680. #define vsli_n_p16(__p0, __p1, __p2) __extension__ ({ \
  25681.   poly16x4_t __ret; \
  25682.   poly16x4_t __s0 = __p0; \
  25683.   poly16x4_t __s1 = __p1; \
  25684.   __ret = (poly16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
  25685.   __ret; \
  25686. })
  25687. #else
  25688. #define vsli_n_p16(__p0, __p1, __p2) __extension__ ({ \
  25689.   poly16x4_t __ret; \
  25690.   poly16x4_t __s0 = __p0; \
  25691.   poly16x4_t __s1 = __p1; \
  25692.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25693.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  25694.   __ret = (poly16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
  25695.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25696.   __ret; \
  25697. })
  25698. #endif
  25699.  
  25700. #ifdef __LITTLE_ENDIAN__
  25701. #define vsliq_n_p8(__p0, __p1, __p2) __extension__ ({ \
  25702.   poly8x16_t __ret; \
  25703.   poly8x16_t __s0 = __p0; \
  25704.   poly8x16_t __s1 = __p1; \
  25705.   __ret = (poly8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
  25706.   __ret; \
  25707. })
  25708. #else
  25709. #define vsliq_n_p8(__p0, __p1, __p2) __extension__ ({ \
  25710.   poly8x16_t __ret; \
  25711.   poly8x16_t __s0 = __p0; \
  25712.   poly8x16_t __s1 = __p1; \
  25713.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25714.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25715.   __ret = (poly8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
  25716.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25717.   __ret; \
  25718. })
  25719. #endif
  25720.  
  25721. #ifdef __LITTLE_ENDIAN__
  25722. #define vsliq_n_p16(__p0, __p1, __p2) __extension__ ({ \
  25723.   poly16x8_t __ret; \
  25724.   poly16x8_t __s0 = __p0; \
  25725.   poly16x8_t __s1 = __p1; \
  25726.   __ret = (poly16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
  25727.   __ret; \
  25728. })
  25729. #else
  25730. #define vsliq_n_p16(__p0, __p1, __p2) __extension__ ({ \
  25731.   poly16x8_t __ret; \
  25732.   poly16x8_t __s0 = __p0; \
  25733.   poly16x8_t __s1 = __p1; \
  25734.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25735.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  25736.   __ret = (poly16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
  25737.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25738.   __ret; \
  25739. })
  25740. #endif
  25741.  
  25742. #ifdef __LITTLE_ENDIAN__
  25743. #define vsliq_n_u8(__p0, __p1, __p2) __extension__ ({ \
  25744.   uint8x16_t __ret; \
  25745.   uint8x16_t __s0 = __p0; \
  25746.   uint8x16_t __s1 = __p1; \
  25747.   __ret = (uint8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
  25748.   __ret; \
  25749. })
  25750. #else
  25751. #define vsliq_n_u8(__p0, __p1, __p2) __extension__ ({ \
  25752.   uint8x16_t __ret; \
  25753.   uint8x16_t __s0 = __p0; \
  25754.   uint8x16_t __s1 = __p1; \
  25755.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25756.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25757.   __ret = (uint8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
  25758.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25759.   __ret; \
  25760. })
  25761. #endif
  25762.  
  25763. #ifdef __LITTLE_ENDIAN__
  25764. #define vsliq_n_u32(__p0, __p1, __p2) __extension__ ({ \
  25765.   uint32x4_t __ret; \
  25766.   uint32x4_t __s0 = __p0; \
  25767.   uint32x4_t __s1 = __p1; \
  25768.   __ret = (uint32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
  25769.   __ret; \
  25770. })
  25771. #else
  25772. #define vsliq_n_u32(__p0, __p1, __p2) __extension__ ({ \
  25773.   uint32x4_t __ret; \
  25774.   uint32x4_t __s0 = __p0; \
  25775.   uint32x4_t __s1 = __p1; \
  25776.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25777.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  25778.   __ret = (uint32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
  25779.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25780.   __ret; \
  25781. })
  25782. #endif
  25783.  
  25784. #ifdef __LITTLE_ENDIAN__
  25785. #define vsliq_n_u64(__p0, __p1, __p2) __extension__ ({ \
  25786.   uint64x2_t __ret; \
  25787.   uint64x2_t __s0 = __p0; \
  25788.   uint64x2_t __s1 = __p1; \
  25789.   __ret = (uint64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
  25790.   __ret; \
  25791. })
  25792. #else
  25793. #define vsliq_n_u64(__p0, __p1, __p2) __extension__ ({ \
  25794.   uint64x2_t __ret; \
  25795.   uint64x2_t __s0 = __p0; \
  25796.   uint64x2_t __s1 = __p1; \
  25797.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25798.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  25799.   __ret = (uint64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
  25800.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25801.   __ret; \
  25802. })
  25803. #endif
  25804.  
  25805. #ifdef __LITTLE_ENDIAN__
  25806. #define vsliq_n_u16(__p0, __p1, __p2) __extension__ ({ \
  25807.   uint16x8_t __ret; \
  25808.   uint16x8_t __s0 = __p0; \
  25809.   uint16x8_t __s1 = __p1; \
  25810.   __ret = (uint16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
  25811.   __ret; \
  25812. })
  25813. #else
  25814. #define vsliq_n_u16(__p0, __p1, __p2) __extension__ ({ \
  25815.   uint16x8_t __ret; \
  25816.   uint16x8_t __s0 = __p0; \
  25817.   uint16x8_t __s1 = __p1; \
  25818.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25819.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  25820.   __ret = (uint16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
  25821.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25822.   __ret; \
  25823. })
  25824. #endif
  25825.  
  25826. #ifdef __LITTLE_ENDIAN__
  25827. #define vsliq_n_s8(__p0, __p1, __p2) __extension__ ({ \
  25828.   int8x16_t __ret; \
  25829.   int8x16_t __s0 = __p0; \
  25830.   int8x16_t __s1 = __p1; \
  25831.   __ret = (int8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
  25832.   __ret; \
  25833. })
  25834. #else
  25835. #define vsliq_n_s8(__p0, __p1, __p2) __extension__ ({ \
  25836.   int8x16_t __ret; \
  25837.   int8x16_t __s0 = __p0; \
  25838.   int8x16_t __s1 = __p1; \
  25839.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25840.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25841.   __ret = (int8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
  25842.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  25843.   __ret; \
  25844. })
  25845. #endif
  25846.  
  25847. #ifdef __LITTLE_ENDIAN__
  25848. #define vsliq_n_s32(__p0, __p1, __p2) __extension__ ({ \
  25849.   int32x4_t __ret; \
  25850.   int32x4_t __s0 = __p0; \
  25851.   int32x4_t __s1 = __p1; \
  25852.   __ret = (int32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
  25853.   __ret; \
  25854. })
  25855. #else
  25856. #define vsliq_n_s32(__p0, __p1, __p2) __extension__ ({ \
  25857.   int32x4_t __ret; \
  25858.   int32x4_t __s0 = __p0; \
  25859.   int32x4_t __s1 = __p1; \
  25860.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25861.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  25862.   __ret = (int32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
  25863.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25864.   __ret; \
  25865. })
  25866. #endif
  25867.  
  25868. #ifdef __LITTLE_ENDIAN__
  25869. #define vsliq_n_s64(__p0, __p1, __p2) __extension__ ({ \
  25870.   int64x2_t __ret; \
  25871.   int64x2_t __s0 = __p0; \
  25872.   int64x2_t __s1 = __p1; \
  25873.   __ret = (int64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
  25874.   __ret; \
  25875. })
  25876. #else
  25877. #define vsliq_n_s64(__p0, __p1, __p2) __extension__ ({ \
  25878.   int64x2_t __ret; \
  25879.   int64x2_t __s0 = __p0; \
  25880.   int64x2_t __s1 = __p1; \
  25881.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25882.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  25883.   __ret = (int64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
  25884.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25885.   __ret; \
  25886. })
  25887. #endif
  25888.  
  25889. #ifdef __LITTLE_ENDIAN__
  25890. #define vsliq_n_s16(__p0, __p1, __p2) __extension__ ({ \
  25891.   int16x8_t __ret; \
  25892.   int16x8_t __s0 = __p0; \
  25893.   int16x8_t __s1 = __p1; \
  25894.   __ret = (int16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
  25895.   __ret; \
  25896. })
  25897. #else
  25898. #define vsliq_n_s16(__p0, __p1, __p2) __extension__ ({ \
  25899.   int16x8_t __ret; \
  25900.   int16x8_t __s0 = __p0; \
  25901.   int16x8_t __s1 = __p1; \
  25902.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25903.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  25904.   __ret = (int16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
  25905.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25906.   __ret; \
  25907. })
  25908. #endif
  25909.  
  25910. #ifdef __LITTLE_ENDIAN__
  25911. #define vsli_n_u8(__p0, __p1, __p2) __extension__ ({ \
  25912.   uint8x8_t __ret; \
  25913.   uint8x8_t __s0 = __p0; \
  25914.   uint8x8_t __s1 = __p1; \
  25915.   __ret = (uint8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
  25916.   __ret; \
  25917. })
  25918. #else
  25919. #define vsli_n_u8(__p0, __p1, __p2) __extension__ ({ \
  25920.   uint8x8_t __ret; \
  25921.   uint8x8_t __s0 = __p0; \
  25922.   uint8x8_t __s1 = __p1; \
  25923.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25924.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  25925.   __ret = (uint8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
  25926.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25927.   __ret; \
  25928. })
  25929. #endif
  25930.  
  25931. #ifdef __LITTLE_ENDIAN__
  25932. #define vsli_n_u32(__p0, __p1, __p2) __extension__ ({ \
  25933.   uint32x2_t __ret; \
  25934.   uint32x2_t __s0 = __p0; \
  25935.   uint32x2_t __s1 = __p1; \
  25936.   __ret = (uint32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
  25937.   __ret; \
  25938. })
  25939. #else
  25940. #define vsli_n_u32(__p0, __p1, __p2) __extension__ ({ \
  25941.   uint32x2_t __ret; \
  25942.   uint32x2_t __s0 = __p0; \
  25943.   uint32x2_t __s1 = __p1; \
  25944.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  25945.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  25946.   __ret = (uint32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
  25947.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  25948.   __ret; \
  25949. })
  25950. #endif
  25951.  
  25952. #define vsli_n_u64(__p0, __p1, __p2) __extension__ ({ \
  25953.   uint64x1_t __ret; \
  25954.   uint64x1_t __s0 = __p0; \
  25955.   uint64x1_t __s1 = __p1; \
  25956.   __ret = (uint64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
  25957.   __ret; \
  25958. })
  25959. #ifdef __LITTLE_ENDIAN__
  25960. #define vsli_n_u16(__p0, __p1, __p2) __extension__ ({ \
  25961.   uint16x4_t __ret; \
  25962.   uint16x4_t __s0 = __p0; \
  25963.   uint16x4_t __s1 = __p1; \
  25964.   __ret = (uint16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
  25965.   __ret; \
  25966. })
  25967. #else
  25968. #define vsli_n_u16(__p0, __p1, __p2) __extension__ ({ \
  25969.   uint16x4_t __ret; \
  25970.   uint16x4_t __s0 = __p0; \
  25971.   uint16x4_t __s1 = __p1; \
  25972.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  25973.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  25974.   __ret = (uint16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
  25975.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  25976.   __ret; \
  25977. })
  25978. #endif
  25979.  
  25980. #ifdef __LITTLE_ENDIAN__
  25981. #define vsli_n_s8(__p0, __p1, __p2) __extension__ ({ \
  25982.   int8x8_t __ret; \
  25983.   int8x8_t __s0 = __p0; \
  25984.   int8x8_t __s1 = __p1; \
  25985.   __ret = (int8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
  25986.   __ret; \
  25987. })
  25988. #else
  25989. #define vsli_n_s8(__p0, __p1, __p2) __extension__ ({ \
  25990.   int8x8_t __ret; \
  25991.   int8x8_t __s0 = __p0; \
  25992.   int8x8_t __s1 = __p1; \
  25993.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  25994.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  25995.   __ret = (int8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
  25996.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  25997.   __ret; \
  25998. })
  25999. #endif
  26000.  
  26001. #ifdef __LITTLE_ENDIAN__
  26002. #define vsli_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26003.   int32x2_t __ret; \
  26004.   int32x2_t __s0 = __p0; \
  26005.   int32x2_t __s1 = __p1; \
  26006.   __ret = (int32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
  26007.   __ret; \
  26008. })
  26009. #else
  26010. #define vsli_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26011.   int32x2_t __ret; \
  26012.   int32x2_t __s0 = __p0; \
  26013.   int32x2_t __s1 = __p1; \
  26014.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  26015.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26016.   __ret = (int32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  26017.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  26018.   __ret; \
  26019. })
  26020. #endif
  26021.  
  26022. #define vsli_n_s64(__p0, __p1, __p2) __extension__ ({ \
  26023.   int64x1_t __ret; \
  26024.   int64x1_t __s0 = __p0; \
  26025.   int64x1_t __s1 = __p1; \
  26026.   __ret = (int64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
  26027.   __ret; \
  26028. })
  26029. #ifdef __LITTLE_ENDIAN__
  26030. #define vsli_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26031.   int16x4_t __ret; \
  26032.   int16x4_t __s0 = __p0; \
  26033.   int16x4_t __s1 = __p1; \
  26034.   __ret = (int16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
  26035.   __ret; \
  26036. })
  26037. #else
  26038. #define vsli_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26039.   int16x4_t __ret; \
  26040.   int16x4_t __s0 = __p0; \
  26041.   int16x4_t __s1 = __p1; \
  26042.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26043.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26044.   __ret = (int16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  26045.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26046.   __ret; \
  26047. })
  26048. #endif
  26049.  
  26050. #ifdef __LITTLE_ENDIAN__
  26051. #define vsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
  26052.   uint8x16_t __ret; \
  26053.   uint8x16_t __s0 = __p0; \
  26054.   uint8x16_t __s1 = __p1; \
  26055.   __ret = (uint8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
  26056.   __ret; \
  26057. })
  26058. #else
  26059. #define vsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
  26060.   uint8x16_t __ret; \
  26061.   uint8x16_t __s0 = __p0; \
  26062.   uint8x16_t __s1 = __p1; \
  26063.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26064.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26065.   __ret = (uint8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
  26066.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26067.   __ret; \
  26068. })
  26069. #endif
  26070.  
  26071. #ifdef __LITTLE_ENDIAN__
  26072. #define vsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
  26073.   uint32x4_t __ret; \
  26074.   uint32x4_t __s0 = __p0; \
  26075.   uint32x4_t __s1 = __p1; \
  26076.   __ret = (uint32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
  26077.   __ret; \
  26078. })
  26079. #else
  26080. #define vsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
  26081.   uint32x4_t __ret; \
  26082.   uint32x4_t __s0 = __p0; \
  26083.   uint32x4_t __s1 = __p1; \
  26084.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26085.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26086.   __ret = (uint32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
  26087.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26088.   __ret; \
  26089. })
  26090. #endif
  26091.  
  26092. #ifdef __LITTLE_ENDIAN__
  26093. #define vsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
  26094.   uint64x2_t __ret; \
  26095.   uint64x2_t __s0 = __p0; \
  26096.   uint64x2_t __s1 = __p1; \
  26097.   __ret = (uint64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
  26098.   __ret; \
  26099. })
  26100. #else
  26101. #define vsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
  26102.   uint64x2_t __ret; \
  26103.   uint64x2_t __s0 = __p0; \
  26104.   uint64x2_t __s1 = __p1; \
  26105.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  26106.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26107.   __ret = (uint64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
  26108.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  26109.   __ret; \
  26110. })
  26111. #endif
  26112.  
  26113. #ifdef __LITTLE_ENDIAN__
  26114. #define vsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
  26115.   uint16x8_t __ret; \
  26116.   uint16x8_t __s0 = __p0; \
  26117.   uint16x8_t __s1 = __p1; \
  26118.   __ret = (uint16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
  26119.   __ret; \
  26120. })
  26121. #else
  26122. #define vsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
  26123.   uint16x8_t __ret; \
  26124.   uint16x8_t __s0 = __p0; \
  26125.   uint16x8_t __s1 = __p1; \
  26126.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26127.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26128.   __ret = (uint16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
  26129.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26130.   __ret; \
  26131. })
  26132. #endif
  26133.  
  26134. #ifdef __LITTLE_ENDIAN__
  26135. #define vsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
  26136.   int8x16_t __ret; \
  26137.   int8x16_t __s0 = __p0; \
  26138.   int8x16_t __s1 = __p1; \
  26139.   __ret = (int8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
  26140.   __ret; \
  26141. })
  26142. #else
  26143. #define vsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
  26144.   int8x16_t __ret; \
  26145.   int8x16_t __s0 = __p0; \
  26146.   int8x16_t __s1 = __p1; \
  26147.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26148.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26149.   __ret = (int8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
  26150.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26151.   __ret; \
  26152. })
  26153. #endif
  26154.  
  26155. #ifdef __LITTLE_ENDIAN__
  26156. #define vsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26157.   int32x4_t __ret; \
  26158.   int32x4_t __s0 = __p0; \
  26159.   int32x4_t __s1 = __p1; \
  26160.   __ret = (int32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
  26161.   __ret; \
  26162. })
  26163. #else
  26164. #define vsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26165.   int32x4_t __ret; \
  26166.   int32x4_t __s0 = __p0; \
  26167.   int32x4_t __s1 = __p1; \
  26168.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26169.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26170.   __ret = (int32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
  26171.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26172.   __ret; \
  26173. })
  26174. #endif
  26175.  
  26176. #ifdef __LITTLE_ENDIAN__
  26177. #define vsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
  26178.   int64x2_t __ret; \
  26179.   int64x2_t __s0 = __p0; \
  26180.   int64x2_t __s1 = __p1; \
  26181.   __ret = (int64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
  26182.   __ret; \
  26183. })
  26184. #else
  26185. #define vsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
  26186.   int64x2_t __ret; \
  26187.   int64x2_t __s0 = __p0; \
  26188.   int64x2_t __s1 = __p1; \
  26189.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  26190.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26191.   __ret = (int64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
  26192.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  26193.   __ret; \
  26194. })
  26195. #endif
  26196.  
  26197. #ifdef __LITTLE_ENDIAN__
  26198. #define vsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26199.   int16x8_t __ret; \
  26200.   int16x8_t __s0 = __p0; \
  26201.   int16x8_t __s1 = __p1; \
  26202.   __ret = (int16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
  26203.   __ret; \
  26204. })
  26205. #else
  26206. #define vsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26207.   int16x8_t __ret; \
  26208.   int16x8_t __s0 = __p0; \
  26209.   int16x8_t __s1 = __p1; \
  26210.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26211.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26212.   __ret = (int16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
  26213.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26214.   __ret; \
  26215. })
  26216. #endif
  26217.  
  26218. #ifdef __LITTLE_ENDIAN__
  26219. #define vsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
  26220.   uint8x8_t __ret; \
  26221.   uint8x8_t __s0 = __p0; \
  26222.   uint8x8_t __s1 = __p1; \
  26223.   __ret = (uint8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
  26224.   __ret; \
  26225. })
  26226. #else
  26227. #define vsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
  26228.   uint8x8_t __ret; \
  26229.   uint8x8_t __s0 = __p0; \
  26230.   uint8x8_t __s1 = __p1; \
  26231.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26232.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26233.   __ret = (uint8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
  26234.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26235.   __ret; \
  26236. })
  26237. #endif
  26238.  
  26239. #ifdef __LITTLE_ENDIAN__
  26240. #define vsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
  26241.   uint32x2_t __ret; \
  26242.   uint32x2_t __s0 = __p0; \
  26243.   uint32x2_t __s1 = __p1; \
  26244.   __ret = (uint32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
  26245.   __ret; \
  26246. })
  26247. #else
  26248. #define vsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
  26249.   uint32x2_t __ret; \
  26250.   uint32x2_t __s0 = __p0; \
  26251.   uint32x2_t __s1 = __p1; \
  26252.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  26253.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26254.   __ret = (uint32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
  26255.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  26256.   __ret; \
  26257. })
  26258. #endif
  26259.  
  26260. #define vsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
  26261.   uint64x1_t __ret; \
  26262.   uint64x1_t __s0 = __p0; \
  26263.   uint64x1_t __s1 = __p1; \
  26264.   __ret = (uint64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
  26265.   __ret; \
  26266. })
  26267. #ifdef __LITTLE_ENDIAN__
  26268. #define vsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
  26269.   uint16x4_t __ret; \
  26270.   uint16x4_t __s0 = __p0; \
  26271.   uint16x4_t __s1 = __p1; \
  26272.   __ret = (uint16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
  26273.   __ret; \
  26274. })
  26275. #else
  26276. #define vsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
  26277.   uint16x4_t __ret; \
  26278.   uint16x4_t __s0 = __p0; \
  26279.   uint16x4_t __s1 = __p1; \
  26280.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26281.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26282.   __ret = (uint16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
  26283.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26284.   __ret; \
  26285. })
  26286. #endif
  26287.  
  26288. #ifdef __LITTLE_ENDIAN__
  26289. #define vsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
  26290.   int8x8_t __ret; \
  26291.   int8x8_t __s0 = __p0; \
  26292.   int8x8_t __s1 = __p1; \
  26293.   __ret = (int8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
  26294.   __ret; \
  26295. })
  26296. #else
  26297. #define vsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
  26298.   int8x8_t __ret; \
  26299.   int8x8_t __s0 = __p0; \
  26300.   int8x8_t __s1 = __p1; \
  26301.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26302.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26303.   __ret = (int8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
  26304.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26305.   __ret; \
  26306. })
  26307. #endif
  26308.  
  26309. #ifdef __LITTLE_ENDIAN__
  26310. #define vsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26311.   int32x2_t __ret; \
  26312.   int32x2_t __s0 = __p0; \
  26313.   int32x2_t __s1 = __p1; \
  26314.   __ret = (int32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
  26315.   __ret; \
  26316. })
  26317. #else
  26318. #define vsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26319.   int32x2_t __ret; \
  26320.   int32x2_t __s0 = __p0; \
  26321.   int32x2_t __s1 = __p1; \
  26322.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  26323.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26324.   __ret = (int32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  26325.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  26326.   __ret; \
  26327. })
  26328. #endif
  26329.  
  26330. #define vsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
  26331.   int64x1_t __ret; \
  26332.   int64x1_t __s0 = __p0; \
  26333.   int64x1_t __s1 = __p1; \
  26334.   __ret = (int64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
  26335.   __ret; \
  26336. })
  26337. #ifdef __LITTLE_ENDIAN__
  26338. #define vsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26339.   int16x4_t __ret; \
  26340.   int16x4_t __s0 = __p0; \
  26341.   int16x4_t __s1 = __p1; \
  26342.   __ret = (int16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
  26343.   __ret; \
  26344. })
  26345. #else
  26346. #define vsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26347.   int16x4_t __ret; \
  26348.   int16x4_t __s0 = __p0; \
  26349.   int16x4_t __s1 = __p1; \
  26350.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26351.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26352.   __ret = (int16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  26353.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26354.   __ret; \
  26355. })
  26356. #endif
  26357.  
  26358. #ifdef __LITTLE_ENDIAN__
  26359. #define vsri_n_p8(__p0, __p1, __p2) __extension__ ({ \
  26360.   poly8x8_t __ret; \
  26361.   poly8x8_t __s0 = __p0; \
  26362.   poly8x8_t __s1 = __p1; \
  26363.   __ret = (poly8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
  26364.   __ret; \
  26365. })
  26366. #else
  26367. #define vsri_n_p8(__p0, __p1, __p2) __extension__ ({ \
  26368.   poly8x8_t __ret; \
  26369.   poly8x8_t __s0 = __p0; \
  26370.   poly8x8_t __s1 = __p1; \
  26371.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26372.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26373.   __ret = (poly8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
  26374.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26375.   __ret; \
  26376. })
  26377. #endif
  26378.  
  26379. #ifdef __LITTLE_ENDIAN__
  26380. #define vsri_n_p16(__p0, __p1, __p2) __extension__ ({ \
  26381.   poly16x4_t __ret; \
  26382.   poly16x4_t __s0 = __p0; \
  26383.   poly16x4_t __s1 = __p1; \
  26384.   __ret = (poly16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
  26385.   __ret; \
  26386. })
  26387. #else
  26388. #define vsri_n_p16(__p0, __p1, __p2) __extension__ ({ \
  26389.   poly16x4_t __ret; \
  26390.   poly16x4_t __s0 = __p0; \
  26391.   poly16x4_t __s1 = __p1; \
  26392.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26393.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26394.   __ret = (poly16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
  26395.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26396.   __ret; \
  26397. })
  26398. #endif
  26399.  
  26400. #ifdef __LITTLE_ENDIAN__
  26401. #define vsriq_n_p8(__p0, __p1, __p2) __extension__ ({ \
  26402.   poly8x16_t __ret; \
  26403.   poly8x16_t __s0 = __p0; \
  26404.   poly8x16_t __s1 = __p1; \
  26405.   __ret = (poly8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
  26406.   __ret; \
  26407. })
  26408. #else
  26409. #define vsriq_n_p8(__p0, __p1, __p2) __extension__ ({ \
  26410.   poly8x16_t __ret; \
  26411.   poly8x16_t __s0 = __p0; \
  26412.   poly8x16_t __s1 = __p1; \
  26413.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26414.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26415.   __ret = (poly8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
  26416.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26417.   __ret; \
  26418. })
  26419. #endif
  26420.  
  26421. #ifdef __LITTLE_ENDIAN__
  26422. #define vsriq_n_p16(__p0, __p1, __p2) __extension__ ({ \
  26423.   poly16x8_t __ret; \
  26424.   poly16x8_t __s0 = __p0; \
  26425.   poly16x8_t __s1 = __p1; \
  26426.   __ret = (poly16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
  26427.   __ret; \
  26428. })
  26429. #else
  26430. #define vsriq_n_p16(__p0, __p1, __p2) __extension__ ({ \
  26431.   poly16x8_t __ret; \
  26432.   poly16x8_t __s0 = __p0; \
  26433.   poly16x8_t __s1 = __p1; \
  26434.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26435.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26436.   __ret = (poly16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
  26437.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26438.   __ret; \
  26439. })
  26440. #endif
  26441.  
  26442. #ifdef __LITTLE_ENDIAN__
  26443. #define vsriq_n_u8(__p0, __p1, __p2) __extension__ ({ \
  26444.   uint8x16_t __ret; \
  26445.   uint8x16_t __s0 = __p0; \
  26446.   uint8x16_t __s1 = __p1; \
  26447.   __ret = (uint8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
  26448.   __ret; \
  26449. })
  26450. #else
  26451. #define vsriq_n_u8(__p0, __p1, __p2) __extension__ ({ \
  26452.   uint8x16_t __ret; \
  26453.   uint8x16_t __s0 = __p0; \
  26454.   uint8x16_t __s1 = __p1; \
  26455.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26456.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26457.   __ret = (uint8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
  26458.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26459.   __ret; \
  26460. })
  26461. #endif
  26462.  
  26463. #ifdef __LITTLE_ENDIAN__
  26464. #define vsriq_n_u32(__p0, __p1, __p2) __extension__ ({ \
  26465.   uint32x4_t __ret; \
  26466.   uint32x4_t __s0 = __p0; \
  26467.   uint32x4_t __s1 = __p1; \
  26468.   __ret = (uint32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
  26469.   __ret; \
  26470. })
  26471. #else
  26472. #define vsriq_n_u32(__p0, __p1, __p2) __extension__ ({ \
  26473.   uint32x4_t __ret; \
  26474.   uint32x4_t __s0 = __p0; \
  26475.   uint32x4_t __s1 = __p1; \
  26476.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26477.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26478.   __ret = (uint32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
  26479.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26480.   __ret; \
  26481. })
  26482. #endif
  26483.  
  26484. #ifdef __LITTLE_ENDIAN__
  26485. #define vsriq_n_u64(__p0, __p1, __p2) __extension__ ({ \
  26486.   uint64x2_t __ret; \
  26487.   uint64x2_t __s0 = __p0; \
  26488.   uint64x2_t __s1 = __p1; \
  26489.   __ret = (uint64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
  26490.   __ret; \
  26491. })
  26492. #else
  26493. #define vsriq_n_u64(__p0, __p1, __p2) __extension__ ({ \
  26494.   uint64x2_t __ret; \
  26495.   uint64x2_t __s0 = __p0; \
  26496.   uint64x2_t __s1 = __p1; \
  26497.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  26498.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26499.   __ret = (uint64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
  26500.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  26501.   __ret; \
  26502. })
  26503. #endif
  26504.  
  26505. #ifdef __LITTLE_ENDIAN__
  26506. #define vsriq_n_u16(__p0, __p1, __p2) __extension__ ({ \
  26507.   uint16x8_t __ret; \
  26508.   uint16x8_t __s0 = __p0; \
  26509.   uint16x8_t __s1 = __p1; \
  26510.   __ret = (uint16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
  26511.   __ret; \
  26512. })
  26513. #else
  26514. #define vsriq_n_u16(__p0, __p1, __p2) __extension__ ({ \
  26515.   uint16x8_t __ret; \
  26516.   uint16x8_t __s0 = __p0; \
  26517.   uint16x8_t __s1 = __p1; \
  26518.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26519.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26520.   __ret = (uint16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
  26521.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26522.   __ret; \
  26523. })
  26524. #endif
  26525.  
  26526. #ifdef __LITTLE_ENDIAN__
  26527. #define vsriq_n_s8(__p0, __p1, __p2) __extension__ ({ \
  26528.   int8x16_t __ret; \
  26529.   int8x16_t __s0 = __p0; \
  26530.   int8x16_t __s1 = __p1; \
  26531.   __ret = (int8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
  26532.   __ret; \
  26533. })
  26534. #else
  26535. #define vsriq_n_s8(__p0, __p1, __p2) __extension__ ({ \
  26536.   int8x16_t __ret; \
  26537.   int8x16_t __s0 = __p0; \
  26538.   int8x16_t __s1 = __p1; \
  26539.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26540.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26541.   __ret = (int8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
  26542.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26543.   __ret; \
  26544. })
  26545. #endif
  26546.  
  26547. #ifdef __LITTLE_ENDIAN__
  26548. #define vsriq_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26549.   int32x4_t __ret; \
  26550.   int32x4_t __s0 = __p0; \
  26551.   int32x4_t __s1 = __p1; \
  26552.   __ret = (int32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
  26553.   __ret; \
  26554. })
  26555. #else
  26556. #define vsriq_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26557.   int32x4_t __ret; \
  26558.   int32x4_t __s0 = __p0; \
  26559.   int32x4_t __s1 = __p1; \
  26560.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26561.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26562.   __ret = (int32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
  26563.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26564.   __ret; \
  26565. })
  26566. #endif
  26567.  
  26568. #ifdef __LITTLE_ENDIAN__
  26569. #define vsriq_n_s64(__p0, __p1, __p2) __extension__ ({ \
  26570.   int64x2_t __ret; \
  26571.   int64x2_t __s0 = __p0; \
  26572.   int64x2_t __s1 = __p1; \
  26573.   __ret = (int64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
  26574.   __ret; \
  26575. })
  26576. #else
  26577. #define vsriq_n_s64(__p0, __p1, __p2) __extension__ ({ \
  26578.   int64x2_t __ret; \
  26579.   int64x2_t __s0 = __p0; \
  26580.   int64x2_t __s1 = __p1; \
  26581.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  26582.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26583.   __ret = (int64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
  26584.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  26585.   __ret; \
  26586. })
  26587. #endif
  26588.  
  26589. #ifdef __LITTLE_ENDIAN__
  26590. #define vsriq_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26591.   int16x8_t __ret; \
  26592.   int16x8_t __s0 = __p0; \
  26593.   int16x8_t __s1 = __p1; \
  26594.   __ret = (int16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
  26595.   __ret; \
  26596. })
  26597. #else
  26598. #define vsriq_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26599.   int16x8_t __ret; \
  26600.   int16x8_t __s0 = __p0; \
  26601.   int16x8_t __s1 = __p1; \
  26602.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26603.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26604.   __ret = (int16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
  26605.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26606.   __ret; \
  26607. })
  26608. #endif
  26609.  
  26610. #ifdef __LITTLE_ENDIAN__
  26611. #define vsri_n_u8(__p0, __p1, __p2) __extension__ ({ \
  26612.   uint8x8_t __ret; \
  26613.   uint8x8_t __s0 = __p0; \
  26614.   uint8x8_t __s1 = __p1; \
  26615.   __ret = (uint8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
  26616.   __ret; \
  26617. })
  26618. #else
  26619. #define vsri_n_u8(__p0, __p1, __p2) __extension__ ({ \
  26620.   uint8x8_t __ret; \
  26621.   uint8x8_t __s0 = __p0; \
  26622.   uint8x8_t __s1 = __p1; \
  26623.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26624.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26625.   __ret = (uint8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
  26626.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26627.   __ret; \
  26628. })
  26629. #endif
  26630.  
  26631. #ifdef __LITTLE_ENDIAN__
  26632. #define vsri_n_u32(__p0, __p1, __p2) __extension__ ({ \
  26633.   uint32x2_t __ret; \
  26634.   uint32x2_t __s0 = __p0; \
  26635.   uint32x2_t __s1 = __p1; \
  26636.   __ret = (uint32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
  26637.   __ret; \
  26638. })
  26639. #else
  26640. #define vsri_n_u32(__p0, __p1, __p2) __extension__ ({ \
  26641.   uint32x2_t __ret; \
  26642.   uint32x2_t __s0 = __p0; \
  26643.   uint32x2_t __s1 = __p1; \
  26644.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  26645.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26646.   __ret = (uint32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
  26647.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  26648.   __ret; \
  26649. })
  26650. #endif
  26651.  
  26652. #define vsri_n_u64(__p0, __p1, __p2) __extension__ ({ \
  26653.   uint64x1_t __ret; \
  26654.   uint64x1_t __s0 = __p0; \
  26655.   uint64x1_t __s1 = __p1; \
  26656.   __ret = (uint64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
  26657.   __ret; \
  26658. })
  26659. #ifdef __LITTLE_ENDIAN__
  26660. #define vsri_n_u16(__p0, __p1, __p2) __extension__ ({ \
  26661.   uint16x4_t __ret; \
  26662.   uint16x4_t __s0 = __p0; \
  26663.   uint16x4_t __s1 = __p1; \
  26664.   __ret = (uint16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
  26665.   __ret; \
  26666. })
  26667. #else
  26668. #define vsri_n_u16(__p0, __p1, __p2) __extension__ ({ \
  26669.   uint16x4_t __ret; \
  26670.   uint16x4_t __s0 = __p0; \
  26671.   uint16x4_t __s1 = __p1; \
  26672.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26673.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26674.   __ret = (uint16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
  26675.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26676.   __ret; \
  26677. })
  26678. #endif
  26679.  
  26680. #ifdef __LITTLE_ENDIAN__
  26681. #define vsri_n_s8(__p0, __p1, __p2) __extension__ ({ \
  26682.   int8x8_t __ret; \
  26683.   int8x8_t __s0 = __p0; \
  26684.   int8x8_t __s1 = __p1; \
  26685.   __ret = (int8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
  26686.   __ret; \
  26687. })
  26688. #else
  26689. #define vsri_n_s8(__p0, __p1, __p2) __extension__ ({ \
  26690.   int8x8_t __ret; \
  26691.   int8x8_t __s0 = __p0; \
  26692.   int8x8_t __s1 = __p1; \
  26693.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  26694.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26695.   __ret = (int8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
  26696.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  26697.   __ret; \
  26698. })
  26699. #endif
  26700.  
  26701. #ifdef __LITTLE_ENDIAN__
  26702. #define vsri_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26703.   int32x2_t __ret; \
  26704.   int32x2_t __s0 = __p0; \
  26705.   int32x2_t __s1 = __p1; \
  26706.   __ret = (int32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
  26707.   __ret; \
  26708. })
  26709. #else
  26710. #define vsri_n_s32(__p0, __p1, __p2) __extension__ ({ \
  26711.   int32x2_t __ret; \
  26712.   int32x2_t __s0 = __p0; \
  26713.   int32x2_t __s1 = __p1; \
  26714.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  26715.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26716.   __ret = (int32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  26717.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  26718.   __ret; \
  26719. })
  26720. #endif
  26721.  
  26722. #define vsri_n_s64(__p0, __p1, __p2) __extension__ ({ \
  26723.   int64x1_t __ret; \
  26724.   int64x1_t __s0 = __p0; \
  26725.   int64x1_t __s1 = __p1; \
  26726.   __ret = (int64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
  26727.   __ret; \
  26728. })
  26729. #ifdef __LITTLE_ENDIAN__
  26730. #define vsri_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26731.   int16x4_t __ret; \
  26732.   int16x4_t __s0 = __p0; \
  26733.   int16x4_t __s1 = __p1; \
  26734.   __ret = (int16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
  26735.   __ret; \
  26736. })
  26737. #else
  26738. #define vsri_n_s16(__p0, __p1, __p2) __extension__ ({ \
  26739.   int16x4_t __ret; \
  26740.   int16x4_t __s0 = __p0; \
  26741.   int16x4_t __s1 = __p1; \
  26742.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  26743.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26744.   __ret = (int16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  26745.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  26746.   __ret; \
  26747. })
  26748. #endif
  26749.  
  26750. #ifdef __LITTLE_ENDIAN__
  26751. #define vst1_p8(__p0, __p1) __extension__ ({ \
  26752.   poly8x8_t __s1 = __p1; \
  26753.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 4); \
  26754. })
  26755. #else
  26756. #define vst1_p8(__p0, __p1) __extension__ ({ \
  26757.   poly8x8_t __s1 = __p1; \
  26758.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26759.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 4); \
  26760. })
  26761. #endif
  26762.  
  26763. #ifdef __LITTLE_ENDIAN__
  26764. #define vst1_p16(__p0, __p1) __extension__ ({ \
  26765.   poly16x4_t __s1 = __p1; \
  26766.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 5); \
  26767. })
  26768. #else
  26769. #define vst1_p16(__p0, __p1) __extension__ ({ \
  26770.   poly16x4_t __s1 = __p1; \
  26771.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26772.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 5); \
  26773. })
  26774. #endif
  26775.  
  26776. #ifdef __LITTLE_ENDIAN__
  26777. #define vst1q_p8(__p0, __p1) __extension__ ({ \
  26778.   poly8x16_t __s1 = __p1; \
  26779.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 36); \
  26780. })
  26781. #else
  26782. #define vst1q_p8(__p0, __p1) __extension__ ({ \
  26783.   poly8x16_t __s1 = __p1; \
  26784.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26785.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 36); \
  26786. })
  26787. #endif
  26788.  
  26789. #ifdef __LITTLE_ENDIAN__
  26790. #define vst1q_p16(__p0, __p1) __extension__ ({ \
  26791.   poly16x8_t __s1 = __p1; \
  26792.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 37); \
  26793. })
  26794. #else
  26795. #define vst1q_p16(__p0, __p1) __extension__ ({ \
  26796.   poly16x8_t __s1 = __p1; \
  26797.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26798.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 37); \
  26799. })
  26800. #endif
  26801.  
  26802. #ifdef __LITTLE_ENDIAN__
  26803. #define vst1q_u8(__p0, __p1) __extension__ ({ \
  26804.   uint8x16_t __s1 = __p1; \
  26805.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 48); \
  26806. })
  26807. #else
  26808. #define vst1q_u8(__p0, __p1) __extension__ ({ \
  26809.   uint8x16_t __s1 = __p1; \
  26810.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26811.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 48); \
  26812. })
  26813. #endif
  26814.  
  26815. #ifdef __LITTLE_ENDIAN__
  26816. #define vst1q_u32(__p0, __p1) __extension__ ({ \
  26817.   uint32x4_t __s1 = __p1; \
  26818.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 50); \
  26819. })
  26820. #else
  26821. #define vst1q_u32(__p0, __p1) __extension__ ({ \
  26822.   uint32x4_t __s1 = __p1; \
  26823.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26824.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 50); \
  26825. })
  26826. #endif
  26827.  
  26828. #ifdef __LITTLE_ENDIAN__
  26829. #define vst1q_u64(__p0, __p1) __extension__ ({ \
  26830.   uint64x2_t __s1 = __p1; \
  26831.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 51); \
  26832. })
  26833. #else
  26834. #define vst1q_u64(__p0, __p1) __extension__ ({ \
  26835.   uint64x2_t __s1 = __p1; \
  26836.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26837.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 51); \
  26838. })
  26839. #endif
  26840.  
  26841. #ifdef __LITTLE_ENDIAN__
  26842. #define vst1q_u16(__p0, __p1) __extension__ ({ \
  26843.   uint16x8_t __s1 = __p1; \
  26844.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 49); \
  26845. })
  26846. #else
  26847. #define vst1q_u16(__p0, __p1) __extension__ ({ \
  26848.   uint16x8_t __s1 = __p1; \
  26849.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26850.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 49); \
  26851. })
  26852. #endif
  26853.  
  26854. #ifdef __LITTLE_ENDIAN__
  26855. #define vst1q_s8(__p0, __p1) __extension__ ({ \
  26856.   int8x16_t __s1 = __p1; \
  26857.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 32); \
  26858. })
  26859. #else
  26860. #define vst1q_s8(__p0, __p1) __extension__ ({ \
  26861.   int8x16_t __s1 = __p1; \
  26862.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  26863.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 32); \
  26864. })
  26865. #endif
  26866.  
  26867. #ifdef __LITTLE_ENDIAN__
  26868. #define vst1q_f32(__p0, __p1) __extension__ ({ \
  26869.   float32x4_t __s1 = __p1; \
  26870.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 41); \
  26871. })
  26872. #else
  26873. #define vst1q_f32(__p0, __p1) __extension__ ({ \
  26874.   float32x4_t __s1 = __p1; \
  26875.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26876.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 41); \
  26877. })
  26878. #endif
  26879.  
  26880. #ifdef __LITTLE_ENDIAN__
  26881. #define vst1q_s32(__p0, __p1) __extension__ ({ \
  26882.   int32x4_t __s1 = __p1; \
  26883.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 34); \
  26884. })
  26885. #else
  26886. #define vst1q_s32(__p0, __p1) __extension__ ({ \
  26887.   int32x4_t __s1 = __p1; \
  26888.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26889.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 34); \
  26890. })
  26891. #endif
  26892.  
  26893. #ifdef __LITTLE_ENDIAN__
  26894. #define vst1q_s64(__p0, __p1) __extension__ ({ \
  26895.   int64x2_t __s1 = __p1; \
  26896.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 35); \
  26897. })
  26898. #else
  26899. #define vst1q_s64(__p0, __p1) __extension__ ({ \
  26900.   int64x2_t __s1 = __p1; \
  26901.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26902.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 35); \
  26903. })
  26904. #endif
  26905.  
  26906. #ifdef __LITTLE_ENDIAN__
  26907. #define vst1q_s16(__p0, __p1) __extension__ ({ \
  26908.   int16x8_t __s1 = __p1; \
  26909.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 33); \
  26910. })
  26911. #else
  26912. #define vst1q_s16(__p0, __p1) __extension__ ({ \
  26913.   int16x8_t __s1 = __p1; \
  26914.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26915.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 33); \
  26916. })
  26917. #endif
  26918.  
  26919. #ifdef __LITTLE_ENDIAN__
  26920. #define vst1_u8(__p0, __p1) __extension__ ({ \
  26921.   uint8x8_t __s1 = __p1; \
  26922.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 16); \
  26923. })
  26924. #else
  26925. #define vst1_u8(__p0, __p1) __extension__ ({ \
  26926.   uint8x8_t __s1 = __p1; \
  26927.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26928.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 16); \
  26929. })
  26930. #endif
  26931.  
  26932. #ifdef __LITTLE_ENDIAN__
  26933. #define vst1_u32(__p0, __p1) __extension__ ({ \
  26934.   uint32x2_t __s1 = __p1; \
  26935.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 18); \
  26936. })
  26937. #else
  26938. #define vst1_u32(__p0, __p1) __extension__ ({ \
  26939.   uint32x2_t __s1 = __p1; \
  26940.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26941.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 18); \
  26942. })
  26943. #endif
  26944.  
  26945. #define vst1_u64(__p0, __p1) __extension__ ({ \
  26946.   uint64x1_t __s1 = __p1; \
  26947.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 19); \
  26948. })
  26949. #ifdef __LITTLE_ENDIAN__
  26950. #define vst1_u16(__p0, __p1) __extension__ ({ \
  26951.   uint16x4_t __s1 = __p1; \
  26952.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 17); \
  26953. })
  26954. #else
  26955. #define vst1_u16(__p0, __p1) __extension__ ({ \
  26956.   uint16x4_t __s1 = __p1; \
  26957.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  26958.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 17); \
  26959. })
  26960. #endif
  26961.  
  26962. #ifdef __LITTLE_ENDIAN__
  26963. #define vst1_s8(__p0, __p1) __extension__ ({ \
  26964.   int8x8_t __s1 = __p1; \
  26965.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 0); \
  26966. })
  26967. #else
  26968. #define vst1_s8(__p0, __p1) __extension__ ({ \
  26969.   int8x8_t __s1 = __p1; \
  26970.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  26971.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 0); \
  26972. })
  26973. #endif
  26974.  
  26975. #ifdef __LITTLE_ENDIAN__
  26976. #define vst1_f32(__p0, __p1) __extension__ ({ \
  26977.   float32x2_t __s1 = __p1; \
  26978.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 9); \
  26979. })
  26980. #else
  26981. #define vst1_f32(__p0, __p1) __extension__ ({ \
  26982.   float32x2_t __s1 = __p1; \
  26983.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26984.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 9); \
  26985. })
  26986. #endif
  26987.  
  26988. #ifdef __LITTLE_ENDIAN__
  26989. #define vst1_s32(__p0, __p1) __extension__ ({ \
  26990.   int32x2_t __s1 = __p1; \
  26991.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 2); \
  26992. })
  26993. #else
  26994. #define vst1_s32(__p0, __p1) __extension__ ({ \
  26995.   int32x2_t __s1 = __p1; \
  26996.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  26997.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 2); \
  26998. })
  26999. #endif
  27000.  
  27001. #define vst1_s64(__p0, __p1) __extension__ ({ \
  27002.   int64x1_t __s1 = __p1; \
  27003.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 3); \
  27004. })
  27005. #ifdef __LITTLE_ENDIAN__
  27006. #define vst1_s16(__p0, __p1) __extension__ ({ \
  27007.   int16x4_t __s1 = __p1; \
  27008.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 1); \
  27009. })
  27010. #else
  27011. #define vst1_s16(__p0, __p1) __extension__ ({ \
  27012.   int16x4_t __s1 = __p1; \
  27013.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  27014.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 1); \
  27015. })
  27016. #endif
  27017.  
  27018. #ifdef __LITTLE_ENDIAN__
  27019. #define vst1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  27020.   poly8x8_t __s1 = __p1; \
  27021.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 4); \
  27022. })
  27023. #else
  27024. #define vst1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  27025.   poly8x8_t __s1 = __p1; \
  27026.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  27027.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 4); \
  27028. })
  27029. #endif
  27030.  
  27031. #ifdef __LITTLE_ENDIAN__
  27032. #define vst1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  27033.   poly16x4_t __s1 = __p1; \
  27034.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 5); \
  27035. })
  27036. #else
  27037. #define vst1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  27038.   poly16x4_t __s1 = __p1; \
  27039.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  27040.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 5); \
  27041. })
  27042. #endif
  27043.  
  27044. #ifdef __LITTLE_ENDIAN__
  27045. #define vst1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  27046.   poly8x16_t __s1 = __p1; \
  27047.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 36); \
  27048. })
  27049. #else
  27050. #define vst1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  27051.   poly8x16_t __s1 = __p1; \
  27052.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27053.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 36); \
  27054. })
  27055. #endif
  27056.  
  27057. #ifdef __LITTLE_ENDIAN__
  27058. #define vst1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  27059.   poly16x8_t __s1 = __p1; \
  27060.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 37); \
  27061. })
  27062. #else
  27063. #define vst1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  27064.   poly16x8_t __s1 = __p1; \
  27065.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  27066.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 37); \
  27067. })
  27068. #endif
  27069.  
  27070. #ifdef __LITTLE_ENDIAN__
  27071. #define vst1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  27072.   uint8x16_t __s1 = __p1; \
  27073.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 48); \
  27074. })
  27075. #else
  27076. #define vst1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  27077.   uint8x16_t __s1 = __p1; \
  27078.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27079.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 48); \
  27080. })
  27081. #endif
  27082.  
  27083. #ifdef __LITTLE_ENDIAN__
  27084. #define vst1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  27085.   uint32x4_t __s1 = __p1; \
  27086.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 50); \
  27087. })
  27088. #else
  27089. #define vst1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  27090.   uint32x4_t __s1 = __p1; \
  27091.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  27092.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 50); \
  27093. })
  27094. #endif
  27095.  
  27096. #ifdef __LITTLE_ENDIAN__
  27097. #define vst1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  27098.   uint64x2_t __s1 = __p1; \
  27099.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 51); \
  27100. })
  27101. #else
  27102. #define vst1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  27103.   uint64x2_t __s1 = __p1; \
  27104.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  27105.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 51); \
  27106. })
  27107. #endif
  27108.  
  27109. #ifdef __LITTLE_ENDIAN__
  27110. #define vst1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  27111.   uint16x8_t __s1 = __p1; \
  27112.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 49); \
  27113. })
  27114. #else
  27115. #define vst1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  27116.   uint16x8_t __s1 = __p1; \
  27117.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  27118.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 49); \
  27119. })
  27120. #endif
  27121.  
  27122. #ifdef __LITTLE_ENDIAN__
  27123. #define vst1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  27124.   int8x16_t __s1 = __p1; \
  27125.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 32); \
  27126. })
  27127. #else
  27128. #define vst1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  27129.   int8x16_t __s1 = __p1; \
  27130.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27131.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 32); \
  27132. })
  27133. #endif
  27134.  
  27135. #ifdef __LITTLE_ENDIAN__
  27136. #define vst1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  27137.   float32x4_t __s1 = __p1; \
  27138.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 41); \
  27139. })
  27140. #else
  27141. #define vst1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  27142.   float32x4_t __s1 = __p1; \
  27143.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  27144.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 41); \
  27145. })
  27146. #endif
  27147.  
  27148. #ifdef __LITTLE_ENDIAN__
  27149. #define vst1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  27150.   int32x4_t __s1 = __p1; \
  27151.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 34); \
  27152. })
  27153. #else
  27154. #define vst1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  27155.   int32x4_t __s1 = __p1; \
  27156.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  27157.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 34); \
  27158. })
  27159. #endif
  27160.  
  27161. #ifdef __LITTLE_ENDIAN__
  27162. #define vst1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  27163.   int64x2_t __s1 = __p1; \
  27164.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 35); \
  27165. })
  27166. #else
  27167. #define vst1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  27168.   int64x2_t __s1 = __p1; \
  27169.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  27170.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 35); \
  27171. })
  27172. #endif
  27173.  
  27174. #ifdef __LITTLE_ENDIAN__
  27175. #define vst1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  27176.   int16x8_t __s1 = __p1; \
  27177.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 33); \
  27178. })
  27179. #else
  27180. #define vst1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  27181.   int16x8_t __s1 = __p1; \
  27182.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  27183.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 33); \
  27184. })
  27185. #endif
  27186.  
  27187. #ifdef __LITTLE_ENDIAN__
  27188. #define vst1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  27189.   uint8x8_t __s1 = __p1; \
  27190.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 16); \
  27191. })
  27192. #else
  27193. #define vst1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  27194.   uint8x8_t __s1 = __p1; \
  27195.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  27196.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 16); \
  27197. })
  27198. #endif
  27199.  
  27200. #ifdef __LITTLE_ENDIAN__
  27201. #define vst1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  27202.   uint32x2_t __s1 = __p1; \
  27203.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 18); \
  27204. })
  27205. #else
  27206. #define vst1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  27207.   uint32x2_t __s1 = __p1; \
  27208.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  27209.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 18); \
  27210. })
  27211. #endif
  27212.  
  27213. #define vst1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  27214.   uint64x1_t __s1 = __p1; \
  27215.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
  27216. })
  27217. #ifdef __LITTLE_ENDIAN__
  27218. #define vst1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  27219.   uint16x4_t __s1 = __p1; \
  27220.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 17); \
  27221. })
  27222. #else
  27223. #define vst1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  27224.   uint16x4_t __s1 = __p1; \
  27225.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  27226.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 17); \
  27227. })
  27228. #endif
  27229.  
  27230. #ifdef __LITTLE_ENDIAN__
  27231. #define vst1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  27232.   int8x8_t __s1 = __p1; \
  27233.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 0); \
  27234. })
  27235. #else
  27236. #define vst1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  27237.   int8x8_t __s1 = __p1; \
  27238.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  27239.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 0); \
  27240. })
  27241. #endif
  27242.  
  27243. #ifdef __LITTLE_ENDIAN__
  27244. #define vst1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  27245.   float32x2_t __s1 = __p1; \
  27246.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 9); \
  27247. })
  27248. #else
  27249. #define vst1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  27250.   float32x2_t __s1 = __p1; \
  27251.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  27252.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 9); \
  27253. })
  27254. #endif
  27255.  
  27256. #ifdef __LITTLE_ENDIAN__
  27257. #define vst1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  27258.   int32x2_t __s1 = __p1; \
  27259.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 2); \
  27260. })
  27261. #else
  27262. #define vst1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  27263.   int32x2_t __s1 = __p1; \
  27264.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  27265.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 2); \
  27266. })
  27267. #endif
  27268.  
  27269. #define vst1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  27270.   int64x1_t __s1 = __p1; \
  27271.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
  27272. })
  27273. #ifdef __LITTLE_ENDIAN__
  27274. #define vst1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  27275.   int16x4_t __s1 = __p1; \
  27276.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 1); \
  27277. })
  27278. #else
  27279. #define vst1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  27280.   int16x4_t __s1 = __p1; \
  27281.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  27282.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 1); \
  27283. })
  27284. #endif
  27285.  
  27286. #ifdef __LITTLE_ENDIAN__
  27287. #define vst1_p8_x2(__p0, __p1) __extension__ ({ \
  27288.   poly8x8x2_t __s1 = __p1; \
  27289.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 4); \
  27290. })
  27291. #else
  27292. #define vst1_p8_x2(__p0, __p1) __extension__ ({ \
  27293.   poly8x8x2_t __s1 = __p1; \
  27294.   poly8x8x2_t __rev1; \
  27295.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27296.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27297.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 4); \
  27298. })
  27299. #endif
  27300.  
  27301. #ifdef __LITTLE_ENDIAN__
  27302. #define vst1_p16_x2(__p0, __p1) __extension__ ({ \
  27303.   poly16x4x2_t __s1 = __p1; \
  27304.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 5); \
  27305. })
  27306. #else
  27307. #define vst1_p16_x2(__p0, __p1) __extension__ ({ \
  27308.   poly16x4x2_t __s1 = __p1; \
  27309.   poly16x4x2_t __rev1; \
  27310.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27311.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27312.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 5); \
  27313. })
  27314. #endif
  27315.  
  27316. #ifdef __LITTLE_ENDIAN__
  27317. #define vst1q_p8_x2(__p0, __p1) __extension__ ({ \
  27318.   poly8x16x2_t __s1 = __p1; \
  27319.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 36); \
  27320. })
  27321. #else
  27322. #define vst1q_p8_x2(__p0, __p1) __extension__ ({ \
  27323.   poly8x16x2_t __s1 = __p1; \
  27324.   poly8x16x2_t __rev1; \
  27325.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27326.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27327.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 36); \
  27328. })
  27329. #endif
  27330.  
  27331. #ifdef __LITTLE_ENDIAN__
  27332. #define vst1q_p16_x2(__p0, __p1) __extension__ ({ \
  27333.   poly16x8x2_t __s1 = __p1; \
  27334.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 37); \
  27335. })
  27336. #else
  27337. #define vst1q_p16_x2(__p0, __p1) __extension__ ({ \
  27338.   poly16x8x2_t __s1 = __p1; \
  27339.   poly16x8x2_t __rev1; \
  27340.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27341.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27342.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 37); \
  27343. })
  27344. #endif
  27345.  
  27346. #ifdef __LITTLE_ENDIAN__
  27347. #define vst1q_u8_x2(__p0, __p1) __extension__ ({ \
  27348.   uint8x16x2_t __s1 = __p1; \
  27349.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 48); \
  27350. })
  27351. #else
  27352. #define vst1q_u8_x2(__p0, __p1) __extension__ ({ \
  27353.   uint8x16x2_t __s1 = __p1; \
  27354.   uint8x16x2_t __rev1; \
  27355.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27356.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27357.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 48); \
  27358. })
  27359. #endif
  27360.  
  27361. #ifdef __LITTLE_ENDIAN__
  27362. #define vst1q_u32_x2(__p0, __p1) __extension__ ({ \
  27363.   uint32x4x2_t __s1 = __p1; \
  27364.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 50); \
  27365. })
  27366. #else
  27367. #define vst1q_u32_x2(__p0, __p1) __extension__ ({ \
  27368.   uint32x4x2_t __s1 = __p1; \
  27369.   uint32x4x2_t __rev1; \
  27370.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27371.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27372.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 50); \
  27373. })
  27374. #endif
  27375.  
  27376. #ifdef __LITTLE_ENDIAN__
  27377. #define vst1q_u64_x2(__p0, __p1) __extension__ ({ \
  27378.   uint64x2x2_t __s1 = __p1; \
  27379.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 51); \
  27380. })
  27381. #else
  27382. #define vst1q_u64_x2(__p0, __p1) __extension__ ({ \
  27383.   uint64x2x2_t __s1 = __p1; \
  27384.   uint64x2x2_t __rev1; \
  27385.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27386.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27387.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 51); \
  27388. })
  27389. #endif
  27390.  
  27391. #ifdef __LITTLE_ENDIAN__
  27392. #define vst1q_u16_x2(__p0, __p1) __extension__ ({ \
  27393.   uint16x8x2_t __s1 = __p1; \
  27394.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 49); \
  27395. })
  27396. #else
  27397. #define vst1q_u16_x2(__p0, __p1) __extension__ ({ \
  27398.   uint16x8x2_t __s1 = __p1; \
  27399.   uint16x8x2_t __rev1; \
  27400.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27401.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27402.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 49); \
  27403. })
  27404. #endif
  27405.  
  27406. #ifdef __LITTLE_ENDIAN__
  27407. #define vst1q_s8_x2(__p0, __p1) __extension__ ({ \
  27408.   int8x16x2_t __s1 = __p1; \
  27409.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 32); \
  27410. })
  27411. #else
  27412. #define vst1q_s8_x2(__p0, __p1) __extension__ ({ \
  27413.   int8x16x2_t __s1 = __p1; \
  27414.   int8x16x2_t __rev1; \
  27415.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27416.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27417.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 32); \
  27418. })
  27419. #endif
  27420.  
  27421. #ifdef __LITTLE_ENDIAN__
  27422. #define vst1q_f32_x2(__p0, __p1) __extension__ ({ \
  27423.   float32x4x2_t __s1 = __p1; \
  27424.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 41); \
  27425. })
  27426. #else
  27427. #define vst1q_f32_x2(__p0, __p1) __extension__ ({ \
  27428.   float32x4x2_t __s1 = __p1; \
  27429.   float32x4x2_t __rev1; \
  27430.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27431.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27432.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 41); \
  27433. })
  27434. #endif
  27435.  
  27436. #ifdef __LITTLE_ENDIAN__
  27437. #define vst1q_s32_x2(__p0, __p1) __extension__ ({ \
  27438.   int32x4x2_t __s1 = __p1; \
  27439.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 34); \
  27440. })
  27441. #else
  27442. #define vst1q_s32_x2(__p0, __p1) __extension__ ({ \
  27443.   int32x4x2_t __s1 = __p1; \
  27444.   int32x4x2_t __rev1; \
  27445.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27446.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27447.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 34); \
  27448. })
  27449. #endif
  27450.  
  27451. #ifdef __LITTLE_ENDIAN__
  27452. #define vst1q_s64_x2(__p0, __p1) __extension__ ({ \
  27453.   int64x2x2_t __s1 = __p1; \
  27454.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 35); \
  27455. })
  27456. #else
  27457. #define vst1q_s64_x2(__p0, __p1) __extension__ ({ \
  27458.   int64x2x2_t __s1 = __p1; \
  27459.   int64x2x2_t __rev1; \
  27460.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27461.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27462.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 35); \
  27463. })
  27464. #endif
  27465.  
  27466. #ifdef __LITTLE_ENDIAN__
  27467. #define vst1q_s16_x2(__p0, __p1) __extension__ ({ \
  27468.   int16x8x2_t __s1 = __p1; \
  27469.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 33); \
  27470. })
  27471. #else
  27472. #define vst1q_s16_x2(__p0, __p1) __extension__ ({ \
  27473.   int16x8x2_t __s1 = __p1; \
  27474.   int16x8x2_t __rev1; \
  27475.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27476.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27477.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 33); \
  27478. })
  27479. #endif
  27480.  
  27481. #ifdef __LITTLE_ENDIAN__
  27482. #define vst1_u8_x2(__p0, __p1) __extension__ ({ \
  27483.   uint8x8x2_t __s1 = __p1; \
  27484.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 16); \
  27485. })
  27486. #else
  27487. #define vst1_u8_x2(__p0, __p1) __extension__ ({ \
  27488.   uint8x8x2_t __s1 = __p1; \
  27489.   uint8x8x2_t __rev1; \
  27490.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27491.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27492.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 16); \
  27493. })
  27494. #endif
  27495.  
  27496. #ifdef __LITTLE_ENDIAN__
  27497. #define vst1_u32_x2(__p0, __p1) __extension__ ({ \
  27498.   uint32x2x2_t __s1 = __p1; \
  27499.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 18); \
  27500. })
  27501. #else
  27502. #define vst1_u32_x2(__p0, __p1) __extension__ ({ \
  27503.   uint32x2x2_t __s1 = __p1; \
  27504.   uint32x2x2_t __rev1; \
  27505.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27506.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27507.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 18); \
  27508. })
  27509. #endif
  27510.  
  27511. #define vst1_u64_x2(__p0, __p1) __extension__ ({ \
  27512.   uint64x1x2_t __s1 = __p1; \
  27513.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
  27514. })
  27515. #ifdef __LITTLE_ENDIAN__
  27516. #define vst1_u16_x2(__p0, __p1) __extension__ ({ \
  27517.   uint16x4x2_t __s1 = __p1; \
  27518.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 17); \
  27519. })
  27520. #else
  27521. #define vst1_u16_x2(__p0, __p1) __extension__ ({ \
  27522.   uint16x4x2_t __s1 = __p1; \
  27523.   uint16x4x2_t __rev1; \
  27524.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27525.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27526.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 17); \
  27527. })
  27528. #endif
  27529.  
  27530. #ifdef __LITTLE_ENDIAN__
  27531. #define vst1_s8_x2(__p0, __p1) __extension__ ({ \
  27532.   int8x8x2_t __s1 = __p1; \
  27533.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 0); \
  27534. })
  27535. #else
  27536. #define vst1_s8_x2(__p0, __p1) __extension__ ({ \
  27537.   int8x8x2_t __s1 = __p1; \
  27538.   int8x8x2_t __rev1; \
  27539.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27540.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27541.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 0); \
  27542. })
  27543. #endif
  27544.  
  27545. #ifdef __LITTLE_ENDIAN__
  27546. #define vst1_f32_x2(__p0, __p1) __extension__ ({ \
  27547.   float32x2x2_t __s1 = __p1; \
  27548.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 9); \
  27549. })
  27550. #else
  27551. #define vst1_f32_x2(__p0, __p1) __extension__ ({ \
  27552.   float32x2x2_t __s1 = __p1; \
  27553.   float32x2x2_t __rev1; \
  27554.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27555.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27556.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 9); \
  27557. })
  27558. #endif
  27559.  
  27560. #ifdef __LITTLE_ENDIAN__
  27561. #define vst1_s32_x2(__p0, __p1) __extension__ ({ \
  27562.   int32x2x2_t __s1 = __p1; \
  27563.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 2); \
  27564. })
  27565. #else
  27566. #define vst1_s32_x2(__p0, __p1) __extension__ ({ \
  27567.   int32x2x2_t __s1 = __p1; \
  27568.   int32x2x2_t __rev1; \
  27569.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27570.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27571.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 2); \
  27572. })
  27573. #endif
  27574.  
  27575. #define vst1_s64_x2(__p0, __p1) __extension__ ({ \
  27576.   int64x1x2_t __s1 = __p1; \
  27577.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 3); \
  27578. })
  27579. #ifdef __LITTLE_ENDIAN__
  27580. #define vst1_s16_x2(__p0, __p1) __extension__ ({ \
  27581.   int16x4x2_t __s1 = __p1; \
  27582.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 1); \
  27583. })
  27584. #else
  27585. #define vst1_s16_x2(__p0, __p1) __extension__ ({ \
  27586.   int16x4x2_t __s1 = __p1; \
  27587.   int16x4x2_t __rev1; \
  27588.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27589.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27590.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 1); \
  27591. })
  27592. #endif
  27593.  
  27594. #ifdef __LITTLE_ENDIAN__
  27595. #define vst1_p8_x3(__p0, __p1) __extension__ ({ \
  27596.   poly8x8x3_t __s1 = __p1; \
  27597.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 4); \
  27598. })
  27599. #else
  27600. #define vst1_p8_x3(__p0, __p1) __extension__ ({ \
  27601.   poly8x8x3_t __s1 = __p1; \
  27602.   poly8x8x3_t __rev1; \
  27603.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27604.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27605.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  27606.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 4); \
  27607. })
  27608. #endif
  27609.  
  27610. #ifdef __LITTLE_ENDIAN__
  27611. #define vst1_p16_x3(__p0, __p1) __extension__ ({ \
  27612.   poly16x4x3_t __s1 = __p1; \
  27613.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 5); \
  27614. })
  27615. #else
  27616. #define vst1_p16_x3(__p0, __p1) __extension__ ({ \
  27617.   poly16x4x3_t __s1 = __p1; \
  27618.   poly16x4x3_t __rev1; \
  27619.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27620.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27621.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  27622.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 5); \
  27623. })
  27624. #endif
  27625.  
  27626. #ifdef __LITTLE_ENDIAN__
  27627. #define vst1q_p8_x3(__p0, __p1) __extension__ ({ \
  27628.   poly8x16x3_t __s1 = __p1; \
  27629.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 36); \
  27630. })
  27631. #else
  27632. #define vst1q_p8_x3(__p0, __p1) __extension__ ({ \
  27633.   poly8x16x3_t __s1 = __p1; \
  27634.   poly8x16x3_t __rev1; \
  27635.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27636.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27637.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27638.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 36); \
  27639. })
  27640. #endif
  27641.  
  27642. #ifdef __LITTLE_ENDIAN__
  27643. #define vst1q_p16_x3(__p0, __p1) __extension__ ({ \
  27644.   poly16x8x3_t __s1 = __p1; \
  27645.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 37); \
  27646. })
  27647. #else
  27648. #define vst1q_p16_x3(__p0, __p1) __extension__ ({ \
  27649.   poly16x8x3_t __s1 = __p1; \
  27650.   poly16x8x3_t __rev1; \
  27651.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27652.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27653.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  27654.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 37); \
  27655. })
  27656. #endif
  27657.  
  27658. #ifdef __LITTLE_ENDIAN__
  27659. #define vst1q_u8_x3(__p0, __p1) __extension__ ({ \
  27660.   uint8x16x3_t __s1 = __p1; \
  27661.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 48); \
  27662. })
  27663. #else
  27664. #define vst1q_u8_x3(__p0, __p1) __extension__ ({ \
  27665.   uint8x16x3_t __s1 = __p1; \
  27666.   uint8x16x3_t __rev1; \
  27667.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27668.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27669.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27670.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 48); \
  27671. })
  27672. #endif
  27673.  
  27674. #ifdef __LITTLE_ENDIAN__
  27675. #define vst1q_u32_x3(__p0, __p1) __extension__ ({ \
  27676.   uint32x4x3_t __s1 = __p1; \
  27677.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 50); \
  27678. })
  27679. #else
  27680. #define vst1q_u32_x3(__p0, __p1) __extension__ ({ \
  27681.   uint32x4x3_t __s1 = __p1; \
  27682.   uint32x4x3_t __rev1; \
  27683.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27684.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27685.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  27686.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 50); \
  27687. })
  27688. #endif
  27689.  
  27690. #ifdef __LITTLE_ENDIAN__
  27691. #define vst1q_u64_x3(__p0, __p1) __extension__ ({ \
  27692.   uint64x2x3_t __s1 = __p1; \
  27693.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 51); \
  27694. })
  27695. #else
  27696. #define vst1q_u64_x3(__p0, __p1) __extension__ ({ \
  27697.   uint64x2x3_t __s1 = __p1; \
  27698.   uint64x2x3_t __rev1; \
  27699.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27700.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27701.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  27702.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 51); \
  27703. })
  27704. #endif
  27705.  
  27706. #ifdef __LITTLE_ENDIAN__
  27707. #define vst1q_u16_x3(__p0, __p1) __extension__ ({ \
  27708.   uint16x8x3_t __s1 = __p1; \
  27709.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 49); \
  27710. })
  27711. #else
  27712. #define vst1q_u16_x3(__p0, __p1) __extension__ ({ \
  27713.   uint16x8x3_t __s1 = __p1; \
  27714.   uint16x8x3_t __rev1; \
  27715.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27716.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27717.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  27718.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 49); \
  27719. })
  27720. #endif
  27721.  
  27722. #ifdef __LITTLE_ENDIAN__
  27723. #define vst1q_s8_x3(__p0, __p1) __extension__ ({ \
  27724.   int8x16x3_t __s1 = __p1; \
  27725.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 32); \
  27726. })
  27727. #else
  27728. #define vst1q_s8_x3(__p0, __p1) __extension__ ({ \
  27729.   int8x16x3_t __s1 = __p1; \
  27730.   int8x16x3_t __rev1; \
  27731.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27732.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27733.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27734.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 32); \
  27735. })
  27736. #endif
  27737.  
  27738. #ifdef __LITTLE_ENDIAN__
  27739. #define vst1q_f32_x3(__p0, __p1) __extension__ ({ \
  27740.   float32x4x3_t __s1 = __p1; \
  27741.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 41); \
  27742. })
  27743. #else
  27744. #define vst1q_f32_x3(__p0, __p1) __extension__ ({ \
  27745.   float32x4x3_t __s1 = __p1; \
  27746.   float32x4x3_t __rev1; \
  27747.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27748.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27749.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  27750.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 41); \
  27751. })
  27752. #endif
  27753.  
  27754. #ifdef __LITTLE_ENDIAN__
  27755. #define vst1q_s32_x3(__p0, __p1) __extension__ ({ \
  27756.   int32x4x3_t __s1 = __p1; \
  27757.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 34); \
  27758. })
  27759. #else
  27760. #define vst1q_s32_x3(__p0, __p1) __extension__ ({ \
  27761.   int32x4x3_t __s1 = __p1; \
  27762.   int32x4x3_t __rev1; \
  27763.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27764.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27765.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  27766.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 34); \
  27767. })
  27768. #endif
  27769.  
  27770. #ifdef __LITTLE_ENDIAN__
  27771. #define vst1q_s64_x3(__p0, __p1) __extension__ ({ \
  27772.   int64x2x3_t __s1 = __p1; \
  27773.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 35); \
  27774. })
  27775. #else
  27776. #define vst1q_s64_x3(__p0, __p1) __extension__ ({ \
  27777.   int64x2x3_t __s1 = __p1; \
  27778.   int64x2x3_t __rev1; \
  27779.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27780.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27781.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  27782.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 35); \
  27783. })
  27784. #endif
  27785.  
  27786. #ifdef __LITTLE_ENDIAN__
  27787. #define vst1q_s16_x3(__p0, __p1) __extension__ ({ \
  27788.   int16x8x3_t __s1 = __p1; \
  27789.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 33); \
  27790. })
  27791. #else
  27792. #define vst1q_s16_x3(__p0, __p1) __extension__ ({ \
  27793.   int16x8x3_t __s1 = __p1; \
  27794.   int16x8x3_t __rev1; \
  27795.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27796.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27797.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  27798.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 33); \
  27799. })
  27800. #endif
  27801.  
  27802. #ifdef __LITTLE_ENDIAN__
  27803. #define vst1_u8_x3(__p0, __p1) __extension__ ({ \
  27804.   uint8x8x3_t __s1 = __p1; \
  27805.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 16); \
  27806. })
  27807. #else
  27808. #define vst1_u8_x3(__p0, __p1) __extension__ ({ \
  27809.   uint8x8x3_t __s1 = __p1; \
  27810.   uint8x8x3_t __rev1; \
  27811.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27812.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27813.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  27814.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 16); \
  27815. })
  27816. #endif
  27817.  
  27818. #ifdef __LITTLE_ENDIAN__
  27819. #define vst1_u32_x3(__p0, __p1) __extension__ ({ \
  27820.   uint32x2x3_t __s1 = __p1; \
  27821.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 18); \
  27822. })
  27823. #else
  27824. #define vst1_u32_x3(__p0, __p1) __extension__ ({ \
  27825.   uint32x2x3_t __s1 = __p1; \
  27826.   uint32x2x3_t __rev1; \
  27827.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27828.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27829.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  27830.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 18); \
  27831. })
  27832. #endif
  27833.  
  27834. #define vst1_u64_x3(__p0, __p1) __extension__ ({ \
  27835.   uint64x1x3_t __s1 = __p1; \
  27836.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
  27837. })
  27838. #ifdef __LITTLE_ENDIAN__
  27839. #define vst1_u16_x3(__p0, __p1) __extension__ ({ \
  27840.   uint16x4x3_t __s1 = __p1; \
  27841.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 17); \
  27842. })
  27843. #else
  27844. #define vst1_u16_x3(__p0, __p1) __extension__ ({ \
  27845.   uint16x4x3_t __s1 = __p1; \
  27846.   uint16x4x3_t __rev1; \
  27847.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27848.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27849.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  27850.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 17); \
  27851. })
  27852. #endif
  27853.  
  27854. #ifdef __LITTLE_ENDIAN__
  27855. #define vst1_s8_x3(__p0, __p1) __extension__ ({ \
  27856.   int8x8x3_t __s1 = __p1; \
  27857.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 0); \
  27858. })
  27859. #else
  27860. #define vst1_s8_x3(__p0, __p1) __extension__ ({ \
  27861.   int8x8x3_t __s1 = __p1; \
  27862.   int8x8x3_t __rev1; \
  27863.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27864.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27865.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  27866.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 0); \
  27867. })
  27868. #endif
  27869.  
  27870. #ifdef __LITTLE_ENDIAN__
  27871. #define vst1_f32_x3(__p0, __p1) __extension__ ({ \
  27872.   float32x2x3_t __s1 = __p1; \
  27873.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 9); \
  27874. })
  27875. #else
  27876. #define vst1_f32_x3(__p0, __p1) __extension__ ({ \
  27877.   float32x2x3_t __s1 = __p1; \
  27878.   float32x2x3_t __rev1; \
  27879.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27880.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27881.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  27882.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 9); \
  27883. })
  27884. #endif
  27885.  
  27886. #ifdef __LITTLE_ENDIAN__
  27887. #define vst1_s32_x3(__p0, __p1) __extension__ ({ \
  27888.   int32x2x3_t __s1 = __p1; \
  27889.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 2); \
  27890. })
  27891. #else
  27892. #define vst1_s32_x3(__p0, __p1) __extension__ ({ \
  27893.   int32x2x3_t __s1 = __p1; \
  27894.   int32x2x3_t __rev1; \
  27895.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  27896.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  27897.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  27898.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 2); \
  27899. })
  27900. #endif
  27901.  
  27902. #define vst1_s64_x3(__p0, __p1) __extension__ ({ \
  27903.   int64x1x3_t __s1 = __p1; \
  27904.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 3); \
  27905. })
  27906. #ifdef __LITTLE_ENDIAN__
  27907. #define vst1_s16_x3(__p0, __p1) __extension__ ({ \
  27908.   int16x4x3_t __s1 = __p1; \
  27909.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 1); \
  27910. })
  27911. #else
  27912. #define vst1_s16_x3(__p0, __p1) __extension__ ({ \
  27913.   int16x4x3_t __s1 = __p1; \
  27914.   int16x4x3_t __rev1; \
  27915.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27916.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27917.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  27918.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 1); \
  27919. })
  27920. #endif
  27921.  
  27922. #ifdef __LITTLE_ENDIAN__
  27923. #define vst1_p8_x4(__p0, __p1) __extension__ ({ \
  27924.   poly8x8x4_t __s1 = __p1; \
  27925.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 4); \
  27926. })
  27927. #else
  27928. #define vst1_p8_x4(__p0, __p1) __extension__ ({ \
  27929.   poly8x8x4_t __s1 = __p1; \
  27930.   poly8x8x4_t __rev1; \
  27931.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27932.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27933.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  27934.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  27935.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 4); \
  27936. })
  27937. #endif
  27938.  
  27939. #ifdef __LITTLE_ENDIAN__
  27940. #define vst1_p16_x4(__p0, __p1) __extension__ ({ \
  27941.   poly16x4x4_t __s1 = __p1; \
  27942.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 5); \
  27943. })
  27944. #else
  27945. #define vst1_p16_x4(__p0, __p1) __extension__ ({ \
  27946.   poly16x4x4_t __s1 = __p1; \
  27947.   poly16x4x4_t __rev1; \
  27948.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  27949.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  27950.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  27951.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  27952.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 5); \
  27953. })
  27954. #endif
  27955.  
  27956. #ifdef __LITTLE_ENDIAN__
  27957. #define vst1q_p8_x4(__p0, __p1) __extension__ ({ \
  27958.   poly8x16x4_t __s1 = __p1; \
  27959.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 36); \
  27960. })
  27961. #else
  27962. #define vst1q_p8_x4(__p0, __p1) __extension__ ({ \
  27963.   poly8x16x4_t __s1 = __p1; \
  27964.   poly8x16x4_t __rev1; \
  27965.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27966.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27967.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27968.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  27969.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 36); \
  27970. })
  27971. #endif
  27972.  
  27973. #ifdef __LITTLE_ENDIAN__
  27974. #define vst1q_p16_x4(__p0, __p1) __extension__ ({ \
  27975.   poly16x8x4_t __s1 = __p1; \
  27976.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 37); \
  27977. })
  27978. #else
  27979. #define vst1q_p16_x4(__p0, __p1) __extension__ ({ \
  27980.   poly16x8x4_t __s1 = __p1; \
  27981.   poly16x8x4_t __rev1; \
  27982.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  27983.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  27984.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  27985.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  27986.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 37); \
  27987. })
  27988. #endif
  27989.  
  27990. #ifdef __LITTLE_ENDIAN__
  27991. #define vst1q_u8_x4(__p0, __p1) __extension__ ({ \
  27992.   uint8x16x4_t __s1 = __p1; \
  27993.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 48); \
  27994. })
  27995. #else
  27996. #define vst1q_u8_x4(__p0, __p1) __extension__ ({ \
  27997.   uint8x16x4_t __s1 = __p1; \
  27998.   uint8x16x4_t __rev1; \
  27999.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28000.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28001.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28002.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28003.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 48); \
  28004. })
  28005. #endif
  28006.  
  28007. #ifdef __LITTLE_ENDIAN__
  28008. #define vst1q_u32_x4(__p0, __p1) __extension__ ({ \
  28009.   uint32x4x4_t __s1 = __p1; \
  28010.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 50); \
  28011. })
  28012. #else
  28013. #define vst1q_u32_x4(__p0, __p1) __extension__ ({ \
  28014.   uint32x4x4_t __s1 = __p1; \
  28015.   uint32x4x4_t __rev1; \
  28016.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28017.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28018.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28019.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  28020.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 50); \
  28021. })
  28022. #endif
  28023.  
  28024. #ifdef __LITTLE_ENDIAN__
  28025. #define vst1q_u64_x4(__p0, __p1) __extension__ ({ \
  28026.   uint64x2x4_t __s1 = __p1; \
  28027.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 51); \
  28028. })
  28029. #else
  28030. #define vst1q_u64_x4(__p0, __p1) __extension__ ({ \
  28031.   uint64x2x4_t __s1 = __p1; \
  28032.   uint64x2x4_t __rev1; \
  28033.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28034.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28035.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  28036.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  28037.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 51); \
  28038. })
  28039. #endif
  28040.  
  28041. #ifdef __LITTLE_ENDIAN__
  28042. #define vst1q_u16_x4(__p0, __p1) __extension__ ({ \
  28043.   uint16x8x4_t __s1 = __p1; \
  28044.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 49); \
  28045. })
  28046. #else
  28047. #define vst1q_u16_x4(__p0, __p1) __extension__ ({ \
  28048.   uint16x8x4_t __s1 = __p1; \
  28049.   uint16x8x4_t __rev1; \
  28050.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28051.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28052.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  28053.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  28054.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 49); \
  28055. })
  28056. #endif
  28057.  
  28058. #ifdef __LITTLE_ENDIAN__
  28059. #define vst1q_s8_x4(__p0, __p1) __extension__ ({ \
  28060.   int8x16x4_t __s1 = __p1; \
  28061.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 32); \
  28062. })
  28063. #else
  28064. #define vst1q_s8_x4(__p0, __p1) __extension__ ({ \
  28065.   int8x16x4_t __s1 = __p1; \
  28066.   int8x16x4_t __rev1; \
  28067.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28068.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28069.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28070.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28071.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 32); \
  28072. })
  28073. #endif
  28074.  
  28075. #ifdef __LITTLE_ENDIAN__
  28076. #define vst1q_f32_x4(__p0, __p1) __extension__ ({ \
  28077.   float32x4x4_t __s1 = __p1; \
  28078.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 41); \
  28079. })
  28080. #else
  28081. #define vst1q_f32_x4(__p0, __p1) __extension__ ({ \
  28082.   float32x4x4_t __s1 = __p1; \
  28083.   float32x4x4_t __rev1; \
  28084.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28085.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28086.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28087.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  28088.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 41); \
  28089. })
  28090. #endif
  28091.  
  28092. #ifdef __LITTLE_ENDIAN__
  28093. #define vst1q_s32_x4(__p0, __p1) __extension__ ({ \
  28094.   int32x4x4_t __s1 = __p1; \
  28095.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 34); \
  28096. })
  28097. #else
  28098. #define vst1q_s32_x4(__p0, __p1) __extension__ ({ \
  28099.   int32x4x4_t __s1 = __p1; \
  28100.   int32x4x4_t __rev1; \
  28101.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28102.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28103.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28104.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  28105.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 34); \
  28106. })
  28107. #endif
  28108.  
  28109. #ifdef __LITTLE_ENDIAN__
  28110. #define vst1q_s64_x4(__p0, __p1) __extension__ ({ \
  28111.   int64x2x4_t __s1 = __p1; \
  28112.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 35); \
  28113. })
  28114. #else
  28115. #define vst1q_s64_x4(__p0, __p1) __extension__ ({ \
  28116.   int64x2x4_t __s1 = __p1; \
  28117.   int64x2x4_t __rev1; \
  28118.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28119.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28120.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  28121.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  28122.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 35); \
  28123. })
  28124. #endif
  28125.  
  28126. #ifdef __LITTLE_ENDIAN__
  28127. #define vst1q_s16_x4(__p0, __p1) __extension__ ({ \
  28128.   int16x8x4_t __s1 = __p1; \
  28129.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 33); \
  28130. })
  28131. #else
  28132. #define vst1q_s16_x4(__p0, __p1) __extension__ ({ \
  28133.   int16x8x4_t __s1 = __p1; \
  28134.   int16x8x4_t __rev1; \
  28135.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28136.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28137.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  28138.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  28139.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 33); \
  28140. })
  28141. #endif
  28142.  
  28143. #ifdef __LITTLE_ENDIAN__
  28144. #define vst1_u8_x4(__p0, __p1) __extension__ ({ \
  28145.   uint8x8x4_t __s1 = __p1; \
  28146.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 16); \
  28147. })
  28148. #else
  28149. #define vst1_u8_x4(__p0, __p1) __extension__ ({ \
  28150.   uint8x8x4_t __s1 = __p1; \
  28151.   uint8x8x4_t __rev1; \
  28152.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28153.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28154.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  28155.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  28156.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 16); \
  28157. })
  28158. #endif
  28159.  
  28160. #ifdef __LITTLE_ENDIAN__
  28161. #define vst1_u32_x4(__p0, __p1) __extension__ ({ \
  28162.   uint32x2x4_t __s1 = __p1; \
  28163.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 18); \
  28164. })
  28165. #else
  28166. #define vst1_u32_x4(__p0, __p1) __extension__ ({ \
  28167.   uint32x2x4_t __s1 = __p1; \
  28168.   uint32x2x4_t __rev1; \
  28169.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28170.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28171.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  28172.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  28173.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 18); \
  28174. })
  28175. #endif
  28176.  
  28177. #define vst1_u64_x4(__p0, __p1) __extension__ ({ \
  28178.   uint64x1x4_t __s1 = __p1; \
  28179.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
  28180. })
  28181. #ifdef __LITTLE_ENDIAN__
  28182. #define vst1_u16_x4(__p0, __p1) __extension__ ({ \
  28183.   uint16x4x4_t __s1 = __p1; \
  28184.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 17); \
  28185. })
  28186. #else
  28187. #define vst1_u16_x4(__p0, __p1) __extension__ ({ \
  28188.   uint16x4x4_t __s1 = __p1; \
  28189.   uint16x4x4_t __rev1; \
  28190.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28191.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28192.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28193.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  28194.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 17); \
  28195. })
  28196. #endif
  28197.  
  28198. #ifdef __LITTLE_ENDIAN__
  28199. #define vst1_s8_x4(__p0, __p1) __extension__ ({ \
  28200.   int8x8x4_t __s1 = __p1; \
  28201.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 0); \
  28202. })
  28203. #else
  28204. #define vst1_s8_x4(__p0, __p1) __extension__ ({ \
  28205.   int8x8x4_t __s1 = __p1; \
  28206.   int8x8x4_t __rev1; \
  28207.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28208.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28209.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  28210.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  28211.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 0); \
  28212. })
  28213. #endif
  28214.  
  28215. #ifdef __LITTLE_ENDIAN__
  28216. #define vst1_f32_x4(__p0, __p1) __extension__ ({ \
  28217.   float32x2x4_t __s1 = __p1; \
  28218.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 9); \
  28219. })
  28220. #else
  28221. #define vst1_f32_x4(__p0, __p1) __extension__ ({ \
  28222.   float32x2x4_t __s1 = __p1; \
  28223.   float32x2x4_t __rev1; \
  28224.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28225.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28226.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  28227.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  28228.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 9); \
  28229. })
  28230. #endif
  28231.  
  28232. #ifdef __LITTLE_ENDIAN__
  28233. #define vst1_s32_x4(__p0, __p1) __extension__ ({ \
  28234.   int32x2x4_t __s1 = __p1; \
  28235.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 2); \
  28236. })
  28237. #else
  28238. #define vst1_s32_x4(__p0, __p1) __extension__ ({ \
  28239.   int32x2x4_t __s1 = __p1; \
  28240.   int32x2x4_t __rev1; \
  28241.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28242.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28243.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  28244.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  28245.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 2); \
  28246. })
  28247. #endif
  28248.  
  28249. #define vst1_s64_x4(__p0, __p1) __extension__ ({ \
  28250.   int64x1x4_t __s1 = __p1; \
  28251.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 3); \
  28252. })
  28253. #ifdef __LITTLE_ENDIAN__
  28254. #define vst1_s16_x4(__p0, __p1) __extension__ ({ \
  28255.   int16x4x4_t __s1 = __p1; \
  28256.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 1); \
  28257. })
  28258. #else
  28259. #define vst1_s16_x4(__p0, __p1) __extension__ ({ \
  28260.   int16x4x4_t __s1 = __p1; \
  28261.   int16x4x4_t __rev1; \
  28262.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28263.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28264.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28265.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  28266.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 1); \
  28267. })
  28268. #endif
  28269.  
  28270. #ifdef __LITTLE_ENDIAN__
  28271. #define vst2_p8(__p0, __p1) __extension__ ({ \
  28272.   poly8x8x2_t __s1 = __p1; \
  28273.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 4); \
  28274. })
  28275. #else
  28276. #define vst2_p8(__p0, __p1) __extension__ ({ \
  28277.   poly8x8x2_t __s1 = __p1; \
  28278.   poly8x8x2_t __rev1; \
  28279.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28280.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28281.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 4); \
  28282. })
  28283. #endif
  28284.  
  28285. #ifdef __LITTLE_ENDIAN__
  28286. #define vst2_p16(__p0, __p1) __extension__ ({ \
  28287.   poly16x4x2_t __s1 = __p1; \
  28288.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 5); \
  28289. })
  28290. #else
  28291. #define vst2_p16(__p0, __p1) __extension__ ({ \
  28292.   poly16x4x2_t __s1 = __p1; \
  28293.   poly16x4x2_t __rev1; \
  28294.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28295.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28296.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 5); \
  28297. })
  28298. #endif
  28299.  
  28300. #ifdef __LITTLE_ENDIAN__
  28301. #define vst2q_p8(__p0, __p1) __extension__ ({ \
  28302.   poly8x16x2_t __s1 = __p1; \
  28303.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 36); \
  28304. })
  28305. #else
  28306. #define vst2q_p8(__p0, __p1) __extension__ ({ \
  28307.   poly8x16x2_t __s1 = __p1; \
  28308.   poly8x16x2_t __rev1; \
  28309.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28310.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28311.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 36); \
  28312. })
  28313. #endif
  28314.  
  28315. #ifdef __LITTLE_ENDIAN__
  28316. #define vst2q_p16(__p0, __p1) __extension__ ({ \
  28317.   poly16x8x2_t __s1 = __p1; \
  28318.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 37); \
  28319. })
  28320. #else
  28321. #define vst2q_p16(__p0, __p1) __extension__ ({ \
  28322.   poly16x8x2_t __s1 = __p1; \
  28323.   poly16x8x2_t __rev1; \
  28324.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28325.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28326.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 37); \
  28327. })
  28328. #endif
  28329.  
  28330. #ifdef __LITTLE_ENDIAN__
  28331. #define vst2q_u8(__p0, __p1) __extension__ ({ \
  28332.   uint8x16x2_t __s1 = __p1; \
  28333.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 48); \
  28334. })
  28335. #else
  28336. #define vst2q_u8(__p0, __p1) __extension__ ({ \
  28337.   uint8x16x2_t __s1 = __p1; \
  28338.   uint8x16x2_t __rev1; \
  28339.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28340.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28341.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 48); \
  28342. })
  28343. #endif
  28344.  
  28345. #ifdef __LITTLE_ENDIAN__
  28346. #define vst2q_u32(__p0, __p1) __extension__ ({ \
  28347.   uint32x4x2_t __s1 = __p1; \
  28348.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 50); \
  28349. })
  28350. #else
  28351. #define vst2q_u32(__p0, __p1) __extension__ ({ \
  28352.   uint32x4x2_t __s1 = __p1; \
  28353.   uint32x4x2_t __rev1; \
  28354.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28355.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28356.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 50); \
  28357. })
  28358. #endif
  28359.  
  28360. #ifdef __LITTLE_ENDIAN__
  28361. #define vst2q_u16(__p0, __p1) __extension__ ({ \
  28362.   uint16x8x2_t __s1 = __p1; \
  28363.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 49); \
  28364. })
  28365. #else
  28366. #define vst2q_u16(__p0, __p1) __extension__ ({ \
  28367.   uint16x8x2_t __s1 = __p1; \
  28368.   uint16x8x2_t __rev1; \
  28369.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28370.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28371.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 49); \
  28372. })
  28373. #endif
  28374.  
  28375. #ifdef __LITTLE_ENDIAN__
  28376. #define vst2q_s8(__p0, __p1) __extension__ ({ \
  28377.   int8x16x2_t __s1 = __p1; \
  28378.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 32); \
  28379. })
  28380. #else
  28381. #define vst2q_s8(__p0, __p1) __extension__ ({ \
  28382.   int8x16x2_t __s1 = __p1; \
  28383.   int8x16x2_t __rev1; \
  28384.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28385.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28386.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 32); \
  28387. })
  28388. #endif
  28389.  
  28390. #ifdef __LITTLE_ENDIAN__
  28391. #define vst2q_f32(__p0, __p1) __extension__ ({ \
  28392.   float32x4x2_t __s1 = __p1; \
  28393.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 41); \
  28394. })
  28395. #else
  28396. #define vst2q_f32(__p0, __p1) __extension__ ({ \
  28397.   float32x4x2_t __s1 = __p1; \
  28398.   float32x4x2_t __rev1; \
  28399.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28400.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28401.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 41); \
  28402. })
  28403. #endif
  28404.  
  28405. #ifdef __LITTLE_ENDIAN__
  28406. #define vst2q_s32(__p0, __p1) __extension__ ({ \
  28407.   int32x4x2_t __s1 = __p1; \
  28408.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 34); \
  28409. })
  28410. #else
  28411. #define vst2q_s32(__p0, __p1) __extension__ ({ \
  28412.   int32x4x2_t __s1 = __p1; \
  28413.   int32x4x2_t __rev1; \
  28414.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28415.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28416.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 34); \
  28417. })
  28418. #endif
  28419.  
  28420. #ifdef __LITTLE_ENDIAN__
  28421. #define vst2q_s16(__p0, __p1) __extension__ ({ \
  28422.   int16x8x2_t __s1 = __p1; \
  28423.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 33); \
  28424. })
  28425. #else
  28426. #define vst2q_s16(__p0, __p1) __extension__ ({ \
  28427.   int16x8x2_t __s1 = __p1; \
  28428.   int16x8x2_t __rev1; \
  28429.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28430.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28431.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 33); \
  28432. })
  28433. #endif
  28434.  
  28435. #ifdef __LITTLE_ENDIAN__
  28436. #define vst2_u8(__p0, __p1) __extension__ ({ \
  28437.   uint8x8x2_t __s1 = __p1; \
  28438.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 16); \
  28439. })
  28440. #else
  28441. #define vst2_u8(__p0, __p1) __extension__ ({ \
  28442.   uint8x8x2_t __s1 = __p1; \
  28443.   uint8x8x2_t __rev1; \
  28444.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28445.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28446.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 16); \
  28447. })
  28448. #endif
  28449.  
  28450. #ifdef __LITTLE_ENDIAN__
  28451. #define vst2_u32(__p0, __p1) __extension__ ({ \
  28452.   uint32x2x2_t __s1 = __p1; \
  28453.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 18); \
  28454. })
  28455. #else
  28456. #define vst2_u32(__p0, __p1) __extension__ ({ \
  28457.   uint32x2x2_t __s1 = __p1; \
  28458.   uint32x2x2_t __rev1; \
  28459.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28460.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28461.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 18); \
  28462. })
  28463. #endif
  28464.  
  28465. #define vst2_u64(__p0, __p1) __extension__ ({ \
  28466.   uint64x1x2_t __s1 = __p1; \
  28467.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
  28468. })
  28469. #ifdef __LITTLE_ENDIAN__
  28470. #define vst2_u16(__p0, __p1) __extension__ ({ \
  28471.   uint16x4x2_t __s1 = __p1; \
  28472.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 17); \
  28473. })
  28474. #else
  28475. #define vst2_u16(__p0, __p1) __extension__ ({ \
  28476.   uint16x4x2_t __s1 = __p1; \
  28477.   uint16x4x2_t __rev1; \
  28478.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28479.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28480.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 17); \
  28481. })
  28482. #endif
  28483.  
  28484. #ifdef __LITTLE_ENDIAN__
  28485. #define vst2_s8(__p0, __p1) __extension__ ({ \
  28486.   int8x8x2_t __s1 = __p1; \
  28487.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 0); \
  28488. })
  28489. #else
  28490. #define vst2_s8(__p0, __p1) __extension__ ({ \
  28491.   int8x8x2_t __s1 = __p1; \
  28492.   int8x8x2_t __rev1; \
  28493.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28494.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28495.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 0); \
  28496. })
  28497. #endif
  28498.  
  28499. #ifdef __LITTLE_ENDIAN__
  28500. #define vst2_f32(__p0, __p1) __extension__ ({ \
  28501.   float32x2x2_t __s1 = __p1; \
  28502.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 9); \
  28503. })
  28504. #else
  28505. #define vst2_f32(__p0, __p1) __extension__ ({ \
  28506.   float32x2x2_t __s1 = __p1; \
  28507.   float32x2x2_t __rev1; \
  28508.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28509.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28510.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 9); \
  28511. })
  28512. #endif
  28513.  
  28514. #ifdef __LITTLE_ENDIAN__
  28515. #define vst2_s32(__p0, __p1) __extension__ ({ \
  28516.   int32x2x2_t __s1 = __p1; \
  28517.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 2); \
  28518. })
  28519. #else
  28520. #define vst2_s32(__p0, __p1) __extension__ ({ \
  28521.   int32x2x2_t __s1 = __p1; \
  28522.   int32x2x2_t __rev1; \
  28523.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28524.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28525.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 2); \
  28526. })
  28527. #endif
  28528.  
  28529. #define vst2_s64(__p0, __p1) __extension__ ({ \
  28530.   int64x1x2_t __s1 = __p1; \
  28531.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 3); \
  28532. })
  28533. #ifdef __LITTLE_ENDIAN__
  28534. #define vst2_s16(__p0, __p1) __extension__ ({ \
  28535.   int16x4x2_t __s1 = __p1; \
  28536.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 1); \
  28537. })
  28538. #else
  28539. #define vst2_s16(__p0, __p1) __extension__ ({ \
  28540.   int16x4x2_t __s1 = __p1; \
  28541.   int16x4x2_t __rev1; \
  28542.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28543.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28544.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 1); \
  28545. })
  28546. #endif
  28547.  
  28548. #ifdef __LITTLE_ENDIAN__
  28549. #define vst2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  28550.   poly8x8x2_t __s1 = __p1; \
  28551.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 4); \
  28552. })
  28553. #else
  28554. #define vst2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  28555.   poly8x8x2_t __s1 = __p1; \
  28556.   poly8x8x2_t __rev1; \
  28557.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28558.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28559.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 4); \
  28560. })
  28561. #endif
  28562.  
  28563. #ifdef __LITTLE_ENDIAN__
  28564. #define vst2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  28565.   poly16x4x2_t __s1 = __p1; \
  28566.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 5); \
  28567. })
  28568. #else
  28569. #define vst2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  28570.   poly16x4x2_t __s1 = __p1; \
  28571.   poly16x4x2_t __rev1; \
  28572.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28573.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28574.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 5); \
  28575. })
  28576. #endif
  28577.  
  28578. #ifdef __LITTLE_ENDIAN__
  28579. #define vst2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  28580.   poly16x8x2_t __s1 = __p1; \
  28581.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 37); \
  28582. })
  28583. #else
  28584. #define vst2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  28585.   poly16x8x2_t __s1 = __p1; \
  28586.   poly16x8x2_t __rev1; \
  28587.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28588.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28589.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 37); \
  28590. })
  28591. #endif
  28592.  
  28593. #ifdef __LITTLE_ENDIAN__
  28594. #define vst2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  28595.   uint32x4x2_t __s1 = __p1; \
  28596.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 50); \
  28597. })
  28598. #else
  28599. #define vst2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  28600.   uint32x4x2_t __s1 = __p1; \
  28601.   uint32x4x2_t __rev1; \
  28602.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28603.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28604.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 50); \
  28605. })
  28606. #endif
  28607.  
  28608. #ifdef __LITTLE_ENDIAN__
  28609. #define vst2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  28610.   uint16x8x2_t __s1 = __p1; \
  28611.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 49); \
  28612. })
  28613. #else
  28614. #define vst2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  28615.   uint16x8x2_t __s1 = __p1; \
  28616.   uint16x8x2_t __rev1; \
  28617.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28618.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28619.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 49); \
  28620. })
  28621. #endif
  28622.  
  28623. #ifdef __LITTLE_ENDIAN__
  28624. #define vst2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  28625.   float32x4x2_t __s1 = __p1; \
  28626.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 41); \
  28627. })
  28628. #else
  28629. #define vst2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  28630.   float32x4x2_t __s1 = __p1; \
  28631.   float32x4x2_t __rev1; \
  28632.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28633.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28634.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 41); \
  28635. })
  28636. #endif
  28637.  
  28638. #ifdef __LITTLE_ENDIAN__
  28639. #define vst2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  28640.   int32x4x2_t __s1 = __p1; \
  28641.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 34); \
  28642. })
  28643. #else
  28644. #define vst2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  28645.   int32x4x2_t __s1 = __p1; \
  28646.   int32x4x2_t __rev1; \
  28647.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28648.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28649.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 34); \
  28650. })
  28651. #endif
  28652.  
  28653. #ifdef __LITTLE_ENDIAN__
  28654. #define vst2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  28655.   int16x8x2_t __s1 = __p1; \
  28656.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 33); \
  28657. })
  28658. #else
  28659. #define vst2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  28660.   int16x8x2_t __s1 = __p1; \
  28661.   int16x8x2_t __rev1; \
  28662.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28663.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28664.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 33); \
  28665. })
  28666. #endif
  28667.  
  28668. #ifdef __LITTLE_ENDIAN__
  28669. #define vst2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  28670.   uint8x8x2_t __s1 = __p1; \
  28671.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 16); \
  28672. })
  28673. #else
  28674. #define vst2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  28675.   uint8x8x2_t __s1 = __p1; \
  28676.   uint8x8x2_t __rev1; \
  28677.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28678.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28679.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 16); \
  28680. })
  28681. #endif
  28682.  
  28683. #ifdef __LITTLE_ENDIAN__
  28684. #define vst2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  28685.   uint32x2x2_t __s1 = __p1; \
  28686.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 18); \
  28687. })
  28688. #else
  28689. #define vst2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  28690.   uint32x2x2_t __s1 = __p1; \
  28691.   uint32x2x2_t __rev1; \
  28692.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28693.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28694.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 18); \
  28695. })
  28696. #endif
  28697.  
  28698. #ifdef __LITTLE_ENDIAN__
  28699. #define vst2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  28700.   uint16x4x2_t __s1 = __p1; \
  28701.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 17); \
  28702. })
  28703. #else
  28704. #define vst2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  28705.   uint16x4x2_t __s1 = __p1; \
  28706.   uint16x4x2_t __rev1; \
  28707.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28708.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28709.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 17); \
  28710. })
  28711. #endif
  28712.  
  28713. #ifdef __LITTLE_ENDIAN__
  28714. #define vst2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  28715.   int8x8x2_t __s1 = __p1; \
  28716.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 0); \
  28717. })
  28718. #else
  28719. #define vst2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  28720.   int8x8x2_t __s1 = __p1; \
  28721.   int8x8x2_t __rev1; \
  28722.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28723.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28724.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 0); \
  28725. })
  28726. #endif
  28727.  
  28728. #ifdef __LITTLE_ENDIAN__
  28729. #define vst2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  28730.   float32x2x2_t __s1 = __p1; \
  28731.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 9); \
  28732. })
  28733. #else
  28734. #define vst2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  28735.   float32x2x2_t __s1 = __p1; \
  28736.   float32x2x2_t __rev1; \
  28737.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28738.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28739.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 9); \
  28740. })
  28741. #endif
  28742.  
  28743. #ifdef __LITTLE_ENDIAN__
  28744. #define vst2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  28745.   int32x2x2_t __s1 = __p1; \
  28746.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 2); \
  28747. })
  28748. #else
  28749. #define vst2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  28750.   int32x2x2_t __s1 = __p1; \
  28751.   int32x2x2_t __rev1; \
  28752.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28753.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28754.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 2); \
  28755. })
  28756. #endif
  28757.  
  28758. #ifdef __LITTLE_ENDIAN__
  28759. #define vst2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  28760.   int16x4x2_t __s1 = __p1; \
  28761.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 1); \
  28762. })
  28763. #else
  28764. #define vst2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  28765.   int16x4x2_t __s1 = __p1; \
  28766.   int16x4x2_t __rev1; \
  28767.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28768.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28769.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 1); \
  28770. })
  28771. #endif
  28772.  
  28773. #ifdef __LITTLE_ENDIAN__
  28774. #define vst3_p8(__p0, __p1) __extension__ ({ \
  28775.   poly8x8x3_t __s1 = __p1; \
  28776.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 4); \
  28777. })
  28778. #else
  28779. #define vst3_p8(__p0, __p1) __extension__ ({ \
  28780.   poly8x8x3_t __s1 = __p1; \
  28781.   poly8x8x3_t __rev1; \
  28782.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28783.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28784.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  28785.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 4); \
  28786. })
  28787. #endif
  28788.  
  28789. #ifdef __LITTLE_ENDIAN__
  28790. #define vst3_p16(__p0, __p1) __extension__ ({ \
  28791.   poly16x4x3_t __s1 = __p1; \
  28792.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 5); \
  28793. })
  28794. #else
  28795. #define vst3_p16(__p0, __p1) __extension__ ({ \
  28796.   poly16x4x3_t __s1 = __p1; \
  28797.   poly16x4x3_t __rev1; \
  28798.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28799.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28800.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28801.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 5); \
  28802. })
  28803. #endif
  28804.  
  28805. #ifdef __LITTLE_ENDIAN__
  28806. #define vst3q_p8(__p0, __p1) __extension__ ({ \
  28807.   poly8x16x3_t __s1 = __p1; \
  28808.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 36); \
  28809. })
  28810. #else
  28811. #define vst3q_p8(__p0, __p1) __extension__ ({ \
  28812.   poly8x16x3_t __s1 = __p1; \
  28813.   poly8x16x3_t __rev1; \
  28814.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28815.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28816.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28817.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 36); \
  28818. })
  28819. #endif
  28820.  
  28821. #ifdef __LITTLE_ENDIAN__
  28822. #define vst3q_p16(__p0, __p1) __extension__ ({ \
  28823.   poly16x8x3_t __s1 = __p1; \
  28824.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 37); \
  28825. })
  28826. #else
  28827. #define vst3q_p16(__p0, __p1) __extension__ ({ \
  28828.   poly16x8x3_t __s1 = __p1; \
  28829.   poly16x8x3_t __rev1; \
  28830.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28831.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28832.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  28833.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 37); \
  28834. })
  28835. #endif
  28836.  
  28837. #ifdef __LITTLE_ENDIAN__
  28838. #define vst3q_u8(__p0, __p1) __extension__ ({ \
  28839.   uint8x16x3_t __s1 = __p1; \
  28840.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 48); \
  28841. })
  28842. #else
  28843. #define vst3q_u8(__p0, __p1) __extension__ ({ \
  28844.   uint8x16x3_t __s1 = __p1; \
  28845.   uint8x16x3_t __rev1; \
  28846.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28847.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28848.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28849.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 48); \
  28850. })
  28851. #endif
  28852.  
  28853. #ifdef __LITTLE_ENDIAN__
  28854. #define vst3q_u32(__p0, __p1) __extension__ ({ \
  28855.   uint32x4x3_t __s1 = __p1; \
  28856.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 50); \
  28857. })
  28858. #else
  28859. #define vst3q_u32(__p0, __p1) __extension__ ({ \
  28860.   uint32x4x3_t __s1 = __p1; \
  28861.   uint32x4x3_t __rev1; \
  28862.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28863.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28864.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28865.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 50); \
  28866. })
  28867. #endif
  28868.  
  28869. #ifdef __LITTLE_ENDIAN__
  28870. #define vst3q_u16(__p0, __p1) __extension__ ({ \
  28871.   uint16x8x3_t __s1 = __p1; \
  28872.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 49); \
  28873. })
  28874. #else
  28875. #define vst3q_u16(__p0, __p1) __extension__ ({ \
  28876.   uint16x8x3_t __s1 = __p1; \
  28877.   uint16x8x3_t __rev1; \
  28878.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28879.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28880.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  28881.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 49); \
  28882. })
  28883. #endif
  28884.  
  28885. #ifdef __LITTLE_ENDIAN__
  28886. #define vst3q_s8(__p0, __p1) __extension__ ({ \
  28887.   int8x16x3_t __s1 = __p1; \
  28888.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 32); \
  28889. })
  28890. #else
  28891. #define vst3q_s8(__p0, __p1) __extension__ ({ \
  28892.   int8x16x3_t __s1 = __p1; \
  28893.   int8x16x3_t __rev1; \
  28894.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28895.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28896.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  28897.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 32); \
  28898. })
  28899. #endif
  28900.  
  28901. #ifdef __LITTLE_ENDIAN__
  28902. #define vst3q_f32(__p0, __p1) __extension__ ({ \
  28903.   float32x4x3_t __s1 = __p1; \
  28904.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 41); \
  28905. })
  28906. #else
  28907. #define vst3q_f32(__p0, __p1) __extension__ ({ \
  28908.   float32x4x3_t __s1 = __p1; \
  28909.   float32x4x3_t __rev1; \
  28910.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28911.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28912.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28913.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 41); \
  28914. })
  28915. #endif
  28916.  
  28917. #ifdef __LITTLE_ENDIAN__
  28918. #define vst3q_s32(__p0, __p1) __extension__ ({ \
  28919.   int32x4x3_t __s1 = __p1; \
  28920.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 34); \
  28921. })
  28922. #else
  28923. #define vst3q_s32(__p0, __p1) __extension__ ({ \
  28924.   int32x4x3_t __s1 = __p1; \
  28925.   int32x4x3_t __rev1; \
  28926.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28927.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28928.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28929.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 34); \
  28930. })
  28931. #endif
  28932.  
  28933. #ifdef __LITTLE_ENDIAN__
  28934. #define vst3q_s16(__p0, __p1) __extension__ ({ \
  28935.   int16x8x3_t __s1 = __p1; \
  28936.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 33); \
  28937. })
  28938. #else
  28939. #define vst3q_s16(__p0, __p1) __extension__ ({ \
  28940.   int16x8x3_t __s1 = __p1; \
  28941.   int16x8x3_t __rev1; \
  28942.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28943.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28944.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  28945.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 33); \
  28946. })
  28947. #endif
  28948.  
  28949. #ifdef __LITTLE_ENDIAN__
  28950. #define vst3_u8(__p0, __p1) __extension__ ({ \
  28951.   uint8x8x3_t __s1 = __p1; \
  28952.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 16); \
  28953. })
  28954. #else
  28955. #define vst3_u8(__p0, __p1) __extension__ ({ \
  28956.   uint8x8x3_t __s1 = __p1; \
  28957.   uint8x8x3_t __rev1; \
  28958.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  28959.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  28960.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  28961.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 16); \
  28962. })
  28963. #endif
  28964.  
  28965. #ifdef __LITTLE_ENDIAN__
  28966. #define vst3_u32(__p0, __p1) __extension__ ({ \
  28967.   uint32x2x3_t __s1 = __p1; \
  28968.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 18); \
  28969. })
  28970. #else
  28971. #define vst3_u32(__p0, __p1) __extension__ ({ \
  28972.   uint32x2x3_t __s1 = __p1; \
  28973.   uint32x2x3_t __rev1; \
  28974.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  28975.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  28976.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  28977.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 18); \
  28978. })
  28979. #endif
  28980.  
  28981. #define vst3_u64(__p0, __p1) __extension__ ({ \
  28982.   uint64x1x3_t __s1 = __p1; \
  28983.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
  28984. })
  28985. #ifdef __LITTLE_ENDIAN__
  28986. #define vst3_u16(__p0, __p1) __extension__ ({ \
  28987.   uint16x4x3_t __s1 = __p1; \
  28988.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 17); \
  28989. })
  28990. #else
  28991. #define vst3_u16(__p0, __p1) __extension__ ({ \
  28992.   uint16x4x3_t __s1 = __p1; \
  28993.   uint16x4x3_t __rev1; \
  28994.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  28995.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  28996.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  28997.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 17); \
  28998. })
  28999. #endif
  29000.  
  29001. #ifdef __LITTLE_ENDIAN__
  29002. #define vst3_s8(__p0, __p1) __extension__ ({ \
  29003.   int8x8x3_t __s1 = __p1; \
  29004.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 0); \
  29005. })
  29006. #else
  29007. #define vst3_s8(__p0, __p1) __extension__ ({ \
  29008.   int8x8x3_t __s1 = __p1; \
  29009.   int8x8x3_t __rev1; \
  29010.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29011.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29012.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29013.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 0); \
  29014. })
  29015. #endif
  29016.  
  29017. #ifdef __LITTLE_ENDIAN__
  29018. #define vst3_f32(__p0, __p1) __extension__ ({ \
  29019.   float32x2x3_t __s1 = __p1; \
  29020.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 9); \
  29021. })
  29022. #else
  29023. #define vst3_f32(__p0, __p1) __extension__ ({ \
  29024.   float32x2x3_t __s1 = __p1; \
  29025.   float32x2x3_t __rev1; \
  29026.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29027.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29028.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29029.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 9); \
  29030. })
  29031. #endif
  29032.  
  29033. #ifdef __LITTLE_ENDIAN__
  29034. #define vst3_s32(__p0, __p1) __extension__ ({ \
  29035.   int32x2x3_t __s1 = __p1; \
  29036.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 2); \
  29037. })
  29038. #else
  29039. #define vst3_s32(__p0, __p1) __extension__ ({ \
  29040.   int32x2x3_t __s1 = __p1; \
  29041.   int32x2x3_t __rev1; \
  29042.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29043.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29044.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29045.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 2); \
  29046. })
  29047. #endif
  29048.  
  29049. #define vst3_s64(__p0, __p1) __extension__ ({ \
  29050.   int64x1x3_t __s1 = __p1; \
  29051.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 3); \
  29052. })
  29053. #ifdef __LITTLE_ENDIAN__
  29054. #define vst3_s16(__p0, __p1) __extension__ ({ \
  29055.   int16x4x3_t __s1 = __p1; \
  29056.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 1); \
  29057. })
  29058. #else
  29059. #define vst3_s16(__p0, __p1) __extension__ ({ \
  29060.   int16x4x3_t __s1 = __p1; \
  29061.   int16x4x3_t __rev1; \
  29062.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29063.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29064.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29065.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 1); \
  29066. })
  29067. #endif
  29068.  
  29069. #ifdef __LITTLE_ENDIAN__
  29070. #define vst3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  29071.   poly8x8x3_t __s1 = __p1; \
  29072.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 4); \
  29073. })
  29074. #else
  29075. #define vst3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  29076.   poly8x8x3_t __s1 = __p1; \
  29077.   poly8x8x3_t __rev1; \
  29078.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29079.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29080.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29081.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 4); \
  29082. })
  29083. #endif
  29084.  
  29085. #ifdef __LITTLE_ENDIAN__
  29086. #define vst3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  29087.   poly16x4x3_t __s1 = __p1; \
  29088.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 5); \
  29089. })
  29090. #else
  29091. #define vst3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  29092.   poly16x4x3_t __s1 = __p1; \
  29093.   poly16x4x3_t __rev1; \
  29094.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29095.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29096.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29097.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 5); \
  29098. })
  29099. #endif
  29100.  
  29101. #ifdef __LITTLE_ENDIAN__
  29102. #define vst3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  29103.   poly16x8x3_t __s1 = __p1; \
  29104.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 37); \
  29105. })
  29106. #else
  29107. #define vst3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  29108.   poly16x8x3_t __s1 = __p1; \
  29109.   poly16x8x3_t __rev1; \
  29110.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29111.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29112.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29113.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 37); \
  29114. })
  29115. #endif
  29116.  
  29117. #ifdef __LITTLE_ENDIAN__
  29118. #define vst3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  29119.   uint32x4x3_t __s1 = __p1; \
  29120.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 50); \
  29121. })
  29122. #else
  29123. #define vst3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  29124.   uint32x4x3_t __s1 = __p1; \
  29125.   uint32x4x3_t __rev1; \
  29126.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29127.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29128.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29129.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 50); \
  29130. })
  29131. #endif
  29132.  
  29133. #ifdef __LITTLE_ENDIAN__
  29134. #define vst3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  29135.   uint16x8x3_t __s1 = __p1; \
  29136.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 49); \
  29137. })
  29138. #else
  29139. #define vst3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  29140.   uint16x8x3_t __s1 = __p1; \
  29141.   uint16x8x3_t __rev1; \
  29142.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29143.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29144.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29145.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 49); \
  29146. })
  29147. #endif
  29148.  
  29149. #ifdef __LITTLE_ENDIAN__
  29150. #define vst3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  29151.   float32x4x3_t __s1 = __p1; \
  29152.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 41); \
  29153. })
  29154. #else
  29155. #define vst3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  29156.   float32x4x3_t __s1 = __p1; \
  29157.   float32x4x3_t __rev1; \
  29158.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29159.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29160.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29161.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 41); \
  29162. })
  29163. #endif
  29164.  
  29165. #ifdef __LITTLE_ENDIAN__
  29166. #define vst3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  29167.   int32x4x3_t __s1 = __p1; \
  29168.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 34); \
  29169. })
  29170. #else
  29171. #define vst3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  29172.   int32x4x3_t __s1 = __p1; \
  29173.   int32x4x3_t __rev1; \
  29174.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29175.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29176.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29177.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 34); \
  29178. })
  29179. #endif
  29180.  
  29181. #ifdef __LITTLE_ENDIAN__
  29182. #define vst3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  29183.   int16x8x3_t __s1 = __p1; \
  29184.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 33); \
  29185. })
  29186. #else
  29187. #define vst3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  29188.   int16x8x3_t __s1 = __p1; \
  29189.   int16x8x3_t __rev1; \
  29190.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29191.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29192.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29193.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 33); \
  29194. })
  29195. #endif
  29196.  
  29197. #ifdef __LITTLE_ENDIAN__
  29198. #define vst3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  29199.   uint8x8x3_t __s1 = __p1; \
  29200.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 16); \
  29201. })
  29202. #else
  29203. #define vst3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  29204.   uint8x8x3_t __s1 = __p1; \
  29205.   uint8x8x3_t __rev1; \
  29206.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29207.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29208.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29209.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 16); \
  29210. })
  29211. #endif
  29212.  
  29213. #ifdef __LITTLE_ENDIAN__
  29214. #define vst3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  29215.   uint32x2x3_t __s1 = __p1; \
  29216.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 18); \
  29217. })
  29218. #else
  29219. #define vst3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  29220.   uint32x2x3_t __s1 = __p1; \
  29221.   uint32x2x3_t __rev1; \
  29222.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29223.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29224.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29225.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 18); \
  29226. })
  29227. #endif
  29228.  
  29229. #ifdef __LITTLE_ENDIAN__
  29230. #define vst3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  29231.   uint16x4x3_t __s1 = __p1; \
  29232.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 17); \
  29233. })
  29234. #else
  29235. #define vst3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  29236.   uint16x4x3_t __s1 = __p1; \
  29237.   uint16x4x3_t __rev1; \
  29238.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29239.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29240.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29241.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 17); \
  29242. })
  29243. #endif
  29244.  
  29245. #ifdef __LITTLE_ENDIAN__
  29246. #define vst3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  29247.   int8x8x3_t __s1 = __p1; \
  29248.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 0); \
  29249. })
  29250. #else
  29251. #define vst3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  29252.   int8x8x3_t __s1 = __p1; \
  29253.   int8x8x3_t __rev1; \
  29254.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29255.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29256.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29257.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 0); \
  29258. })
  29259. #endif
  29260.  
  29261. #ifdef __LITTLE_ENDIAN__
  29262. #define vst3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  29263.   float32x2x3_t __s1 = __p1; \
  29264.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 9); \
  29265. })
  29266. #else
  29267. #define vst3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  29268.   float32x2x3_t __s1 = __p1; \
  29269.   float32x2x3_t __rev1; \
  29270.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29271.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29272.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29273.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 9); \
  29274. })
  29275. #endif
  29276.  
  29277. #ifdef __LITTLE_ENDIAN__
  29278. #define vst3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  29279.   int32x2x3_t __s1 = __p1; \
  29280.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 2); \
  29281. })
  29282. #else
  29283. #define vst3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  29284.   int32x2x3_t __s1 = __p1; \
  29285.   int32x2x3_t __rev1; \
  29286.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29287.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29288.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29289.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 2); \
  29290. })
  29291. #endif
  29292.  
  29293. #ifdef __LITTLE_ENDIAN__
  29294. #define vst3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  29295.   int16x4x3_t __s1 = __p1; \
  29296.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 1); \
  29297. })
  29298. #else
  29299. #define vst3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  29300.   int16x4x3_t __s1 = __p1; \
  29301.   int16x4x3_t __rev1; \
  29302.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29303.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29304.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29305.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 1); \
  29306. })
  29307. #endif
  29308.  
  29309. #ifdef __LITTLE_ENDIAN__
  29310. #define vst4_p8(__p0, __p1) __extension__ ({ \
  29311.   poly8x8x4_t __s1 = __p1; \
  29312.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 4); \
  29313. })
  29314. #else
  29315. #define vst4_p8(__p0, __p1) __extension__ ({ \
  29316.   poly8x8x4_t __s1 = __p1; \
  29317.   poly8x8x4_t __rev1; \
  29318.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29319.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29320.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29321.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29322.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 4); \
  29323. })
  29324. #endif
  29325.  
  29326. #ifdef __LITTLE_ENDIAN__
  29327. #define vst4_p16(__p0, __p1) __extension__ ({ \
  29328.   poly16x4x4_t __s1 = __p1; \
  29329.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 5); \
  29330. })
  29331. #else
  29332. #define vst4_p16(__p0, __p1) __extension__ ({ \
  29333.   poly16x4x4_t __s1 = __p1; \
  29334.   poly16x4x4_t __rev1; \
  29335.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29336.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29337.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29338.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29339.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 5); \
  29340. })
  29341. #endif
  29342.  
  29343. #ifdef __LITTLE_ENDIAN__
  29344. #define vst4q_p8(__p0, __p1) __extension__ ({ \
  29345.   poly8x16x4_t __s1 = __p1; \
  29346.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 36); \
  29347. })
  29348. #else
  29349. #define vst4q_p8(__p0, __p1) __extension__ ({ \
  29350.   poly8x16x4_t __s1 = __p1; \
  29351.   poly8x16x4_t __rev1; \
  29352.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29353.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29354.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29355.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29356.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 36); \
  29357. })
  29358. #endif
  29359.  
  29360. #ifdef __LITTLE_ENDIAN__
  29361. #define vst4q_p16(__p0, __p1) __extension__ ({ \
  29362.   poly16x8x4_t __s1 = __p1; \
  29363.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 37); \
  29364. })
  29365. #else
  29366. #define vst4q_p16(__p0, __p1) __extension__ ({ \
  29367.   poly16x8x4_t __s1 = __p1; \
  29368.   poly16x8x4_t __rev1; \
  29369.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29370.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29371.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29372.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29373.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 37); \
  29374. })
  29375. #endif
  29376.  
  29377. #ifdef __LITTLE_ENDIAN__
  29378. #define vst4q_u8(__p0, __p1) __extension__ ({ \
  29379.   uint8x16x4_t __s1 = __p1; \
  29380.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 48); \
  29381. })
  29382. #else
  29383. #define vst4q_u8(__p0, __p1) __extension__ ({ \
  29384.   uint8x16x4_t __s1 = __p1; \
  29385.   uint8x16x4_t __rev1; \
  29386.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29387.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29388.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29389.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29390.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 48); \
  29391. })
  29392. #endif
  29393.  
  29394. #ifdef __LITTLE_ENDIAN__
  29395. #define vst4q_u32(__p0, __p1) __extension__ ({ \
  29396.   uint32x4x4_t __s1 = __p1; \
  29397.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 50); \
  29398. })
  29399. #else
  29400. #define vst4q_u32(__p0, __p1) __extension__ ({ \
  29401.   uint32x4x4_t __s1 = __p1; \
  29402.   uint32x4x4_t __rev1; \
  29403.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29404.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29405.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29406.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29407.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 50); \
  29408. })
  29409. #endif
  29410.  
  29411. #ifdef __LITTLE_ENDIAN__
  29412. #define vst4q_u16(__p0, __p1) __extension__ ({ \
  29413.   uint16x8x4_t __s1 = __p1; \
  29414.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 49); \
  29415. })
  29416. #else
  29417. #define vst4q_u16(__p0, __p1) __extension__ ({ \
  29418.   uint16x8x4_t __s1 = __p1; \
  29419.   uint16x8x4_t __rev1; \
  29420.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29421.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29422.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29423.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29424.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 49); \
  29425. })
  29426. #endif
  29427.  
  29428. #ifdef __LITTLE_ENDIAN__
  29429. #define vst4q_s8(__p0, __p1) __extension__ ({ \
  29430.   int8x16x4_t __s1 = __p1; \
  29431.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 32); \
  29432. })
  29433. #else
  29434. #define vst4q_s8(__p0, __p1) __extension__ ({ \
  29435.   int8x16x4_t __s1 = __p1; \
  29436.   int8x16x4_t __rev1; \
  29437.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29438.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29439.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29440.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  29441.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 32); \
  29442. })
  29443. #endif
  29444.  
  29445. #ifdef __LITTLE_ENDIAN__
  29446. #define vst4q_f32(__p0, __p1) __extension__ ({ \
  29447.   float32x4x4_t __s1 = __p1; \
  29448.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 41); \
  29449. })
  29450. #else
  29451. #define vst4q_f32(__p0, __p1) __extension__ ({ \
  29452.   float32x4x4_t __s1 = __p1; \
  29453.   float32x4x4_t __rev1; \
  29454.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29455.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29456.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29457.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29458.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 41); \
  29459. })
  29460. #endif
  29461.  
  29462. #ifdef __LITTLE_ENDIAN__
  29463. #define vst4q_s32(__p0, __p1) __extension__ ({ \
  29464.   int32x4x4_t __s1 = __p1; \
  29465.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 34); \
  29466. })
  29467. #else
  29468. #define vst4q_s32(__p0, __p1) __extension__ ({ \
  29469.   int32x4x4_t __s1 = __p1; \
  29470.   int32x4x4_t __rev1; \
  29471.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29472.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29473.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29474.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29475.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 34); \
  29476. })
  29477. #endif
  29478.  
  29479. #ifdef __LITTLE_ENDIAN__
  29480. #define vst4q_s16(__p0, __p1) __extension__ ({ \
  29481.   int16x8x4_t __s1 = __p1; \
  29482.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 33); \
  29483. })
  29484. #else
  29485. #define vst4q_s16(__p0, __p1) __extension__ ({ \
  29486.   int16x8x4_t __s1 = __p1; \
  29487.   int16x8x4_t __rev1; \
  29488.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29489.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29490.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29491.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29492.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 33); \
  29493. })
  29494. #endif
  29495.  
  29496. #ifdef __LITTLE_ENDIAN__
  29497. #define vst4_u8(__p0, __p1) __extension__ ({ \
  29498.   uint8x8x4_t __s1 = __p1; \
  29499.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 16); \
  29500. })
  29501. #else
  29502. #define vst4_u8(__p0, __p1) __extension__ ({ \
  29503.   uint8x8x4_t __s1 = __p1; \
  29504.   uint8x8x4_t __rev1; \
  29505.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29506.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29507.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29508.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29509.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 16); \
  29510. })
  29511. #endif
  29512.  
  29513. #ifdef __LITTLE_ENDIAN__
  29514. #define vst4_u32(__p0, __p1) __extension__ ({ \
  29515.   uint32x2x4_t __s1 = __p1; \
  29516.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 18); \
  29517. })
  29518. #else
  29519. #define vst4_u32(__p0, __p1) __extension__ ({ \
  29520.   uint32x2x4_t __s1 = __p1; \
  29521.   uint32x2x4_t __rev1; \
  29522.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29523.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29524.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29525.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  29526.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 18); \
  29527. })
  29528. #endif
  29529.  
  29530. #define vst4_u64(__p0, __p1) __extension__ ({ \
  29531.   uint64x1x4_t __s1 = __p1; \
  29532.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
  29533. })
  29534. #ifdef __LITTLE_ENDIAN__
  29535. #define vst4_u16(__p0, __p1) __extension__ ({ \
  29536.   uint16x4x4_t __s1 = __p1; \
  29537.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 17); \
  29538. })
  29539. #else
  29540. #define vst4_u16(__p0, __p1) __extension__ ({ \
  29541.   uint16x4x4_t __s1 = __p1; \
  29542.   uint16x4x4_t __rev1; \
  29543.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29544.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29545.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29546.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29547.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 17); \
  29548. })
  29549. #endif
  29550.  
  29551. #ifdef __LITTLE_ENDIAN__
  29552. #define vst4_s8(__p0, __p1) __extension__ ({ \
  29553.   int8x8x4_t __s1 = __p1; \
  29554.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 0); \
  29555. })
  29556. #else
  29557. #define vst4_s8(__p0, __p1) __extension__ ({ \
  29558.   int8x8x4_t __s1 = __p1; \
  29559.   int8x8x4_t __rev1; \
  29560.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29561.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29562.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29563.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29564.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 0); \
  29565. })
  29566. #endif
  29567.  
  29568. #ifdef __LITTLE_ENDIAN__
  29569. #define vst4_f32(__p0, __p1) __extension__ ({ \
  29570.   float32x2x4_t __s1 = __p1; \
  29571.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 9); \
  29572. })
  29573. #else
  29574. #define vst4_f32(__p0, __p1) __extension__ ({ \
  29575.   float32x2x4_t __s1 = __p1; \
  29576.   float32x2x4_t __rev1; \
  29577.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29578.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29579.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29580.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  29581.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 9); \
  29582. })
  29583. #endif
  29584.  
  29585. #ifdef __LITTLE_ENDIAN__
  29586. #define vst4_s32(__p0, __p1) __extension__ ({ \
  29587.   int32x2x4_t __s1 = __p1; \
  29588.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 2); \
  29589. })
  29590. #else
  29591. #define vst4_s32(__p0, __p1) __extension__ ({ \
  29592.   int32x2x4_t __s1 = __p1; \
  29593.   int32x2x4_t __rev1; \
  29594.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29595.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29596.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29597.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  29598.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 2); \
  29599. })
  29600. #endif
  29601.  
  29602. #define vst4_s64(__p0, __p1) __extension__ ({ \
  29603.   int64x1x4_t __s1 = __p1; \
  29604.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 3); \
  29605. })
  29606. #ifdef __LITTLE_ENDIAN__
  29607. #define vst4_s16(__p0, __p1) __extension__ ({ \
  29608.   int16x4x4_t __s1 = __p1; \
  29609.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 1); \
  29610. })
  29611. #else
  29612. #define vst4_s16(__p0, __p1) __extension__ ({ \
  29613.   int16x4x4_t __s1 = __p1; \
  29614.   int16x4x4_t __rev1; \
  29615.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29616.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29617.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29618.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29619.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 1); \
  29620. })
  29621. #endif
  29622.  
  29623. #ifdef __LITTLE_ENDIAN__
  29624. #define vst4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  29625.   poly8x8x4_t __s1 = __p1; \
  29626.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 4); \
  29627. })
  29628. #else
  29629. #define vst4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  29630.   poly8x8x4_t __s1 = __p1; \
  29631.   poly8x8x4_t __rev1; \
  29632.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29633.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29634.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29635.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29636.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 4); \
  29637. })
  29638. #endif
  29639.  
  29640. #ifdef __LITTLE_ENDIAN__
  29641. #define vst4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  29642.   poly16x4x4_t __s1 = __p1; \
  29643.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 5); \
  29644. })
  29645. #else
  29646. #define vst4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  29647.   poly16x4x4_t __s1 = __p1; \
  29648.   poly16x4x4_t __rev1; \
  29649.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29650.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29651.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29652.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29653.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 5); \
  29654. })
  29655. #endif
  29656.  
  29657. #ifdef __LITTLE_ENDIAN__
  29658. #define vst4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  29659.   poly16x8x4_t __s1 = __p1; \
  29660.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 37); \
  29661. })
  29662. #else
  29663. #define vst4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
  29664.   poly16x8x4_t __s1 = __p1; \
  29665.   poly16x8x4_t __rev1; \
  29666.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29667.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29668.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29669.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29670.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 37); \
  29671. })
  29672. #endif
  29673.  
  29674. #ifdef __LITTLE_ENDIAN__
  29675. #define vst4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  29676.   uint32x4x4_t __s1 = __p1; \
  29677.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 50); \
  29678. })
  29679. #else
  29680. #define vst4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  29681.   uint32x4x4_t __s1 = __p1; \
  29682.   uint32x4x4_t __rev1; \
  29683.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29684.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29685.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29686.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29687.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 50); \
  29688. })
  29689. #endif
  29690.  
  29691. #ifdef __LITTLE_ENDIAN__
  29692. #define vst4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  29693.   uint16x8x4_t __s1 = __p1; \
  29694.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 49); \
  29695. })
  29696. #else
  29697. #define vst4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  29698.   uint16x8x4_t __s1 = __p1; \
  29699.   uint16x8x4_t __rev1; \
  29700.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29701.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29702.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29703.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29704.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 49); \
  29705. })
  29706. #endif
  29707.  
  29708. #ifdef __LITTLE_ENDIAN__
  29709. #define vst4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  29710.   float32x4x4_t __s1 = __p1; \
  29711.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 41); \
  29712. })
  29713. #else
  29714. #define vst4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  29715.   float32x4x4_t __s1 = __p1; \
  29716.   float32x4x4_t __rev1; \
  29717.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29718.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29719.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29720.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29721.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 41); \
  29722. })
  29723. #endif
  29724.  
  29725. #ifdef __LITTLE_ENDIAN__
  29726. #define vst4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  29727.   int32x4x4_t __s1 = __p1; \
  29728.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 34); \
  29729. })
  29730. #else
  29731. #define vst4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  29732.   int32x4x4_t __s1 = __p1; \
  29733.   int32x4x4_t __rev1; \
  29734.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29735.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29736.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29737.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29738.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 34); \
  29739. })
  29740. #endif
  29741.  
  29742. #ifdef __LITTLE_ENDIAN__
  29743. #define vst4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  29744.   int16x8x4_t __s1 = __p1; \
  29745.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 33); \
  29746. })
  29747. #else
  29748. #define vst4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  29749.   int16x8x4_t __s1 = __p1; \
  29750.   int16x8x4_t __rev1; \
  29751.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29752.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29753.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29754.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29755.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 33); \
  29756. })
  29757. #endif
  29758.  
  29759. #ifdef __LITTLE_ENDIAN__
  29760. #define vst4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  29761.   uint8x8x4_t __s1 = __p1; \
  29762.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 16); \
  29763. })
  29764. #else
  29765. #define vst4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  29766.   uint8x8x4_t __s1 = __p1; \
  29767.   uint8x8x4_t __rev1; \
  29768.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29769.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29770.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29771.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29772.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 16); \
  29773. })
  29774. #endif
  29775.  
  29776. #ifdef __LITTLE_ENDIAN__
  29777. #define vst4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  29778.   uint32x2x4_t __s1 = __p1; \
  29779.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 18); \
  29780. })
  29781. #else
  29782. #define vst4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
  29783.   uint32x2x4_t __s1 = __p1; \
  29784.   uint32x2x4_t __rev1; \
  29785.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29786.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29787.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29788.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  29789.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 18); \
  29790. })
  29791. #endif
  29792.  
  29793. #ifdef __LITTLE_ENDIAN__
  29794. #define vst4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  29795.   uint16x4x4_t __s1 = __p1; \
  29796.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 17); \
  29797. })
  29798. #else
  29799. #define vst4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
  29800.   uint16x4x4_t __s1 = __p1; \
  29801.   uint16x4x4_t __rev1; \
  29802.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29803.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29804.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29805.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29806.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 17); \
  29807. })
  29808. #endif
  29809.  
  29810. #ifdef __LITTLE_ENDIAN__
  29811. #define vst4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  29812.   int8x8x4_t __s1 = __p1; \
  29813.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 0); \
  29814. })
  29815. #else
  29816. #define vst4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  29817.   int8x8x4_t __s1 = __p1; \
  29818.   int8x8x4_t __rev1; \
  29819.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  29820.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  29821.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  29822.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  29823.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 0); \
  29824. })
  29825. #endif
  29826.  
  29827. #ifdef __LITTLE_ENDIAN__
  29828. #define vst4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  29829.   float32x2x4_t __s1 = __p1; \
  29830.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 9); \
  29831. })
  29832. #else
  29833. #define vst4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
  29834.   float32x2x4_t __s1 = __p1; \
  29835.   float32x2x4_t __rev1; \
  29836.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29837.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29838.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29839.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  29840.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 9); \
  29841. })
  29842. #endif
  29843.  
  29844. #ifdef __LITTLE_ENDIAN__
  29845. #define vst4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  29846.   int32x2x4_t __s1 = __p1; \
  29847.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 2); \
  29848. })
  29849. #else
  29850. #define vst4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  29851.   int32x2x4_t __s1 = __p1; \
  29852.   int32x2x4_t __rev1; \
  29853.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  29854.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  29855.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  29856.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  29857.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 2); \
  29858. })
  29859. #endif
  29860.  
  29861. #ifdef __LITTLE_ENDIAN__
  29862. #define vst4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  29863.   int16x4x4_t __s1 = __p1; \
  29864.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 1); \
  29865. })
  29866. #else
  29867. #define vst4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  29868.   int16x4x4_t __s1 = __p1; \
  29869.   int16x4x4_t __rev1; \
  29870.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  29871.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  29872.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  29873.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  29874.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 1); \
  29875. })
  29876. #endif
  29877.  
  29878. #ifdef __LITTLE_ENDIAN__
  29879. __ai uint8x16_t vsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  29880.   uint8x16_t __ret;
  29881.   __ret = __p0 - __p1;
  29882.   return __ret;
  29883. }
  29884. #else
  29885. __ai uint8x16_t vsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  29886.   uint8x16_t __ret;
  29887.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  29888.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  29889.   __ret = __rev0 - __rev1;
  29890.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  29891.   return __ret;
  29892. }
  29893. #endif
  29894.  
  29895. #ifdef __LITTLE_ENDIAN__
  29896. __ai uint32x4_t vsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  29897.   uint32x4_t __ret;
  29898.   __ret = __p0 - __p1;
  29899.   return __ret;
  29900. }
  29901. #else
  29902. __ai uint32x4_t vsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  29903.   uint32x4_t __ret;
  29904.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  29905.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  29906.   __ret = __rev0 - __rev1;
  29907.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  29908.   return __ret;
  29909. }
  29910. #endif
  29911.  
  29912. #ifdef __LITTLE_ENDIAN__
  29913. __ai uint64x2_t vsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  29914.   uint64x2_t __ret;
  29915.   __ret = __p0 - __p1;
  29916.   return __ret;
  29917. }
  29918. #else
  29919. __ai uint64x2_t vsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  29920.   uint64x2_t __ret;
  29921.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  29922.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  29923.   __ret = __rev0 - __rev1;
  29924.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  29925.   return __ret;
  29926. }
  29927. #endif
  29928.  
  29929. #ifdef __LITTLE_ENDIAN__
  29930. __ai uint16x8_t vsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  29931.   uint16x8_t __ret;
  29932.   __ret = __p0 - __p1;
  29933.   return __ret;
  29934. }
  29935. #else
  29936. __ai uint16x8_t vsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  29937.   uint16x8_t __ret;
  29938.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  29939.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  29940.   __ret = __rev0 - __rev1;
  29941.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  29942.   return __ret;
  29943. }
  29944. #endif
  29945.  
  29946. #ifdef __LITTLE_ENDIAN__
  29947. __ai int8x16_t vsubq_s8(int8x16_t __p0, int8x16_t __p1) {
  29948.   int8x16_t __ret;
  29949.   __ret = __p0 - __p1;
  29950.   return __ret;
  29951. }
  29952. #else
  29953. __ai int8x16_t vsubq_s8(int8x16_t __p0, int8x16_t __p1) {
  29954.   int8x16_t __ret;
  29955.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  29956.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  29957.   __ret = __rev0 - __rev1;
  29958.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  29959.   return __ret;
  29960. }
  29961. #endif
  29962.  
  29963. #ifdef __LITTLE_ENDIAN__
  29964. __ai float32x4_t vsubq_f32(float32x4_t __p0, float32x4_t __p1) {
  29965.   float32x4_t __ret;
  29966.   __ret = __p0 - __p1;
  29967.   return __ret;
  29968. }
  29969. #else
  29970. __ai float32x4_t vsubq_f32(float32x4_t __p0, float32x4_t __p1) {
  29971.   float32x4_t __ret;
  29972.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  29973.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  29974.   __ret = __rev0 - __rev1;
  29975.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  29976.   return __ret;
  29977. }
  29978. #endif
  29979.  
  29980. #ifdef __LITTLE_ENDIAN__
  29981. __ai int32x4_t vsubq_s32(int32x4_t __p0, int32x4_t __p1) {
  29982.   int32x4_t __ret;
  29983.   __ret = __p0 - __p1;
  29984.   return __ret;
  29985. }
  29986. #else
  29987. __ai int32x4_t vsubq_s32(int32x4_t __p0, int32x4_t __p1) {
  29988.   int32x4_t __ret;
  29989.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  29990.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  29991.   __ret = __rev0 - __rev1;
  29992.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  29993.   return __ret;
  29994. }
  29995. #endif
  29996.  
  29997. #ifdef __LITTLE_ENDIAN__
  29998. __ai int64x2_t vsubq_s64(int64x2_t __p0, int64x2_t __p1) {
  29999.   int64x2_t __ret;
  30000.   __ret = __p0 - __p1;
  30001.   return __ret;
  30002. }
  30003. #else
  30004. __ai int64x2_t vsubq_s64(int64x2_t __p0, int64x2_t __p1) {
  30005.   int64x2_t __ret;
  30006.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30007.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30008.   __ret = __rev0 - __rev1;
  30009.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30010.   return __ret;
  30011. }
  30012. #endif
  30013.  
  30014. #ifdef __LITTLE_ENDIAN__
  30015. __ai int16x8_t vsubq_s16(int16x8_t __p0, int16x8_t __p1) {
  30016.   int16x8_t __ret;
  30017.   __ret = __p0 - __p1;
  30018.   return __ret;
  30019. }
  30020. #else
  30021. __ai int16x8_t vsubq_s16(int16x8_t __p0, int16x8_t __p1) {
  30022.   int16x8_t __ret;
  30023.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30024.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30025.   __ret = __rev0 - __rev1;
  30026.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30027.   return __ret;
  30028. }
  30029. #endif
  30030.  
  30031. #ifdef __LITTLE_ENDIAN__
  30032. __ai uint8x8_t vsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
  30033.   uint8x8_t __ret;
  30034.   __ret = __p0 - __p1;
  30035.   return __ret;
  30036. }
  30037. #else
  30038. __ai uint8x8_t vsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
  30039.   uint8x8_t __ret;
  30040.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30041.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30042.   __ret = __rev0 - __rev1;
  30043.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30044.   return __ret;
  30045. }
  30046. #endif
  30047.  
  30048. #ifdef __LITTLE_ENDIAN__
  30049. __ai uint32x2_t vsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
  30050.   uint32x2_t __ret;
  30051.   __ret = __p0 - __p1;
  30052.   return __ret;
  30053. }
  30054. #else
  30055. __ai uint32x2_t vsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
  30056.   uint32x2_t __ret;
  30057.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30058.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30059.   __ret = __rev0 - __rev1;
  30060.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30061.   return __ret;
  30062. }
  30063. #endif
  30064.  
  30065. __ai uint64x1_t vsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
  30066.   uint64x1_t __ret;
  30067.   __ret = __p0 - __p1;
  30068.   return __ret;
  30069. }
  30070. #ifdef __LITTLE_ENDIAN__
  30071. __ai uint16x4_t vsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
  30072.   uint16x4_t __ret;
  30073.   __ret = __p0 - __p1;
  30074.   return __ret;
  30075. }
  30076. #else
  30077. __ai uint16x4_t vsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
  30078.   uint16x4_t __ret;
  30079.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  30080.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  30081.   __ret = __rev0 - __rev1;
  30082.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  30083.   return __ret;
  30084. }
  30085. #endif
  30086.  
  30087. #ifdef __LITTLE_ENDIAN__
  30088. __ai int8x8_t vsub_s8(int8x8_t __p0, int8x8_t __p1) {
  30089.   int8x8_t __ret;
  30090.   __ret = __p0 - __p1;
  30091.   return __ret;
  30092. }
  30093. #else
  30094. __ai int8x8_t vsub_s8(int8x8_t __p0, int8x8_t __p1) {
  30095.   int8x8_t __ret;
  30096.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30097.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30098.   __ret = __rev0 - __rev1;
  30099.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30100.   return __ret;
  30101. }
  30102. #endif
  30103.  
  30104. #ifdef __LITTLE_ENDIAN__
  30105. __ai float32x2_t vsub_f32(float32x2_t __p0, float32x2_t __p1) {
  30106.   float32x2_t __ret;
  30107.   __ret = __p0 - __p1;
  30108.   return __ret;
  30109. }
  30110. #else
  30111. __ai float32x2_t vsub_f32(float32x2_t __p0, float32x2_t __p1) {
  30112.   float32x2_t __ret;
  30113.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30114.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30115.   __ret = __rev0 - __rev1;
  30116.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30117.   return __ret;
  30118. }
  30119. #endif
  30120.  
  30121. #ifdef __LITTLE_ENDIAN__
  30122. __ai int32x2_t vsub_s32(int32x2_t __p0, int32x2_t __p1) {
  30123.   int32x2_t __ret;
  30124.   __ret = __p0 - __p1;
  30125.   return __ret;
  30126. }
  30127. #else
  30128. __ai int32x2_t vsub_s32(int32x2_t __p0, int32x2_t __p1) {
  30129.   int32x2_t __ret;
  30130.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30131.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30132.   __ret = __rev0 - __rev1;
  30133.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30134.   return __ret;
  30135. }
  30136. #endif
  30137.  
  30138. __ai int64x1_t vsub_s64(int64x1_t __p0, int64x1_t __p1) {
  30139.   int64x1_t __ret;
  30140.   __ret = __p0 - __p1;
  30141.   return __ret;
  30142. }
  30143. #ifdef __LITTLE_ENDIAN__
  30144. __ai int16x4_t vsub_s16(int16x4_t __p0, int16x4_t __p1) {
  30145.   int16x4_t __ret;
  30146.   __ret = __p0 - __p1;
  30147.   return __ret;
  30148. }
  30149. #else
  30150. __ai int16x4_t vsub_s16(int16x4_t __p0, int16x4_t __p1) {
  30151.   int16x4_t __ret;
  30152.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  30153.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  30154.   __ret = __rev0 - __rev1;
  30155.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  30156.   return __ret;
  30157. }
  30158. #endif
  30159.  
  30160. #ifdef __LITTLE_ENDIAN__
  30161. __ai uint16x4_t vsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  30162.   uint16x4_t __ret;
  30163.   __ret = (uint16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  30164.   return __ret;
  30165. }
  30166. #else
  30167. __ai uint16x4_t vsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  30168.   uint16x4_t __ret;
  30169.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  30170.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  30171.   __ret = (uint16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
  30172.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  30173.   return __ret;
  30174. }
  30175. __ai uint16x4_t __noswap_vsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
  30176.   uint16x4_t __ret;
  30177.   __ret = (uint16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
  30178.   return __ret;
  30179. }
  30180. #endif
  30181.  
  30182. #ifdef __LITTLE_ENDIAN__
  30183. __ai uint32x2_t vsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  30184.   uint32x2_t __ret;
  30185.   __ret = (uint32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  30186.   return __ret;
  30187. }
  30188. #else
  30189. __ai uint32x2_t vsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  30190.   uint32x2_t __ret;
  30191.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30192.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30193.   __ret = (uint32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
  30194.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30195.   return __ret;
  30196. }
  30197. __ai uint32x2_t __noswap_vsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
  30198.   uint32x2_t __ret;
  30199.   __ret = (uint32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
  30200.   return __ret;
  30201. }
  30202. #endif
  30203.  
  30204. #ifdef __LITTLE_ENDIAN__
  30205. __ai uint8x8_t vsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  30206.   uint8x8_t __ret;
  30207.   __ret = (uint8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  30208.   return __ret;
  30209. }
  30210. #else
  30211. __ai uint8x8_t vsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  30212.   uint8x8_t __ret;
  30213.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30214.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30215.   __ret = (uint8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
  30216.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30217.   return __ret;
  30218. }
  30219. __ai uint8x8_t __noswap_vsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
  30220.   uint8x8_t __ret;
  30221.   __ret = (uint8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
  30222.   return __ret;
  30223. }
  30224. #endif
  30225.  
  30226. #ifdef __LITTLE_ENDIAN__
  30227. __ai int16x4_t vsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
  30228.   int16x4_t __ret;
  30229.   __ret = (int16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  30230.   return __ret;
  30231. }
  30232. #else
  30233. __ai int16x4_t vsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
  30234.   int16x4_t __ret;
  30235.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  30236.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  30237.   __ret = (int16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
  30238.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  30239.   return __ret;
  30240. }
  30241. __ai int16x4_t __noswap_vsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
  30242.   int16x4_t __ret;
  30243.   __ret = (int16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
  30244.   return __ret;
  30245. }
  30246. #endif
  30247.  
  30248. #ifdef __LITTLE_ENDIAN__
  30249. __ai int32x2_t vsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
  30250.   int32x2_t __ret;
  30251.   __ret = (int32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  30252.   return __ret;
  30253. }
  30254. #else
  30255. __ai int32x2_t vsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
  30256.   int32x2_t __ret;
  30257.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30258.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30259.   __ret = (int32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
  30260.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30261.   return __ret;
  30262. }
  30263. __ai int32x2_t __noswap_vsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
  30264.   int32x2_t __ret;
  30265.   __ret = (int32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
  30266.   return __ret;
  30267. }
  30268. #endif
  30269.  
  30270. #ifdef __LITTLE_ENDIAN__
  30271. __ai int8x8_t vsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
  30272.   int8x8_t __ret;
  30273.   __ret = (int8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  30274.   return __ret;
  30275. }
  30276. #else
  30277. __ai int8x8_t vsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
  30278.   int8x8_t __ret;
  30279.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30280.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30281.   __ret = (int8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
  30282.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30283.   return __ret;
  30284. }
  30285. __ai int8x8_t __noswap_vsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
  30286.   int8x8_t __ret;
  30287.   __ret = (int8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
  30288.   return __ret;
  30289. }
  30290. #endif
  30291.  
  30292. #ifdef __LITTLE_ENDIAN__
  30293. __ai uint16x8_t vsubl_u8(uint8x8_t __p0, uint8x8_t __p1) {
  30294.   uint16x8_t __ret;
  30295.   __ret = vmovl_u8(__p0) - vmovl_u8(__p1);
  30296.   return __ret;
  30297. }
  30298. #else
  30299. __ai uint16x8_t vsubl_u8(uint8x8_t __p0, uint8x8_t __p1) {
  30300.   uint16x8_t __ret;
  30301.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30302.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30303.   __ret = __noswap_vmovl_u8(__rev0) - __noswap_vmovl_u8(__rev1);
  30304.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30305.   return __ret;
  30306. }
  30307. #endif
  30308.  
  30309. #ifdef __LITTLE_ENDIAN__
  30310. __ai uint64x2_t vsubl_u32(uint32x2_t __p0, uint32x2_t __p1) {
  30311.   uint64x2_t __ret;
  30312.   __ret = vmovl_u32(__p0) - vmovl_u32(__p1);
  30313.   return __ret;
  30314. }
  30315. #else
  30316. __ai uint64x2_t vsubl_u32(uint32x2_t __p0, uint32x2_t __p1) {
  30317.   uint64x2_t __ret;
  30318.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30319.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30320.   __ret = __noswap_vmovl_u32(__rev0) - __noswap_vmovl_u32(__rev1);
  30321.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30322.   return __ret;
  30323. }
  30324. #endif
  30325.  
  30326. #ifdef __LITTLE_ENDIAN__
  30327. __ai uint32x4_t vsubl_u16(uint16x4_t __p0, uint16x4_t __p1) {
  30328.   uint32x4_t __ret;
  30329.   __ret = vmovl_u16(__p0) - vmovl_u16(__p1);
  30330.   return __ret;
  30331. }
  30332. #else
  30333. __ai uint32x4_t vsubl_u16(uint16x4_t __p0, uint16x4_t __p1) {
  30334.   uint32x4_t __ret;
  30335.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  30336.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  30337.   __ret = __noswap_vmovl_u16(__rev0) - __noswap_vmovl_u16(__rev1);
  30338.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  30339.   return __ret;
  30340. }
  30341. #endif
  30342.  
  30343. #ifdef __LITTLE_ENDIAN__
  30344. __ai int16x8_t vsubl_s8(int8x8_t __p0, int8x8_t __p1) {
  30345.   int16x8_t __ret;
  30346.   __ret = vmovl_s8(__p0) - vmovl_s8(__p1);
  30347.   return __ret;
  30348. }
  30349. #else
  30350. __ai int16x8_t vsubl_s8(int8x8_t __p0, int8x8_t __p1) {
  30351.   int16x8_t __ret;
  30352.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30353.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30354.   __ret = __noswap_vmovl_s8(__rev0) - __noswap_vmovl_s8(__rev1);
  30355.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30356.   return __ret;
  30357. }
  30358. #endif
  30359.  
  30360. #ifdef __LITTLE_ENDIAN__
  30361. __ai int64x2_t vsubl_s32(int32x2_t __p0, int32x2_t __p1) {
  30362.   int64x2_t __ret;
  30363.   __ret = vmovl_s32(__p0) - vmovl_s32(__p1);
  30364.   return __ret;
  30365. }
  30366. #else
  30367. __ai int64x2_t vsubl_s32(int32x2_t __p0, int32x2_t __p1) {
  30368.   int64x2_t __ret;
  30369.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30370.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30371.   __ret = __noswap_vmovl_s32(__rev0) - __noswap_vmovl_s32(__rev1);
  30372.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30373.   return __ret;
  30374. }
  30375. #endif
  30376.  
  30377. #ifdef __LITTLE_ENDIAN__
  30378. __ai int32x4_t vsubl_s16(int16x4_t __p0, int16x4_t __p1) {
  30379.   int32x4_t __ret;
  30380.   __ret = vmovl_s16(__p0) - vmovl_s16(__p1);
  30381.   return __ret;
  30382. }
  30383. #else
  30384. __ai int32x4_t vsubl_s16(int16x4_t __p0, int16x4_t __p1) {
  30385.   int32x4_t __ret;
  30386.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  30387.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  30388.   __ret = __noswap_vmovl_s16(__rev0) - __noswap_vmovl_s16(__rev1);
  30389.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  30390.   return __ret;
  30391. }
  30392. #endif
  30393.  
  30394. #ifdef __LITTLE_ENDIAN__
  30395. __ai uint16x8_t vsubw_u8(uint16x8_t __p0, uint8x8_t __p1) {
  30396.   uint16x8_t __ret;
  30397.   __ret = __p0 - vmovl_u8(__p1);
  30398.   return __ret;
  30399. }
  30400. #else
  30401. __ai uint16x8_t vsubw_u8(uint16x8_t __p0, uint8x8_t __p1) {
  30402.   uint16x8_t __ret;
  30403.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30404.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30405.   __ret = __rev0 - __noswap_vmovl_u8(__rev1);
  30406.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30407.   return __ret;
  30408. }
  30409. #endif
  30410.  
  30411. #ifdef __LITTLE_ENDIAN__
  30412. __ai uint64x2_t vsubw_u32(uint64x2_t __p0, uint32x2_t __p1) {
  30413.   uint64x2_t __ret;
  30414.   __ret = __p0 - vmovl_u32(__p1);
  30415.   return __ret;
  30416. }
  30417. #else
  30418. __ai uint64x2_t vsubw_u32(uint64x2_t __p0, uint32x2_t __p1) {
  30419.   uint64x2_t __ret;
  30420.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30421.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30422.   __ret = __rev0 - __noswap_vmovl_u32(__rev1);
  30423.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30424.   return __ret;
  30425. }
  30426. #endif
  30427.  
  30428. #ifdef __LITTLE_ENDIAN__
  30429. __ai uint32x4_t vsubw_u16(uint32x4_t __p0, uint16x4_t __p1) {
  30430.   uint32x4_t __ret;
  30431.   __ret = __p0 - vmovl_u16(__p1);
  30432.   return __ret;
  30433. }
  30434. #else
  30435. __ai uint32x4_t vsubw_u16(uint32x4_t __p0, uint16x4_t __p1) {
  30436.   uint32x4_t __ret;
  30437.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  30438.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  30439.   __ret = __rev0 - __noswap_vmovl_u16(__rev1);
  30440.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  30441.   return __ret;
  30442. }
  30443. #endif
  30444.  
  30445. #ifdef __LITTLE_ENDIAN__
  30446. __ai int16x8_t vsubw_s8(int16x8_t __p0, int8x8_t __p1) {
  30447.   int16x8_t __ret;
  30448.   __ret = __p0 - vmovl_s8(__p1);
  30449.   return __ret;
  30450. }
  30451. #else
  30452. __ai int16x8_t vsubw_s8(int16x8_t __p0, int8x8_t __p1) {
  30453.   int16x8_t __ret;
  30454.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30455.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30456.   __ret = __rev0 - __noswap_vmovl_s8(__rev1);
  30457.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30458.   return __ret;
  30459. }
  30460. #endif
  30461.  
  30462. #ifdef __LITTLE_ENDIAN__
  30463. __ai int64x2_t vsubw_s32(int64x2_t __p0, int32x2_t __p1) {
  30464.   int64x2_t __ret;
  30465.   __ret = __p0 - vmovl_s32(__p1);
  30466.   return __ret;
  30467. }
  30468. #else
  30469. __ai int64x2_t vsubw_s32(int64x2_t __p0, int32x2_t __p1) {
  30470.   int64x2_t __ret;
  30471.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  30472.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  30473.   __ret = __rev0 - __noswap_vmovl_s32(__rev1);
  30474.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  30475.   return __ret;
  30476. }
  30477. #endif
  30478.  
  30479. #ifdef __LITTLE_ENDIAN__
  30480. __ai int32x4_t vsubw_s16(int32x4_t __p0, int16x4_t __p1) {
  30481.   int32x4_t __ret;
  30482.   __ret = __p0 - vmovl_s16(__p1);
  30483.   return __ret;
  30484. }
  30485. #else
  30486. __ai int32x4_t vsubw_s16(int32x4_t __p0, int16x4_t __p1) {
  30487.   int32x4_t __ret;
  30488.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  30489.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  30490.   __ret = __rev0 - __noswap_vmovl_s16(__rev1);
  30491.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  30492.   return __ret;
  30493. }
  30494. #endif
  30495.  
  30496. #ifdef __LITTLE_ENDIAN__
  30497. __ai poly8x8_t vtbl1_p8(poly8x8_t __p0, uint8x8_t __p1) {
  30498.   poly8x8_t __ret;
  30499.   __ret = (poly8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__p0, (int8x8_t)__p1, 4);
  30500.   return __ret;
  30501. }
  30502. #else
  30503. __ai poly8x8_t vtbl1_p8(poly8x8_t __p0, uint8x8_t __p1) {
  30504.   poly8x8_t __ret;
  30505.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30506.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30507.   __ret = (poly8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__rev0, (int8x8_t)__rev1, 4);
  30508.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30509.   return __ret;
  30510. }
  30511. #endif
  30512.  
  30513. #ifdef __LITTLE_ENDIAN__
  30514. __ai uint8x8_t vtbl1_u8(uint8x8_t __p0, uint8x8_t __p1) {
  30515.   uint8x8_t __ret;
  30516.   __ret = (uint8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  30517.   return __ret;
  30518. }
  30519. #else
  30520. __ai uint8x8_t vtbl1_u8(uint8x8_t __p0, uint8x8_t __p1) {
  30521.   uint8x8_t __ret;
  30522.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30523.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30524.   __ret = (uint8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  30525.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30526.   return __ret;
  30527. }
  30528. #endif
  30529.  
  30530. #ifdef __LITTLE_ENDIAN__
  30531. __ai int8x8_t vtbl1_s8(int8x8_t __p0, int8x8_t __p1) {
  30532.   int8x8_t __ret;
  30533.   __ret = (int8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  30534.   return __ret;
  30535. }
  30536. #else
  30537. __ai int8x8_t vtbl1_s8(int8x8_t __p0, int8x8_t __p1) {
  30538.   int8x8_t __ret;
  30539.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30540.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30541.   __ret = (int8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  30542.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30543.   return __ret;
  30544. }
  30545. #endif
  30546.  
  30547. #ifdef __LITTLE_ENDIAN__
  30548. __ai poly8x8_t vtbl2_p8(poly8x8x2_t __p0, uint8x8_t __p1) {
  30549.   poly8x8_t __ret;
  30550.   __ret = (poly8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p1, 4);
  30551.   return __ret;
  30552. }
  30553. #else
  30554. __ai poly8x8_t vtbl2_p8(poly8x8x2_t __p0, uint8x8_t __p1) {
  30555.   poly8x8_t __ret;
  30556.   poly8x8x2_t __rev0;
  30557.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30558.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30559.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30560.   __ret = (poly8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev1, 4);
  30561.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30562.   return __ret;
  30563. }
  30564. #endif
  30565.  
  30566. #ifdef __LITTLE_ENDIAN__
  30567. __ai uint8x8_t vtbl2_u8(uint8x8x2_t __p0, uint8x8_t __p1) {
  30568.   uint8x8_t __ret;
  30569.   __ret = (uint8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p1, 16);
  30570.   return __ret;
  30571. }
  30572. #else
  30573. __ai uint8x8_t vtbl2_u8(uint8x8x2_t __p0, uint8x8_t __p1) {
  30574.   uint8x8_t __ret;
  30575.   uint8x8x2_t __rev0;
  30576.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30577.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30578.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30579.   __ret = (uint8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev1, 16);
  30580.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30581.   return __ret;
  30582. }
  30583. #endif
  30584.  
  30585. #ifdef __LITTLE_ENDIAN__
  30586. __ai int8x8_t vtbl2_s8(int8x8x2_t __p0, int8x8_t __p1) {
  30587.   int8x8_t __ret;
  30588.   __ret = (int8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p1, 0);
  30589.   return __ret;
  30590. }
  30591. #else
  30592. __ai int8x8_t vtbl2_s8(int8x8x2_t __p0, int8x8_t __p1) {
  30593.   int8x8_t __ret;
  30594.   int8x8x2_t __rev0;
  30595.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30596.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30597.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30598.   __ret = (int8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev1, 0);
  30599.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30600.   return __ret;
  30601. }
  30602. #endif
  30603.  
  30604. #ifdef __LITTLE_ENDIAN__
  30605. __ai poly8x8_t vtbl3_p8(poly8x8x3_t __p0, uint8x8_t __p1) {
  30606.   poly8x8_t __ret;
  30607.   __ret = (poly8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p1, 4);
  30608.   return __ret;
  30609. }
  30610. #else
  30611. __ai poly8x8_t vtbl3_p8(poly8x8x3_t __p0, uint8x8_t __p1) {
  30612.   poly8x8_t __ret;
  30613.   poly8x8x3_t __rev0;
  30614.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30615.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30616.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30617.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30618.   __ret = (poly8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev1, 4);
  30619.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30620.   return __ret;
  30621. }
  30622. #endif
  30623.  
  30624. #ifdef __LITTLE_ENDIAN__
  30625. __ai uint8x8_t vtbl3_u8(uint8x8x3_t __p0, uint8x8_t __p1) {
  30626.   uint8x8_t __ret;
  30627.   __ret = (uint8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p1, 16);
  30628.   return __ret;
  30629. }
  30630. #else
  30631. __ai uint8x8_t vtbl3_u8(uint8x8x3_t __p0, uint8x8_t __p1) {
  30632.   uint8x8_t __ret;
  30633.   uint8x8x3_t __rev0;
  30634.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30635.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30636.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30637.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30638.   __ret = (uint8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev1, 16);
  30639.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30640.   return __ret;
  30641. }
  30642. #endif
  30643.  
  30644. #ifdef __LITTLE_ENDIAN__
  30645. __ai int8x8_t vtbl3_s8(int8x8x3_t __p0, int8x8_t __p1) {
  30646.   int8x8_t __ret;
  30647.   __ret = (int8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p1, 0);
  30648.   return __ret;
  30649. }
  30650. #else
  30651. __ai int8x8_t vtbl3_s8(int8x8x3_t __p0, int8x8_t __p1) {
  30652.   int8x8_t __ret;
  30653.   int8x8x3_t __rev0;
  30654.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30655.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30656.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30657.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30658.   __ret = (int8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev1, 0);
  30659.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30660.   return __ret;
  30661. }
  30662. #endif
  30663.  
  30664. #ifdef __LITTLE_ENDIAN__
  30665. __ai poly8x8_t vtbl4_p8(poly8x8x4_t __p0, uint8x8_t __p1) {
  30666.   poly8x8_t __ret;
  30667.   __ret = (poly8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p0.val[3], (int8x8_t)__p1, 4);
  30668.   return __ret;
  30669. }
  30670. #else
  30671. __ai poly8x8_t vtbl4_p8(poly8x8x4_t __p0, uint8x8_t __p1) {
  30672.   poly8x8_t __ret;
  30673.   poly8x8x4_t __rev0;
  30674.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30675.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30676.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30677.   __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
  30678.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30679.   __ret = (poly8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev0.val[3], (int8x8_t)__rev1, 4);
  30680.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30681.   return __ret;
  30682. }
  30683. #endif
  30684.  
  30685. #ifdef __LITTLE_ENDIAN__
  30686. __ai uint8x8_t vtbl4_u8(uint8x8x4_t __p0, uint8x8_t __p1) {
  30687.   uint8x8_t __ret;
  30688.   __ret = (uint8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p0.val[3], (int8x8_t)__p1, 16);
  30689.   return __ret;
  30690. }
  30691. #else
  30692. __ai uint8x8_t vtbl4_u8(uint8x8x4_t __p0, uint8x8_t __p1) {
  30693.   uint8x8_t __ret;
  30694.   uint8x8x4_t __rev0;
  30695.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30696.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30697.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30698.   __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
  30699.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30700.   __ret = (uint8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev0.val[3], (int8x8_t)__rev1, 16);
  30701.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30702.   return __ret;
  30703. }
  30704. #endif
  30705.  
  30706. #ifdef __LITTLE_ENDIAN__
  30707. __ai int8x8_t vtbl4_s8(int8x8x4_t __p0, int8x8_t __p1) {
  30708.   int8x8_t __ret;
  30709.   __ret = (int8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p0.val[3], (int8x8_t)__p1, 0);
  30710.   return __ret;
  30711. }
  30712. #else
  30713. __ai int8x8_t vtbl4_s8(int8x8x4_t __p0, int8x8_t __p1) {
  30714.   int8x8_t __ret;
  30715.   int8x8x4_t __rev0;
  30716.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30717.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30718.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30719.   __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
  30720.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30721.   __ret = (int8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev0.val[3], (int8x8_t)__rev1, 0);
  30722.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30723.   return __ret;
  30724. }
  30725. #endif
  30726.  
  30727. #ifdef __LITTLE_ENDIAN__
  30728. __ai poly8x8_t vtbx1_p8(poly8x8_t __p0, poly8x8_t __p1, uint8x8_t __p2) {
  30729.   poly8x8_t __ret;
  30730.   __ret = (poly8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 4);
  30731.   return __ret;
  30732. }
  30733. #else
  30734. __ai poly8x8_t vtbx1_p8(poly8x8_t __p0, poly8x8_t __p1, uint8x8_t __p2) {
  30735.   poly8x8_t __ret;
  30736.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30737.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30738.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30739.   __ret = (poly8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 4);
  30740.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30741.   return __ret;
  30742. }
  30743. #endif
  30744.  
  30745. #ifdef __LITTLE_ENDIAN__
  30746. __ai uint8x8_t vtbx1_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  30747.   uint8x8_t __ret;
  30748.   __ret = (uint8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 16);
  30749.   return __ret;
  30750. }
  30751. #else
  30752. __ai uint8x8_t vtbx1_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  30753.   uint8x8_t __ret;
  30754.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30755.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30756.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30757.   __ret = (uint8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 16);
  30758.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30759.   return __ret;
  30760. }
  30761. #endif
  30762.  
  30763. #ifdef __LITTLE_ENDIAN__
  30764. __ai int8x8_t vtbx1_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  30765.   int8x8_t __ret;
  30766.   __ret = (int8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 0);
  30767.   return __ret;
  30768. }
  30769. #else
  30770. __ai int8x8_t vtbx1_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  30771.   int8x8_t __ret;
  30772.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30773.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30774.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30775.   __ret = (int8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 0);
  30776.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30777.   return __ret;
  30778. }
  30779. #endif
  30780.  
  30781. #ifdef __LITTLE_ENDIAN__
  30782. __ai poly8x8_t vtbx2_p8(poly8x8_t __p0, poly8x8x2_t __p1, uint8x8_t __p2) {
  30783.   poly8x8_t __ret;
  30784.   __ret = (poly8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p2, 4);
  30785.   return __ret;
  30786. }
  30787. #else
  30788. __ai poly8x8_t vtbx2_p8(poly8x8_t __p0, poly8x8x2_t __p1, uint8x8_t __p2) {
  30789.   poly8x8_t __ret;
  30790.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30791.   poly8x8x2_t __rev1;
  30792.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30793.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30794.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30795.   __ret = (poly8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev2, 4);
  30796.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30797.   return __ret;
  30798. }
  30799. #endif
  30800.  
  30801. #ifdef __LITTLE_ENDIAN__
  30802. __ai uint8x8_t vtbx2_u8(uint8x8_t __p0, uint8x8x2_t __p1, uint8x8_t __p2) {
  30803.   uint8x8_t __ret;
  30804.   __ret = (uint8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p2, 16);
  30805.   return __ret;
  30806. }
  30807. #else
  30808. __ai uint8x8_t vtbx2_u8(uint8x8_t __p0, uint8x8x2_t __p1, uint8x8_t __p2) {
  30809.   uint8x8_t __ret;
  30810.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30811.   uint8x8x2_t __rev1;
  30812.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30813.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30814.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30815.   __ret = (uint8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev2, 16);
  30816.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30817.   return __ret;
  30818. }
  30819. #endif
  30820.  
  30821. #ifdef __LITTLE_ENDIAN__
  30822. __ai int8x8_t vtbx2_s8(int8x8_t __p0, int8x8x2_t __p1, int8x8_t __p2) {
  30823.   int8x8_t __ret;
  30824.   __ret = (int8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p2, 0);
  30825.   return __ret;
  30826. }
  30827. #else
  30828. __ai int8x8_t vtbx2_s8(int8x8_t __p0, int8x8x2_t __p1, int8x8_t __p2) {
  30829.   int8x8_t __ret;
  30830.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30831.   int8x8x2_t __rev1;
  30832.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30833.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30834.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30835.   __ret = (int8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev2, 0);
  30836.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30837.   return __ret;
  30838. }
  30839. #endif
  30840.  
  30841. #ifdef __LITTLE_ENDIAN__
  30842. __ai poly8x8_t vtbx3_p8(poly8x8_t __p0, poly8x8x3_t __p1, uint8x8_t __p2) {
  30843.   poly8x8_t __ret;
  30844.   __ret = (poly8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p2, 4);
  30845.   return __ret;
  30846. }
  30847. #else
  30848. __ai poly8x8_t vtbx3_p8(poly8x8_t __p0, poly8x8x3_t __p1, uint8x8_t __p2) {
  30849.   poly8x8_t __ret;
  30850.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30851.   poly8x8x3_t __rev1;
  30852.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30853.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30854.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30855.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30856.   __ret = (poly8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev2, 4);
  30857.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30858.   return __ret;
  30859. }
  30860. #endif
  30861.  
  30862. #ifdef __LITTLE_ENDIAN__
  30863. __ai uint8x8_t vtbx3_u8(uint8x8_t __p0, uint8x8x3_t __p1, uint8x8_t __p2) {
  30864.   uint8x8_t __ret;
  30865.   __ret = (uint8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p2, 16);
  30866.   return __ret;
  30867. }
  30868. #else
  30869. __ai uint8x8_t vtbx3_u8(uint8x8_t __p0, uint8x8x3_t __p1, uint8x8_t __p2) {
  30870.   uint8x8_t __ret;
  30871.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30872.   uint8x8x3_t __rev1;
  30873.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30874.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30875.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30876.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30877.   __ret = (uint8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev2, 16);
  30878.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30879.   return __ret;
  30880. }
  30881. #endif
  30882.  
  30883. #ifdef __LITTLE_ENDIAN__
  30884. __ai int8x8_t vtbx3_s8(int8x8_t __p0, int8x8x3_t __p1, int8x8_t __p2) {
  30885.   int8x8_t __ret;
  30886.   __ret = (int8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p2, 0);
  30887.   return __ret;
  30888. }
  30889. #else
  30890. __ai int8x8_t vtbx3_s8(int8x8_t __p0, int8x8x3_t __p1, int8x8_t __p2) {
  30891.   int8x8_t __ret;
  30892.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30893.   int8x8x3_t __rev1;
  30894.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30895.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30896.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30897.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30898.   __ret = (int8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev2, 0);
  30899.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30900.   return __ret;
  30901. }
  30902. #endif
  30903.  
  30904. #ifdef __LITTLE_ENDIAN__
  30905. __ai poly8x8_t vtbx4_p8(poly8x8_t __p0, poly8x8x4_t __p1, uint8x8_t __p2) {
  30906.   poly8x8_t __ret;
  30907.   __ret = (poly8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p1.val[3], (int8x8_t)__p2, 4);
  30908.   return __ret;
  30909. }
  30910. #else
  30911. __ai poly8x8_t vtbx4_p8(poly8x8_t __p0, poly8x8x4_t __p1, uint8x8_t __p2) {
  30912.   poly8x8_t __ret;
  30913.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30914.   poly8x8x4_t __rev1;
  30915.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30916.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30917.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30918.   __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
  30919.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30920.   __ret = (poly8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], (int8x8_t)__rev2, 4);
  30921.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30922.   return __ret;
  30923. }
  30924. #endif
  30925.  
  30926. #ifdef __LITTLE_ENDIAN__
  30927. __ai uint8x8_t vtbx4_u8(uint8x8_t __p0, uint8x8x4_t __p1, uint8x8_t __p2) {
  30928.   uint8x8_t __ret;
  30929.   __ret = (uint8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p1.val[3], (int8x8_t)__p2, 16);
  30930.   return __ret;
  30931. }
  30932. #else
  30933. __ai uint8x8_t vtbx4_u8(uint8x8_t __p0, uint8x8x4_t __p1, uint8x8_t __p2) {
  30934.   uint8x8_t __ret;
  30935.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30936.   uint8x8x4_t __rev1;
  30937.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30938.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30939.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30940.   __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
  30941.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30942.   __ret = (uint8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], (int8x8_t)__rev2, 16);
  30943.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30944.   return __ret;
  30945. }
  30946. #endif
  30947.  
  30948. #ifdef __LITTLE_ENDIAN__
  30949. __ai int8x8_t vtbx4_s8(int8x8_t __p0, int8x8x4_t __p1, int8x8_t __p2) {
  30950.   int8x8_t __ret;
  30951.   __ret = (int8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p1.val[3], (int8x8_t)__p2, 0);
  30952.   return __ret;
  30953. }
  30954. #else
  30955. __ai int8x8_t vtbx4_s8(int8x8_t __p0, int8x8x4_t __p1, int8x8_t __p2) {
  30956.   int8x8_t __ret;
  30957.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30958.   int8x8x4_t __rev1;
  30959.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30960.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30961.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
  30962.   __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
  30963.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  30964.   __ret = (int8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], (int8x8_t)__rev2, 0);
  30965.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  30966.   return __ret;
  30967. }
  30968. #endif
  30969.  
  30970. #ifdef __LITTLE_ENDIAN__
  30971. __ai poly8x8x2_t vtrn_p8(poly8x8_t __p0, poly8x8_t __p1) {
  30972.   poly8x8x2_t __ret;
  30973.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 4);
  30974.   return __ret;
  30975. }
  30976. #else
  30977. __ai poly8x8x2_t vtrn_p8(poly8x8_t __p0, poly8x8_t __p1) {
  30978.   poly8x8x2_t __ret;
  30979.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  30980.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  30981.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 4);
  30982.  
  30983.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  30984.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  30985.   return __ret;
  30986. }
  30987. #endif
  30988.  
  30989. #ifdef __LITTLE_ENDIAN__
  30990. __ai poly16x4x2_t vtrn_p16(poly16x4_t __p0, poly16x4_t __p1) {
  30991.   poly16x4x2_t __ret;
  30992.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 5);
  30993.   return __ret;
  30994. }
  30995. #else
  30996. __ai poly16x4x2_t vtrn_p16(poly16x4_t __p0, poly16x4_t __p1) {
  30997.   poly16x4x2_t __ret;
  30998.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  30999.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31000.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 5);
  31001.  
  31002.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31003.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31004.   return __ret;
  31005. }
  31006. #endif
  31007.  
  31008. #ifdef __LITTLE_ENDIAN__
  31009. __ai poly8x16x2_t vtrnq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  31010.   poly8x16x2_t __ret;
  31011.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 36);
  31012.   return __ret;
  31013. }
  31014. #else
  31015. __ai poly8x16x2_t vtrnq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  31016.   poly8x16x2_t __ret;
  31017.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31018.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31019.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 36);
  31020.  
  31021.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31022.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31023.   return __ret;
  31024. }
  31025. #endif
  31026.  
  31027. #ifdef __LITTLE_ENDIAN__
  31028. __ai poly16x8x2_t vtrnq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  31029.   poly16x8x2_t __ret;
  31030.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 37);
  31031.   return __ret;
  31032. }
  31033. #else
  31034. __ai poly16x8x2_t vtrnq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  31035.   poly16x8x2_t __ret;
  31036.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31037.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31038.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 37);
  31039.  
  31040.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31041.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31042.   return __ret;
  31043. }
  31044. #endif
  31045.  
  31046. #ifdef __LITTLE_ENDIAN__
  31047. __ai uint8x16x2_t vtrnq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  31048.   uint8x16x2_t __ret;
  31049.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 48);
  31050.   return __ret;
  31051. }
  31052. #else
  31053. __ai uint8x16x2_t vtrnq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  31054.   uint8x16x2_t __ret;
  31055.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31056.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31057.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  31058.  
  31059.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31060.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31061.   return __ret;
  31062. }
  31063. #endif
  31064.  
  31065. #ifdef __LITTLE_ENDIAN__
  31066. __ai uint32x4x2_t vtrnq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  31067.   uint32x4x2_t __ret;
  31068.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 50);
  31069.   return __ret;
  31070. }
  31071. #else
  31072. __ai uint32x4x2_t vtrnq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  31073.   uint32x4x2_t __ret;
  31074.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31075.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31076.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  31077.  
  31078.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31079.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31080.   return __ret;
  31081. }
  31082. #endif
  31083.  
  31084. #ifdef __LITTLE_ENDIAN__
  31085. __ai uint16x8x2_t vtrnq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  31086.   uint16x8x2_t __ret;
  31087.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 49);
  31088.   return __ret;
  31089. }
  31090. #else
  31091. __ai uint16x8x2_t vtrnq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  31092.   uint16x8x2_t __ret;
  31093.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31094.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31095.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  31096.  
  31097.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31098.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31099.   return __ret;
  31100. }
  31101. #endif
  31102.  
  31103. #ifdef __LITTLE_ENDIAN__
  31104. __ai int8x16x2_t vtrnq_s8(int8x16_t __p0, int8x16_t __p1) {
  31105.   int8x16x2_t __ret;
  31106.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 32);
  31107.   return __ret;
  31108. }
  31109. #else
  31110. __ai int8x16x2_t vtrnq_s8(int8x16_t __p0, int8x16_t __p1) {
  31111.   int8x16x2_t __ret;
  31112.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31113.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31114.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  31115.  
  31116.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31117.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31118.   return __ret;
  31119. }
  31120. #endif
  31121.  
  31122. #ifdef __LITTLE_ENDIAN__
  31123. __ai float32x4x2_t vtrnq_f32(float32x4_t __p0, float32x4_t __p1) {
  31124.   float32x4x2_t __ret;
  31125.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 41);
  31126.   return __ret;
  31127. }
  31128. #else
  31129. __ai float32x4x2_t vtrnq_f32(float32x4_t __p0, float32x4_t __p1) {
  31130.   float32x4x2_t __ret;
  31131.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31132.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31133.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  31134.  
  31135.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31136.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31137.   return __ret;
  31138. }
  31139. #endif
  31140.  
  31141. #ifdef __LITTLE_ENDIAN__
  31142. __ai int32x4x2_t vtrnq_s32(int32x4_t __p0, int32x4_t __p1) {
  31143.   int32x4x2_t __ret;
  31144.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 34);
  31145.   return __ret;
  31146. }
  31147. #else
  31148. __ai int32x4x2_t vtrnq_s32(int32x4_t __p0, int32x4_t __p1) {
  31149.   int32x4x2_t __ret;
  31150.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31151.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31152.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  31153.  
  31154.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31155.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31156.   return __ret;
  31157. }
  31158. #endif
  31159.  
  31160. #ifdef __LITTLE_ENDIAN__
  31161. __ai int16x8x2_t vtrnq_s16(int16x8_t __p0, int16x8_t __p1) {
  31162.   int16x8x2_t __ret;
  31163.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 33);
  31164.   return __ret;
  31165. }
  31166. #else
  31167. __ai int16x8x2_t vtrnq_s16(int16x8_t __p0, int16x8_t __p1) {
  31168.   int16x8x2_t __ret;
  31169.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31170.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31171.   __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  31172.  
  31173.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31174.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31175.   return __ret;
  31176. }
  31177. #endif
  31178.  
  31179. #ifdef __LITTLE_ENDIAN__
  31180. __ai uint8x8x2_t vtrn_u8(uint8x8_t __p0, uint8x8_t __p1) {
  31181.   uint8x8x2_t __ret;
  31182.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 16);
  31183.   return __ret;
  31184. }
  31185. #else
  31186. __ai uint8x8x2_t vtrn_u8(uint8x8_t __p0, uint8x8_t __p1) {
  31187.   uint8x8x2_t __ret;
  31188.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31189.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31190.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  31191.  
  31192.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31193.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31194.   return __ret;
  31195. }
  31196. #endif
  31197.  
  31198. #ifdef __LITTLE_ENDIAN__
  31199. __ai uint32x2x2_t vtrn_u32(uint32x2_t __p0, uint32x2_t __p1) {
  31200.   uint32x2x2_t __ret;
  31201.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 18);
  31202.   return __ret;
  31203. }
  31204. #else
  31205. __ai uint32x2x2_t vtrn_u32(uint32x2_t __p0, uint32x2_t __p1) {
  31206.   uint32x2x2_t __ret;
  31207.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  31208.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  31209.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  31210.  
  31211.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
  31212.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
  31213.   return __ret;
  31214. }
  31215. #endif
  31216.  
  31217. #ifdef __LITTLE_ENDIAN__
  31218. __ai uint16x4x2_t vtrn_u16(uint16x4_t __p0, uint16x4_t __p1) {
  31219.   uint16x4x2_t __ret;
  31220.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 17);
  31221.   return __ret;
  31222. }
  31223. #else
  31224. __ai uint16x4x2_t vtrn_u16(uint16x4_t __p0, uint16x4_t __p1) {
  31225.   uint16x4x2_t __ret;
  31226.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31227.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31228.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  31229.  
  31230.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31231.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31232.   return __ret;
  31233. }
  31234. #endif
  31235.  
  31236. #ifdef __LITTLE_ENDIAN__
  31237. __ai int8x8x2_t vtrn_s8(int8x8_t __p0, int8x8_t __p1) {
  31238.   int8x8x2_t __ret;
  31239.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 0);
  31240.   return __ret;
  31241. }
  31242. #else
  31243. __ai int8x8x2_t vtrn_s8(int8x8_t __p0, int8x8_t __p1) {
  31244.   int8x8x2_t __ret;
  31245.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31246.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31247.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  31248.  
  31249.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31250.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31251.   return __ret;
  31252. }
  31253. #endif
  31254.  
  31255. #ifdef __LITTLE_ENDIAN__
  31256. __ai float32x2x2_t vtrn_f32(float32x2_t __p0, float32x2_t __p1) {
  31257.   float32x2x2_t __ret;
  31258.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 9);
  31259.   return __ret;
  31260. }
  31261. #else
  31262. __ai float32x2x2_t vtrn_f32(float32x2_t __p0, float32x2_t __p1) {
  31263.   float32x2x2_t __ret;
  31264.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  31265.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  31266.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  31267.  
  31268.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
  31269.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
  31270.   return __ret;
  31271. }
  31272. #endif
  31273.  
  31274. #ifdef __LITTLE_ENDIAN__
  31275. __ai int32x2x2_t vtrn_s32(int32x2_t __p0, int32x2_t __p1) {
  31276.   int32x2x2_t __ret;
  31277.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 2);
  31278.   return __ret;
  31279. }
  31280. #else
  31281. __ai int32x2x2_t vtrn_s32(int32x2_t __p0, int32x2_t __p1) {
  31282.   int32x2x2_t __ret;
  31283.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  31284.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  31285.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  31286.  
  31287.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
  31288.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
  31289.   return __ret;
  31290. }
  31291. #endif
  31292.  
  31293. #ifdef __LITTLE_ENDIAN__
  31294. __ai int16x4x2_t vtrn_s16(int16x4_t __p0, int16x4_t __p1) {
  31295.   int16x4x2_t __ret;
  31296.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 1);
  31297.   return __ret;
  31298. }
  31299. #else
  31300. __ai int16x4x2_t vtrn_s16(int16x4_t __p0, int16x4_t __p1) {
  31301.   int16x4x2_t __ret;
  31302.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31303.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31304.   __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  31305.  
  31306.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31307.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31308.   return __ret;
  31309. }
  31310. #endif
  31311.  
  31312. #ifdef __LITTLE_ENDIAN__
  31313. __ai uint8x8_t vtst_p8(poly8x8_t __p0, poly8x8_t __p1) {
  31314.   uint8x8_t __ret;
  31315.   __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  31316.   return __ret;
  31317. }
  31318. #else
  31319. __ai uint8x8_t vtst_p8(poly8x8_t __p0, poly8x8_t __p1) {
  31320.   uint8x8_t __ret;
  31321.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31322.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31323.   __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  31324.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  31325.   return __ret;
  31326. }
  31327. #endif
  31328.  
  31329. #ifdef __LITTLE_ENDIAN__
  31330. __ai uint16x4_t vtst_p16(poly16x4_t __p0, poly16x4_t __p1) {
  31331.   uint16x4_t __ret;
  31332.   __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  31333.   return __ret;
  31334. }
  31335. #else
  31336. __ai uint16x4_t vtst_p16(poly16x4_t __p0, poly16x4_t __p1) {
  31337.   uint16x4_t __ret;
  31338.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31339.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31340.   __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  31341.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  31342.   return __ret;
  31343. }
  31344. #endif
  31345.  
  31346. #ifdef __LITTLE_ENDIAN__
  31347. __ai uint8x16_t vtstq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  31348.   uint8x16_t __ret;
  31349.   __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  31350.   return __ret;
  31351. }
  31352. #else
  31353. __ai uint8x16_t vtstq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  31354.   uint8x16_t __ret;
  31355.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31356.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31357.   __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  31358.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31359.   return __ret;
  31360. }
  31361. #endif
  31362.  
  31363. #ifdef __LITTLE_ENDIAN__
  31364. __ai uint16x8_t vtstq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  31365.   uint16x8_t __ret;
  31366.   __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  31367.   return __ret;
  31368. }
  31369. #else
  31370. __ai uint16x8_t vtstq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  31371.   uint16x8_t __ret;
  31372.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31373.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31374.   __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  31375.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  31376.   return __ret;
  31377. }
  31378. #endif
  31379.  
  31380. #ifdef __LITTLE_ENDIAN__
  31381. __ai uint8x16_t vtstq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  31382.   uint8x16_t __ret;
  31383.   __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  31384.   return __ret;
  31385. }
  31386. #else
  31387. __ai uint8x16_t vtstq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  31388.   uint8x16_t __ret;
  31389.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31390.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31391.   __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  31392.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31393.   return __ret;
  31394. }
  31395. #endif
  31396.  
  31397. #ifdef __LITTLE_ENDIAN__
  31398. __ai uint32x4_t vtstq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  31399.   uint32x4_t __ret;
  31400.   __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  31401.   return __ret;
  31402. }
  31403. #else
  31404. __ai uint32x4_t vtstq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  31405.   uint32x4_t __ret;
  31406.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31407.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31408.   __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  31409.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  31410.   return __ret;
  31411. }
  31412. #endif
  31413.  
  31414. #ifdef __LITTLE_ENDIAN__
  31415. __ai uint16x8_t vtstq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  31416.   uint16x8_t __ret;
  31417.   __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  31418.   return __ret;
  31419. }
  31420. #else
  31421. __ai uint16x8_t vtstq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  31422.   uint16x8_t __ret;
  31423.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31424.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31425.   __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  31426.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  31427.   return __ret;
  31428. }
  31429. #endif
  31430.  
  31431. #ifdef __LITTLE_ENDIAN__
  31432. __ai uint8x16_t vtstq_s8(int8x16_t __p0, int8x16_t __p1) {
  31433.   uint8x16_t __ret;
  31434.   __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  31435.   return __ret;
  31436. }
  31437. #else
  31438. __ai uint8x16_t vtstq_s8(int8x16_t __p0, int8x16_t __p1) {
  31439.   uint8x16_t __ret;
  31440.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31441.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31442.   __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  31443.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31444.   return __ret;
  31445. }
  31446. #endif
  31447.  
  31448. #ifdef __LITTLE_ENDIAN__
  31449. __ai uint32x4_t vtstq_s32(int32x4_t __p0, int32x4_t __p1) {
  31450.   uint32x4_t __ret;
  31451.   __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  31452.   return __ret;
  31453. }
  31454. #else
  31455. __ai uint32x4_t vtstq_s32(int32x4_t __p0, int32x4_t __p1) {
  31456.   uint32x4_t __ret;
  31457.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31458.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31459.   __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  31460.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  31461.   return __ret;
  31462. }
  31463. #endif
  31464.  
  31465. #ifdef __LITTLE_ENDIAN__
  31466. __ai uint16x8_t vtstq_s16(int16x8_t __p0, int16x8_t __p1) {
  31467.   uint16x8_t __ret;
  31468.   __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  31469.   return __ret;
  31470. }
  31471. #else
  31472. __ai uint16x8_t vtstq_s16(int16x8_t __p0, int16x8_t __p1) {
  31473.   uint16x8_t __ret;
  31474.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31475.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31476.   __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  31477.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  31478.   return __ret;
  31479. }
  31480. #endif
  31481.  
  31482. #ifdef __LITTLE_ENDIAN__
  31483. __ai uint8x8_t vtst_u8(uint8x8_t __p0, uint8x8_t __p1) {
  31484.   uint8x8_t __ret;
  31485.   __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  31486.   return __ret;
  31487. }
  31488. #else
  31489. __ai uint8x8_t vtst_u8(uint8x8_t __p0, uint8x8_t __p1) {
  31490.   uint8x8_t __ret;
  31491.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31492.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31493.   __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  31494.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  31495.   return __ret;
  31496. }
  31497. #endif
  31498.  
  31499. #ifdef __LITTLE_ENDIAN__
  31500. __ai uint32x2_t vtst_u32(uint32x2_t __p0, uint32x2_t __p1) {
  31501.   uint32x2_t __ret;
  31502.   __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  31503.   return __ret;
  31504. }
  31505. #else
  31506. __ai uint32x2_t vtst_u32(uint32x2_t __p0, uint32x2_t __p1) {
  31507.   uint32x2_t __ret;
  31508.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  31509.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  31510.   __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  31511.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  31512.   return __ret;
  31513. }
  31514. #endif
  31515.  
  31516. #ifdef __LITTLE_ENDIAN__
  31517. __ai uint16x4_t vtst_u16(uint16x4_t __p0, uint16x4_t __p1) {
  31518.   uint16x4_t __ret;
  31519.   __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  31520.   return __ret;
  31521. }
  31522. #else
  31523. __ai uint16x4_t vtst_u16(uint16x4_t __p0, uint16x4_t __p1) {
  31524.   uint16x4_t __ret;
  31525.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31526.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31527.   __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  31528.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  31529.   return __ret;
  31530. }
  31531. #endif
  31532.  
  31533. #ifdef __LITTLE_ENDIAN__
  31534. __ai uint8x8_t vtst_s8(int8x8_t __p0, int8x8_t __p1) {
  31535.   uint8x8_t __ret;
  31536.   __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  31537.   return __ret;
  31538. }
  31539. #else
  31540. __ai uint8x8_t vtst_s8(int8x8_t __p0, int8x8_t __p1) {
  31541.   uint8x8_t __ret;
  31542.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31543.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31544.   __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  31545.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  31546.   return __ret;
  31547. }
  31548. #endif
  31549.  
  31550. #ifdef __LITTLE_ENDIAN__
  31551. __ai uint32x2_t vtst_s32(int32x2_t __p0, int32x2_t __p1) {
  31552.   uint32x2_t __ret;
  31553.   __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  31554.   return __ret;
  31555. }
  31556. #else
  31557. __ai uint32x2_t vtst_s32(int32x2_t __p0, int32x2_t __p1) {
  31558.   uint32x2_t __ret;
  31559.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  31560.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  31561.   __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  31562.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  31563.   return __ret;
  31564. }
  31565. #endif
  31566.  
  31567. #ifdef __LITTLE_ENDIAN__
  31568. __ai uint16x4_t vtst_s16(int16x4_t __p0, int16x4_t __p1) {
  31569.   uint16x4_t __ret;
  31570.   __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  31571.   return __ret;
  31572. }
  31573. #else
  31574. __ai uint16x4_t vtst_s16(int16x4_t __p0, int16x4_t __p1) {
  31575.   uint16x4_t __ret;
  31576.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31577.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31578.   __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  31579.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  31580.   return __ret;
  31581. }
  31582. #endif
  31583.  
  31584. #ifdef __LITTLE_ENDIAN__
  31585. __ai poly8x8x2_t vuzp_p8(poly8x8_t __p0, poly8x8_t __p1) {
  31586.   poly8x8x2_t __ret;
  31587.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 4);
  31588.   return __ret;
  31589. }
  31590. #else
  31591. __ai poly8x8x2_t vuzp_p8(poly8x8_t __p0, poly8x8_t __p1) {
  31592.   poly8x8x2_t __ret;
  31593.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31594.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31595.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 4);
  31596.  
  31597.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31598.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31599.   return __ret;
  31600. }
  31601. #endif
  31602.  
  31603. #ifdef __LITTLE_ENDIAN__
  31604. __ai poly16x4x2_t vuzp_p16(poly16x4_t __p0, poly16x4_t __p1) {
  31605.   poly16x4x2_t __ret;
  31606.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 5);
  31607.   return __ret;
  31608. }
  31609. #else
  31610. __ai poly16x4x2_t vuzp_p16(poly16x4_t __p0, poly16x4_t __p1) {
  31611.   poly16x4x2_t __ret;
  31612.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31613.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31614.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 5);
  31615.  
  31616.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31617.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31618.   return __ret;
  31619. }
  31620. #endif
  31621.  
  31622. #ifdef __LITTLE_ENDIAN__
  31623. __ai poly8x16x2_t vuzpq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  31624.   poly8x16x2_t __ret;
  31625.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 36);
  31626.   return __ret;
  31627. }
  31628. #else
  31629. __ai poly8x16x2_t vuzpq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  31630.   poly8x16x2_t __ret;
  31631.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31632.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31633.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 36);
  31634.  
  31635.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31636.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31637.   return __ret;
  31638. }
  31639. #endif
  31640.  
  31641. #ifdef __LITTLE_ENDIAN__
  31642. __ai poly16x8x2_t vuzpq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  31643.   poly16x8x2_t __ret;
  31644.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 37);
  31645.   return __ret;
  31646. }
  31647. #else
  31648. __ai poly16x8x2_t vuzpq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  31649.   poly16x8x2_t __ret;
  31650.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31651.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31652.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 37);
  31653.  
  31654.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31655.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31656.   return __ret;
  31657. }
  31658. #endif
  31659.  
  31660. #ifdef __LITTLE_ENDIAN__
  31661. __ai uint8x16x2_t vuzpq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  31662.   uint8x16x2_t __ret;
  31663.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 48);
  31664.   return __ret;
  31665. }
  31666. #else
  31667. __ai uint8x16x2_t vuzpq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  31668.   uint8x16x2_t __ret;
  31669.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31670.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31671.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  31672.  
  31673.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31674.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31675.   return __ret;
  31676. }
  31677. #endif
  31678.  
  31679. #ifdef __LITTLE_ENDIAN__
  31680. __ai uint32x4x2_t vuzpq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  31681.   uint32x4x2_t __ret;
  31682.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 50);
  31683.   return __ret;
  31684. }
  31685. #else
  31686. __ai uint32x4x2_t vuzpq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  31687.   uint32x4x2_t __ret;
  31688.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31689.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31690.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  31691.  
  31692.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31693.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31694.   return __ret;
  31695. }
  31696. #endif
  31697.  
  31698. #ifdef __LITTLE_ENDIAN__
  31699. __ai uint16x8x2_t vuzpq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  31700.   uint16x8x2_t __ret;
  31701.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 49);
  31702.   return __ret;
  31703. }
  31704. #else
  31705. __ai uint16x8x2_t vuzpq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  31706.   uint16x8x2_t __ret;
  31707.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31708.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31709.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  31710.  
  31711.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31712.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31713.   return __ret;
  31714. }
  31715. #endif
  31716.  
  31717. #ifdef __LITTLE_ENDIAN__
  31718. __ai int8x16x2_t vuzpq_s8(int8x16_t __p0, int8x16_t __p1) {
  31719.   int8x16x2_t __ret;
  31720.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 32);
  31721.   return __ret;
  31722. }
  31723. #else
  31724. __ai int8x16x2_t vuzpq_s8(int8x16_t __p0, int8x16_t __p1) {
  31725.   int8x16x2_t __ret;
  31726.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31727.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31728.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  31729.  
  31730.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31731.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31732.   return __ret;
  31733. }
  31734. #endif
  31735.  
  31736. #ifdef __LITTLE_ENDIAN__
  31737. __ai float32x4x2_t vuzpq_f32(float32x4_t __p0, float32x4_t __p1) {
  31738.   float32x4x2_t __ret;
  31739.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 41);
  31740.   return __ret;
  31741. }
  31742. #else
  31743. __ai float32x4x2_t vuzpq_f32(float32x4_t __p0, float32x4_t __p1) {
  31744.   float32x4x2_t __ret;
  31745.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31746.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31747.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  31748.  
  31749.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31750.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31751.   return __ret;
  31752. }
  31753. #endif
  31754.  
  31755. #ifdef __LITTLE_ENDIAN__
  31756. __ai int32x4x2_t vuzpq_s32(int32x4_t __p0, int32x4_t __p1) {
  31757.   int32x4x2_t __ret;
  31758.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 34);
  31759.   return __ret;
  31760. }
  31761. #else
  31762. __ai int32x4x2_t vuzpq_s32(int32x4_t __p0, int32x4_t __p1) {
  31763.   int32x4x2_t __ret;
  31764.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31765.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31766.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  31767.  
  31768.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31769.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31770.   return __ret;
  31771. }
  31772. #endif
  31773.  
  31774. #ifdef __LITTLE_ENDIAN__
  31775. __ai int16x8x2_t vuzpq_s16(int16x8_t __p0, int16x8_t __p1) {
  31776.   int16x8x2_t __ret;
  31777.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 33);
  31778.   return __ret;
  31779. }
  31780. #else
  31781. __ai int16x8x2_t vuzpq_s16(int16x8_t __p0, int16x8_t __p1) {
  31782.   int16x8x2_t __ret;
  31783.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31784.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31785.   __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  31786.  
  31787.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31788.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31789.   return __ret;
  31790. }
  31791. #endif
  31792.  
  31793. #ifdef __LITTLE_ENDIAN__
  31794. __ai uint8x8x2_t vuzp_u8(uint8x8_t __p0, uint8x8_t __p1) {
  31795.   uint8x8x2_t __ret;
  31796.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 16);
  31797.   return __ret;
  31798. }
  31799. #else
  31800. __ai uint8x8x2_t vuzp_u8(uint8x8_t __p0, uint8x8_t __p1) {
  31801.   uint8x8x2_t __ret;
  31802.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31803.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31804.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  31805.  
  31806.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31807.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31808.   return __ret;
  31809. }
  31810. #endif
  31811.  
  31812. #ifdef __LITTLE_ENDIAN__
  31813. __ai uint32x2x2_t vuzp_u32(uint32x2_t __p0, uint32x2_t __p1) {
  31814.   uint32x2x2_t __ret;
  31815.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 18);
  31816.   return __ret;
  31817. }
  31818. #else
  31819. __ai uint32x2x2_t vuzp_u32(uint32x2_t __p0, uint32x2_t __p1) {
  31820.   uint32x2x2_t __ret;
  31821.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  31822.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  31823.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  31824.  
  31825.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
  31826.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
  31827.   return __ret;
  31828. }
  31829. #endif
  31830.  
  31831. #ifdef __LITTLE_ENDIAN__
  31832. __ai uint16x4x2_t vuzp_u16(uint16x4_t __p0, uint16x4_t __p1) {
  31833.   uint16x4x2_t __ret;
  31834.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 17);
  31835.   return __ret;
  31836. }
  31837. #else
  31838. __ai uint16x4x2_t vuzp_u16(uint16x4_t __p0, uint16x4_t __p1) {
  31839.   uint16x4x2_t __ret;
  31840.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31841.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31842.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  31843.  
  31844.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31845.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31846.   return __ret;
  31847. }
  31848. #endif
  31849.  
  31850. #ifdef __LITTLE_ENDIAN__
  31851. __ai int8x8x2_t vuzp_s8(int8x8_t __p0, int8x8_t __p1) {
  31852.   int8x8x2_t __ret;
  31853.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 0);
  31854.   return __ret;
  31855. }
  31856. #else
  31857. __ai int8x8x2_t vuzp_s8(int8x8_t __p0, int8x8_t __p1) {
  31858.   int8x8x2_t __ret;
  31859.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31860.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31861.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  31862.  
  31863.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31864.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31865.   return __ret;
  31866. }
  31867. #endif
  31868.  
  31869. #ifdef __LITTLE_ENDIAN__
  31870. __ai float32x2x2_t vuzp_f32(float32x2_t __p0, float32x2_t __p1) {
  31871.   float32x2x2_t __ret;
  31872.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 9);
  31873.   return __ret;
  31874. }
  31875. #else
  31876. __ai float32x2x2_t vuzp_f32(float32x2_t __p0, float32x2_t __p1) {
  31877.   float32x2x2_t __ret;
  31878.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  31879.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  31880.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  31881.  
  31882.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
  31883.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
  31884.   return __ret;
  31885. }
  31886. #endif
  31887.  
  31888. #ifdef __LITTLE_ENDIAN__
  31889. __ai int32x2x2_t vuzp_s32(int32x2_t __p0, int32x2_t __p1) {
  31890.   int32x2x2_t __ret;
  31891.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 2);
  31892.   return __ret;
  31893. }
  31894. #else
  31895. __ai int32x2x2_t vuzp_s32(int32x2_t __p0, int32x2_t __p1) {
  31896.   int32x2x2_t __ret;
  31897.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  31898.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  31899.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  31900.  
  31901.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
  31902.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
  31903.   return __ret;
  31904. }
  31905. #endif
  31906.  
  31907. #ifdef __LITTLE_ENDIAN__
  31908. __ai int16x4x2_t vuzp_s16(int16x4_t __p0, int16x4_t __p1) {
  31909.   int16x4x2_t __ret;
  31910.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 1);
  31911.   return __ret;
  31912. }
  31913. #else
  31914. __ai int16x4x2_t vuzp_s16(int16x4_t __p0, int16x4_t __p1) {
  31915.   int16x4x2_t __ret;
  31916.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31917.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31918.   __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  31919.  
  31920.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31921.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31922.   return __ret;
  31923. }
  31924. #endif
  31925.  
  31926. #ifdef __LITTLE_ENDIAN__
  31927. __ai poly8x8x2_t vzip_p8(poly8x8_t __p0, poly8x8_t __p1) {
  31928.   poly8x8x2_t __ret;
  31929.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 4);
  31930.   return __ret;
  31931. }
  31932. #else
  31933. __ai poly8x8x2_t vzip_p8(poly8x8_t __p0, poly8x8_t __p1) {
  31934.   poly8x8x2_t __ret;
  31935.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31936.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31937.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 4);
  31938.  
  31939.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31940.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31941.   return __ret;
  31942. }
  31943. #endif
  31944.  
  31945. #ifdef __LITTLE_ENDIAN__
  31946. __ai poly16x4x2_t vzip_p16(poly16x4_t __p0, poly16x4_t __p1) {
  31947.   poly16x4x2_t __ret;
  31948.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 5);
  31949.   return __ret;
  31950. }
  31951. #else
  31952. __ai poly16x4x2_t vzip_p16(poly16x4_t __p0, poly16x4_t __p1) {
  31953.   poly16x4x2_t __ret;
  31954.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  31955.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  31956.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 5);
  31957.  
  31958.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  31959.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  31960.   return __ret;
  31961. }
  31962. #endif
  31963.  
  31964. #ifdef __LITTLE_ENDIAN__
  31965. __ai poly8x16x2_t vzipq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  31966.   poly8x16x2_t __ret;
  31967.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 36);
  31968.   return __ret;
  31969. }
  31970. #else
  31971. __ai poly8x16x2_t vzipq_p8(poly8x16_t __p0, poly8x16_t __p1) {
  31972.   poly8x16x2_t __ret;
  31973.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31974.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31975.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 36);
  31976.  
  31977.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31978.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  31979.   return __ret;
  31980. }
  31981. #endif
  31982.  
  31983. #ifdef __LITTLE_ENDIAN__
  31984. __ai poly16x8x2_t vzipq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  31985.   poly16x8x2_t __ret;
  31986.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 37);
  31987.   return __ret;
  31988. }
  31989. #else
  31990. __ai poly16x8x2_t vzipq_p16(poly16x8_t __p0, poly16x8_t __p1) {
  31991.   poly16x8x2_t __ret;
  31992.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  31993.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  31994.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 37);
  31995.  
  31996.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  31997.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  31998.   return __ret;
  31999. }
  32000. #endif
  32001.  
  32002. #ifdef __LITTLE_ENDIAN__
  32003. __ai uint8x16x2_t vzipq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  32004.   uint8x16x2_t __ret;
  32005.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 48);
  32006.   return __ret;
  32007. }
  32008. #else
  32009. __ai uint8x16x2_t vzipq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  32010.   uint8x16x2_t __ret;
  32011.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  32012.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  32013.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  32014.  
  32015.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  32016.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  32017.   return __ret;
  32018. }
  32019. #endif
  32020.  
  32021. #ifdef __LITTLE_ENDIAN__
  32022. __ai uint32x4x2_t vzipq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  32023.   uint32x4x2_t __ret;
  32024.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 50);
  32025.   return __ret;
  32026. }
  32027. #else
  32028. __ai uint32x4x2_t vzipq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  32029.   uint32x4x2_t __ret;
  32030.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32031.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  32032.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  32033.  
  32034.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  32035.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  32036.   return __ret;
  32037. }
  32038. #endif
  32039.  
  32040. #ifdef __LITTLE_ENDIAN__
  32041. __ai uint16x8x2_t vzipq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  32042.   uint16x8x2_t __ret;
  32043.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 49);
  32044.   return __ret;
  32045. }
  32046. #else
  32047. __ai uint16x8x2_t vzipq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  32048.   uint16x8x2_t __ret;
  32049.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  32050.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  32051.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  32052.  
  32053.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  32054.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  32055.   return __ret;
  32056. }
  32057. #endif
  32058.  
  32059. #ifdef __LITTLE_ENDIAN__
  32060. __ai int8x16x2_t vzipq_s8(int8x16_t __p0, int8x16_t __p1) {
  32061.   int8x16x2_t __ret;
  32062.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 32);
  32063.   return __ret;
  32064. }
  32065. #else
  32066. __ai int8x16x2_t vzipq_s8(int8x16_t __p0, int8x16_t __p1) {
  32067.   int8x16x2_t __ret;
  32068.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  32069.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  32070.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  32071.  
  32072.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  32073.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  32074.   return __ret;
  32075. }
  32076. #endif
  32077.  
  32078. #ifdef __LITTLE_ENDIAN__
  32079. __ai float32x4x2_t vzipq_f32(float32x4_t __p0, float32x4_t __p1) {
  32080.   float32x4x2_t __ret;
  32081.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 41);
  32082.   return __ret;
  32083. }
  32084. #else
  32085. __ai float32x4x2_t vzipq_f32(float32x4_t __p0, float32x4_t __p1) {
  32086.   float32x4x2_t __ret;
  32087.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32088.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  32089.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  32090.  
  32091.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  32092.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  32093.   return __ret;
  32094. }
  32095. #endif
  32096.  
  32097. #ifdef __LITTLE_ENDIAN__
  32098. __ai int32x4x2_t vzipq_s32(int32x4_t __p0, int32x4_t __p1) {
  32099.   int32x4x2_t __ret;
  32100.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 34);
  32101.   return __ret;
  32102. }
  32103. #else
  32104. __ai int32x4x2_t vzipq_s32(int32x4_t __p0, int32x4_t __p1) {
  32105.   int32x4x2_t __ret;
  32106.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32107.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  32108.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  32109.  
  32110.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  32111.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  32112.   return __ret;
  32113. }
  32114. #endif
  32115.  
  32116. #ifdef __LITTLE_ENDIAN__
  32117. __ai int16x8x2_t vzipq_s16(int16x8_t __p0, int16x8_t __p1) {
  32118.   int16x8x2_t __ret;
  32119.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 33);
  32120.   return __ret;
  32121. }
  32122. #else
  32123. __ai int16x8x2_t vzipq_s16(int16x8_t __p0, int16x8_t __p1) {
  32124.   int16x8x2_t __ret;
  32125.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  32126.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  32127.   __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  32128.  
  32129.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  32130.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  32131.   return __ret;
  32132. }
  32133. #endif
  32134.  
  32135. #ifdef __LITTLE_ENDIAN__
  32136. __ai uint8x8x2_t vzip_u8(uint8x8_t __p0, uint8x8_t __p1) {
  32137.   uint8x8x2_t __ret;
  32138.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 16);
  32139.   return __ret;
  32140. }
  32141. #else
  32142. __ai uint8x8x2_t vzip_u8(uint8x8_t __p0, uint8x8_t __p1) {
  32143.   uint8x8x2_t __ret;
  32144.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  32145.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  32146.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  32147.  
  32148.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  32149.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  32150.   return __ret;
  32151. }
  32152. #endif
  32153.  
  32154. #ifdef __LITTLE_ENDIAN__
  32155. __ai uint32x2x2_t vzip_u32(uint32x2_t __p0, uint32x2_t __p1) {
  32156.   uint32x2x2_t __ret;
  32157.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 18);
  32158.   return __ret;
  32159. }
  32160. #else
  32161. __ai uint32x2x2_t vzip_u32(uint32x2_t __p0, uint32x2_t __p1) {
  32162.   uint32x2x2_t __ret;
  32163.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  32164.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  32165.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  32166.  
  32167.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
  32168.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
  32169.   return __ret;
  32170. }
  32171. #endif
  32172.  
  32173. #ifdef __LITTLE_ENDIAN__
  32174. __ai uint16x4x2_t vzip_u16(uint16x4_t __p0, uint16x4_t __p1) {
  32175.   uint16x4x2_t __ret;
  32176.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 17);
  32177.   return __ret;
  32178. }
  32179. #else
  32180. __ai uint16x4x2_t vzip_u16(uint16x4_t __p0, uint16x4_t __p1) {
  32181.   uint16x4x2_t __ret;
  32182.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32183.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  32184.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  32185.  
  32186.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  32187.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  32188.   return __ret;
  32189. }
  32190. #endif
  32191.  
  32192. #ifdef __LITTLE_ENDIAN__
  32193. __ai int8x8x2_t vzip_s8(int8x8_t __p0, int8x8_t __p1) {
  32194.   int8x8x2_t __ret;
  32195.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 0);
  32196.   return __ret;
  32197. }
  32198. #else
  32199. __ai int8x8x2_t vzip_s8(int8x8_t __p0, int8x8_t __p1) {
  32200.   int8x8x2_t __ret;
  32201.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  32202.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  32203.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  32204.  
  32205.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  32206.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  32207.   return __ret;
  32208. }
  32209. #endif
  32210.  
  32211. #ifdef __LITTLE_ENDIAN__
  32212. __ai float32x2x2_t vzip_f32(float32x2_t __p0, float32x2_t __p1) {
  32213.   float32x2x2_t __ret;
  32214.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 9);
  32215.   return __ret;
  32216. }
  32217. #else
  32218. __ai float32x2x2_t vzip_f32(float32x2_t __p0, float32x2_t __p1) {
  32219.   float32x2x2_t __ret;
  32220.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  32221.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  32222.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  32223.  
  32224.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
  32225.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
  32226.   return __ret;
  32227. }
  32228. #endif
  32229.  
  32230. #ifdef __LITTLE_ENDIAN__
  32231. __ai int32x2x2_t vzip_s32(int32x2_t __p0, int32x2_t __p1) {
  32232.   int32x2x2_t __ret;
  32233.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 2);
  32234.   return __ret;
  32235. }
  32236. #else
  32237. __ai int32x2x2_t vzip_s32(int32x2_t __p0, int32x2_t __p1) {
  32238.   int32x2x2_t __ret;
  32239.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  32240.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  32241.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  32242.  
  32243.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
  32244.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
  32245.   return __ret;
  32246. }
  32247. #endif
  32248.  
  32249. #ifdef __LITTLE_ENDIAN__
  32250. __ai int16x4x2_t vzip_s16(int16x4_t __p0, int16x4_t __p1) {
  32251.   int16x4x2_t __ret;
  32252.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 1);
  32253.   return __ret;
  32254. }
  32255. #else
  32256. __ai int16x4x2_t vzip_s16(int16x4_t __p0, int16x4_t __p1) {
  32257.   int16x4x2_t __ret;
  32258.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32259.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  32260.   __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  32261.  
  32262.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  32263.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  32264.   return __ret;
  32265. }
  32266. #endif
  32267.  
  32268. #ifdef __LITTLE_ENDIAN__
  32269. #define splatq_lane_bf16(__p0, __p1) __extension__ ({ \
  32270.   bfloat16x8_t __ret; \
  32271.   bfloat16x4_t __s0 = __p0; \
  32272.   __ret = (bfloat16x8_t) __builtin_neon_splatq_lane_bf16((int8x8_t)__s0, __p1, 11); \
  32273.   __ret; \
  32274. })
  32275. #else
  32276. #define splatq_lane_bf16(__p0, __p1) __extension__ ({ \
  32277.   bfloat16x8_t __ret; \
  32278.   bfloat16x4_t __s0 = __p0; \
  32279.   bfloat16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  32280.   __ret = (bfloat16x8_t) __builtin_neon_splatq_lane_bf16((int8x8_t)__rev0, __p1, 11); \
  32281.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  32282.   __ret; \
  32283. })
  32284. #define __noswap_splatq_lane_bf16(__p0, __p1) __extension__ ({ \
  32285.   bfloat16x8_t __ret; \
  32286.   bfloat16x4_t __s0 = __p0; \
  32287.   __ret = (bfloat16x8_t) __builtin_neon_splatq_lane_bf16((int8x8_t)__s0, __p1, 11); \
  32288.   __ret; \
  32289. })
  32290. #endif
  32291.  
  32292. #ifdef __LITTLE_ENDIAN__
  32293. #define splat_lane_bf16(__p0, __p1) __extension__ ({ \
  32294.   bfloat16x4_t __ret; \
  32295.   bfloat16x4_t __s0 = __p0; \
  32296.   __ret = (bfloat16x4_t) __builtin_neon_splat_lane_bf16((int8x8_t)__s0, __p1, 11); \
  32297.   __ret; \
  32298. })
  32299. #else
  32300. #define splat_lane_bf16(__p0, __p1) __extension__ ({ \
  32301.   bfloat16x4_t __ret; \
  32302.   bfloat16x4_t __s0 = __p0; \
  32303.   bfloat16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  32304.   __ret = (bfloat16x4_t) __builtin_neon_splat_lane_bf16((int8x8_t)__rev0, __p1, 11); \
  32305.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  32306.   __ret; \
  32307. })
  32308. #define __noswap_splat_lane_bf16(__p0, __p1) __extension__ ({ \
  32309.   bfloat16x4_t __ret; \
  32310.   bfloat16x4_t __s0 = __p0; \
  32311.   __ret = (bfloat16x4_t) __builtin_neon_splat_lane_bf16((int8x8_t)__s0, __p1, 11); \
  32312.   __ret; \
  32313. })
  32314. #endif
  32315.  
  32316. #ifdef __LITTLE_ENDIAN__
  32317. #define splatq_laneq_bf16(__p0, __p1) __extension__ ({ \
  32318.   bfloat16x8_t __ret; \
  32319.   bfloat16x8_t __s0 = __p0; \
  32320.   __ret = (bfloat16x8_t) __builtin_neon_splatq_laneq_bf16((int8x16_t)__s0, __p1, 43); \
  32321.   __ret; \
  32322. })
  32323. #else
  32324. #define splatq_laneq_bf16(__p0, __p1) __extension__ ({ \
  32325.   bfloat16x8_t __ret; \
  32326.   bfloat16x8_t __s0 = __p0; \
  32327.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  32328.   __ret = (bfloat16x8_t) __builtin_neon_splatq_laneq_bf16((int8x16_t)__rev0, __p1, 43); \
  32329.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  32330.   __ret; \
  32331. })
  32332. #define __noswap_splatq_laneq_bf16(__p0, __p1) __extension__ ({ \
  32333.   bfloat16x8_t __ret; \
  32334.   bfloat16x8_t __s0 = __p0; \
  32335.   __ret = (bfloat16x8_t) __builtin_neon_splatq_laneq_bf16((int8x16_t)__s0, __p1, 43); \
  32336.   __ret; \
  32337. })
  32338. #endif
  32339.  
  32340. #ifdef __LITTLE_ENDIAN__
  32341. #define splat_laneq_bf16(__p0, __p1) __extension__ ({ \
  32342.   bfloat16x4_t __ret; \
  32343.   bfloat16x8_t __s0 = __p0; \
  32344.   __ret = (bfloat16x4_t) __builtin_neon_splat_laneq_bf16((int8x16_t)__s0, __p1, 43); \
  32345.   __ret; \
  32346. })
  32347. #else
  32348. #define splat_laneq_bf16(__p0, __p1) __extension__ ({ \
  32349.   bfloat16x4_t __ret; \
  32350.   bfloat16x8_t __s0 = __p0; \
  32351.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  32352.   __ret = (bfloat16x4_t) __builtin_neon_splat_laneq_bf16((int8x16_t)__rev0, __p1, 43); \
  32353.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  32354.   __ret; \
  32355. })
  32356. #define __noswap_splat_laneq_bf16(__p0, __p1) __extension__ ({ \
  32357.   bfloat16x4_t __ret; \
  32358.   bfloat16x8_t __s0 = __p0; \
  32359.   __ret = (bfloat16x4_t) __builtin_neon_splat_laneq_bf16((int8x16_t)__s0, __p1, 43); \
  32360.   __ret; \
  32361. })
  32362. #endif
  32363.  
  32364. #ifdef __LITTLE_ENDIAN__
  32365. __ai __attribute__((target("bf16"))) float32x4_t vbfdotq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32366.   float32x4_t __ret;
  32367.   __ret = (float32x4_t) __builtin_neon_vbfdotq_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  32368.   return __ret;
  32369. }
  32370. #else
  32371. __ai __attribute__((target("bf16"))) float32x4_t vbfdotq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32372.   float32x4_t __ret;
  32373.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32374.   bfloat16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  32375.   bfloat16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  32376.   __ret = (float32x4_t) __builtin_neon_vbfdotq_f32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  32377.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  32378.   return __ret;
  32379. }
  32380. __ai __attribute__((target("bf16"))) float32x4_t __noswap_vbfdotq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32381.   float32x4_t __ret;
  32382.   __ret = (float32x4_t) __builtin_neon_vbfdotq_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  32383.   return __ret;
  32384. }
  32385. #endif
  32386.  
  32387. #ifdef __LITTLE_ENDIAN__
  32388. __ai __attribute__((target("bf16"))) float32x2_t vbfdot_f32(float32x2_t __p0, bfloat16x4_t __p1, bfloat16x4_t __p2) {
  32389.   float32x2_t __ret;
  32390.   __ret = (float32x2_t) __builtin_neon_vbfdot_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  32391.   return __ret;
  32392. }
  32393. #else
  32394. __ai __attribute__((target("bf16"))) float32x2_t vbfdot_f32(float32x2_t __p0, bfloat16x4_t __p1, bfloat16x4_t __p2) {
  32395.   float32x2_t __ret;
  32396.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  32397.   bfloat16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  32398.   bfloat16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  32399.   __ret = (float32x2_t) __builtin_neon_vbfdot_f32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  32400.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  32401.   return __ret;
  32402. }
  32403. __ai __attribute__((target("bf16"))) float32x2_t __noswap_vbfdot_f32(float32x2_t __p0, bfloat16x4_t __p1, bfloat16x4_t __p2) {
  32404.   float32x2_t __ret;
  32405.   __ret = (float32x2_t) __builtin_neon_vbfdot_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  32406.   return __ret;
  32407. }
  32408. #endif
  32409.  
  32410. #ifdef __LITTLE_ENDIAN__
  32411. #define vbfdotq_lane_f32(__p0_126, __p1_126, __p2_126, __p3_126) __extension__ ({ \
  32412.   float32x4_t __ret_126; \
  32413.   float32x4_t __s0_126 = __p0_126; \
  32414.   bfloat16x8_t __s1_126 = __p1_126; \
  32415.   bfloat16x4_t __s2_126 = __p2_126; \
  32416. bfloat16x4_t __reint_126 = __s2_126; \
  32417. float32x4_t __reint1_126 = splatq_lane_f32(*(float32x2_t *) &__reint_126, __p3_126); \
  32418.   __ret_126 = vbfdotq_f32(__s0_126, __s1_126, *(bfloat16x8_t *) &__reint1_126); \
  32419.   __ret_126; \
  32420. })
  32421. #else
  32422. #define vbfdotq_lane_f32(__p0_127, __p1_127, __p2_127, __p3_127) __extension__ ({ \
  32423.   float32x4_t __ret_127; \
  32424.   float32x4_t __s0_127 = __p0_127; \
  32425.   bfloat16x8_t __s1_127 = __p1_127; \
  32426.   bfloat16x4_t __s2_127 = __p2_127; \
  32427.   float32x4_t __rev0_127;  __rev0_127 = __builtin_shufflevector(__s0_127, __s0_127, 3, 2, 1, 0); \
  32428.   bfloat16x8_t __rev1_127;  __rev1_127 = __builtin_shufflevector(__s1_127, __s1_127, 7, 6, 5, 4, 3, 2, 1, 0); \
  32429.   bfloat16x4_t __rev2_127;  __rev2_127 = __builtin_shufflevector(__s2_127, __s2_127, 3, 2, 1, 0); \
  32430. bfloat16x4_t __reint_127 = __rev2_127; \
  32431. float32x4_t __reint1_127 = __noswap_splatq_lane_f32(*(float32x2_t *) &__reint_127, __p3_127); \
  32432.   __ret_127 = __noswap_vbfdotq_f32(__rev0_127, __rev1_127, *(bfloat16x8_t *) &__reint1_127); \
  32433.   __ret_127 = __builtin_shufflevector(__ret_127, __ret_127, 3, 2, 1, 0); \
  32434.   __ret_127; \
  32435. })
  32436. #endif
  32437.  
  32438. #ifdef __LITTLE_ENDIAN__
  32439. #define vbfdot_lane_f32(__p0_128, __p1_128, __p2_128, __p3_128) __extension__ ({ \
  32440.   float32x2_t __ret_128; \
  32441.   float32x2_t __s0_128 = __p0_128; \
  32442.   bfloat16x4_t __s1_128 = __p1_128; \
  32443.   bfloat16x4_t __s2_128 = __p2_128; \
  32444. bfloat16x4_t __reint_128 = __s2_128; \
  32445. float32x2_t __reint1_128 = splat_lane_f32(*(float32x2_t *) &__reint_128, __p3_128); \
  32446.   __ret_128 = vbfdot_f32(__s0_128, __s1_128, *(bfloat16x4_t *) &__reint1_128); \
  32447.   __ret_128; \
  32448. })
  32449. #else
  32450. #define vbfdot_lane_f32(__p0_129, __p1_129, __p2_129, __p3_129) __extension__ ({ \
  32451.   float32x2_t __ret_129; \
  32452.   float32x2_t __s0_129 = __p0_129; \
  32453.   bfloat16x4_t __s1_129 = __p1_129; \
  32454.   bfloat16x4_t __s2_129 = __p2_129; \
  32455.   float32x2_t __rev0_129;  __rev0_129 = __builtin_shufflevector(__s0_129, __s0_129, 1, 0); \
  32456.   bfloat16x4_t __rev1_129;  __rev1_129 = __builtin_shufflevector(__s1_129, __s1_129, 3, 2, 1, 0); \
  32457.   bfloat16x4_t __rev2_129;  __rev2_129 = __builtin_shufflevector(__s2_129, __s2_129, 3, 2, 1, 0); \
  32458. bfloat16x4_t __reint_129 = __rev2_129; \
  32459. float32x2_t __reint1_129 = __noswap_splat_lane_f32(*(float32x2_t *) &__reint_129, __p3_129); \
  32460.   __ret_129 = __noswap_vbfdot_f32(__rev0_129, __rev1_129, *(bfloat16x4_t *) &__reint1_129); \
  32461.   __ret_129 = __builtin_shufflevector(__ret_129, __ret_129, 1, 0); \
  32462.   __ret_129; \
  32463. })
  32464. #endif
  32465.  
  32466. #ifdef __LITTLE_ENDIAN__
  32467. #define vbfdotq_laneq_f32(__p0_130, __p1_130, __p2_130, __p3_130) __extension__ ({ \
  32468.   float32x4_t __ret_130; \
  32469.   float32x4_t __s0_130 = __p0_130; \
  32470.   bfloat16x8_t __s1_130 = __p1_130; \
  32471.   bfloat16x8_t __s2_130 = __p2_130; \
  32472. bfloat16x8_t __reint_130 = __s2_130; \
  32473. float32x4_t __reint1_130 = splatq_laneq_f32(*(float32x4_t *) &__reint_130, __p3_130); \
  32474.   __ret_130 = vbfdotq_f32(__s0_130, __s1_130, *(bfloat16x8_t *) &__reint1_130); \
  32475.   __ret_130; \
  32476. })
  32477. #else
  32478. #define vbfdotq_laneq_f32(__p0_131, __p1_131, __p2_131, __p3_131) __extension__ ({ \
  32479.   float32x4_t __ret_131; \
  32480.   float32x4_t __s0_131 = __p0_131; \
  32481.   bfloat16x8_t __s1_131 = __p1_131; \
  32482.   bfloat16x8_t __s2_131 = __p2_131; \
  32483.   float32x4_t __rev0_131;  __rev0_131 = __builtin_shufflevector(__s0_131, __s0_131, 3, 2, 1, 0); \
  32484.   bfloat16x8_t __rev1_131;  __rev1_131 = __builtin_shufflevector(__s1_131, __s1_131, 7, 6, 5, 4, 3, 2, 1, 0); \
  32485.   bfloat16x8_t __rev2_131;  __rev2_131 = __builtin_shufflevector(__s2_131, __s2_131, 7, 6, 5, 4, 3, 2, 1, 0); \
  32486. bfloat16x8_t __reint_131 = __rev2_131; \
  32487. float32x4_t __reint1_131 = __noswap_splatq_laneq_f32(*(float32x4_t *) &__reint_131, __p3_131); \
  32488.   __ret_131 = __noswap_vbfdotq_f32(__rev0_131, __rev1_131, *(bfloat16x8_t *) &__reint1_131); \
  32489.   __ret_131 = __builtin_shufflevector(__ret_131, __ret_131, 3, 2, 1, 0); \
  32490.   __ret_131; \
  32491. })
  32492. #endif
  32493.  
  32494. #ifdef __LITTLE_ENDIAN__
  32495. #define vbfdot_laneq_f32(__p0_132, __p1_132, __p2_132, __p3_132) __extension__ ({ \
  32496.   float32x2_t __ret_132; \
  32497.   float32x2_t __s0_132 = __p0_132; \
  32498.   bfloat16x4_t __s1_132 = __p1_132; \
  32499.   bfloat16x8_t __s2_132 = __p2_132; \
  32500. bfloat16x8_t __reint_132 = __s2_132; \
  32501. float32x2_t __reint1_132 = splat_laneq_f32(*(float32x4_t *) &__reint_132, __p3_132); \
  32502.   __ret_132 = vbfdot_f32(__s0_132, __s1_132, *(bfloat16x4_t *) &__reint1_132); \
  32503.   __ret_132; \
  32504. })
  32505. #else
  32506. #define vbfdot_laneq_f32(__p0_133, __p1_133, __p2_133, __p3_133) __extension__ ({ \
  32507.   float32x2_t __ret_133; \
  32508.   float32x2_t __s0_133 = __p0_133; \
  32509.   bfloat16x4_t __s1_133 = __p1_133; \
  32510.   bfloat16x8_t __s2_133 = __p2_133; \
  32511.   float32x2_t __rev0_133;  __rev0_133 = __builtin_shufflevector(__s0_133, __s0_133, 1, 0); \
  32512.   bfloat16x4_t __rev1_133;  __rev1_133 = __builtin_shufflevector(__s1_133, __s1_133, 3, 2, 1, 0); \
  32513.   bfloat16x8_t __rev2_133;  __rev2_133 = __builtin_shufflevector(__s2_133, __s2_133, 7, 6, 5, 4, 3, 2, 1, 0); \
  32514. bfloat16x8_t __reint_133 = __rev2_133; \
  32515. float32x2_t __reint1_133 = __noswap_splat_laneq_f32(*(float32x4_t *) &__reint_133, __p3_133); \
  32516.   __ret_133 = __noswap_vbfdot_f32(__rev0_133, __rev1_133, *(bfloat16x4_t *) &__reint1_133); \
  32517.   __ret_133 = __builtin_shufflevector(__ret_133, __ret_133, 1, 0); \
  32518.   __ret_133; \
  32519. })
  32520. #endif
  32521.  
  32522. #ifdef __LITTLE_ENDIAN__
  32523. __ai __attribute__((target("bf16"))) float32x4_t vbfmlalbq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32524.   float32x4_t __ret;
  32525.   __ret = (float32x4_t) __builtin_neon_vbfmlalbq_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  32526.   return __ret;
  32527. }
  32528. #else
  32529. __ai __attribute__((target("bf16"))) float32x4_t vbfmlalbq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32530.   float32x4_t __ret;
  32531.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32532.   bfloat16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  32533.   bfloat16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  32534.   __ret = (float32x4_t) __builtin_neon_vbfmlalbq_f32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  32535.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  32536.   return __ret;
  32537. }
  32538. __ai __attribute__((target("bf16"))) float32x4_t __noswap_vbfmlalbq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32539.   float32x4_t __ret;
  32540.   __ret = (float32x4_t) __builtin_neon_vbfmlalbq_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  32541.   return __ret;
  32542. }
  32543. #endif
  32544.  
  32545. #ifdef __LITTLE_ENDIAN__
  32546. __ai __attribute__((target("bf16"))) float32x4_t vbfmlaltq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32547.   float32x4_t __ret;
  32548.   __ret = (float32x4_t) __builtin_neon_vbfmlaltq_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  32549.   return __ret;
  32550. }
  32551. #else
  32552. __ai __attribute__((target("bf16"))) float32x4_t vbfmlaltq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32553.   float32x4_t __ret;
  32554.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32555.   bfloat16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  32556.   bfloat16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  32557.   __ret = (float32x4_t) __builtin_neon_vbfmlaltq_f32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  32558.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  32559.   return __ret;
  32560. }
  32561. __ai __attribute__((target("bf16"))) float32x4_t __noswap_vbfmlaltq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32562.   float32x4_t __ret;
  32563.   __ret = (float32x4_t) __builtin_neon_vbfmlaltq_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  32564.   return __ret;
  32565. }
  32566. #endif
  32567.  
  32568. #ifdef __LITTLE_ENDIAN__
  32569. __ai __attribute__((target("bf16"))) float32x4_t vbfmmlaq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32570.   float32x4_t __ret;
  32571.   __ret = (float32x4_t) __builtin_neon_vbfmmlaq_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  32572.   return __ret;
  32573. }
  32574. #else
  32575. __ai __attribute__((target("bf16"))) float32x4_t vbfmmlaq_f32(float32x4_t __p0, bfloat16x8_t __p1, bfloat16x8_t __p2) {
  32576.   float32x4_t __ret;
  32577.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32578.   bfloat16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  32579.   bfloat16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  32580.   __ret = (float32x4_t) __builtin_neon_vbfmmlaq_f32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  32581.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  32582.   return __ret;
  32583. }
  32584. #endif
  32585.  
  32586. #ifdef __LITTLE_ENDIAN__
  32587. __ai __attribute__((target("bf16"))) bfloat16x8_t vcombine_bf16(bfloat16x4_t __p0, bfloat16x4_t __p1) {
  32588.   bfloat16x8_t __ret;
  32589.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  32590.   return __ret;
  32591. }
  32592. #else
  32593. __ai __attribute__((target("bf16"))) bfloat16x8_t vcombine_bf16(bfloat16x4_t __p0, bfloat16x4_t __p1) {
  32594.   bfloat16x8_t __ret;
  32595.   bfloat16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  32596.   bfloat16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  32597.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
  32598.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  32599.   return __ret;
  32600. }
  32601. __ai __attribute__((target("bf16"))) bfloat16x8_t __noswap_vcombine_bf16(bfloat16x4_t __p0, bfloat16x4_t __p1) {
  32602.   bfloat16x8_t __ret;
  32603.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
  32604.   return __ret;
  32605. }
  32606. #endif
  32607.  
  32608. #define vcreate_bf16(__p0) __extension__ ({ \
  32609.   bfloat16x4_t __ret; \
  32610.   uint64_t __promote = __p0; \
  32611.   __ret = (bfloat16x4_t)(__promote); \
  32612.   __ret; \
  32613. })
  32614. #ifdef __LITTLE_ENDIAN__
  32615. __ai __attribute__((target("bf16"))) float32x4_t vcvt_f32_bf16(bfloat16x4_t __p0_134) {
  32616.   float32x4_t __ret_134;
  32617. bfloat16x4_t __reint_134 = __p0_134;
  32618. int32x4_t __reint1_134 = vshll_n_s16(*(int16x4_t *) &__reint_134, 16);
  32619.   __ret_134 = *(float32x4_t *) &__reint1_134;
  32620.   return __ret_134;
  32621. }
  32622. #else
  32623. __ai __attribute__((target("bf16"))) float32x4_t vcvt_f32_bf16(bfloat16x4_t __p0_135) {
  32624.   float32x4_t __ret_135;
  32625.   bfloat16x4_t __rev0_135;  __rev0_135 = __builtin_shufflevector(__p0_135, __p0_135, 3, 2, 1, 0);
  32626. bfloat16x4_t __reint_135 = __rev0_135;
  32627. int32x4_t __reint1_135 = __noswap_vshll_n_s16(*(int16x4_t *) &__reint_135, 16);
  32628.   __ret_135 = *(float32x4_t *) &__reint1_135;
  32629.   __ret_135 = __builtin_shufflevector(__ret_135, __ret_135, 3, 2, 1, 0);
  32630.   return __ret_135;
  32631. }
  32632. __ai __attribute__((target("bf16"))) float32x4_t __noswap_vcvt_f32_bf16(bfloat16x4_t __p0_136) {
  32633.   float32x4_t __ret_136;
  32634. bfloat16x4_t __reint_136 = __p0_136;
  32635. int32x4_t __reint1_136 = __noswap_vshll_n_s16(*(int16x4_t *) &__reint_136, 16);
  32636.   __ret_136 = *(float32x4_t *) &__reint1_136;
  32637.   return __ret_136;
  32638. }
  32639. #endif
  32640.  
  32641. __ai __attribute__((target("bf16"))) float32_t vcvtah_f32_bf16(bfloat16_t __p0) {
  32642.   float32_t __ret;
  32643. bfloat16_t __reint = __p0;
  32644. int32_t __reint1 = *(int32_t *) &__reint << 16;
  32645.   __ret = *(float32_t *) &__reint1;
  32646.   return __ret;
  32647. }
  32648. __ai __attribute__((target("bf16"))) bfloat16_t vcvth_bf16_f32(float32_t __p0) {
  32649.   bfloat16_t __ret;
  32650.   __ret = (bfloat16_t) __builtin_neon_vcvth_bf16_f32(__p0);
  32651.   return __ret;
  32652. }
  32653. #ifdef __LITTLE_ENDIAN__
  32654. #define vduph_lane_bf16(__p0, __p1) __extension__ ({ \
  32655.   bfloat16_t __ret; \
  32656.   bfloat16x4_t __s0 = __p0; \
  32657.   __ret = (bfloat16_t) __builtin_neon_vduph_lane_bf16((bfloat16x4_t)__s0, __p1); \
  32658.   __ret; \
  32659. })
  32660. #else
  32661. #define vduph_lane_bf16(__p0, __p1) __extension__ ({ \
  32662.   bfloat16_t __ret; \
  32663.   bfloat16x4_t __s0 = __p0; \
  32664.   bfloat16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  32665.   __ret = (bfloat16_t) __builtin_neon_vduph_lane_bf16((bfloat16x4_t)__rev0, __p1); \
  32666.   __ret; \
  32667. })
  32668. #endif
  32669.  
  32670. #ifdef __LITTLE_ENDIAN__
  32671. #define vdupq_lane_bf16(__p0_137, __p1_137) __extension__ ({ \
  32672.   bfloat16x8_t __ret_137; \
  32673.   bfloat16x4_t __s0_137 = __p0_137; \
  32674.   __ret_137 = splatq_lane_bf16(__s0_137, __p1_137); \
  32675.   __ret_137; \
  32676. })
  32677. #else
  32678. #define vdupq_lane_bf16(__p0_138, __p1_138) __extension__ ({ \
  32679.   bfloat16x8_t __ret_138; \
  32680.   bfloat16x4_t __s0_138 = __p0_138; \
  32681.   bfloat16x4_t __rev0_138;  __rev0_138 = __builtin_shufflevector(__s0_138, __s0_138, 3, 2, 1, 0); \
  32682.   __ret_138 = __noswap_splatq_lane_bf16(__rev0_138, __p1_138); \
  32683.   __ret_138 = __builtin_shufflevector(__ret_138, __ret_138, 7, 6, 5, 4, 3, 2, 1, 0); \
  32684.   __ret_138; \
  32685. })
  32686. #endif
  32687.  
  32688. #ifdef __LITTLE_ENDIAN__
  32689. #define vdup_lane_bf16(__p0_139, __p1_139) __extension__ ({ \
  32690.   bfloat16x4_t __ret_139; \
  32691.   bfloat16x4_t __s0_139 = __p0_139; \
  32692.   __ret_139 = splat_lane_bf16(__s0_139, __p1_139); \
  32693.   __ret_139; \
  32694. })
  32695. #else
  32696. #define vdup_lane_bf16(__p0_140, __p1_140) __extension__ ({ \
  32697.   bfloat16x4_t __ret_140; \
  32698.   bfloat16x4_t __s0_140 = __p0_140; \
  32699.   bfloat16x4_t __rev0_140;  __rev0_140 = __builtin_shufflevector(__s0_140, __s0_140, 3, 2, 1, 0); \
  32700.   __ret_140 = __noswap_splat_lane_bf16(__rev0_140, __p1_140); \
  32701.   __ret_140 = __builtin_shufflevector(__ret_140, __ret_140, 3, 2, 1, 0); \
  32702.   __ret_140; \
  32703. })
  32704. #endif
  32705.  
  32706. #ifdef __LITTLE_ENDIAN__
  32707. #define vduph_laneq_bf16(__p0, __p1) __extension__ ({ \
  32708.   bfloat16_t __ret; \
  32709.   bfloat16x8_t __s0 = __p0; \
  32710.   __ret = (bfloat16_t) __builtin_neon_vduph_laneq_bf16((bfloat16x8_t)__s0, __p1); \
  32711.   __ret; \
  32712. })
  32713. #else
  32714. #define vduph_laneq_bf16(__p0, __p1) __extension__ ({ \
  32715.   bfloat16_t __ret; \
  32716.   bfloat16x8_t __s0 = __p0; \
  32717.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  32718.   __ret = (bfloat16_t) __builtin_neon_vduph_laneq_bf16((bfloat16x8_t)__rev0, __p1); \
  32719.   __ret; \
  32720. })
  32721. #endif
  32722.  
  32723. #ifdef __LITTLE_ENDIAN__
  32724. #define vdupq_laneq_bf16(__p0_141, __p1_141) __extension__ ({ \
  32725.   bfloat16x8_t __ret_141; \
  32726.   bfloat16x8_t __s0_141 = __p0_141; \
  32727.   __ret_141 = splatq_laneq_bf16(__s0_141, __p1_141); \
  32728.   __ret_141; \
  32729. })
  32730. #else
  32731. #define vdupq_laneq_bf16(__p0_142, __p1_142) __extension__ ({ \
  32732.   bfloat16x8_t __ret_142; \
  32733.   bfloat16x8_t __s0_142 = __p0_142; \
  32734.   bfloat16x8_t __rev0_142;  __rev0_142 = __builtin_shufflevector(__s0_142, __s0_142, 7, 6, 5, 4, 3, 2, 1, 0); \
  32735.   __ret_142 = __noswap_splatq_laneq_bf16(__rev0_142, __p1_142); \
  32736.   __ret_142 = __builtin_shufflevector(__ret_142, __ret_142, 7, 6, 5, 4, 3, 2, 1, 0); \
  32737.   __ret_142; \
  32738. })
  32739. #endif
  32740.  
  32741. #ifdef __LITTLE_ENDIAN__
  32742. #define vdup_laneq_bf16(__p0_143, __p1_143) __extension__ ({ \
  32743.   bfloat16x4_t __ret_143; \
  32744.   bfloat16x8_t __s0_143 = __p0_143; \
  32745.   __ret_143 = splat_laneq_bf16(__s0_143, __p1_143); \
  32746.   __ret_143; \
  32747. })
  32748. #else
  32749. #define vdup_laneq_bf16(__p0_144, __p1_144) __extension__ ({ \
  32750.   bfloat16x4_t __ret_144; \
  32751.   bfloat16x8_t __s0_144 = __p0_144; \
  32752.   bfloat16x8_t __rev0_144;  __rev0_144 = __builtin_shufflevector(__s0_144, __s0_144, 7, 6, 5, 4, 3, 2, 1, 0); \
  32753.   __ret_144 = __noswap_splat_laneq_bf16(__rev0_144, __p1_144); \
  32754.   __ret_144 = __builtin_shufflevector(__ret_144, __ret_144, 3, 2, 1, 0); \
  32755.   __ret_144; \
  32756. })
  32757. #endif
  32758.  
  32759. #ifdef __LITTLE_ENDIAN__
  32760. __ai __attribute__((target("bf16"))) bfloat16x8_t vdupq_n_bf16(bfloat16_t __p0) {
  32761.   bfloat16x8_t __ret;
  32762.   __ret = (bfloat16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  32763.   return __ret;
  32764. }
  32765. #else
  32766. __ai __attribute__((target("bf16"))) bfloat16x8_t vdupq_n_bf16(bfloat16_t __p0) {
  32767.   bfloat16x8_t __ret;
  32768.   __ret = (bfloat16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
  32769.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  32770.   return __ret;
  32771. }
  32772. #endif
  32773.  
  32774. #ifdef __LITTLE_ENDIAN__
  32775. __ai __attribute__((target("bf16"))) bfloat16x4_t vdup_n_bf16(bfloat16_t __p0) {
  32776.   bfloat16x4_t __ret;
  32777.   __ret = (bfloat16x4_t) {__p0, __p0, __p0, __p0};
  32778.   return __ret;
  32779. }
  32780. #else
  32781. __ai __attribute__((target("bf16"))) bfloat16x4_t vdup_n_bf16(bfloat16_t __p0) {
  32782.   bfloat16x4_t __ret;
  32783.   __ret = (bfloat16x4_t) {__p0, __p0, __p0, __p0};
  32784.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  32785.   return __ret;
  32786. }
  32787. #endif
  32788.  
  32789. #ifdef __LITTLE_ENDIAN__
  32790. __ai __attribute__((target("bf16"))) bfloat16x4_t vget_high_bf16(bfloat16x8_t __p0) {
  32791.   bfloat16x4_t __ret;
  32792.   __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  32793.   return __ret;
  32794. }
  32795. #else
  32796. __ai __attribute__((target("bf16"))) bfloat16x4_t vget_high_bf16(bfloat16x8_t __p0) {
  32797.   bfloat16x4_t __ret;
  32798.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  32799.   __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
  32800.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  32801.   return __ret;
  32802. }
  32803. __ai __attribute__((target("bf16"))) bfloat16x4_t __noswap_vget_high_bf16(bfloat16x8_t __p0) {
  32804.   bfloat16x4_t __ret;
  32805.   __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
  32806.   return __ret;
  32807. }
  32808. #endif
  32809.  
  32810. #ifdef __LITTLE_ENDIAN__
  32811. #define vgetq_lane_bf16(__p0, __p1) __extension__ ({ \
  32812.   bfloat16_t __ret; \
  32813.   bfloat16x8_t __s0 = __p0; \
  32814.   __ret = (bfloat16_t) __builtin_neon_vgetq_lane_bf16((bfloat16x8_t)__s0, __p1); \
  32815.   __ret; \
  32816. })
  32817. #else
  32818. #define vgetq_lane_bf16(__p0, __p1) __extension__ ({ \
  32819.   bfloat16_t __ret; \
  32820.   bfloat16x8_t __s0 = __p0; \
  32821.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  32822.   __ret = (bfloat16_t) __builtin_neon_vgetq_lane_bf16((bfloat16x8_t)__rev0, __p1); \
  32823.   __ret; \
  32824. })
  32825. #define __noswap_vgetq_lane_bf16(__p0, __p1) __extension__ ({ \
  32826.   bfloat16_t __ret; \
  32827.   bfloat16x8_t __s0 = __p0; \
  32828.   __ret = (bfloat16_t) __builtin_neon_vgetq_lane_bf16((bfloat16x8_t)__s0, __p1); \
  32829.   __ret; \
  32830. })
  32831. #endif
  32832.  
  32833. #ifdef __LITTLE_ENDIAN__
  32834. #define vget_lane_bf16(__p0, __p1) __extension__ ({ \
  32835.   bfloat16_t __ret; \
  32836.   bfloat16x4_t __s0 = __p0; \
  32837.   __ret = (bfloat16_t) __builtin_neon_vget_lane_bf16((bfloat16x4_t)__s0, __p1); \
  32838.   __ret; \
  32839. })
  32840. #else
  32841. #define vget_lane_bf16(__p0, __p1) __extension__ ({ \
  32842.   bfloat16_t __ret; \
  32843.   bfloat16x4_t __s0 = __p0; \
  32844.   bfloat16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  32845.   __ret = (bfloat16_t) __builtin_neon_vget_lane_bf16((bfloat16x4_t)__rev0, __p1); \
  32846.   __ret; \
  32847. })
  32848. #define __noswap_vget_lane_bf16(__p0, __p1) __extension__ ({ \
  32849.   bfloat16_t __ret; \
  32850.   bfloat16x4_t __s0 = __p0; \
  32851.   __ret = (bfloat16_t) __builtin_neon_vget_lane_bf16((bfloat16x4_t)__s0, __p1); \
  32852.   __ret; \
  32853. })
  32854. #endif
  32855.  
  32856. #ifdef __LITTLE_ENDIAN__
  32857. __ai __attribute__((target("bf16"))) bfloat16x4_t vget_low_bf16(bfloat16x8_t __p0) {
  32858.   bfloat16x4_t __ret;
  32859.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  32860.   return __ret;
  32861. }
  32862. #else
  32863. __ai __attribute__((target("bf16"))) bfloat16x4_t vget_low_bf16(bfloat16x8_t __p0) {
  32864.   bfloat16x4_t __ret;
  32865.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  32866.   __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
  32867.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  32868.   return __ret;
  32869. }
  32870. __ai __attribute__((target("bf16"))) bfloat16x4_t __noswap_vget_low_bf16(bfloat16x8_t __p0) {
  32871.   bfloat16x4_t __ret;
  32872.   __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
  32873.   return __ret;
  32874. }
  32875. #endif
  32876.  
  32877. #ifdef __LITTLE_ENDIAN__
  32878. #define vld1q_bf16(__p0) __extension__ ({ \
  32879.   bfloat16x8_t __ret; \
  32880.   __ret = (bfloat16x8_t) __builtin_neon_vld1q_bf16(__p0, 43); \
  32881.   __ret; \
  32882. })
  32883. #else
  32884. #define vld1q_bf16(__p0) __extension__ ({ \
  32885.   bfloat16x8_t __ret; \
  32886.   __ret = (bfloat16x8_t) __builtin_neon_vld1q_bf16(__p0, 43); \
  32887.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  32888.   __ret; \
  32889. })
  32890. #endif
  32891.  
  32892. #ifdef __LITTLE_ENDIAN__
  32893. #define vld1_bf16(__p0) __extension__ ({ \
  32894.   bfloat16x4_t __ret; \
  32895.   __ret = (bfloat16x4_t) __builtin_neon_vld1_bf16(__p0, 11); \
  32896.   __ret; \
  32897. })
  32898. #else
  32899. #define vld1_bf16(__p0) __extension__ ({ \
  32900.   bfloat16x4_t __ret; \
  32901.   __ret = (bfloat16x4_t) __builtin_neon_vld1_bf16(__p0, 11); \
  32902.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  32903.   __ret; \
  32904. })
  32905. #endif
  32906.  
  32907. #ifdef __LITTLE_ENDIAN__
  32908. #define vld1q_dup_bf16(__p0) __extension__ ({ \
  32909.   bfloat16x8_t __ret; \
  32910.   __ret = (bfloat16x8_t) __builtin_neon_vld1q_dup_bf16(__p0, 43); \
  32911.   __ret; \
  32912. })
  32913. #else
  32914. #define vld1q_dup_bf16(__p0) __extension__ ({ \
  32915.   bfloat16x8_t __ret; \
  32916.   __ret = (bfloat16x8_t) __builtin_neon_vld1q_dup_bf16(__p0, 43); \
  32917.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  32918.   __ret; \
  32919. })
  32920. #endif
  32921.  
  32922. #ifdef __LITTLE_ENDIAN__
  32923. #define vld1_dup_bf16(__p0) __extension__ ({ \
  32924.   bfloat16x4_t __ret; \
  32925.   __ret = (bfloat16x4_t) __builtin_neon_vld1_dup_bf16(__p0, 11); \
  32926.   __ret; \
  32927. })
  32928. #else
  32929. #define vld1_dup_bf16(__p0) __extension__ ({ \
  32930.   bfloat16x4_t __ret; \
  32931.   __ret = (bfloat16x4_t) __builtin_neon_vld1_dup_bf16(__p0, 11); \
  32932.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  32933.   __ret; \
  32934. })
  32935. #endif
  32936.  
  32937. #ifdef __LITTLE_ENDIAN__
  32938. #define vld1q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  32939.   bfloat16x8_t __ret; \
  32940.   bfloat16x8_t __s1 = __p1; \
  32941.   __ret = (bfloat16x8_t) __builtin_neon_vld1q_lane_bf16(__p0, (int8x16_t)__s1, __p2, 43); \
  32942.   __ret; \
  32943. })
  32944. #else
  32945. #define vld1q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  32946.   bfloat16x8_t __ret; \
  32947.   bfloat16x8_t __s1 = __p1; \
  32948.   bfloat16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  32949.   __ret = (bfloat16x8_t) __builtin_neon_vld1q_lane_bf16(__p0, (int8x16_t)__rev1, __p2, 43); \
  32950.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  32951.   __ret; \
  32952. })
  32953. #endif
  32954.  
  32955. #ifdef __LITTLE_ENDIAN__
  32956. #define vld1_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  32957.   bfloat16x4_t __ret; \
  32958.   bfloat16x4_t __s1 = __p1; \
  32959.   __ret = (bfloat16x4_t) __builtin_neon_vld1_lane_bf16(__p0, (int8x8_t)__s1, __p2, 11); \
  32960.   __ret; \
  32961. })
  32962. #else
  32963. #define vld1_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  32964.   bfloat16x4_t __ret; \
  32965.   bfloat16x4_t __s1 = __p1; \
  32966.   bfloat16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  32967.   __ret = (bfloat16x4_t) __builtin_neon_vld1_lane_bf16(__p0, (int8x8_t)__rev1, __p2, 11); \
  32968.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  32969.   __ret; \
  32970. })
  32971. #endif
  32972.  
  32973. #ifdef __LITTLE_ENDIAN__
  32974. #define vld1q_bf16_x2(__p0) __extension__ ({ \
  32975.   bfloat16x8x2_t __ret; \
  32976.   __builtin_neon_vld1q_bf16_x2(&__ret, __p0, 43); \
  32977.   __ret; \
  32978. })
  32979. #else
  32980. #define vld1q_bf16_x2(__p0) __extension__ ({ \
  32981.   bfloat16x8x2_t __ret; \
  32982.   __builtin_neon_vld1q_bf16_x2(&__ret, __p0, 43); \
  32983.  \
  32984.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  32985.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  32986.   __ret; \
  32987. })
  32988. #endif
  32989.  
  32990. #ifdef __LITTLE_ENDIAN__
  32991. #define vld1_bf16_x2(__p0) __extension__ ({ \
  32992.   bfloat16x4x2_t __ret; \
  32993.   __builtin_neon_vld1_bf16_x2(&__ret, __p0, 11); \
  32994.   __ret; \
  32995. })
  32996. #else
  32997. #define vld1_bf16_x2(__p0) __extension__ ({ \
  32998.   bfloat16x4x2_t __ret; \
  32999.   __builtin_neon_vld1_bf16_x2(&__ret, __p0, 11); \
  33000.  \
  33001.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33002.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33003.   __ret; \
  33004. })
  33005. #endif
  33006.  
  33007. #ifdef __LITTLE_ENDIAN__
  33008. #define vld1q_bf16_x3(__p0) __extension__ ({ \
  33009.   bfloat16x8x3_t __ret; \
  33010.   __builtin_neon_vld1q_bf16_x3(&__ret, __p0, 43); \
  33011.   __ret; \
  33012. })
  33013. #else
  33014. #define vld1q_bf16_x3(__p0) __extension__ ({ \
  33015.   bfloat16x8x3_t __ret; \
  33016.   __builtin_neon_vld1q_bf16_x3(&__ret, __p0, 43); \
  33017.  \
  33018.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33019.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33020.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33021.   __ret; \
  33022. })
  33023. #endif
  33024.  
  33025. #ifdef __LITTLE_ENDIAN__
  33026. #define vld1_bf16_x3(__p0) __extension__ ({ \
  33027.   bfloat16x4x3_t __ret; \
  33028.   __builtin_neon_vld1_bf16_x3(&__ret, __p0, 11); \
  33029.   __ret; \
  33030. })
  33031. #else
  33032. #define vld1_bf16_x3(__p0) __extension__ ({ \
  33033.   bfloat16x4x3_t __ret; \
  33034.   __builtin_neon_vld1_bf16_x3(&__ret, __p0, 11); \
  33035.  \
  33036.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33037.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33038.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  33039.   __ret; \
  33040. })
  33041. #endif
  33042.  
  33043. #ifdef __LITTLE_ENDIAN__
  33044. #define vld1q_bf16_x4(__p0) __extension__ ({ \
  33045.   bfloat16x8x4_t __ret; \
  33046.   __builtin_neon_vld1q_bf16_x4(&__ret, __p0, 43); \
  33047.   __ret; \
  33048. })
  33049. #else
  33050. #define vld1q_bf16_x4(__p0) __extension__ ({ \
  33051.   bfloat16x8x4_t __ret; \
  33052.   __builtin_neon_vld1q_bf16_x4(&__ret, __p0, 43); \
  33053.  \
  33054.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33055.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33056.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33057.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  33058.   __ret; \
  33059. })
  33060. #endif
  33061.  
  33062. #ifdef __LITTLE_ENDIAN__
  33063. #define vld1_bf16_x4(__p0) __extension__ ({ \
  33064.   bfloat16x4x4_t __ret; \
  33065.   __builtin_neon_vld1_bf16_x4(&__ret, __p0, 11); \
  33066.   __ret; \
  33067. })
  33068. #else
  33069. #define vld1_bf16_x4(__p0) __extension__ ({ \
  33070.   bfloat16x4x4_t __ret; \
  33071.   __builtin_neon_vld1_bf16_x4(&__ret, __p0, 11); \
  33072.  \
  33073.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33074.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33075.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  33076.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  33077.   __ret; \
  33078. })
  33079. #endif
  33080.  
  33081. #ifdef __LITTLE_ENDIAN__
  33082. #define vld2q_bf16(__p0) __extension__ ({ \
  33083.   bfloat16x8x2_t __ret; \
  33084.   __builtin_neon_vld2q_bf16(&__ret, __p0, 43); \
  33085.   __ret; \
  33086. })
  33087. #else
  33088. #define vld2q_bf16(__p0) __extension__ ({ \
  33089.   bfloat16x8x2_t __ret; \
  33090.   __builtin_neon_vld2q_bf16(&__ret, __p0, 43); \
  33091.  \
  33092.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33093.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33094.   __ret; \
  33095. })
  33096. #endif
  33097.  
  33098. #ifdef __LITTLE_ENDIAN__
  33099. #define vld2_bf16(__p0) __extension__ ({ \
  33100.   bfloat16x4x2_t __ret; \
  33101.   __builtin_neon_vld2_bf16(&__ret, __p0, 11); \
  33102.   __ret; \
  33103. })
  33104. #else
  33105. #define vld2_bf16(__p0) __extension__ ({ \
  33106.   bfloat16x4x2_t __ret; \
  33107.   __builtin_neon_vld2_bf16(&__ret, __p0, 11); \
  33108.  \
  33109.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33110.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33111.   __ret; \
  33112. })
  33113. #endif
  33114.  
  33115. #ifdef __LITTLE_ENDIAN__
  33116. #define vld2q_dup_bf16(__p0) __extension__ ({ \
  33117.   bfloat16x8x2_t __ret; \
  33118.   __builtin_neon_vld2q_dup_bf16(&__ret, __p0, 43); \
  33119.   __ret; \
  33120. })
  33121. #else
  33122. #define vld2q_dup_bf16(__p0) __extension__ ({ \
  33123.   bfloat16x8x2_t __ret; \
  33124.   __builtin_neon_vld2q_dup_bf16(&__ret, __p0, 43); \
  33125.  \
  33126.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33127.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33128.   __ret; \
  33129. })
  33130. #endif
  33131.  
  33132. #ifdef __LITTLE_ENDIAN__
  33133. #define vld2_dup_bf16(__p0) __extension__ ({ \
  33134.   bfloat16x4x2_t __ret; \
  33135.   __builtin_neon_vld2_dup_bf16(&__ret, __p0, 11); \
  33136.   __ret; \
  33137. })
  33138. #else
  33139. #define vld2_dup_bf16(__p0) __extension__ ({ \
  33140.   bfloat16x4x2_t __ret; \
  33141.   __builtin_neon_vld2_dup_bf16(&__ret, __p0, 11); \
  33142.  \
  33143.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33144.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33145.   __ret; \
  33146. })
  33147. #endif
  33148.  
  33149. #ifdef __LITTLE_ENDIAN__
  33150. #define vld2q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33151.   bfloat16x8x2_t __ret; \
  33152.   bfloat16x8x2_t __s1 = __p1; \
  33153.   __builtin_neon_vld2q_lane_bf16(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 43); \
  33154.   __ret; \
  33155. })
  33156. #else
  33157. #define vld2q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33158.   bfloat16x8x2_t __ret; \
  33159.   bfloat16x8x2_t __s1 = __p1; \
  33160.   bfloat16x8x2_t __rev1; \
  33161.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33162.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33163.   __builtin_neon_vld2q_lane_bf16(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 43); \
  33164.  \
  33165.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33166.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33167.   __ret; \
  33168. })
  33169. #endif
  33170.  
  33171. #ifdef __LITTLE_ENDIAN__
  33172. #define vld2_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33173.   bfloat16x4x2_t __ret; \
  33174.   bfloat16x4x2_t __s1 = __p1; \
  33175.   __builtin_neon_vld2_lane_bf16(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 11); \
  33176.   __ret; \
  33177. })
  33178. #else
  33179. #define vld2_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33180.   bfloat16x4x2_t __ret; \
  33181.   bfloat16x4x2_t __s1 = __p1; \
  33182.   bfloat16x4x2_t __rev1; \
  33183.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33184.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33185.   __builtin_neon_vld2_lane_bf16(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 11); \
  33186.  \
  33187.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33188.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33189.   __ret; \
  33190. })
  33191. #endif
  33192.  
  33193. #ifdef __LITTLE_ENDIAN__
  33194. #define vld3q_bf16(__p0) __extension__ ({ \
  33195.   bfloat16x8x3_t __ret; \
  33196.   __builtin_neon_vld3q_bf16(&__ret, __p0, 43); \
  33197.   __ret; \
  33198. })
  33199. #else
  33200. #define vld3q_bf16(__p0) __extension__ ({ \
  33201.   bfloat16x8x3_t __ret; \
  33202.   __builtin_neon_vld3q_bf16(&__ret, __p0, 43); \
  33203.  \
  33204.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33205.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33206.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33207.   __ret; \
  33208. })
  33209. #endif
  33210.  
  33211. #ifdef __LITTLE_ENDIAN__
  33212. #define vld3_bf16(__p0) __extension__ ({ \
  33213.   bfloat16x4x3_t __ret; \
  33214.   __builtin_neon_vld3_bf16(&__ret, __p0, 11); \
  33215.   __ret; \
  33216. })
  33217. #else
  33218. #define vld3_bf16(__p0) __extension__ ({ \
  33219.   bfloat16x4x3_t __ret; \
  33220.   __builtin_neon_vld3_bf16(&__ret, __p0, 11); \
  33221.  \
  33222.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33223.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33224.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  33225.   __ret; \
  33226. })
  33227. #endif
  33228.  
  33229. #ifdef __LITTLE_ENDIAN__
  33230. #define vld3q_dup_bf16(__p0) __extension__ ({ \
  33231.   bfloat16x8x3_t __ret; \
  33232.   __builtin_neon_vld3q_dup_bf16(&__ret, __p0, 43); \
  33233.   __ret; \
  33234. })
  33235. #else
  33236. #define vld3q_dup_bf16(__p0) __extension__ ({ \
  33237.   bfloat16x8x3_t __ret; \
  33238.   __builtin_neon_vld3q_dup_bf16(&__ret, __p0, 43); \
  33239.  \
  33240.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33241.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33242.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33243.   __ret; \
  33244. })
  33245. #endif
  33246.  
  33247. #ifdef __LITTLE_ENDIAN__
  33248. #define vld3_dup_bf16(__p0) __extension__ ({ \
  33249.   bfloat16x4x3_t __ret; \
  33250.   __builtin_neon_vld3_dup_bf16(&__ret, __p0, 11); \
  33251.   __ret; \
  33252. })
  33253. #else
  33254. #define vld3_dup_bf16(__p0) __extension__ ({ \
  33255.   bfloat16x4x3_t __ret; \
  33256.   __builtin_neon_vld3_dup_bf16(&__ret, __p0, 11); \
  33257.  \
  33258.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33259.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33260.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  33261.   __ret; \
  33262. })
  33263. #endif
  33264.  
  33265. #ifdef __LITTLE_ENDIAN__
  33266. #define vld3q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33267.   bfloat16x8x3_t __ret; \
  33268.   bfloat16x8x3_t __s1 = __p1; \
  33269.   __builtin_neon_vld3q_lane_bf16(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 43); \
  33270.   __ret; \
  33271. })
  33272. #else
  33273. #define vld3q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33274.   bfloat16x8x3_t __ret; \
  33275.   bfloat16x8x3_t __s1 = __p1; \
  33276.   bfloat16x8x3_t __rev1; \
  33277.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33278.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33279.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33280.   __builtin_neon_vld3q_lane_bf16(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 43); \
  33281.  \
  33282.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33283.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33284.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33285.   __ret; \
  33286. })
  33287. #endif
  33288.  
  33289. #ifdef __LITTLE_ENDIAN__
  33290. #define vld3_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33291.   bfloat16x4x3_t __ret; \
  33292.   bfloat16x4x3_t __s1 = __p1; \
  33293.   __builtin_neon_vld3_lane_bf16(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 11); \
  33294.   __ret; \
  33295. })
  33296. #else
  33297. #define vld3_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33298.   bfloat16x4x3_t __ret; \
  33299.   bfloat16x4x3_t __s1 = __p1; \
  33300.   bfloat16x4x3_t __rev1; \
  33301.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33302.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33303.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  33304.   __builtin_neon_vld3_lane_bf16(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 11); \
  33305.  \
  33306.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33307.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33308.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  33309.   __ret; \
  33310. })
  33311. #endif
  33312.  
  33313. #ifdef __LITTLE_ENDIAN__
  33314. #define vld4q_bf16(__p0) __extension__ ({ \
  33315.   bfloat16x8x4_t __ret; \
  33316.   __builtin_neon_vld4q_bf16(&__ret, __p0, 43); \
  33317.   __ret; \
  33318. })
  33319. #else
  33320. #define vld4q_bf16(__p0) __extension__ ({ \
  33321.   bfloat16x8x4_t __ret; \
  33322.   __builtin_neon_vld4q_bf16(&__ret, __p0, 43); \
  33323.  \
  33324.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33325.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33326.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33327.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  33328.   __ret; \
  33329. })
  33330. #endif
  33331.  
  33332. #ifdef __LITTLE_ENDIAN__
  33333. #define vld4_bf16(__p0) __extension__ ({ \
  33334.   bfloat16x4x4_t __ret; \
  33335.   __builtin_neon_vld4_bf16(&__ret, __p0, 11); \
  33336.   __ret; \
  33337. })
  33338. #else
  33339. #define vld4_bf16(__p0) __extension__ ({ \
  33340.   bfloat16x4x4_t __ret; \
  33341.   __builtin_neon_vld4_bf16(&__ret, __p0, 11); \
  33342.  \
  33343.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33344.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33345.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  33346.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  33347.   __ret; \
  33348. })
  33349. #endif
  33350.  
  33351. #ifdef __LITTLE_ENDIAN__
  33352. #define vld4q_dup_bf16(__p0) __extension__ ({ \
  33353.   bfloat16x8x4_t __ret; \
  33354.   __builtin_neon_vld4q_dup_bf16(&__ret, __p0, 43); \
  33355.   __ret; \
  33356. })
  33357. #else
  33358. #define vld4q_dup_bf16(__p0) __extension__ ({ \
  33359.   bfloat16x8x4_t __ret; \
  33360.   __builtin_neon_vld4q_dup_bf16(&__ret, __p0, 43); \
  33361.  \
  33362.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33363.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33364.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33365.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  33366.   __ret; \
  33367. })
  33368. #endif
  33369.  
  33370. #ifdef __LITTLE_ENDIAN__
  33371. #define vld4_dup_bf16(__p0) __extension__ ({ \
  33372.   bfloat16x4x4_t __ret; \
  33373.   __builtin_neon_vld4_dup_bf16(&__ret, __p0, 11); \
  33374.   __ret; \
  33375. })
  33376. #else
  33377. #define vld4_dup_bf16(__p0) __extension__ ({ \
  33378.   bfloat16x4x4_t __ret; \
  33379.   __builtin_neon_vld4_dup_bf16(&__ret, __p0, 11); \
  33380.  \
  33381.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33382.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33383.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  33384.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  33385.   __ret; \
  33386. })
  33387. #endif
  33388.  
  33389. #ifdef __LITTLE_ENDIAN__
  33390. #define vld4q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33391.   bfloat16x8x4_t __ret; \
  33392.   bfloat16x8x4_t __s1 = __p1; \
  33393.   __builtin_neon_vld4q_lane_bf16(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 43); \
  33394.   __ret; \
  33395. })
  33396. #else
  33397. #define vld4q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33398.   bfloat16x8x4_t __ret; \
  33399.   bfloat16x8x4_t __s1 = __p1; \
  33400.   bfloat16x8x4_t __rev1; \
  33401.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33402.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33403.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33404.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  33405.   __builtin_neon_vld4q_lane_bf16(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 43); \
  33406.  \
  33407.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33408.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33409.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33410.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  33411.   __ret; \
  33412. })
  33413. #endif
  33414.  
  33415. #ifdef __LITTLE_ENDIAN__
  33416. #define vld4_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33417.   bfloat16x4x4_t __ret; \
  33418.   bfloat16x4x4_t __s1 = __p1; \
  33419.   __builtin_neon_vld4_lane_bf16(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 11); \
  33420.   __ret; \
  33421. })
  33422. #else
  33423. #define vld4_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33424.   bfloat16x4x4_t __ret; \
  33425.   bfloat16x4x4_t __s1 = __p1; \
  33426.   bfloat16x4x4_t __rev1; \
  33427.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33428.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33429.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  33430.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  33431.   __builtin_neon_vld4_lane_bf16(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 11); \
  33432.  \
  33433.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  33434.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  33435.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  33436.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  33437.   __ret; \
  33438. })
  33439. #endif
  33440.  
  33441. #ifdef __LITTLE_ENDIAN__
  33442. #define vsetq_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33443.   bfloat16x8_t __ret; \
  33444.   bfloat16_t __s0 = __p0; \
  33445.   bfloat16x8_t __s1 = __p1; \
  33446.   __ret = (bfloat16x8_t) __builtin_neon_vsetq_lane_bf16(__s0, (bfloat16x8_t)__s1, __p2); \
  33447.   __ret; \
  33448. })
  33449. #else
  33450. #define vsetq_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33451.   bfloat16x8_t __ret; \
  33452.   bfloat16_t __s0 = __p0; \
  33453.   bfloat16x8_t __s1 = __p1; \
  33454.   bfloat16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  33455.   __ret = (bfloat16x8_t) __builtin_neon_vsetq_lane_bf16(__s0, (bfloat16x8_t)__rev1, __p2); \
  33456.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  33457.   __ret; \
  33458. })
  33459. #define __noswap_vsetq_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33460.   bfloat16x8_t __ret; \
  33461.   bfloat16_t __s0 = __p0; \
  33462.   bfloat16x8_t __s1 = __p1; \
  33463.   __ret = (bfloat16x8_t) __builtin_neon_vsetq_lane_bf16(__s0, (bfloat16x8_t)__s1, __p2); \
  33464.   __ret; \
  33465. })
  33466. #endif
  33467.  
  33468. #ifdef __LITTLE_ENDIAN__
  33469. #define vset_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33470.   bfloat16x4_t __ret; \
  33471.   bfloat16_t __s0 = __p0; \
  33472.   bfloat16x4_t __s1 = __p1; \
  33473.   __ret = (bfloat16x4_t) __builtin_neon_vset_lane_bf16(__s0, (bfloat16x4_t)__s1, __p2); \
  33474.   __ret; \
  33475. })
  33476. #else
  33477. #define vset_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33478.   bfloat16x4_t __ret; \
  33479.   bfloat16_t __s0 = __p0; \
  33480.   bfloat16x4_t __s1 = __p1; \
  33481.   bfloat16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  33482.   __ret = (bfloat16x4_t) __builtin_neon_vset_lane_bf16(__s0, (bfloat16x4_t)__rev1, __p2); \
  33483.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  33484.   __ret; \
  33485. })
  33486. #define __noswap_vset_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33487.   bfloat16x4_t __ret; \
  33488.   bfloat16_t __s0 = __p0; \
  33489.   bfloat16x4_t __s1 = __p1; \
  33490.   __ret = (bfloat16x4_t) __builtin_neon_vset_lane_bf16(__s0, (bfloat16x4_t)__s1, __p2); \
  33491.   __ret; \
  33492. })
  33493. #endif
  33494.  
  33495. #ifdef __LITTLE_ENDIAN__
  33496. #define vst1q_bf16(__p0, __p1) __extension__ ({ \
  33497.   bfloat16x8_t __s1 = __p1; \
  33498.   __builtin_neon_vst1q_bf16(__p0, (int8x16_t)__s1, 43); \
  33499. })
  33500. #else
  33501. #define vst1q_bf16(__p0, __p1) __extension__ ({ \
  33502.   bfloat16x8_t __s1 = __p1; \
  33503.   bfloat16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  33504.   __builtin_neon_vst1q_bf16(__p0, (int8x16_t)__rev1, 43); \
  33505. })
  33506. #endif
  33507.  
  33508. #ifdef __LITTLE_ENDIAN__
  33509. #define vst1_bf16(__p0, __p1) __extension__ ({ \
  33510.   bfloat16x4_t __s1 = __p1; \
  33511.   __builtin_neon_vst1_bf16(__p0, (int8x8_t)__s1, 11); \
  33512. })
  33513. #else
  33514. #define vst1_bf16(__p0, __p1) __extension__ ({ \
  33515.   bfloat16x4_t __s1 = __p1; \
  33516.   bfloat16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  33517.   __builtin_neon_vst1_bf16(__p0, (int8x8_t)__rev1, 11); \
  33518. })
  33519. #endif
  33520.  
  33521. #ifdef __LITTLE_ENDIAN__
  33522. #define vst1q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33523.   bfloat16x8_t __s1 = __p1; \
  33524.   __builtin_neon_vst1q_lane_bf16(__p0, (int8x16_t)__s1, __p2, 43); \
  33525. })
  33526. #else
  33527. #define vst1q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33528.   bfloat16x8_t __s1 = __p1; \
  33529.   bfloat16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  33530.   __builtin_neon_vst1q_lane_bf16(__p0, (int8x16_t)__rev1, __p2, 43); \
  33531. })
  33532. #endif
  33533.  
  33534. #ifdef __LITTLE_ENDIAN__
  33535. #define vst1_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33536.   bfloat16x4_t __s1 = __p1; \
  33537.   __builtin_neon_vst1_lane_bf16(__p0, (int8x8_t)__s1, __p2, 11); \
  33538. })
  33539. #else
  33540. #define vst1_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33541.   bfloat16x4_t __s1 = __p1; \
  33542.   bfloat16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  33543.   __builtin_neon_vst1_lane_bf16(__p0, (int8x8_t)__rev1, __p2, 11); \
  33544. })
  33545. #endif
  33546.  
  33547. #ifdef __LITTLE_ENDIAN__
  33548. #define vst1q_bf16_x2(__p0, __p1) __extension__ ({ \
  33549.   bfloat16x8x2_t __s1 = __p1; \
  33550.   __builtin_neon_vst1q_bf16_x2(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 43); \
  33551. })
  33552. #else
  33553. #define vst1q_bf16_x2(__p0, __p1) __extension__ ({ \
  33554.   bfloat16x8x2_t __s1 = __p1; \
  33555.   bfloat16x8x2_t __rev1; \
  33556.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33557.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33558.   __builtin_neon_vst1q_bf16_x2(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 43); \
  33559. })
  33560. #endif
  33561.  
  33562. #ifdef __LITTLE_ENDIAN__
  33563. #define vst1_bf16_x2(__p0, __p1) __extension__ ({ \
  33564.   bfloat16x4x2_t __s1 = __p1; \
  33565.   __builtin_neon_vst1_bf16_x2(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 11); \
  33566. })
  33567. #else
  33568. #define vst1_bf16_x2(__p0, __p1) __extension__ ({ \
  33569.   bfloat16x4x2_t __s1 = __p1; \
  33570.   bfloat16x4x2_t __rev1; \
  33571.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33572.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33573.   __builtin_neon_vst1_bf16_x2(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 11); \
  33574. })
  33575. #endif
  33576.  
  33577. #ifdef __LITTLE_ENDIAN__
  33578. #define vst1q_bf16_x3(__p0, __p1) __extension__ ({ \
  33579.   bfloat16x8x3_t __s1 = __p1; \
  33580.   __builtin_neon_vst1q_bf16_x3(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 43); \
  33581. })
  33582. #else
  33583. #define vst1q_bf16_x3(__p0, __p1) __extension__ ({ \
  33584.   bfloat16x8x3_t __s1 = __p1; \
  33585.   bfloat16x8x3_t __rev1; \
  33586.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33587.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33588.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33589.   __builtin_neon_vst1q_bf16_x3(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 43); \
  33590. })
  33591. #endif
  33592.  
  33593. #ifdef __LITTLE_ENDIAN__
  33594. #define vst1_bf16_x3(__p0, __p1) __extension__ ({ \
  33595.   bfloat16x4x3_t __s1 = __p1; \
  33596.   __builtin_neon_vst1_bf16_x3(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 11); \
  33597. })
  33598. #else
  33599. #define vst1_bf16_x3(__p0, __p1) __extension__ ({ \
  33600.   bfloat16x4x3_t __s1 = __p1; \
  33601.   bfloat16x4x3_t __rev1; \
  33602.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33603.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33604.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  33605.   __builtin_neon_vst1_bf16_x3(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 11); \
  33606. })
  33607. #endif
  33608.  
  33609. #ifdef __LITTLE_ENDIAN__
  33610. #define vst1q_bf16_x4(__p0, __p1) __extension__ ({ \
  33611.   bfloat16x8x4_t __s1 = __p1; \
  33612.   __builtin_neon_vst1q_bf16_x4(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 43); \
  33613. })
  33614. #else
  33615. #define vst1q_bf16_x4(__p0, __p1) __extension__ ({ \
  33616.   bfloat16x8x4_t __s1 = __p1; \
  33617.   bfloat16x8x4_t __rev1; \
  33618.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33619.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33620.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33621.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  33622.   __builtin_neon_vst1q_bf16_x4(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 43); \
  33623. })
  33624. #endif
  33625.  
  33626. #ifdef __LITTLE_ENDIAN__
  33627. #define vst1_bf16_x4(__p0, __p1) __extension__ ({ \
  33628.   bfloat16x4x4_t __s1 = __p1; \
  33629.   __builtin_neon_vst1_bf16_x4(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 11); \
  33630. })
  33631. #else
  33632. #define vst1_bf16_x4(__p0, __p1) __extension__ ({ \
  33633.   bfloat16x4x4_t __s1 = __p1; \
  33634.   bfloat16x4x4_t __rev1; \
  33635.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33636.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33637.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  33638.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  33639.   __builtin_neon_vst1_bf16_x4(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 11); \
  33640. })
  33641. #endif
  33642.  
  33643. #ifdef __LITTLE_ENDIAN__
  33644. #define vst2q_bf16(__p0, __p1) __extension__ ({ \
  33645.   bfloat16x8x2_t __s1 = __p1; \
  33646.   __builtin_neon_vst2q_bf16(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 43); \
  33647. })
  33648. #else
  33649. #define vst2q_bf16(__p0, __p1) __extension__ ({ \
  33650.   bfloat16x8x2_t __s1 = __p1; \
  33651.   bfloat16x8x2_t __rev1; \
  33652.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33653.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33654.   __builtin_neon_vst2q_bf16(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 43); \
  33655. })
  33656. #endif
  33657.  
  33658. #ifdef __LITTLE_ENDIAN__
  33659. #define vst2_bf16(__p0, __p1) __extension__ ({ \
  33660.   bfloat16x4x2_t __s1 = __p1; \
  33661.   __builtin_neon_vst2_bf16(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 11); \
  33662. })
  33663. #else
  33664. #define vst2_bf16(__p0, __p1) __extension__ ({ \
  33665.   bfloat16x4x2_t __s1 = __p1; \
  33666.   bfloat16x4x2_t __rev1; \
  33667.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33668.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33669.   __builtin_neon_vst2_bf16(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 11); \
  33670. })
  33671. #endif
  33672.  
  33673. #ifdef __LITTLE_ENDIAN__
  33674. #define vst2q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33675.   bfloat16x8x2_t __s1 = __p1; \
  33676.   __builtin_neon_vst2q_lane_bf16(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 43); \
  33677. })
  33678. #else
  33679. #define vst2q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33680.   bfloat16x8x2_t __s1 = __p1; \
  33681.   bfloat16x8x2_t __rev1; \
  33682.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33683.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33684.   __builtin_neon_vst2q_lane_bf16(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 43); \
  33685. })
  33686. #endif
  33687.  
  33688. #ifdef __LITTLE_ENDIAN__
  33689. #define vst2_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33690.   bfloat16x4x2_t __s1 = __p1; \
  33691.   __builtin_neon_vst2_lane_bf16(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 11); \
  33692. })
  33693. #else
  33694. #define vst2_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33695.   bfloat16x4x2_t __s1 = __p1; \
  33696.   bfloat16x4x2_t __rev1; \
  33697.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33698.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33699.   __builtin_neon_vst2_lane_bf16(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 11); \
  33700. })
  33701. #endif
  33702.  
  33703. #ifdef __LITTLE_ENDIAN__
  33704. #define vst3q_bf16(__p0, __p1) __extension__ ({ \
  33705.   bfloat16x8x3_t __s1 = __p1; \
  33706.   __builtin_neon_vst3q_bf16(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 43); \
  33707. })
  33708. #else
  33709. #define vst3q_bf16(__p0, __p1) __extension__ ({ \
  33710.   bfloat16x8x3_t __s1 = __p1; \
  33711.   bfloat16x8x3_t __rev1; \
  33712.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33713.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33714.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33715.   __builtin_neon_vst3q_bf16(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 43); \
  33716. })
  33717. #endif
  33718.  
  33719. #ifdef __LITTLE_ENDIAN__
  33720. #define vst3_bf16(__p0, __p1) __extension__ ({ \
  33721.   bfloat16x4x3_t __s1 = __p1; \
  33722.   __builtin_neon_vst3_bf16(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 11); \
  33723. })
  33724. #else
  33725. #define vst3_bf16(__p0, __p1) __extension__ ({ \
  33726.   bfloat16x4x3_t __s1 = __p1; \
  33727.   bfloat16x4x3_t __rev1; \
  33728.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33729.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33730.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  33731.   __builtin_neon_vst3_bf16(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 11); \
  33732. })
  33733. #endif
  33734.  
  33735. #ifdef __LITTLE_ENDIAN__
  33736. #define vst3q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33737.   bfloat16x8x3_t __s1 = __p1; \
  33738.   __builtin_neon_vst3q_lane_bf16(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 43); \
  33739. })
  33740. #else
  33741. #define vst3q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33742.   bfloat16x8x3_t __s1 = __p1; \
  33743.   bfloat16x8x3_t __rev1; \
  33744.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33745.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33746.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33747.   __builtin_neon_vst3q_lane_bf16(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 43); \
  33748. })
  33749. #endif
  33750.  
  33751. #ifdef __LITTLE_ENDIAN__
  33752. #define vst3_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33753.   bfloat16x4x3_t __s1 = __p1; \
  33754.   __builtin_neon_vst3_lane_bf16(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 11); \
  33755. })
  33756. #else
  33757. #define vst3_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33758.   bfloat16x4x3_t __s1 = __p1; \
  33759.   bfloat16x4x3_t __rev1; \
  33760.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33761.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33762.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  33763.   __builtin_neon_vst3_lane_bf16(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 11); \
  33764. })
  33765. #endif
  33766.  
  33767. #ifdef __LITTLE_ENDIAN__
  33768. #define vst4q_bf16(__p0, __p1) __extension__ ({ \
  33769.   bfloat16x8x4_t __s1 = __p1; \
  33770.   __builtin_neon_vst4q_bf16(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 43); \
  33771. })
  33772. #else
  33773. #define vst4q_bf16(__p0, __p1) __extension__ ({ \
  33774.   bfloat16x8x4_t __s1 = __p1; \
  33775.   bfloat16x8x4_t __rev1; \
  33776.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33777.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33778.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33779.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  33780.   __builtin_neon_vst4q_bf16(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 43); \
  33781. })
  33782. #endif
  33783.  
  33784. #ifdef __LITTLE_ENDIAN__
  33785. #define vst4_bf16(__p0, __p1) __extension__ ({ \
  33786.   bfloat16x4x4_t __s1 = __p1; \
  33787.   __builtin_neon_vst4_bf16(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 11); \
  33788. })
  33789. #else
  33790. #define vst4_bf16(__p0, __p1) __extension__ ({ \
  33791.   bfloat16x4x4_t __s1 = __p1; \
  33792.   bfloat16x4x4_t __rev1; \
  33793.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33794.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33795.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  33796.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  33797.   __builtin_neon_vst4_bf16(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 11); \
  33798. })
  33799. #endif
  33800.  
  33801. #ifdef __LITTLE_ENDIAN__
  33802. #define vst4q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33803.   bfloat16x8x4_t __s1 = __p1; \
  33804.   __builtin_neon_vst4q_lane_bf16(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 43); \
  33805. })
  33806. #else
  33807. #define vst4q_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33808.   bfloat16x8x4_t __s1 = __p1; \
  33809.   bfloat16x8x4_t __rev1; \
  33810.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  33811.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  33812.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  33813.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  33814.   __builtin_neon_vst4q_lane_bf16(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 43); \
  33815. })
  33816. #endif
  33817.  
  33818. #ifdef __LITTLE_ENDIAN__
  33819. #define vst4_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33820.   bfloat16x4x4_t __s1 = __p1; \
  33821.   __builtin_neon_vst4_lane_bf16(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 11); \
  33822. })
  33823. #else
  33824. #define vst4_lane_bf16(__p0, __p1, __p2) __extension__ ({ \
  33825.   bfloat16x4x4_t __s1 = __p1; \
  33826.   bfloat16x4x4_t __rev1; \
  33827.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  33828.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  33829.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  33830.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  33831.   __builtin_neon_vst4_lane_bf16(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 11); \
  33832. })
  33833. #endif
  33834.  
  33835. #ifdef __LITTLE_ENDIAN__
  33836. __ai __attribute__((target("dotprod"))) uint32x4_t vdotq_u32(uint32x4_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  33837.   uint32x4_t __ret;
  33838.   __ret = (uint32x4_t) __builtin_neon_vdotq_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  33839.   return __ret;
  33840. }
  33841. #else
  33842. __ai __attribute__((target("dotprod"))) uint32x4_t vdotq_u32(uint32x4_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  33843.   uint32x4_t __ret;
  33844.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  33845.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  33846.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  33847.   __ret = (uint32x4_t) __builtin_neon_vdotq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  33848.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  33849.   return __ret;
  33850. }
  33851. __ai __attribute__((target("dotprod"))) uint32x4_t __noswap_vdotq_u32(uint32x4_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  33852.   uint32x4_t __ret;
  33853.   __ret = (uint32x4_t) __builtin_neon_vdotq_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  33854.   return __ret;
  33855. }
  33856. #endif
  33857.  
  33858. #ifdef __LITTLE_ENDIAN__
  33859. __ai __attribute__((target("dotprod"))) int32x4_t vdotq_s32(int32x4_t __p0, int8x16_t __p1, int8x16_t __p2) {
  33860.   int32x4_t __ret;
  33861.   __ret = (int32x4_t) __builtin_neon_vdotq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  33862.   return __ret;
  33863. }
  33864. #else
  33865. __ai __attribute__((target("dotprod"))) int32x4_t vdotq_s32(int32x4_t __p0, int8x16_t __p1, int8x16_t __p2) {
  33866.   int32x4_t __ret;
  33867.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  33868.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  33869.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  33870.   __ret = (int32x4_t) __builtin_neon_vdotq_s32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  33871.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  33872.   return __ret;
  33873. }
  33874. __ai __attribute__((target("dotprod"))) int32x4_t __noswap_vdotq_s32(int32x4_t __p0, int8x16_t __p1, int8x16_t __p2) {
  33875.   int32x4_t __ret;
  33876.   __ret = (int32x4_t) __builtin_neon_vdotq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  33877.   return __ret;
  33878. }
  33879. #endif
  33880.  
  33881. #ifdef __LITTLE_ENDIAN__
  33882. __ai __attribute__((target("dotprod"))) uint32x2_t vdot_u32(uint32x2_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  33883.   uint32x2_t __ret;
  33884.   __ret = (uint32x2_t) __builtin_neon_vdot_u32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 18);
  33885.   return __ret;
  33886. }
  33887. #else
  33888. __ai __attribute__((target("dotprod"))) uint32x2_t vdot_u32(uint32x2_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  33889.   uint32x2_t __ret;
  33890.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  33891.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  33892.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  33893.   __ret = (uint32x2_t) __builtin_neon_vdot_u32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 18);
  33894.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  33895.   return __ret;
  33896. }
  33897. __ai __attribute__((target("dotprod"))) uint32x2_t __noswap_vdot_u32(uint32x2_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  33898.   uint32x2_t __ret;
  33899.   __ret = (uint32x2_t) __builtin_neon_vdot_u32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 18);
  33900.   return __ret;
  33901. }
  33902. #endif
  33903.  
  33904. #ifdef __LITTLE_ENDIAN__
  33905. __ai __attribute__((target("dotprod"))) int32x2_t vdot_s32(int32x2_t __p0, int8x8_t __p1, int8x8_t __p2) {
  33906.   int32x2_t __ret;
  33907.   __ret = (int32x2_t) __builtin_neon_vdot_s32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  33908.   return __ret;
  33909. }
  33910. #else
  33911. __ai __attribute__((target("dotprod"))) int32x2_t vdot_s32(int32x2_t __p0, int8x8_t __p1, int8x8_t __p2) {
  33912.   int32x2_t __ret;
  33913.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  33914.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  33915.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  33916.   __ret = (int32x2_t) __builtin_neon_vdot_s32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 2);
  33917.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  33918.   return __ret;
  33919. }
  33920. __ai __attribute__((target("dotprod"))) int32x2_t __noswap_vdot_s32(int32x2_t __p0, int8x8_t __p1, int8x8_t __p2) {
  33921.   int32x2_t __ret;
  33922.   __ret = (int32x2_t) __builtin_neon_vdot_s32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  33923.   return __ret;
  33924. }
  33925. #endif
  33926.  
  33927. #ifdef __LITTLE_ENDIAN__
  33928. #define vdotq_lane_u32(__p0_145, __p1_145, __p2_145, __p3_145) __extension__ ({ \
  33929.   uint32x4_t __ret_145; \
  33930.   uint32x4_t __s0_145 = __p0_145; \
  33931.   uint8x16_t __s1_145 = __p1_145; \
  33932.   uint8x8_t __s2_145 = __p2_145; \
  33933. uint8x8_t __reint_145 = __s2_145; \
  33934. uint32x4_t __reint1_145 = splatq_lane_u32(*(uint32x2_t *) &__reint_145, __p3_145); \
  33935.   __ret_145 = vdotq_u32(__s0_145, __s1_145, *(uint8x16_t *) &__reint1_145); \
  33936.   __ret_145; \
  33937. })
  33938. #else
  33939. #define vdotq_lane_u32(__p0_146, __p1_146, __p2_146, __p3_146) __extension__ ({ \
  33940.   uint32x4_t __ret_146; \
  33941.   uint32x4_t __s0_146 = __p0_146; \
  33942.   uint8x16_t __s1_146 = __p1_146; \
  33943.   uint8x8_t __s2_146 = __p2_146; \
  33944.   uint32x4_t __rev0_146;  __rev0_146 = __builtin_shufflevector(__s0_146, __s0_146, 3, 2, 1, 0); \
  33945.   uint8x16_t __rev1_146;  __rev1_146 = __builtin_shufflevector(__s1_146, __s1_146, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  33946.   uint8x8_t __rev2_146;  __rev2_146 = __builtin_shufflevector(__s2_146, __s2_146, 7, 6, 5, 4, 3, 2, 1, 0); \
  33947. uint8x8_t __reint_146 = __rev2_146; \
  33948. uint32x4_t __reint1_146 = __noswap_splatq_lane_u32(*(uint32x2_t *) &__reint_146, __p3_146); \
  33949.   __ret_146 = __noswap_vdotq_u32(__rev0_146, __rev1_146, *(uint8x16_t *) &__reint1_146); \
  33950.   __ret_146 = __builtin_shufflevector(__ret_146, __ret_146, 3, 2, 1, 0); \
  33951.   __ret_146; \
  33952. })
  33953. #endif
  33954.  
  33955. #ifdef __LITTLE_ENDIAN__
  33956. #define vdotq_lane_s32(__p0_147, __p1_147, __p2_147, __p3_147) __extension__ ({ \
  33957.   int32x4_t __ret_147; \
  33958.   int32x4_t __s0_147 = __p0_147; \
  33959.   int8x16_t __s1_147 = __p1_147; \
  33960.   int8x8_t __s2_147 = __p2_147; \
  33961. int8x8_t __reint_147 = __s2_147; \
  33962. int32x4_t __reint1_147 = splatq_lane_s32(*(int32x2_t *) &__reint_147, __p3_147); \
  33963.   __ret_147 = vdotq_s32(__s0_147, __s1_147, *(int8x16_t *) &__reint1_147); \
  33964.   __ret_147; \
  33965. })
  33966. #else
  33967. #define vdotq_lane_s32(__p0_148, __p1_148, __p2_148, __p3_148) __extension__ ({ \
  33968.   int32x4_t __ret_148; \
  33969.   int32x4_t __s0_148 = __p0_148; \
  33970.   int8x16_t __s1_148 = __p1_148; \
  33971.   int8x8_t __s2_148 = __p2_148; \
  33972.   int32x4_t __rev0_148;  __rev0_148 = __builtin_shufflevector(__s0_148, __s0_148, 3, 2, 1, 0); \
  33973.   int8x16_t __rev1_148;  __rev1_148 = __builtin_shufflevector(__s1_148, __s1_148, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  33974.   int8x8_t __rev2_148;  __rev2_148 = __builtin_shufflevector(__s2_148, __s2_148, 7, 6, 5, 4, 3, 2, 1, 0); \
  33975. int8x8_t __reint_148 = __rev2_148; \
  33976. int32x4_t __reint1_148 = __noswap_splatq_lane_s32(*(int32x2_t *) &__reint_148, __p3_148); \
  33977.   __ret_148 = __noswap_vdotq_s32(__rev0_148, __rev1_148, *(int8x16_t *) &__reint1_148); \
  33978.   __ret_148 = __builtin_shufflevector(__ret_148, __ret_148, 3, 2, 1, 0); \
  33979.   __ret_148; \
  33980. })
  33981. #endif
  33982.  
  33983. #ifdef __LITTLE_ENDIAN__
  33984. #define vdot_lane_u32(__p0_149, __p1_149, __p2_149, __p3_149) __extension__ ({ \
  33985.   uint32x2_t __ret_149; \
  33986.   uint32x2_t __s0_149 = __p0_149; \
  33987.   uint8x8_t __s1_149 = __p1_149; \
  33988.   uint8x8_t __s2_149 = __p2_149; \
  33989. uint8x8_t __reint_149 = __s2_149; \
  33990. uint32x2_t __reint1_149 = splat_lane_u32(*(uint32x2_t *) &__reint_149, __p3_149); \
  33991.   __ret_149 = vdot_u32(__s0_149, __s1_149, *(uint8x8_t *) &__reint1_149); \
  33992.   __ret_149; \
  33993. })
  33994. #else
  33995. #define vdot_lane_u32(__p0_150, __p1_150, __p2_150, __p3_150) __extension__ ({ \
  33996.   uint32x2_t __ret_150; \
  33997.   uint32x2_t __s0_150 = __p0_150; \
  33998.   uint8x8_t __s1_150 = __p1_150; \
  33999.   uint8x8_t __s2_150 = __p2_150; \
  34000.   uint32x2_t __rev0_150;  __rev0_150 = __builtin_shufflevector(__s0_150, __s0_150, 1, 0); \
  34001.   uint8x8_t __rev1_150;  __rev1_150 = __builtin_shufflevector(__s1_150, __s1_150, 7, 6, 5, 4, 3, 2, 1, 0); \
  34002.   uint8x8_t __rev2_150;  __rev2_150 = __builtin_shufflevector(__s2_150, __s2_150, 7, 6, 5, 4, 3, 2, 1, 0); \
  34003. uint8x8_t __reint_150 = __rev2_150; \
  34004. uint32x2_t __reint1_150 = __noswap_splat_lane_u32(*(uint32x2_t *) &__reint_150, __p3_150); \
  34005.   __ret_150 = __noswap_vdot_u32(__rev0_150, __rev1_150, *(uint8x8_t *) &__reint1_150); \
  34006.   __ret_150 = __builtin_shufflevector(__ret_150, __ret_150, 1, 0); \
  34007.   __ret_150; \
  34008. })
  34009. #endif
  34010.  
  34011. #ifdef __LITTLE_ENDIAN__
  34012. #define vdot_lane_s32(__p0_151, __p1_151, __p2_151, __p3_151) __extension__ ({ \
  34013.   int32x2_t __ret_151; \
  34014.   int32x2_t __s0_151 = __p0_151; \
  34015.   int8x8_t __s1_151 = __p1_151; \
  34016.   int8x8_t __s2_151 = __p2_151; \
  34017. int8x8_t __reint_151 = __s2_151; \
  34018. int32x2_t __reint1_151 = splat_lane_s32(*(int32x2_t *) &__reint_151, __p3_151); \
  34019.   __ret_151 = vdot_s32(__s0_151, __s1_151, *(int8x8_t *) &__reint1_151); \
  34020.   __ret_151; \
  34021. })
  34022. #else
  34023. #define vdot_lane_s32(__p0_152, __p1_152, __p2_152, __p3_152) __extension__ ({ \
  34024.   int32x2_t __ret_152; \
  34025.   int32x2_t __s0_152 = __p0_152; \
  34026.   int8x8_t __s1_152 = __p1_152; \
  34027.   int8x8_t __s2_152 = __p2_152; \
  34028.   int32x2_t __rev0_152;  __rev0_152 = __builtin_shufflevector(__s0_152, __s0_152, 1, 0); \
  34029.   int8x8_t __rev1_152;  __rev1_152 = __builtin_shufflevector(__s1_152, __s1_152, 7, 6, 5, 4, 3, 2, 1, 0); \
  34030.   int8x8_t __rev2_152;  __rev2_152 = __builtin_shufflevector(__s2_152, __s2_152, 7, 6, 5, 4, 3, 2, 1, 0); \
  34031. int8x8_t __reint_152 = __rev2_152; \
  34032. int32x2_t __reint1_152 = __noswap_splat_lane_s32(*(int32x2_t *) &__reint_152, __p3_152); \
  34033.   __ret_152 = __noswap_vdot_s32(__rev0_152, __rev1_152, *(int8x8_t *) &__reint1_152); \
  34034.   __ret_152 = __builtin_shufflevector(__ret_152, __ret_152, 1, 0); \
  34035.   __ret_152; \
  34036. })
  34037. #endif
  34038.  
  34039. #ifdef __LITTLE_ENDIAN__
  34040. __ai __attribute__((target("fullfp16"))) float16x8_t vabdq_f16(float16x8_t __p0, float16x8_t __p1) {
  34041.   float16x8_t __ret;
  34042.   __ret = (float16x8_t) __builtin_neon_vabdq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  34043.   return __ret;
  34044. }
  34045. #else
  34046. __ai __attribute__((target("fullfp16"))) float16x8_t vabdq_f16(float16x8_t __p0, float16x8_t __p1) {
  34047.   float16x8_t __ret;
  34048.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34049.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34050.   __ret = (float16x8_t) __builtin_neon_vabdq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  34051.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34052.   return __ret;
  34053. }
  34054. #endif
  34055.  
  34056. #ifdef __LITTLE_ENDIAN__
  34057. __ai __attribute__((target("fullfp16"))) float16x4_t vabd_f16(float16x4_t __p0, float16x4_t __p1) {
  34058.   float16x4_t __ret;
  34059.   __ret = (float16x4_t) __builtin_neon_vabd_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  34060.   return __ret;
  34061. }
  34062. #else
  34063. __ai __attribute__((target("fullfp16"))) float16x4_t vabd_f16(float16x4_t __p0, float16x4_t __p1) {
  34064.   float16x4_t __ret;
  34065.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34066.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34067.   __ret = (float16x4_t) __builtin_neon_vabd_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  34068.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34069.   return __ret;
  34070. }
  34071. #endif
  34072.  
  34073. #ifdef __LITTLE_ENDIAN__
  34074. __ai __attribute__((target("fullfp16"))) float16x8_t vabsq_f16(float16x8_t __p0) {
  34075.   float16x8_t __ret;
  34076.   __ret = (float16x8_t) __builtin_neon_vabsq_f16((int8x16_t)__p0, 40);
  34077.   return __ret;
  34078. }
  34079. #else
  34080. __ai __attribute__((target("fullfp16"))) float16x8_t vabsq_f16(float16x8_t __p0) {
  34081.   float16x8_t __ret;
  34082.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34083.   __ret = (float16x8_t) __builtin_neon_vabsq_f16((int8x16_t)__rev0, 40);
  34084.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34085.   return __ret;
  34086. }
  34087. #endif
  34088.  
  34089. #ifdef __LITTLE_ENDIAN__
  34090. __ai __attribute__((target("fullfp16"))) float16x4_t vabs_f16(float16x4_t __p0) {
  34091.   float16x4_t __ret;
  34092.   __ret = (float16x4_t) __builtin_neon_vabs_f16((int8x8_t)__p0, 8);
  34093.   return __ret;
  34094. }
  34095. #else
  34096. __ai __attribute__((target("fullfp16"))) float16x4_t vabs_f16(float16x4_t __p0) {
  34097.   float16x4_t __ret;
  34098.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34099.   __ret = (float16x4_t) __builtin_neon_vabs_f16((int8x8_t)__rev0, 8);
  34100.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34101.   return __ret;
  34102. }
  34103. #endif
  34104.  
  34105. #ifdef __LITTLE_ENDIAN__
  34106. __ai __attribute__((target("fullfp16"))) float16x8_t vaddq_f16(float16x8_t __p0, float16x8_t __p1) {
  34107.   float16x8_t __ret;
  34108.   __ret = __p0 + __p1;
  34109.   return __ret;
  34110. }
  34111. #else
  34112. __ai __attribute__((target("fullfp16"))) float16x8_t vaddq_f16(float16x8_t __p0, float16x8_t __p1) {
  34113.   float16x8_t __ret;
  34114.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34115.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34116.   __ret = __rev0 + __rev1;
  34117.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34118.   return __ret;
  34119. }
  34120. #endif
  34121.  
  34122. #ifdef __LITTLE_ENDIAN__
  34123. __ai __attribute__((target("fullfp16"))) float16x4_t vadd_f16(float16x4_t __p0, float16x4_t __p1) {
  34124.   float16x4_t __ret;
  34125.   __ret = __p0 + __p1;
  34126.   return __ret;
  34127. }
  34128. #else
  34129. __ai __attribute__((target("fullfp16"))) float16x4_t vadd_f16(float16x4_t __p0, float16x4_t __p1) {
  34130.   float16x4_t __ret;
  34131.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34132.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34133.   __ret = __rev0 + __rev1;
  34134.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34135.   return __ret;
  34136. }
  34137. #endif
  34138.  
  34139. #ifdef __LITTLE_ENDIAN__
  34140. __ai __attribute__((target("fullfp16"))) float16x8_t vbslq_f16(uint16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  34141.   float16x8_t __ret;
  34142.   __ret = (float16x8_t) __builtin_neon_vbslq_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  34143.   return __ret;
  34144. }
  34145. #else
  34146. __ai __attribute__((target("fullfp16"))) float16x8_t vbslq_f16(uint16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  34147.   float16x8_t __ret;
  34148.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34149.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34150.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  34151.   __ret = (float16x8_t) __builtin_neon_vbslq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40);
  34152.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34153.   return __ret;
  34154. }
  34155. #endif
  34156.  
  34157. #ifdef __LITTLE_ENDIAN__
  34158. __ai __attribute__((target("fullfp16"))) float16x4_t vbsl_f16(uint16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  34159.   float16x4_t __ret;
  34160.   __ret = (float16x4_t) __builtin_neon_vbsl_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  34161.   return __ret;
  34162. }
  34163. #else
  34164. __ai __attribute__((target("fullfp16"))) float16x4_t vbsl_f16(uint16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  34165.   float16x4_t __ret;
  34166.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34167.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34168.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  34169.   __ret = (float16x4_t) __builtin_neon_vbsl_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8);
  34170.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34171.   return __ret;
  34172. }
  34173. #endif
  34174.  
  34175. #ifdef __LITTLE_ENDIAN__
  34176. __ai __attribute__((target("fullfp16"))) uint16x8_t vcageq_f16(float16x8_t __p0, float16x8_t __p1) {
  34177.   uint16x8_t __ret;
  34178.   __ret = (uint16x8_t) __builtin_neon_vcageq_f16((int8x16_t)__p0, (int8x16_t)__p1, 49);
  34179.   return __ret;
  34180. }
  34181. #else
  34182. __ai __attribute__((target("fullfp16"))) uint16x8_t vcageq_f16(float16x8_t __p0, float16x8_t __p1) {
  34183.   uint16x8_t __ret;
  34184.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34185.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34186.   __ret = (uint16x8_t) __builtin_neon_vcageq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  34187.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34188.   return __ret;
  34189. }
  34190. #endif
  34191.  
  34192. #ifdef __LITTLE_ENDIAN__
  34193. __ai __attribute__((target("fullfp16"))) uint16x4_t vcage_f16(float16x4_t __p0, float16x4_t __p1) {
  34194.   uint16x4_t __ret;
  34195.   __ret = (uint16x4_t) __builtin_neon_vcage_f16((int8x8_t)__p0, (int8x8_t)__p1, 17);
  34196.   return __ret;
  34197. }
  34198. #else
  34199. __ai __attribute__((target("fullfp16"))) uint16x4_t vcage_f16(float16x4_t __p0, float16x4_t __p1) {
  34200.   uint16x4_t __ret;
  34201.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34202.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34203.   __ret = (uint16x4_t) __builtin_neon_vcage_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  34204.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34205.   return __ret;
  34206. }
  34207. #endif
  34208.  
  34209. #ifdef __LITTLE_ENDIAN__
  34210. __ai __attribute__((target("fullfp16"))) uint16x8_t vcagtq_f16(float16x8_t __p0, float16x8_t __p1) {
  34211.   uint16x8_t __ret;
  34212.   __ret = (uint16x8_t) __builtin_neon_vcagtq_f16((int8x16_t)__p0, (int8x16_t)__p1, 49);
  34213.   return __ret;
  34214. }
  34215. #else
  34216. __ai __attribute__((target("fullfp16"))) uint16x8_t vcagtq_f16(float16x8_t __p0, float16x8_t __p1) {
  34217.   uint16x8_t __ret;
  34218.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34219.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34220.   __ret = (uint16x8_t) __builtin_neon_vcagtq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  34221.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34222.   return __ret;
  34223. }
  34224. #endif
  34225.  
  34226. #ifdef __LITTLE_ENDIAN__
  34227. __ai __attribute__((target("fullfp16"))) uint16x4_t vcagt_f16(float16x4_t __p0, float16x4_t __p1) {
  34228.   uint16x4_t __ret;
  34229.   __ret = (uint16x4_t) __builtin_neon_vcagt_f16((int8x8_t)__p0, (int8x8_t)__p1, 17);
  34230.   return __ret;
  34231. }
  34232. #else
  34233. __ai __attribute__((target("fullfp16"))) uint16x4_t vcagt_f16(float16x4_t __p0, float16x4_t __p1) {
  34234.   uint16x4_t __ret;
  34235.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34236.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34237.   __ret = (uint16x4_t) __builtin_neon_vcagt_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  34238.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34239.   return __ret;
  34240. }
  34241. #endif
  34242.  
  34243. #ifdef __LITTLE_ENDIAN__
  34244. __ai __attribute__((target("fullfp16"))) uint16x8_t vcaleq_f16(float16x8_t __p0, float16x8_t __p1) {
  34245.   uint16x8_t __ret;
  34246.   __ret = (uint16x8_t) __builtin_neon_vcaleq_f16((int8x16_t)__p0, (int8x16_t)__p1, 49);
  34247.   return __ret;
  34248. }
  34249. #else
  34250. __ai __attribute__((target("fullfp16"))) uint16x8_t vcaleq_f16(float16x8_t __p0, float16x8_t __p1) {
  34251.   uint16x8_t __ret;
  34252.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34253.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34254.   __ret = (uint16x8_t) __builtin_neon_vcaleq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  34255.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34256.   return __ret;
  34257. }
  34258. #endif
  34259.  
  34260. #ifdef __LITTLE_ENDIAN__
  34261. __ai __attribute__((target("fullfp16"))) uint16x4_t vcale_f16(float16x4_t __p0, float16x4_t __p1) {
  34262.   uint16x4_t __ret;
  34263.   __ret = (uint16x4_t) __builtin_neon_vcale_f16((int8x8_t)__p0, (int8x8_t)__p1, 17);
  34264.   return __ret;
  34265. }
  34266. #else
  34267. __ai __attribute__((target("fullfp16"))) uint16x4_t vcale_f16(float16x4_t __p0, float16x4_t __p1) {
  34268.   uint16x4_t __ret;
  34269.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34270.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34271.   __ret = (uint16x4_t) __builtin_neon_vcale_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  34272.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34273.   return __ret;
  34274. }
  34275. #endif
  34276.  
  34277. #ifdef __LITTLE_ENDIAN__
  34278. __ai __attribute__((target("fullfp16"))) uint16x8_t vcaltq_f16(float16x8_t __p0, float16x8_t __p1) {
  34279.   uint16x8_t __ret;
  34280.   __ret = (uint16x8_t) __builtin_neon_vcaltq_f16((int8x16_t)__p0, (int8x16_t)__p1, 49);
  34281.   return __ret;
  34282. }
  34283. #else
  34284. __ai __attribute__((target("fullfp16"))) uint16x8_t vcaltq_f16(float16x8_t __p0, float16x8_t __p1) {
  34285.   uint16x8_t __ret;
  34286.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34287.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34288.   __ret = (uint16x8_t) __builtin_neon_vcaltq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  34289.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34290.   return __ret;
  34291. }
  34292. #endif
  34293.  
  34294. #ifdef __LITTLE_ENDIAN__
  34295. __ai __attribute__((target("fullfp16"))) uint16x4_t vcalt_f16(float16x4_t __p0, float16x4_t __p1) {
  34296.   uint16x4_t __ret;
  34297.   __ret = (uint16x4_t) __builtin_neon_vcalt_f16((int8x8_t)__p0, (int8x8_t)__p1, 17);
  34298.   return __ret;
  34299. }
  34300. #else
  34301. __ai __attribute__((target("fullfp16"))) uint16x4_t vcalt_f16(float16x4_t __p0, float16x4_t __p1) {
  34302.   uint16x4_t __ret;
  34303.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34304.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34305.   __ret = (uint16x4_t) __builtin_neon_vcalt_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  34306.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34307.   return __ret;
  34308. }
  34309. #endif
  34310.  
  34311. #ifdef __LITTLE_ENDIAN__
  34312. __ai __attribute__((target("fullfp16"))) uint16x8_t vceqq_f16(float16x8_t __p0, float16x8_t __p1) {
  34313.   uint16x8_t __ret;
  34314.   __ret = (uint16x8_t)(__p0 == __p1);
  34315.   return __ret;
  34316. }
  34317. #else
  34318. __ai __attribute__((target("fullfp16"))) uint16x8_t vceqq_f16(float16x8_t __p0, float16x8_t __p1) {
  34319.   uint16x8_t __ret;
  34320.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34321.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34322.   __ret = (uint16x8_t)(__rev0 == __rev1);
  34323.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34324.   return __ret;
  34325. }
  34326. #endif
  34327.  
  34328. #ifdef __LITTLE_ENDIAN__
  34329. __ai __attribute__((target("fullfp16"))) uint16x4_t vceq_f16(float16x4_t __p0, float16x4_t __p1) {
  34330.   uint16x4_t __ret;
  34331.   __ret = (uint16x4_t)(__p0 == __p1);
  34332.   return __ret;
  34333. }
  34334. #else
  34335. __ai __attribute__((target("fullfp16"))) uint16x4_t vceq_f16(float16x4_t __p0, float16x4_t __p1) {
  34336.   uint16x4_t __ret;
  34337.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34338.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34339.   __ret = (uint16x4_t)(__rev0 == __rev1);
  34340.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34341.   return __ret;
  34342. }
  34343. #endif
  34344.  
  34345. #ifdef __LITTLE_ENDIAN__
  34346. __ai __attribute__((target("fullfp16"))) uint16x8_t vceqzq_f16(float16x8_t __p0) {
  34347.   uint16x8_t __ret;
  34348.   __ret = (uint16x8_t) __builtin_neon_vceqzq_f16((int8x16_t)__p0, 49);
  34349.   return __ret;
  34350. }
  34351. #else
  34352. __ai __attribute__((target("fullfp16"))) uint16x8_t vceqzq_f16(float16x8_t __p0) {
  34353.   uint16x8_t __ret;
  34354.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34355.   __ret = (uint16x8_t) __builtin_neon_vceqzq_f16((int8x16_t)__rev0, 49);
  34356.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34357.   return __ret;
  34358. }
  34359. #endif
  34360.  
  34361. #ifdef __LITTLE_ENDIAN__
  34362. __ai __attribute__((target("fullfp16"))) uint16x4_t vceqz_f16(float16x4_t __p0) {
  34363.   uint16x4_t __ret;
  34364.   __ret = (uint16x4_t) __builtin_neon_vceqz_f16((int8x8_t)__p0, 17);
  34365.   return __ret;
  34366. }
  34367. #else
  34368. __ai __attribute__((target("fullfp16"))) uint16x4_t vceqz_f16(float16x4_t __p0) {
  34369.   uint16x4_t __ret;
  34370.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34371.   __ret = (uint16x4_t) __builtin_neon_vceqz_f16((int8x8_t)__rev0, 17);
  34372.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34373.   return __ret;
  34374. }
  34375. #endif
  34376.  
  34377. #ifdef __LITTLE_ENDIAN__
  34378. __ai __attribute__((target("fullfp16"))) uint16x8_t vcgeq_f16(float16x8_t __p0, float16x8_t __p1) {
  34379.   uint16x8_t __ret;
  34380.   __ret = (uint16x8_t)(__p0 >= __p1);
  34381.   return __ret;
  34382. }
  34383. #else
  34384. __ai __attribute__((target("fullfp16"))) uint16x8_t vcgeq_f16(float16x8_t __p0, float16x8_t __p1) {
  34385.   uint16x8_t __ret;
  34386.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34387.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34388.   __ret = (uint16x8_t)(__rev0 >= __rev1);
  34389.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34390.   return __ret;
  34391. }
  34392. #endif
  34393.  
  34394. #ifdef __LITTLE_ENDIAN__
  34395. __ai __attribute__((target("fullfp16"))) uint16x4_t vcge_f16(float16x4_t __p0, float16x4_t __p1) {
  34396.   uint16x4_t __ret;
  34397.   __ret = (uint16x4_t)(__p0 >= __p1);
  34398.   return __ret;
  34399. }
  34400. #else
  34401. __ai __attribute__((target("fullfp16"))) uint16x4_t vcge_f16(float16x4_t __p0, float16x4_t __p1) {
  34402.   uint16x4_t __ret;
  34403.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34404.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34405.   __ret = (uint16x4_t)(__rev0 >= __rev1);
  34406.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34407.   return __ret;
  34408. }
  34409. #endif
  34410.  
  34411. #ifdef __LITTLE_ENDIAN__
  34412. __ai __attribute__((target("fullfp16"))) uint16x8_t vcgezq_f16(float16x8_t __p0) {
  34413.   uint16x8_t __ret;
  34414.   __ret = (uint16x8_t) __builtin_neon_vcgezq_f16((int8x16_t)__p0, 49);
  34415.   return __ret;
  34416. }
  34417. #else
  34418. __ai __attribute__((target("fullfp16"))) uint16x8_t vcgezq_f16(float16x8_t __p0) {
  34419.   uint16x8_t __ret;
  34420.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34421.   __ret = (uint16x8_t) __builtin_neon_vcgezq_f16((int8x16_t)__rev0, 49);
  34422.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34423.   return __ret;
  34424. }
  34425. #endif
  34426.  
  34427. #ifdef __LITTLE_ENDIAN__
  34428. __ai __attribute__((target("fullfp16"))) uint16x4_t vcgez_f16(float16x4_t __p0) {
  34429.   uint16x4_t __ret;
  34430.   __ret = (uint16x4_t) __builtin_neon_vcgez_f16((int8x8_t)__p0, 17);
  34431.   return __ret;
  34432. }
  34433. #else
  34434. __ai __attribute__((target("fullfp16"))) uint16x4_t vcgez_f16(float16x4_t __p0) {
  34435.   uint16x4_t __ret;
  34436.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34437.   __ret = (uint16x4_t) __builtin_neon_vcgez_f16((int8x8_t)__rev0, 17);
  34438.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34439.   return __ret;
  34440. }
  34441. #endif
  34442.  
  34443. #ifdef __LITTLE_ENDIAN__
  34444. __ai __attribute__((target("fullfp16"))) uint16x8_t vcgtq_f16(float16x8_t __p0, float16x8_t __p1) {
  34445.   uint16x8_t __ret;
  34446.   __ret = (uint16x8_t)(__p0 > __p1);
  34447.   return __ret;
  34448. }
  34449. #else
  34450. __ai __attribute__((target("fullfp16"))) uint16x8_t vcgtq_f16(float16x8_t __p0, float16x8_t __p1) {
  34451.   uint16x8_t __ret;
  34452.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34453.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34454.   __ret = (uint16x8_t)(__rev0 > __rev1);
  34455.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34456.   return __ret;
  34457. }
  34458. #endif
  34459.  
  34460. #ifdef __LITTLE_ENDIAN__
  34461. __ai __attribute__((target("fullfp16"))) uint16x4_t vcgt_f16(float16x4_t __p0, float16x4_t __p1) {
  34462.   uint16x4_t __ret;
  34463.   __ret = (uint16x4_t)(__p0 > __p1);
  34464.   return __ret;
  34465. }
  34466. #else
  34467. __ai __attribute__((target("fullfp16"))) uint16x4_t vcgt_f16(float16x4_t __p0, float16x4_t __p1) {
  34468.   uint16x4_t __ret;
  34469.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34470.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34471.   __ret = (uint16x4_t)(__rev0 > __rev1);
  34472.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34473.   return __ret;
  34474. }
  34475. #endif
  34476.  
  34477. #ifdef __LITTLE_ENDIAN__
  34478. __ai __attribute__((target("fullfp16"))) uint16x8_t vcgtzq_f16(float16x8_t __p0) {
  34479.   uint16x8_t __ret;
  34480.   __ret = (uint16x8_t) __builtin_neon_vcgtzq_f16((int8x16_t)__p0, 49);
  34481.   return __ret;
  34482. }
  34483. #else
  34484. __ai __attribute__((target("fullfp16"))) uint16x8_t vcgtzq_f16(float16x8_t __p0) {
  34485.   uint16x8_t __ret;
  34486.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34487.   __ret = (uint16x8_t) __builtin_neon_vcgtzq_f16((int8x16_t)__rev0, 49);
  34488.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34489.   return __ret;
  34490. }
  34491. #endif
  34492.  
  34493. #ifdef __LITTLE_ENDIAN__
  34494. __ai __attribute__((target("fullfp16"))) uint16x4_t vcgtz_f16(float16x4_t __p0) {
  34495.   uint16x4_t __ret;
  34496.   __ret = (uint16x4_t) __builtin_neon_vcgtz_f16((int8x8_t)__p0, 17);
  34497.   return __ret;
  34498. }
  34499. #else
  34500. __ai __attribute__((target("fullfp16"))) uint16x4_t vcgtz_f16(float16x4_t __p0) {
  34501.   uint16x4_t __ret;
  34502.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34503.   __ret = (uint16x4_t) __builtin_neon_vcgtz_f16((int8x8_t)__rev0, 17);
  34504.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34505.   return __ret;
  34506. }
  34507. #endif
  34508.  
  34509. #ifdef __LITTLE_ENDIAN__
  34510. __ai __attribute__((target("fullfp16"))) uint16x8_t vcleq_f16(float16x8_t __p0, float16x8_t __p1) {
  34511.   uint16x8_t __ret;
  34512.   __ret = (uint16x8_t)(__p0 <= __p1);
  34513.   return __ret;
  34514. }
  34515. #else
  34516. __ai __attribute__((target("fullfp16"))) uint16x8_t vcleq_f16(float16x8_t __p0, float16x8_t __p1) {
  34517.   uint16x8_t __ret;
  34518.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34519.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34520.   __ret = (uint16x8_t)(__rev0 <= __rev1);
  34521.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34522.   return __ret;
  34523. }
  34524. #endif
  34525.  
  34526. #ifdef __LITTLE_ENDIAN__
  34527. __ai __attribute__((target("fullfp16"))) uint16x4_t vcle_f16(float16x4_t __p0, float16x4_t __p1) {
  34528.   uint16x4_t __ret;
  34529.   __ret = (uint16x4_t)(__p0 <= __p1);
  34530.   return __ret;
  34531. }
  34532. #else
  34533. __ai __attribute__((target("fullfp16"))) uint16x4_t vcle_f16(float16x4_t __p0, float16x4_t __p1) {
  34534.   uint16x4_t __ret;
  34535.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34536.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34537.   __ret = (uint16x4_t)(__rev0 <= __rev1);
  34538.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34539.   return __ret;
  34540. }
  34541. #endif
  34542.  
  34543. #ifdef __LITTLE_ENDIAN__
  34544. __ai __attribute__((target("fullfp16"))) uint16x8_t vclezq_f16(float16x8_t __p0) {
  34545.   uint16x8_t __ret;
  34546.   __ret = (uint16x8_t) __builtin_neon_vclezq_f16((int8x16_t)__p0, 49);
  34547.   return __ret;
  34548. }
  34549. #else
  34550. __ai __attribute__((target("fullfp16"))) uint16x8_t vclezq_f16(float16x8_t __p0) {
  34551.   uint16x8_t __ret;
  34552.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34553.   __ret = (uint16x8_t) __builtin_neon_vclezq_f16((int8x16_t)__rev0, 49);
  34554.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34555.   return __ret;
  34556. }
  34557. #endif
  34558.  
  34559. #ifdef __LITTLE_ENDIAN__
  34560. __ai __attribute__((target("fullfp16"))) uint16x4_t vclez_f16(float16x4_t __p0) {
  34561.   uint16x4_t __ret;
  34562.   __ret = (uint16x4_t) __builtin_neon_vclez_f16((int8x8_t)__p0, 17);
  34563.   return __ret;
  34564. }
  34565. #else
  34566. __ai __attribute__((target("fullfp16"))) uint16x4_t vclez_f16(float16x4_t __p0) {
  34567.   uint16x4_t __ret;
  34568.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34569.   __ret = (uint16x4_t) __builtin_neon_vclez_f16((int8x8_t)__rev0, 17);
  34570.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34571.   return __ret;
  34572. }
  34573. #endif
  34574.  
  34575. #ifdef __LITTLE_ENDIAN__
  34576. __ai __attribute__((target("fullfp16"))) uint16x8_t vcltq_f16(float16x8_t __p0, float16x8_t __p1) {
  34577.   uint16x8_t __ret;
  34578.   __ret = (uint16x8_t)(__p0 < __p1);
  34579.   return __ret;
  34580. }
  34581. #else
  34582. __ai __attribute__((target("fullfp16"))) uint16x8_t vcltq_f16(float16x8_t __p0, float16x8_t __p1) {
  34583.   uint16x8_t __ret;
  34584.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34585.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  34586.   __ret = (uint16x8_t)(__rev0 < __rev1);
  34587.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34588.   return __ret;
  34589. }
  34590. #endif
  34591.  
  34592. #ifdef __LITTLE_ENDIAN__
  34593. __ai __attribute__((target("fullfp16"))) uint16x4_t vclt_f16(float16x4_t __p0, float16x4_t __p1) {
  34594.   uint16x4_t __ret;
  34595.   __ret = (uint16x4_t)(__p0 < __p1);
  34596.   return __ret;
  34597. }
  34598. #else
  34599. __ai __attribute__((target("fullfp16"))) uint16x4_t vclt_f16(float16x4_t __p0, float16x4_t __p1) {
  34600.   uint16x4_t __ret;
  34601.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34602.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  34603.   __ret = (uint16x4_t)(__rev0 < __rev1);
  34604.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34605.   return __ret;
  34606. }
  34607. #endif
  34608.  
  34609. #ifdef __LITTLE_ENDIAN__
  34610. __ai __attribute__((target("fullfp16"))) uint16x8_t vcltzq_f16(float16x8_t __p0) {
  34611.   uint16x8_t __ret;
  34612.   __ret = (uint16x8_t) __builtin_neon_vcltzq_f16((int8x16_t)__p0, 49);
  34613.   return __ret;
  34614. }
  34615. #else
  34616. __ai __attribute__((target("fullfp16"))) uint16x8_t vcltzq_f16(float16x8_t __p0) {
  34617.   uint16x8_t __ret;
  34618.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34619.   __ret = (uint16x8_t) __builtin_neon_vcltzq_f16((int8x16_t)__rev0, 49);
  34620.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34621.   return __ret;
  34622. }
  34623. #endif
  34624.  
  34625. #ifdef __LITTLE_ENDIAN__
  34626. __ai __attribute__((target("fullfp16"))) uint16x4_t vcltz_f16(float16x4_t __p0) {
  34627.   uint16x4_t __ret;
  34628.   __ret = (uint16x4_t) __builtin_neon_vcltz_f16((int8x8_t)__p0, 17);
  34629.   return __ret;
  34630. }
  34631. #else
  34632. __ai __attribute__((target("fullfp16"))) uint16x4_t vcltz_f16(float16x4_t __p0) {
  34633.   uint16x4_t __ret;
  34634.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34635.   __ret = (uint16x4_t) __builtin_neon_vcltz_f16((int8x8_t)__rev0, 17);
  34636.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34637.   return __ret;
  34638. }
  34639. #endif
  34640.  
  34641. #ifdef __LITTLE_ENDIAN__
  34642. __ai __attribute__((target("fullfp16"))) float16x8_t vcvtq_f16_u16(uint16x8_t __p0) {
  34643.   float16x8_t __ret;
  34644.   __ret = (float16x8_t) __builtin_neon_vcvtq_f16_u16((int8x16_t)__p0, 49);
  34645.   return __ret;
  34646. }
  34647. #else
  34648. __ai __attribute__((target("fullfp16"))) float16x8_t vcvtq_f16_u16(uint16x8_t __p0) {
  34649.   float16x8_t __ret;
  34650.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34651.   __ret = (float16x8_t) __builtin_neon_vcvtq_f16_u16((int8x16_t)__rev0, 49);
  34652.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34653.   return __ret;
  34654. }
  34655. #endif
  34656.  
  34657. #ifdef __LITTLE_ENDIAN__
  34658. __ai __attribute__((target("fullfp16"))) float16x8_t vcvtq_f16_s16(int16x8_t __p0) {
  34659.   float16x8_t __ret;
  34660.   __ret = (float16x8_t) __builtin_neon_vcvtq_f16_s16((int8x16_t)__p0, 33);
  34661.   return __ret;
  34662. }
  34663. #else
  34664. __ai __attribute__((target("fullfp16"))) float16x8_t vcvtq_f16_s16(int16x8_t __p0) {
  34665.   float16x8_t __ret;
  34666.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34667.   __ret = (float16x8_t) __builtin_neon_vcvtq_f16_s16((int8x16_t)__rev0, 33);
  34668.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34669.   return __ret;
  34670. }
  34671. #endif
  34672.  
  34673. #ifdef __LITTLE_ENDIAN__
  34674. __ai __attribute__((target("fullfp16"))) float16x4_t vcvt_f16_u16(uint16x4_t __p0) {
  34675.   float16x4_t __ret;
  34676.   __ret = (float16x4_t) __builtin_neon_vcvt_f16_u16((int8x8_t)__p0, 17);
  34677.   return __ret;
  34678. }
  34679. #else
  34680. __ai __attribute__((target("fullfp16"))) float16x4_t vcvt_f16_u16(uint16x4_t __p0) {
  34681.   float16x4_t __ret;
  34682.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34683.   __ret = (float16x4_t) __builtin_neon_vcvt_f16_u16((int8x8_t)__rev0, 17);
  34684.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34685.   return __ret;
  34686. }
  34687. #endif
  34688.  
  34689. #ifdef __LITTLE_ENDIAN__
  34690. __ai __attribute__((target("fullfp16"))) float16x4_t vcvt_f16_s16(int16x4_t __p0) {
  34691.   float16x4_t __ret;
  34692.   __ret = (float16x4_t) __builtin_neon_vcvt_f16_s16((int8x8_t)__p0, 1);
  34693.   return __ret;
  34694. }
  34695. #else
  34696. __ai __attribute__((target("fullfp16"))) float16x4_t vcvt_f16_s16(int16x4_t __p0) {
  34697.   float16x4_t __ret;
  34698.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34699.   __ret = (float16x4_t) __builtin_neon_vcvt_f16_s16((int8x8_t)__rev0, 1);
  34700.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34701.   return __ret;
  34702. }
  34703. #endif
  34704.  
  34705. #ifdef __LITTLE_ENDIAN__
  34706. #define vcvtq_n_f16_u16(__p0, __p1) __extension__ ({ \
  34707.   float16x8_t __ret; \
  34708.   uint16x8_t __s0 = __p0; \
  34709.   __ret = (float16x8_t) __builtin_neon_vcvtq_n_f16_u16((int8x16_t)__s0, __p1, 49); \
  34710.   __ret; \
  34711. })
  34712. #else
  34713. #define vcvtq_n_f16_u16(__p0, __p1) __extension__ ({ \
  34714.   float16x8_t __ret; \
  34715.   uint16x8_t __s0 = __p0; \
  34716.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  34717.   __ret = (float16x8_t) __builtin_neon_vcvtq_n_f16_u16((int8x16_t)__rev0, __p1, 49); \
  34718.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  34719.   __ret; \
  34720. })
  34721. #endif
  34722.  
  34723. #ifdef __LITTLE_ENDIAN__
  34724. #define vcvtq_n_f16_s16(__p0, __p1) __extension__ ({ \
  34725.   float16x8_t __ret; \
  34726.   int16x8_t __s0 = __p0; \
  34727.   __ret = (float16x8_t) __builtin_neon_vcvtq_n_f16_s16((int8x16_t)__s0, __p1, 33); \
  34728.   __ret; \
  34729. })
  34730. #else
  34731. #define vcvtq_n_f16_s16(__p0, __p1) __extension__ ({ \
  34732.   float16x8_t __ret; \
  34733.   int16x8_t __s0 = __p0; \
  34734.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  34735.   __ret = (float16x8_t) __builtin_neon_vcvtq_n_f16_s16((int8x16_t)__rev0, __p1, 33); \
  34736.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  34737.   __ret; \
  34738. })
  34739. #endif
  34740.  
  34741. #ifdef __LITTLE_ENDIAN__
  34742. #define vcvt_n_f16_u16(__p0, __p1) __extension__ ({ \
  34743.   float16x4_t __ret; \
  34744.   uint16x4_t __s0 = __p0; \
  34745.   __ret = (float16x4_t) __builtin_neon_vcvt_n_f16_u16((int8x8_t)__s0, __p1, 17); \
  34746.   __ret; \
  34747. })
  34748. #else
  34749. #define vcvt_n_f16_u16(__p0, __p1) __extension__ ({ \
  34750.   float16x4_t __ret; \
  34751.   uint16x4_t __s0 = __p0; \
  34752.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  34753.   __ret = (float16x4_t) __builtin_neon_vcvt_n_f16_u16((int8x8_t)__rev0, __p1, 17); \
  34754.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  34755.   __ret; \
  34756. })
  34757. #endif
  34758.  
  34759. #ifdef __LITTLE_ENDIAN__
  34760. #define vcvt_n_f16_s16(__p0, __p1) __extension__ ({ \
  34761.   float16x4_t __ret; \
  34762.   int16x4_t __s0 = __p0; \
  34763.   __ret = (float16x4_t) __builtin_neon_vcvt_n_f16_s16((int8x8_t)__s0, __p1, 1); \
  34764.   __ret; \
  34765. })
  34766. #else
  34767. #define vcvt_n_f16_s16(__p0, __p1) __extension__ ({ \
  34768.   float16x4_t __ret; \
  34769.   int16x4_t __s0 = __p0; \
  34770.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  34771.   __ret = (float16x4_t) __builtin_neon_vcvt_n_f16_s16((int8x8_t)__rev0, __p1, 1); \
  34772.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  34773.   __ret; \
  34774. })
  34775. #endif
  34776.  
  34777. #ifdef __LITTLE_ENDIAN__
  34778. #define vcvtq_n_s16_f16(__p0, __p1) __extension__ ({ \
  34779.   int16x8_t __ret; \
  34780.   float16x8_t __s0 = __p0; \
  34781.   __ret = (int16x8_t) __builtin_neon_vcvtq_n_s16_f16((int8x16_t)__s0, __p1, 33); \
  34782.   __ret; \
  34783. })
  34784. #else
  34785. #define vcvtq_n_s16_f16(__p0, __p1) __extension__ ({ \
  34786.   int16x8_t __ret; \
  34787.   float16x8_t __s0 = __p0; \
  34788.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  34789.   __ret = (int16x8_t) __builtin_neon_vcvtq_n_s16_f16((int8x16_t)__rev0, __p1, 33); \
  34790.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  34791.   __ret; \
  34792. })
  34793. #endif
  34794.  
  34795. #ifdef __LITTLE_ENDIAN__
  34796. #define vcvt_n_s16_f16(__p0, __p1) __extension__ ({ \
  34797.   int16x4_t __ret; \
  34798.   float16x4_t __s0 = __p0; \
  34799.   __ret = (int16x4_t) __builtin_neon_vcvt_n_s16_f16((int8x8_t)__s0, __p1, 1); \
  34800.   __ret; \
  34801. })
  34802. #else
  34803. #define vcvt_n_s16_f16(__p0, __p1) __extension__ ({ \
  34804.   int16x4_t __ret; \
  34805.   float16x4_t __s0 = __p0; \
  34806.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  34807.   __ret = (int16x4_t) __builtin_neon_vcvt_n_s16_f16((int8x8_t)__rev0, __p1, 1); \
  34808.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  34809.   __ret; \
  34810. })
  34811. #endif
  34812.  
  34813. #ifdef __LITTLE_ENDIAN__
  34814. #define vcvtq_n_u16_f16(__p0, __p1) __extension__ ({ \
  34815.   uint16x8_t __ret; \
  34816.   float16x8_t __s0 = __p0; \
  34817.   __ret = (uint16x8_t) __builtin_neon_vcvtq_n_u16_f16((int8x16_t)__s0, __p1, 49); \
  34818.   __ret; \
  34819. })
  34820. #else
  34821. #define vcvtq_n_u16_f16(__p0, __p1) __extension__ ({ \
  34822.   uint16x8_t __ret; \
  34823.   float16x8_t __s0 = __p0; \
  34824.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  34825.   __ret = (uint16x8_t) __builtin_neon_vcvtq_n_u16_f16((int8x16_t)__rev0, __p1, 49); \
  34826.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  34827.   __ret; \
  34828. })
  34829. #endif
  34830.  
  34831. #ifdef __LITTLE_ENDIAN__
  34832. #define vcvt_n_u16_f16(__p0, __p1) __extension__ ({ \
  34833.   uint16x4_t __ret; \
  34834.   float16x4_t __s0 = __p0; \
  34835.   __ret = (uint16x4_t) __builtin_neon_vcvt_n_u16_f16((int8x8_t)__s0, __p1, 17); \
  34836.   __ret; \
  34837. })
  34838. #else
  34839. #define vcvt_n_u16_f16(__p0, __p1) __extension__ ({ \
  34840.   uint16x4_t __ret; \
  34841.   float16x4_t __s0 = __p0; \
  34842.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  34843.   __ret = (uint16x4_t) __builtin_neon_vcvt_n_u16_f16((int8x8_t)__rev0, __p1, 17); \
  34844.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  34845.   __ret; \
  34846. })
  34847. #endif
  34848.  
  34849. #ifdef __LITTLE_ENDIAN__
  34850. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtq_s16_f16(float16x8_t __p0) {
  34851.   int16x8_t __ret;
  34852.   __ret = (int16x8_t) __builtin_neon_vcvtq_s16_f16((int8x16_t)__p0, 33);
  34853.   return __ret;
  34854. }
  34855. #else
  34856. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtq_s16_f16(float16x8_t __p0) {
  34857.   int16x8_t __ret;
  34858.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34859.   __ret = (int16x8_t) __builtin_neon_vcvtq_s16_f16((int8x16_t)__rev0, 33);
  34860.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34861.   return __ret;
  34862. }
  34863. #endif
  34864.  
  34865. #ifdef __LITTLE_ENDIAN__
  34866. __ai __attribute__((target("fullfp16"))) int16x4_t vcvt_s16_f16(float16x4_t __p0) {
  34867.   int16x4_t __ret;
  34868.   __ret = (int16x4_t) __builtin_neon_vcvt_s16_f16((int8x8_t)__p0, 1);
  34869.   return __ret;
  34870. }
  34871. #else
  34872. __ai __attribute__((target("fullfp16"))) int16x4_t vcvt_s16_f16(float16x4_t __p0) {
  34873.   int16x4_t __ret;
  34874.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34875.   __ret = (int16x4_t) __builtin_neon_vcvt_s16_f16((int8x8_t)__rev0, 1);
  34876.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34877.   return __ret;
  34878. }
  34879. #endif
  34880.  
  34881. #ifdef __LITTLE_ENDIAN__
  34882. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtq_u16_f16(float16x8_t __p0) {
  34883.   uint16x8_t __ret;
  34884.   __ret = (uint16x8_t) __builtin_neon_vcvtq_u16_f16((int8x16_t)__p0, 49);
  34885.   return __ret;
  34886. }
  34887. #else
  34888. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtq_u16_f16(float16x8_t __p0) {
  34889.   uint16x8_t __ret;
  34890.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34891.   __ret = (uint16x8_t) __builtin_neon_vcvtq_u16_f16((int8x16_t)__rev0, 49);
  34892.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34893.   return __ret;
  34894. }
  34895. #endif
  34896.  
  34897. #ifdef __LITTLE_ENDIAN__
  34898. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvt_u16_f16(float16x4_t __p0) {
  34899.   uint16x4_t __ret;
  34900.   __ret = (uint16x4_t) __builtin_neon_vcvt_u16_f16((int8x8_t)__p0, 17);
  34901.   return __ret;
  34902. }
  34903. #else
  34904. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvt_u16_f16(float16x4_t __p0) {
  34905.   uint16x4_t __ret;
  34906.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34907.   __ret = (uint16x4_t) __builtin_neon_vcvt_u16_f16((int8x8_t)__rev0, 17);
  34908.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34909.   return __ret;
  34910. }
  34911. #endif
  34912.  
  34913. #ifdef __LITTLE_ENDIAN__
  34914. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtaq_s16_f16(float16x8_t __p0) {
  34915.   int16x8_t __ret;
  34916.   __ret = (int16x8_t) __builtin_neon_vcvtaq_s16_f16((int8x16_t)__p0, 33);
  34917.   return __ret;
  34918. }
  34919. #else
  34920. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtaq_s16_f16(float16x8_t __p0) {
  34921.   int16x8_t __ret;
  34922.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34923.   __ret = (int16x8_t) __builtin_neon_vcvtaq_s16_f16((int8x16_t)__rev0, 33);
  34924.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34925.   return __ret;
  34926. }
  34927. #endif
  34928.  
  34929. #ifdef __LITTLE_ENDIAN__
  34930. __ai __attribute__((target("fullfp16"))) int16x4_t vcvta_s16_f16(float16x4_t __p0) {
  34931.   int16x4_t __ret;
  34932.   __ret = (int16x4_t) __builtin_neon_vcvta_s16_f16((int8x8_t)__p0, 1);
  34933.   return __ret;
  34934. }
  34935. #else
  34936. __ai __attribute__((target("fullfp16"))) int16x4_t vcvta_s16_f16(float16x4_t __p0) {
  34937.   int16x4_t __ret;
  34938.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34939.   __ret = (int16x4_t) __builtin_neon_vcvta_s16_f16((int8x8_t)__rev0, 1);
  34940.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34941.   return __ret;
  34942. }
  34943. #endif
  34944.  
  34945. #ifdef __LITTLE_ENDIAN__
  34946. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtaq_u16_f16(float16x8_t __p0) {
  34947.   uint16x8_t __ret;
  34948.   __ret = (uint16x8_t) __builtin_neon_vcvtaq_u16_f16((int8x16_t)__p0, 49);
  34949.   return __ret;
  34950. }
  34951. #else
  34952. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtaq_u16_f16(float16x8_t __p0) {
  34953.   uint16x8_t __ret;
  34954.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34955.   __ret = (uint16x8_t) __builtin_neon_vcvtaq_u16_f16((int8x16_t)__rev0, 49);
  34956.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34957.   return __ret;
  34958. }
  34959. #endif
  34960.  
  34961. #ifdef __LITTLE_ENDIAN__
  34962. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvta_u16_f16(float16x4_t __p0) {
  34963.   uint16x4_t __ret;
  34964.   __ret = (uint16x4_t) __builtin_neon_vcvta_u16_f16((int8x8_t)__p0, 17);
  34965.   return __ret;
  34966. }
  34967. #else
  34968. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvta_u16_f16(float16x4_t __p0) {
  34969.   uint16x4_t __ret;
  34970.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  34971.   __ret = (uint16x4_t) __builtin_neon_vcvta_u16_f16((int8x8_t)__rev0, 17);
  34972.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  34973.   return __ret;
  34974. }
  34975. #endif
  34976.  
  34977. #ifdef __LITTLE_ENDIAN__
  34978. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtmq_s16_f16(float16x8_t __p0) {
  34979.   int16x8_t __ret;
  34980.   __ret = (int16x8_t) __builtin_neon_vcvtmq_s16_f16((int8x16_t)__p0, 33);
  34981.   return __ret;
  34982. }
  34983. #else
  34984. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtmq_s16_f16(float16x8_t __p0) {
  34985.   int16x8_t __ret;
  34986.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  34987.   __ret = (int16x8_t) __builtin_neon_vcvtmq_s16_f16((int8x16_t)__rev0, 33);
  34988.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  34989.   return __ret;
  34990. }
  34991. #endif
  34992.  
  34993. #ifdef __LITTLE_ENDIAN__
  34994. __ai __attribute__((target("fullfp16"))) int16x4_t vcvtm_s16_f16(float16x4_t __p0) {
  34995.   int16x4_t __ret;
  34996.   __ret = (int16x4_t) __builtin_neon_vcvtm_s16_f16((int8x8_t)__p0, 1);
  34997.   return __ret;
  34998. }
  34999. #else
  35000. __ai __attribute__((target("fullfp16"))) int16x4_t vcvtm_s16_f16(float16x4_t __p0) {
  35001.   int16x4_t __ret;
  35002.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35003.   __ret = (int16x4_t) __builtin_neon_vcvtm_s16_f16((int8x8_t)__rev0, 1);
  35004.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35005.   return __ret;
  35006. }
  35007. #endif
  35008.  
  35009. #ifdef __LITTLE_ENDIAN__
  35010. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtmq_u16_f16(float16x8_t __p0) {
  35011.   uint16x8_t __ret;
  35012.   __ret = (uint16x8_t) __builtin_neon_vcvtmq_u16_f16((int8x16_t)__p0, 49);
  35013.   return __ret;
  35014. }
  35015. #else
  35016. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtmq_u16_f16(float16x8_t __p0) {
  35017.   uint16x8_t __ret;
  35018.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35019.   __ret = (uint16x8_t) __builtin_neon_vcvtmq_u16_f16((int8x16_t)__rev0, 49);
  35020.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35021.   return __ret;
  35022. }
  35023. #endif
  35024.  
  35025. #ifdef __LITTLE_ENDIAN__
  35026. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvtm_u16_f16(float16x4_t __p0) {
  35027.   uint16x4_t __ret;
  35028.   __ret = (uint16x4_t) __builtin_neon_vcvtm_u16_f16((int8x8_t)__p0, 17);
  35029.   return __ret;
  35030. }
  35031. #else
  35032. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvtm_u16_f16(float16x4_t __p0) {
  35033.   uint16x4_t __ret;
  35034.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35035.   __ret = (uint16x4_t) __builtin_neon_vcvtm_u16_f16((int8x8_t)__rev0, 17);
  35036.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35037.   return __ret;
  35038. }
  35039. #endif
  35040.  
  35041. #ifdef __LITTLE_ENDIAN__
  35042. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtnq_s16_f16(float16x8_t __p0) {
  35043.   int16x8_t __ret;
  35044.   __ret = (int16x8_t) __builtin_neon_vcvtnq_s16_f16((int8x16_t)__p0, 33);
  35045.   return __ret;
  35046. }
  35047. #else
  35048. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtnq_s16_f16(float16x8_t __p0) {
  35049.   int16x8_t __ret;
  35050.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35051.   __ret = (int16x8_t) __builtin_neon_vcvtnq_s16_f16((int8x16_t)__rev0, 33);
  35052.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35053.   return __ret;
  35054. }
  35055. #endif
  35056.  
  35057. #ifdef __LITTLE_ENDIAN__
  35058. __ai __attribute__((target("fullfp16"))) int16x4_t vcvtn_s16_f16(float16x4_t __p0) {
  35059.   int16x4_t __ret;
  35060.   __ret = (int16x4_t) __builtin_neon_vcvtn_s16_f16((int8x8_t)__p0, 1);
  35061.   return __ret;
  35062. }
  35063. #else
  35064. __ai __attribute__((target("fullfp16"))) int16x4_t vcvtn_s16_f16(float16x4_t __p0) {
  35065.   int16x4_t __ret;
  35066.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35067.   __ret = (int16x4_t) __builtin_neon_vcvtn_s16_f16((int8x8_t)__rev0, 1);
  35068.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35069.   return __ret;
  35070. }
  35071. #endif
  35072.  
  35073. #ifdef __LITTLE_ENDIAN__
  35074. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtnq_u16_f16(float16x8_t __p0) {
  35075.   uint16x8_t __ret;
  35076.   __ret = (uint16x8_t) __builtin_neon_vcvtnq_u16_f16((int8x16_t)__p0, 49);
  35077.   return __ret;
  35078. }
  35079. #else
  35080. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtnq_u16_f16(float16x8_t __p0) {
  35081.   uint16x8_t __ret;
  35082.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35083.   __ret = (uint16x8_t) __builtin_neon_vcvtnq_u16_f16((int8x16_t)__rev0, 49);
  35084.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35085.   return __ret;
  35086. }
  35087. #endif
  35088.  
  35089. #ifdef __LITTLE_ENDIAN__
  35090. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvtn_u16_f16(float16x4_t __p0) {
  35091.   uint16x4_t __ret;
  35092.   __ret = (uint16x4_t) __builtin_neon_vcvtn_u16_f16((int8x8_t)__p0, 17);
  35093.   return __ret;
  35094. }
  35095. #else
  35096. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvtn_u16_f16(float16x4_t __p0) {
  35097.   uint16x4_t __ret;
  35098.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35099.   __ret = (uint16x4_t) __builtin_neon_vcvtn_u16_f16((int8x8_t)__rev0, 17);
  35100.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35101.   return __ret;
  35102. }
  35103. #endif
  35104.  
  35105. #ifdef __LITTLE_ENDIAN__
  35106. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtpq_s16_f16(float16x8_t __p0) {
  35107.   int16x8_t __ret;
  35108.   __ret = (int16x8_t) __builtin_neon_vcvtpq_s16_f16((int8x16_t)__p0, 33);
  35109.   return __ret;
  35110. }
  35111. #else
  35112. __ai __attribute__((target("fullfp16"))) int16x8_t vcvtpq_s16_f16(float16x8_t __p0) {
  35113.   int16x8_t __ret;
  35114.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35115.   __ret = (int16x8_t) __builtin_neon_vcvtpq_s16_f16((int8x16_t)__rev0, 33);
  35116.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35117.   return __ret;
  35118. }
  35119. #endif
  35120.  
  35121. #ifdef __LITTLE_ENDIAN__
  35122. __ai __attribute__((target("fullfp16"))) int16x4_t vcvtp_s16_f16(float16x4_t __p0) {
  35123.   int16x4_t __ret;
  35124.   __ret = (int16x4_t) __builtin_neon_vcvtp_s16_f16((int8x8_t)__p0, 1);
  35125.   return __ret;
  35126. }
  35127. #else
  35128. __ai __attribute__((target("fullfp16"))) int16x4_t vcvtp_s16_f16(float16x4_t __p0) {
  35129.   int16x4_t __ret;
  35130.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35131.   __ret = (int16x4_t) __builtin_neon_vcvtp_s16_f16((int8x8_t)__rev0, 1);
  35132.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35133.   return __ret;
  35134. }
  35135. #endif
  35136.  
  35137. #ifdef __LITTLE_ENDIAN__
  35138. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtpq_u16_f16(float16x8_t __p0) {
  35139.   uint16x8_t __ret;
  35140.   __ret = (uint16x8_t) __builtin_neon_vcvtpq_u16_f16((int8x16_t)__p0, 49);
  35141.   return __ret;
  35142. }
  35143. #else
  35144. __ai __attribute__((target("fullfp16"))) uint16x8_t vcvtpq_u16_f16(float16x8_t __p0) {
  35145.   uint16x8_t __ret;
  35146.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35147.   __ret = (uint16x8_t) __builtin_neon_vcvtpq_u16_f16((int8x16_t)__rev0, 49);
  35148.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35149.   return __ret;
  35150. }
  35151. #endif
  35152.  
  35153. #ifdef __LITTLE_ENDIAN__
  35154. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvtp_u16_f16(float16x4_t __p0) {
  35155.   uint16x4_t __ret;
  35156.   __ret = (uint16x4_t) __builtin_neon_vcvtp_u16_f16((int8x8_t)__p0, 17);
  35157.   return __ret;
  35158. }
  35159. #else
  35160. __ai __attribute__((target("fullfp16"))) uint16x4_t vcvtp_u16_f16(float16x4_t __p0) {
  35161.   uint16x4_t __ret;
  35162.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35163.   __ret = (uint16x4_t) __builtin_neon_vcvtp_u16_f16((int8x8_t)__rev0, 17);
  35164.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35165.   return __ret;
  35166. }
  35167. #endif
  35168.  
  35169. #ifdef __LITTLE_ENDIAN__
  35170. #define vextq_f16(__p0, __p1, __p2) __extension__ ({ \
  35171.   float16x8_t __ret; \
  35172.   float16x8_t __s0 = __p0; \
  35173.   float16x8_t __s1 = __p1; \
  35174.   __ret = (float16x8_t) __builtin_neon_vextq_f16((int8x16_t)__s0, (int8x16_t)__s1, __p2, 40); \
  35175.   __ret; \
  35176. })
  35177. #else
  35178. #define vextq_f16(__p0, __p1, __p2) __extension__ ({ \
  35179.   float16x8_t __ret; \
  35180.   float16x8_t __s0 = __p0; \
  35181.   float16x8_t __s1 = __p1; \
  35182.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  35183.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  35184.   __ret = (float16x8_t) __builtin_neon_vextq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 40); \
  35185.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  35186.   __ret; \
  35187. })
  35188. #endif
  35189.  
  35190. #ifdef __LITTLE_ENDIAN__
  35191. #define vext_f16(__p0, __p1, __p2) __extension__ ({ \
  35192.   float16x4_t __ret; \
  35193.   float16x4_t __s0 = __p0; \
  35194.   float16x4_t __s1 = __p1; \
  35195.   __ret = (float16x4_t) __builtin_neon_vext_f16((int8x8_t)__s0, (int8x8_t)__s1, __p2, 8); \
  35196.   __ret; \
  35197. })
  35198. #else
  35199. #define vext_f16(__p0, __p1, __p2) __extension__ ({ \
  35200.   float16x4_t __ret; \
  35201.   float16x4_t __s0 = __p0; \
  35202.   float16x4_t __s1 = __p1; \
  35203.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  35204.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  35205.   __ret = (float16x4_t) __builtin_neon_vext_f16((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 8); \
  35206.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  35207.   __ret; \
  35208. })
  35209. #endif
  35210.  
  35211. #ifdef __LITTLE_ENDIAN__
  35212. __ai __attribute__((target("fullfp16"))) float16x8_t vfmaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  35213.   float16x8_t __ret;
  35214.   __ret = (float16x8_t) __builtin_neon_vfmaq_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  35215.   return __ret;
  35216. }
  35217. #else
  35218. __ai __attribute__((target("fullfp16"))) float16x8_t vfmaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  35219.   float16x8_t __ret;
  35220.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35221.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35222.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  35223.   __ret = (float16x8_t) __builtin_neon_vfmaq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40);
  35224.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35225.   return __ret;
  35226. }
  35227. __ai __attribute__((target("fullfp16"))) float16x8_t __noswap_vfmaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  35228.   float16x8_t __ret;
  35229.   __ret = (float16x8_t) __builtin_neon_vfmaq_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  35230.   return __ret;
  35231. }
  35232. #endif
  35233.  
  35234. #ifdef __LITTLE_ENDIAN__
  35235. __ai __attribute__((target("fullfp16"))) float16x4_t vfma_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  35236.   float16x4_t __ret;
  35237.   __ret = (float16x4_t) __builtin_neon_vfma_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  35238.   return __ret;
  35239. }
  35240. #else
  35241. __ai __attribute__((target("fullfp16"))) float16x4_t vfma_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  35242.   float16x4_t __ret;
  35243.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35244.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35245.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  35246.   __ret = (float16x4_t) __builtin_neon_vfma_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8);
  35247.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35248.   return __ret;
  35249. }
  35250. __ai __attribute__((target("fullfp16"))) float16x4_t __noswap_vfma_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  35251.   float16x4_t __ret;
  35252.   __ret = (float16x4_t) __builtin_neon_vfma_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  35253.   return __ret;
  35254. }
  35255. #endif
  35256.  
  35257. #ifdef __LITTLE_ENDIAN__
  35258. __ai __attribute__((target("fullfp16"))) float16x8_t vfmsq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  35259.   float16x8_t __ret;
  35260.   __ret = vfmaq_f16(__p0, -__p1, __p2);
  35261.   return __ret;
  35262. }
  35263. #else
  35264. __ai __attribute__((target("fullfp16"))) float16x8_t vfmsq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  35265.   float16x8_t __ret;
  35266.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35267.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35268.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  35269.   __ret = __noswap_vfmaq_f16(__rev0, -__rev1, __rev2);
  35270.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35271.   return __ret;
  35272. }
  35273. #endif
  35274.  
  35275. #ifdef __LITTLE_ENDIAN__
  35276. __ai __attribute__((target("fullfp16"))) float16x4_t vfms_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  35277.   float16x4_t __ret;
  35278.   __ret = vfma_f16(__p0, -__p1, __p2);
  35279.   return __ret;
  35280. }
  35281. #else
  35282. __ai __attribute__((target("fullfp16"))) float16x4_t vfms_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  35283.   float16x4_t __ret;
  35284.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35285.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35286.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  35287.   __ret = __noswap_vfma_f16(__rev0, -__rev1, __rev2);
  35288.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35289.   return __ret;
  35290. }
  35291. #endif
  35292.  
  35293. #ifdef __LITTLE_ENDIAN__
  35294. __ai __attribute__((target("fullfp16"))) float16x8_t vmaxq_f16(float16x8_t __p0, float16x8_t __p1) {
  35295.   float16x8_t __ret;
  35296.   __ret = (float16x8_t) __builtin_neon_vmaxq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  35297.   return __ret;
  35298. }
  35299. #else
  35300. __ai __attribute__((target("fullfp16"))) float16x8_t vmaxq_f16(float16x8_t __p0, float16x8_t __p1) {
  35301.   float16x8_t __ret;
  35302.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35303.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35304.   __ret = (float16x8_t) __builtin_neon_vmaxq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  35305.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35306.   return __ret;
  35307. }
  35308. #endif
  35309.  
  35310. #ifdef __LITTLE_ENDIAN__
  35311. __ai __attribute__((target("fullfp16"))) float16x4_t vmax_f16(float16x4_t __p0, float16x4_t __p1) {
  35312.   float16x4_t __ret;
  35313.   __ret = (float16x4_t) __builtin_neon_vmax_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  35314.   return __ret;
  35315. }
  35316. #else
  35317. __ai __attribute__((target("fullfp16"))) float16x4_t vmax_f16(float16x4_t __p0, float16x4_t __p1) {
  35318.   float16x4_t __ret;
  35319.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35320.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35321.   __ret = (float16x4_t) __builtin_neon_vmax_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35322.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35323.   return __ret;
  35324. }
  35325. #endif
  35326.  
  35327. #ifdef __LITTLE_ENDIAN__
  35328. __ai __attribute__((target("fullfp16"))) float16x8_t vminq_f16(float16x8_t __p0, float16x8_t __p1) {
  35329.   float16x8_t __ret;
  35330.   __ret = (float16x8_t) __builtin_neon_vminq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  35331.   return __ret;
  35332. }
  35333. #else
  35334. __ai __attribute__((target("fullfp16"))) float16x8_t vminq_f16(float16x8_t __p0, float16x8_t __p1) {
  35335.   float16x8_t __ret;
  35336.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35337.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35338.   __ret = (float16x8_t) __builtin_neon_vminq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  35339.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35340.   return __ret;
  35341. }
  35342. #endif
  35343.  
  35344. #ifdef __LITTLE_ENDIAN__
  35345. __ai __attribute__((target("fullfp16"))) float16x4_t vmin_f16(float16x4_t __p0, float16x4_t __p1) {
  35346.   float16x4_t __ret;
  35347.   __ret = (float16x4_t) __builtin_neon_vmin_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  35348.   return __ret;
  35349. }
  35350. #else
  35351. __ai __attribute__((target("fullfp16"))) float16x4_t vmin_f16(float16x4_t __p0, float16x4_t __p1) {
  35352.   float16x4_t __ret;
  35353.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35354.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35355.   __ret = (float16x4_t) __builtin_neon_vmin_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35356.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35357.   return __ret;
  35358. }
  35359. #endif
  35360.  
  35361. #ifdef __LITTLE_ENDIAN__
  35362. __ai __attribute__((target("fullfp16"))) float16x8_t vmulq_f16(float16x8_t __p0, float16x8_t __p1) {
  35363.   float16x8_t __ret;
  35364.   __ret = __p0 * __p1;
  35365.   return __ret;
  35366. }
  35367. #else
  35368. __ai __attribute__((target("fullfp16"))) float16x8_t vmulq_f16(float16x8_t __p0, float16x8_t __p1) {
  35369.   float16x8_t __ret;
  35370.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35371.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35372.   __ret = __rev0 * __rev1;
  35373.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35374.   return __ret;
  35375. }
  35376. #endif
  35377.  
  35378. #ifdef __LITTLE_ENDIAN__
  35379. __ai __attribute__((target("fullfp16"))) float16x4_t vmul_f16(float16x4_t __p0, float16x4_t __p1) {
  35380.   float16x4_t __ret;
  35381.   __ret = __p0 * __p1;
  35382.   return __ret;
  35383. }
  35384. #else
  35385. __ai __attribute__((target("fullfp16"))) float16x4_t vmul_f16(float16x4_t __p0, float16x4_t __p1) {
  35386.   float16x4_t __ret;
  35387.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35388.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35389.   __ret = __rev0 * __rev1;
  35390.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35391.   return __ret;
  35392. }
  35393. #endif
  35394.  
  35395. #ifdef __LITTLE_ENDIAN__
  35396. #define vmulq_lane_f16(__p0_153, __p1_153, __p2_153) __extension__ ({ \
  35397.   float16x8_t __ret_153; \
  35398.   float16x8_t __s0_153 = __p0_153; \
  35399.   float16x4_t __s1_153 = __p1_153; \
  35400.   __ret_153 = __s0_153 * splatq_lane_f16(__s1_153, __p2_153); \
  35401.   __ret_153; \
  35402. })
  35403. #else
  35404. #define vmulq_lane_f16(__p0_154, __p1_154, __p2_154) __extension__ ({ \
  35405.   float16x8_t __ret_154; \
  35406.   float16x8_t __s0_154 = __p0_154; \
  35407.   float16x4_t __s1_154 = __p1_154; \
  35408.   float16x8_t __rev0_154;  __rev0_154 = __builtin_shufflevector(__s0_154, __s0_154, 7, 6, 5, 4, 3, 2, 1, 0); \
  35409.   float16x4_t __rev1_154;  __rev1_154 = __builtin_shufflevector(__s1_154, __s1_154, 3, 2, 1, 0); \
  35410.   __ret_154 = __rev0_154 * __noswap_splatq_lane_f16(__rev1_154, __p2_154); \
  35411.   __ret_154 = __builtin_shufflevector(__ret_154, __ret_154, 7, 6, 5, 4, 3, 2, 1, 0); \
  35412.   __ret_154; \
  35413. })
  35414. #endif
  35415.  
  35416. #ifdef __LITTLE_ENDIAN__
  35417. #define vmul_lane_f16(__p0_155, __p1_155, __p2_155) __extension__ ({ \
  35418.   float16x4_t __ret_155; \
  35419.   float16x4_t __s0_155 = __p0_155; \
  35420.   float16x4_t __s1_155 = __p1_155; \
  35421.   __ret_155 = __s0_155 * splat_lane_f16(__s1_155, __p2_155); \
  35422.   __ret_155; \
  35423. })
  35424. #else
  35425. #define vmul_lane_f16(__p0_156, __p1_156, __p2_156) __extension__ ({ \
  35426.   float16x4_t __ret_156; \
  35427.   float16x4_t __s0_156 = __p0_156; \
  35428.   float16x4_t __s1_156 = __p1_156; \
  35429.   float16x4_t __rev0_156;  __rev0_156 = __builtin_shufflevector(__s0_156, __s0_156, 3, 2, 1, 0); \
  35430.   float16x4_t __rev1_156;  __rev1_156 = __builtin_shufflevector(__s1_156, __s1_156, 3, 2, 1, 0); \
  35431.   __ret_156 = __rev0_156 * __noswap_splat_lane_f16(__rev1_156, __p2_156); \
  35432.   __ret_156 = __builtin_shufflevector(__ret_156, __ret_156, 3, 2, 1, 0); \
  35433.   __ret_156; \
  35434. })
  35435. #endif
  35436.  
  35437. #ifdef __LITTLE_ENDIAN__
  35438. #define vmulq_n_f16(__p0, __p1) __extension__ ({ \
  35439.   float16x8_t __ret; \
  35440.   float16x8_t __s0 = __p0; \
  35441.   float16_t __s1 = __p1; \
  35442.   __ret = __s0 * (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1}; \
  35443.   __ret; \
  35444. })
  35445. #else
  35446. #define vmulq_n_f16(__p0, __p1) __extension__ ({ \
  35447.   float16x8_t __ret; \
  35448.   float16x8_t __s0 = __p0; \
  35449.   float16_t __s1 = __p1; \
  35450.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  35451.   __ret = __rev0 * (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1}; \
  35452.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  35453.   __ret; \
  35454. })
  35455. #endif
  35456.  
  35457. #ifdef __LITTLE_ENDIAN__
  35458. #define vmul_n_f16(__p0, __p1) __extension__ ({ \
  35459.   float16x4_t __ret; \
  35460.   float16x4_t __s0 = __p0; \
  35461.   float16_t __s1 = __p1; \
  35462.   __ret = __s0 * (float16x4_t) {__s1, __s1, __s1, __s1}; \
  35463.   __ret; \
  35464. })
  35465. #else
  35466. #define vmul_n_f16(__p0, __p1) __extension__ ({ \
  35467.   float16x4_t __ret; \
  35468.   float16x4_t __s0 = __p0; \
  35469.   float16_t __s1 = __p1; \
  35470.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  35471.   __ret = __rev0 * (float16x4_t) {__s1, __s1, __s1, __s1}; \
  35472.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  35473.   __ret; \
  35474. })
  35475. #endif
  35476.  
  35477. #ifdef __LITTLE_ENDIAN__
  35478. __ai __attribute__((target("fullfp16"))) float16x8_t vnegq_f16(float16x8_t __p0) {
  35479.   float16x8_t __ret;
  35480.   __ret = -__p0;
  35481.   return __ret;
  35482. }
  35483. #else
  35484. __ai __attribute__((target("fullfp16"))) float16x8_t vnegq_f16(float16x8_t __p0) {
  35485.   float16x8_t __ret;
  35486.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35487.   __ret = -__rev0;
  35488.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35489.   return __ret;
  35490. }
  35491. #endif
  35492.  
  35493. #ifdef __LITTLE_ENDIAN__
  35494. __ai __attribute__((target("fullfp16"))) float16x4_t vneg_f16(float16x4_t __p0) {
  35495.   float16x4_t __ret;
  35496.   __ret = -__p0;
  35497.   return __ret;
  35498. }
  35499. #else
  35500. __ai __attribute__((target("fullfp16"))) float16x4_t vneg_f16(float16x4_t __p0) {
  35501.   float16x4_t __ret;
  35502.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35503.   __ret = -__rev0;
  35504.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35505.   return __ret;
  35506. }
  35507. #endif
  35508.  
  35509. #ifdef __LITTLE_ENDIAN__
  35510. __ai __attribute__((target("fullfp16"))) float16x4_t vpadd_f16(float16x4_t __p0, float16x4_t __p1) {
  35511.   float16x4_t __ret;
  35512.   __ret = (float16x4_t) __builtin_neon_vpadd_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  35513.   return __ret;
  35514. }
  35515. #else
  35516. __ai __attribute__((target("fullfp16"))) float16x4_t vpadd_f16(float16x4_t __p0, float16x4_t __p1) {
  35517.   float16x4_t __ret;
  35518.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35519.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35520.   __ret = (float16x4_t) __builtin_neon_vpadd_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35521.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35522.   return __ret;
  35523. }
  35524. #endif
  35525.  
  35526. #ifdef __LITTLE_ENDIAN__
  35527. __ai __attribute__((target("fullfp16"))) float16x4_t vpmax_f16(float16x4_t __p0, float16x4_t __p1) {
  35528.   float16x4_t __ret;
  35529.   __ret = (float16x4_t) __builtin_neon_vpmax_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  35530.   return __ret;
  35531. }
  35532. #else
  35533. __ai __attribute__((target("fullfp16"))) float16x4_t vpmax_f16(float16x4_t __p0, float16x4_t __p1) {
  35534.   float16x4_t __ret;
  35535.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35536.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35537.   __ret = (float16x4_t) __builtin_neon_vpmax_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35538.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35539.   return __ret;
  35540. }
  35541. #endif
  35542.  
  35543. #ifdef __LITTLE_ENDIAN__
  35544. __ai __attribute__((target("fullfp16"))) float16x4_t vpmin_f16(float16x4_t __p0, float16x4_t __p1) {
  35545.   float16x4_t __ret;
  35546.   __ret = (float16x4_t) __builtin_neon_vpmin_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  35547.   return __ret;
  35548. }
  35549. #else
  35550. __ai __attribute__((target("fullfp16"))) float16x4_t vpmin_f16(float16x4_t __p0, float16x4_t __p1) {
  35551.   float16x4_t __ret;
  35552.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35553.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35554.   __ret = (float16x4_t) __builtin_neon_vpmin_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35555.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35556.   return __ret;
  35557. }
  35558. #endif
  35559.  
  35560. #ifdef __LITTLE_ENDIAN__
  35561. __ai __attribute__((target("fullfp16"))) float16x8_t vrecpeq_f16(float16x8_t __p0) {
  35562.   float16x8_t __ret;
  35563.   __ret = (float16x8_t) __builtin_neon_vrecpeq_f16((int8x16_t)__p0, 40);
  35564.   return __ret;
  35565. }
  35566. #else
  35567. __ai __attribute__((target("fullfp16"))) float16x8_t vrecpeq_f16(float16x8_t __p0) {
  35568.   float16x8_t __ret;
  35569.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35570.   __ret = (float16x8_t) __builtin_neon_vrecpeq_f16((int8x16_t)__rev0, 40);
  35571.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35572.   return __ret;
  35573. }
  35574. #endif
  35575.  
  35576. #ifdef __LITTLE_ENDIAN__
  35577. __ai __attribute__((target("fullfp16"))) float16x4_t vrecpe_f16(float16x4_t __p0) {
  35578.   float16x4_t __ret;
  35579.   __ret = (float16x4_t) __builtin_neon_vrecpe_f16((int8x8_t)__p0, 8);
  35580.   return __ret;
  35581. }
  35582. #else
  35583. __ai __attribute__((target("fullfp16"))) float16x4_t vrecpe_f16(float16x4_t __p0) {
  35584.   float16x4_t __ret;
  35585.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35586.   __ret = (float16x4_t) __builtin_neon_vrecpe_f16((int8x8_t)__rev0, 8);
  35587.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35588.   return __ret;
  35589. }
  35590. #endif
  35591.  
  35592. #ifdef __LITTLE_ENDIAN__
  35593. __ai __attribute__((target("fullfp16"))) float16x8_t vrecpsq_f16(float16x8_t __p0, float16x8_t __p1) {
  35594.   float16x8_t __ret;
  35595.   __ret = (float16x8_t) __builtin_neon_vrecpsq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  35596.   return __ret;
  35597. }
  35598. #else
  35599. __ai __attribute__((target("fullfp16"))) float16x8_t vrecpsq_f16(float16x8_t __p0, float16x8_t __p1) {
  35600.   float16x8_t __ret;
  35601.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35602.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35603.   __ret = (float16x8_t) __builtin_neon_vrecpsq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  35604.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35605.   return __ret;
  35606. }
  35607. #endif
  35608.  
  35609. #ifdef __LITTLE_ENDIAN__
  35610. __ai __attribute__((target("fullfp16"))) float16x4_t vrecps_f16(float16x4_t __p0, float16x4_t __p1) {
  35611.   float16x4_t __ret;
  35612.   __ret = (float16x4_t) __builtin_neon_vrecps_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  35613.   return __ret;
  35614. }
  35615. #else
  35616. __ai __attribute__((target("fullfp16"))) float16x4_t vrecps_f16(float16x4_t __p0, float16x4_t __p1) {
  35617.   float16x4_t __ret;
  35618.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35619.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35620.   __ret = (float16x4_t) __builtin_neon_vrecps_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35621.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35622.   return __ret;
  35623. }
  35624. #endif
  35625.  
  35626. #ifdef __LITTLE_ENDIAN__
  35627. __ai __attribute__((target("fullfp16"))) float16x8_t vrev64q_f16(float16x8_t __p0) {
  35628.   float16x8_t __ret;
  35629.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
  35630.   return __ret;
  35631. }
  35632. #else
  35633. __ai __attribute__((target("fullfp16"))) float16x8_t vrev64q_f16(float16x8_t __p0) {
  35634.   float16x8_t __ret;
  35635.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35636.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
  35637.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35638.   return __ret;
  35639. }
  35640. #endif
  35641.  
  35642. #ifdef __LITTLE_ENDIAN__
  35643. __ai __attribute__((target("fullfp16"))) float16x4_t vrev64_f16(float16x4_t __p0) {
  35644.   float16x4_t __ret;
  35645.   __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35646.   return __ret;
  35647. }
  35648. #else
  35649. __ai __attribute__((target("fullfp16"))) float16x4_t vrev64_f16(float16x4_t __p0) {
  35650.   float16x4_t __ret;
  35651.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35652.   __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
  35653.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35654.   return __ret;
  35655. }
  35656. #endif
  35657.  
  35658. #ifdef __LITTLE_ENDIAN__
  35659. __ai __attribute__((target("fullfp16"))) float16x8_t vrsqrteq_f16(float16x8_t __p0) {
  35660.   float16x8_t __ret;
  35661.   __ret = (float16x8_t) __builtin_neon_vrsqrteq_f16((int8x16_t)__p0, 40);
  35662.   return __ret;
  35663. }
  35664. #else
  35665. __ai __attribute__((target("fullfp16"))) float16x8_t vrsqrteq_f16(float16x8_t __p0) {
  35666.   float16x8_t __ret;
  35667.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35668.   __ret = (float16x8_t) __builtin_neon_vrsqrteq_f16((int8x16_t)__rev0, 40);
  35669.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35670.   return __ret;
  35671. }
  35672. #endif
  35673.  
  35674. #ifdef __LITTLE_ENDIAN__
  35675. __ai __attribute__((target("fullfp16"))) float16x4_t vrsqrte_f16(float16x4_t __p0) {
  35676.   float16x4_t __ret;
  35677.   __ret = (float16x4_t) __builtin_neon_vrsqrte_f16((int8x8_t)__p0, 8);
  35678.   return __ret;
  35679. }
  35680. #else
  35681. __ai __attribute__((target("fullfp16"))) float16x4_t vrsqrte_f16(float16x4_t __p0) {
  35682.   float16x4_t __ret;
  35683.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35684.   __ret = (float16x4_t) __builtin_neon_vrsqrte_f16((int8x8_t)__rev0, 8);
  35685.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35686.   return __ret;
  35687. }
  35688. #endif
  35689.  
  35690. #ifdef __LITTLE_ENDIAN__
  35691. __ai __attribute__((target("fullfp16"))) float16x8_t vrsqrtsq_f16(float16x8_t __p0, float16x8_t __p1) {
  35692.   float16x8_t __ret;
  35693.   __ret = (float16x8_t) __builtin_neon_vrsqrtsq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  35694.   return __ret;
  35695. }
  35696. #else
  35697. __ai __attribute__((target("fullfp16"))) float16x8_t vrsqrtsq_f16(float16x8_t __p0, float16x8_t __p1) {
  35698.   float16x8_t __ret;
  35699.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35700.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35701.   __ret = (float16x8_t) __builtin_neon_vrsqrtsq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  35702.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35703.   return __ret;
  35704. }
  35705. #endif
  35706.  
  35707. #ifdef __LITTLE_ENDIAN__
  35708. __ai __attribute__((target("fullfp16"))) float16x4_t vrsqrts_f16(float16x4_t __p0, float16x4_t __p1) {
  35709.   float16x4_t __ret;
  35710.   __ret = (float16x4_t) __builtin_neon_vrsqrts_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  35711.   return __ret;
  35712. }
  35713. #else
  35714. __ai __attribute__((target("fullfp16"))) float16x4_t vrsqrts_f16(float16x4_t __p0, float16x4_t __p1) {
  35715.   float16x4_t __ret;
  35716.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35717.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35718.   __ret = (float16x4_t) __builtin_neon_vrsqrts_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35719.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35720.   return __ret;
  35721. }
  35722. #endif
  35723.  
  35724. #ifdef __LITTLE_ENDIAN__
  35725. __ai __attribute__((target("fullfp16"))) float16x8_t vsubq_f16(float16x8_t __p0, float16x8_t __p1) {
  35726.   float16x8_t __ret;
  35727.   __ret = __p0 - __p1;
  35728.   return __ret;
  35729. }
  35730. #else
  35731. __ai __attribute__((target("fullfp16"))) float16x8_t vsubq_f16(float16x8_t __p0, float16x8_t __p1) {
  35732.   float16x8_t __ret;
  35733.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35734.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35735.   __ret = __rev0 - __rev1;
  35736.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  35737.   return __ret;
  35738. }
  35739. #endif
  35740.  
  35741. #ifdef __LITTLE_ENDIAN__
  35742. __ai __attribute__((target("fullfp16"))) float16x4_t vsub_f16(float16x4_t __p0, float16x4_t __p1) {
  35743.   float16x4_t __ret;
  35744.   __ret = __p0 - __p1;
  35745.   return __ret;
  35746. }
  35747. #else
  35748. __ai __attribute__((target("fullfp16"))) float16x4_t vsub_f16(float16x4_t __p0, float16x4_t __p1) {
  35749.   float16x4_t __ret;
  35750.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35751.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35752.   __ret = __rev0 - __rev1;
  35753.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35754.   return __ret;
  35755. }
  35756. #endif
  35757.  
  35758. #ifdef __LITTLE_ENDIAN__
  35759. __ai __attribute__((target("fullfp16"))) float16x8x2_t vtrnq_f16(float16x8_t __p0, float16x8_t __p1) {
  35760.   float16x8x2_t __ret;
  35761.   __builtin_neon_vtrnq_f16(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 40);
  35762.   return __ret;
  35763. }
  35764. #else
  35765. __ai __attribute__((target("fullfp16"))) float16x8x2_t vtrnq_f16(float16x8_t __p0, float16x8_t __p1) {
  35766.   float16x8x2_t __ret;
  35767.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35768.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35769.   __builtin_neon_vtrnq_f16(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  35770.  
  35771.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  35772.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  35773.   return __ret;
  35774. }
  35775. #endif
  35776.  
  35777. #ifdef __LITTLE_ENDIAN__
  35778. __ai __attribute__((target("fullfp16"))) float16x4x2_t vtrn_f16(float16x4_t __p0, float16x4_t __p1) {
  35779.   float16x4x2_t __ret;
  35780.   __builtin_neon_vtrn_f16(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 8);
  35781.   return __ret;
  35782. }
  35783. #else
  35784. __ai __attribute__((target("fullfp16"))) float16x4x2_t vtrn_f16(float16x4_t __p0, float16x4_t __p1) {
  35785.   float16x4x2_t __ret;
  35786.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35787.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35788.   __builtin_neon_vtrn_f16(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35789.  
  35790.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  35791.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  35792.   return __ret;
  35793. }
  35794. #endif
  35795.  
  35796. #ifdef __LITTLE_ENDIAN__
  35797. __ai __attribute__((target("fullfp16"))) float16x8x2_t vuzpq_f16(float16x8_t __p0, float16x8_t __p1) {
  35798.   float16x8x2_t __ret;
  35799.   __builtin_neon_vuzpq_f16(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 40);
  35800.   return __ret;
  35801. }
  35802. #else
  35803. __ai __attribute__((target("fullfp16"))) float16x8x2_t vuzpq_f16(float16x8_t __p0, float16x8_t __p1) {
  35804.   float16x8x2_t __ret;
  35805.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35806.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35807.   __builtin_neon_vuzpq_f16(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  35808.  
  35809.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  35810.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  35811.   return __ret;
  35812. }
  35813. #endif
  35814.  
  35815. #ifdef __LITTLE_ENDIAN__
  35816. __ai __attribute__((target("fullfp16"))) float16x4x2_t vuzp_f16(float16x4_t __p0, float16x4_t __p1) {
  35817.   float16x4x2_t __ret;
  35818.   __builtin_neon_vuzp_f16(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 8);
  35819.   return __ret;
  35820. }
  35821. #else
  35822. __ai __attribute__((target("fullfp16"))) float16x4x2_t vuzp_f16(float16x4_t __p0, float16x4_t __p1) {
  35823.   float16x4x2_t __ret;
  35824.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35825.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35826.   __builtin_neon_vuzp_f16(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35827.  
  35828.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  35829.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  35830.   return __ret;
  35831. }
  35832. #endif
  35833.  
  35834. #ifdef __LITTLE_ENDIAN__
  35835. __ai __attribute__((target("fullfp16"))) float16x8x2_t vzipq_f16(float16x8_t __p0, float16x8_t __p1) {
  35836.   float16x8x2_t __ret;
  35837.   __builtin_neon_vzipq_f16(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 40);
  35838.   return __ret;
  35839. }
  35840. #else
  35841. __ai __attribute__((target("fullfp16"))) float16x8x2_t vzipq_f16(float16x8_t __p0, float16x8_t __p1) {
  35842.   float16x8x2_t __ret;
  35843.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  35844.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35845.   __builtin_neon_vzipq_f16(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  35846.  
  35847.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
  35848.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
  35849.   return __ret;
  35850. }
  35851. #endif
  35852.  
  35853. #ifdef __LITTLE_ENDIAN__
  35854. __ai __attribute__((target("fullfp16"))) float16x4x2_t vzip_f16(float16x4_t __p0, float16x4_t __p1) {
  35855.   float16x4x2_t __ret;
  35856.   __builtin_neon_vzip_f16(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 8);
  35857.   return __ret;
  35858. }
  35859. #else
  35860. __ai __attribute__((target("fullfp16"))) float16x4x2_t vzip_f16(float16x4_t __p0, float16x4_t __p1) {
  35861.   float16x4x2_t __ret;
  35862.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35863.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  35864.   __builtin_neon_vzip_f16(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  35865.  
  35866.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
  35867.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
  35868.   return __ret;
  35869. }
  35870. #endif
  35871.  
  35872. #ifdef __LITTLE_ENDIAN__
  35873. __ai __attribute__((target("i8mm"))) uint32x4_t vmmlaq_u32(uint32x4_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  35874.   uint32x4_t __ret;
  35875.   __ret = (uint32x4_t) __builtin_neon_vmmlaq_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  35876.   return __ret;
  35877. }
  35878. #else
  35879. __ai __attribute__((target("i8mm"))) uint32x4_t vmmlaq_u32(uint32x4_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  35880.   uint32x4_t __ret;
  35881.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35882.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  35883.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  35884.   __ret = (uint32x4_t) __builtin_neon_vmmlaq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  35885.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35886.   return __ret;
  35887. }
  35888. #endif
  35889.  
  35890. #ifdef __LITTLE_ENDIAN__
  35891. __ai __attribute__((target("i8mm"))) int32x4_t vmmlaq_s32(int32x4_t __p0, int8x16_t __p1, int8x16_t __p2) {
  35892.   int32x4_t __ret;
  35893.   __ret = (int32x4_t) __builtin_neon_vmmlaq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  35894.   return __ret;
  35895. }
  35896. #else
  35897. __ai __attribute__((target("i8mm"))) int32x4_t vmmlaq_s32(int32x4_t __p0, int8x16_t __p1, int8x16_t __p2) {
  35898.   int32x4_t __ret;
  35899.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35900.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  35901.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  35902.   __ret = (int32x4_t) __builtin_neon_vmmlaq_s32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  35903.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35904.   return __ret;
  35905. }
  35906. #endif
  35907.  
  35908. #ifdef __LITTLE_ENDIAN__
  35909. __ai __attribute__((target("i8mm"))) int32x4_t vusdotq_s32(int32x4_t __p0, uint8x16_t __p1, int8x16_t __p2) {
  35910.   int32x4_t __ret;
  35911.   __ret = (int32x4_t) __builtin_neon_vusdotq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  35912.   return __ret;
  35913. }
  35914. #else
  35915. __ai __attribute__((target("i8mm"))) int32x4_t vusdotq_s32(int32x4_t __p0, uint8x16_t __p1, int8x16_t __p2) {
  35916.   int32x4_t __ret;
  35917.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  35918.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  35919.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  35920.   __ret = (int32x4_t) __builtin_neon_vusdotq_s32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  35921.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  35922.   return __ret;
  35923. }
  35924. __ai __attribute__((target("i8mm"))) int32x4_t __noswap_vusdotq_s32(int32x4_t __p0, uint8x16_t __p1, int8x16_t __p2) {
  35925.   int32x4_t __ret;
  35926.   __ret = (int32x4_t) __builtin_neon_vusdotq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  35927.   return __ret;
  35928. }
  35929. #endif
  35930.  
  35931. #ifdef __LITTLE_ENDIAN__
  35932. __ai __attribute__((target("i8mm"))) int32x2_t vusdot_s32(int32x2_t __p0, uint8x8_t __p1, int8x8_t __p2) {
  35933.   int32x2_t __ret;
  35934.   __ret = (int32x2_t) __builtin_neon_vusdot_s32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  35935.   return __ret;
  35936. }
  35937. #else
  35938. __ai __attribute__((target("i8mm"))) int32x2_t vusdot_s32(int32x2_t __p0, uint8x8_t __p1, int8x8_t __p2) {
  35939.   int32x2_t __ret;
  35940.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  35941.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  35942.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  35943.   __ret = (int32x2_t) __builtin_neon_vusdot_s32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 2);
  35944.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  35945.   return __ret;
  35946. }
  35947. __ai __attribute__((target("i8mm"))) int32x2_t __noswap_vusdot_s32(int32x2_t __p0, uint8x8_t __p1, int8x8_t __p2) {
  35948.   int32x2_t __ret;
  35949.   __ret = (int32x2_t) __builtin_neon_vusdot_s32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  35950.   return __ret;
  35951. }
  35952. #endif
  35953.  
  35954. #ifdef __LITTLE_ENDIAN__
  35955. #define vusdotq_lane_s32(__p0_157, __p1_157, __p2_157, __p3_157) __extension__ ({ \
  35956.   int32x4_t __ret_157; \
  35957.   int32x4_t __s0_157 = __p0_157; \
  35958.   uint8x16_t __s1_157 = __p1_157; \
  35959.   int8x8_t __s2_157 = __p2_157; \
  35960. int8x8_t __reint_157 = __s2_157; \
  35961.   __ret_157 = vusdotq_s32(__s0_157, __s1_157, (int8x16_t)(splatq_lane_s32(*(int32x2_t *) &__reint_157, __p3_157))); \
  35962.   __ret_157; \
  35963. })
  35964. #else
  35965. #define vusdotq_lane_s32(__p0_158, __p1_158, __p2_158, __p3_158) __extension__ ({ \
  35966.   int32x4_t __ret_158; \
  35967.   int32x4_t __s0_158 = __p0_158; \
  35968.   uint8x16_t __s1_158 = __p1_158; \
  35969.   int8x8_t __s2_158 = __p2_158; \
  35970.   int32x4_t __rev0_158;  __rev0_158 = __builtin_shufflevector(__s0_158, __s0_158, 3, 2, 1, 0); \
  35971.   uint8x16_t __rev1_158;  __rev1_158 = __builtin_shufflevector(__s1_158, __s1_158, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  35972.   int8x8_t __rev2_158;  __rev2_158 = __builtin_shufflevector(__s2_158, __s2_158, 7, 6, 5, 4, 3, 2, 1, 0); \
  35973. int8x8_t __reint_158 = __rev2_158; \
  35974.   __ret_158 = __noswap_vusdotq_s32(__rev0_158, __rev1_158, (int8x16_t)(__noswap_splatq_lane_s32(*(int32x2_t *) &__reint_158, __p3_158))); \
  35975.   __ret_158 = __builtin_shufflevector(__ret_158, __ret_158, 3, 2, 1, 0); \
  35976.   __ret_158; \
  35977. })
  35978. #endif
  35979.  
  35980. #ifdef __LITTLE_ENDIAN__
  35981. #define vusdot_lane_s32(__p0_159, __p1_159, __p2_159, __p3_159) __extension__ ({ \
  35982.   int32x2_t __ret_159; \
  35983.   int32x2_t __s0_159 = __p0_159; \
  35984.   uint8x8_t __s1_159 = __p1_159; \
  35985.   int8x8_t __s2_159 = __p2_159; \
  35986. int8x8_t __reint_159 = __s2_159; \
  35987.   __ret_159 = vusdot_s32(__s0_159, __s1_159, (int8x8_t)(splat_lane_s32(*(int32x2_t *) &__reint_159, __p3_159))); \
  35988.   __ret_159; \
  35989. })
  35990. #else
  35991. #define vusdot_lane_s32(__p0_160, __p1_160, __p2_160, __p3_160) __extension__ ({ \
  35992.   int32x2_t __ret_160; \
  35993.   int32x2_t __s0_160 = __p0_160; \
  35994.   uint8x8_t __s1_160 = __p1_160; \
  35995.   int8x8_t __s2_160 = __p2_160; \
  35996.   int32x2_t __rev0_160;  __rev0_160 = __builtin_shufflevector(__s0_160, __s0_160, 1, 0); \
  35997.   uint8x8_t __rev1_160;  __rev1_160 = __builtin_shufflevector(__s1_160, __s1_160, 7, 6, 5, 4, 3, 2, 1, 0); \
  35998.   int8x8_t __rev2_160;  __rev2_160 = __builtin_shufflevector(__s2_160, __s2_160, 7, 6, 5, 4, 3, 2, 1, 0); \
  35999. int8x8_t __reint_160 = __rev2_160; \
  36000.   __ret_160 = __noswap_vusdot_s32(__rev0_160, __rev1_160, (int8x8_t)(__noswap_splat_lane_s32(*(int32x2_t *) &__reint_160, __p3_160))); \
  36001.   __ret_160 = __builtin_shufflevector(__ret_160, __ret_160, 1, 0); \
  36002.   __ret_160; \
  36003. })
  36004. #endif
  36005.  
  36006. #ifdef __LITTLE_ENDIAN__
  36007. __ai __attribute__((target("i8mm"))) int32x4_t vusmmlaq_s32(int32x4_t __p0, uint8x16_t __p1, int8x16_t __p2) {
  36008.   int32x4_t __ret;
  36009.   __ret = (int32x4_t) __builtin_neon_vusmmlaq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  36010.   return __ret;
  36011. }
  36012. #else
  36013. __ai __attribute__((target("i8mm"))) int32x4_t vusmmlaq_s32(int32x4_t __p0, uint8x16_t __p1, int8x16_t __p2) {
  36014.   int32x4_t __ret;
  36015.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36016.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  36017.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  36018.   __ret = (int32x4_t) __builtin_neon_vusmmlaq_s32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  36019.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36020.   return __ret;
  36021. }
  36022. #endif
  36023.  
  36024. #ifdef __LITTLE_ENDIAN__
  36025. __ai __attribute__((target("v8.1a"))) int32x4_t vqrdmlahq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  36026.   int32x4_t __ret;
  36027.   __ret = (int32x4_t) __builtin_neon_vqrdmlahq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  36028.   return __ret;
  36029. }
  36030. #else
  36031. __ai __attribute__((target("v8.1a"))) int32x4_t vqrdmlahq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  36032.   int32x4_t __ret;
  36033.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36034.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36035.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  36036.   __ret = (int32x4_t) __builtin_neon_vqrdmlahq_s32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  36037.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36038.   return __ret;
  36039. }
  36040. __ai __attribute__((target("v8.1a"))) int32x4_t __noswap_vqrdmlahq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  36041.   int32x4_t __ret;
  36042.   __ret = (int32x4_t) __builtin_neon_vqrdmlahq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  36043.   return __ret;
  36044. }
  36045. #endif
  36046.  
  36047. #ifdef __LITTLE_ENDIAN__
  36048. __ai __attribute__((target("v8.1a"))) int16x8_t vqrdmlahq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  36049.   int16x8_t __ret;
  36050.   __ret = (int16x8_t) __builtin_neon_vqrdmlahq_s16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
  36051.   return __ret;
  36052. }
  36053. #else
  36054. __ai __attribute__((target("v8.1a"))) int16x8_t vqrdmlahq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  36055.   int16x8_t __ret;
  36056.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  36057.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  36058.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  36059.   __ret = (int16x8_t) __builtin_neon_vqrdmlahq_s16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 33);
  36060.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  36061.   return __ret;
  36062. }
  36063. __ai __attribute__((target("v8.1a"))) int16x8_t __noswap_vqrdmlahq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  36064.   int16x8_t __ret;
  36065.   __ret = (int16x8_t) __builtin_neon_vqrdmlahq_s16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
  36066.   return __ret;
  36067. }
  36068. #endif
  36069.  
  36070. #ifdef __LITTLE_ENDIAN__
  36071. __ai __attribute__((target("v8.1a"))) int32x2_t vqrdmlah_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  36072.   int32x2_t __ret;
  36073.   __ret = (int32x2_t) __builtin_neon_vqrdmlah_s32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  36074.   return __ret;
  36075. }
  36076. #else
  36077. __ai __attribute__((target("v8.1a"))) int32x2_t vqrdmlah_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  36078.   int32x2_t __ret;
  36079.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  36080.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  36081.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  36082.   __ret = (int32x2_t) __builtin_neon_vqrdmlah_s32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 2);
  36083.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  36084.   return __ret;
  36085. }
  36086. __ai __attribute__((target("v8.1a"))) int32x2_t __noswap_vqrdmlah_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  36087.   int32x2_t __ret;
  36088.   __ret = (int32x2_t) __builtin_neon_vqrdmlah_s32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  36089.   return __ret;
  36090. }
  36091. #endif
  36092.  
  36093. #ifdef __LITTLE_ENDIAN__
  36094. __ai __attribute__((target("v8.1a"))) int16x4_t vqrdmlah_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  36095.   int16x4_t __ret;
  36096.   __ret = (int16x4_t) __builtin_neon_vqrdmlah_s16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 1);
  36097.   return __ret;
  36098. }
  36099. #else
  36100. __ai __attribute__((target("v8.1a"))) int16x4_t vqrdmlah_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  36101.   int16x4_t __ret;
  36102.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36103.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36104.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  36105.   __ret = (int16x4_t) __builtin_neon_vqrdmlah_s16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 1);
  36106.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36107.   return __ret;
  36108. }
  36109. __ai __attribute__((target("v8.1a"))) int16x4_t __noswap_vqrdmlah_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  36110.   int16x4_t __ret;
  36111.   __ret = (int16x4_t) __builtin_neon_vqrdmlah_s16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 1);
  36112.   return __ret;
  36113. }
  36114. #endif
  36115.  
  36116. #ifdef __LITTLE_ENDIAN__
  36117. #define vqrdmlahq_lane_s32(__p0_161, __p1_161, __p2_161, __p3_161) __extension__ ({ \
  36118.   int32x4_t __ret_161; \
  36119.   int32x4_t __s0_161 = __p0_161; \
  36120.   int32x4_t __s1_161 = __p1_161; \
  36121.   int32x2_t __s2_161 = __p2_161; \
  36122.   __ret_161 = vqrdmlahq_s32(__s0_161, __s1_161, splatq_lane_s32(__s2_161, __p3_161)); \
  36123.   __ret_161; \
  36124. })
  36125. #else
  36126. #define vqrdmlahq_lane_s32(__p0_162, __p1_162, __p2_162, __p3_162) __extension__ ({ \
  36127.   int32x4_t __ret_162; \
  36128.   int32x4_t __s0_162 = __p0_162; \
  36129.   int32x4_t __s1_162 = __p1_162; \
  36130.   int32x2_t __s2_162 = __p2_162; \
  36131.   int32x4_t __rev0_162;  __rev0_162 = __builtin_shufflevector(__s0_162, __s0_162, 3, 2, 1, 0); \
  36132.   int32x4_t __rev1_162;  __rev1_162 = __builtin_shufflevector(__s1_162, __s1_162, 3, 2, 1, 0); \
  36133.   int32x2_t __rev2_162;  __rev2_162 = __builtin_shufflevector(__s2_162, __s2_162, 1, 0); \
  36134.   __ret_162 = __noswap_vqrdmlahq_s32(__rev0_162, __rev1_162, __noswap_splatq_lane_s32(__rev2_162, __p3_162)); \
  36135.   __ret_162 = __builtin_shufflevector(__ret_162, __ret_162, 3, 2, 1, 0); \
  36136.   __ret_162; \
  36137. })
  36138. #endif
  36139.  
  36140. #ifdef __LITTLE_ENDIAN__
  36141. #define vqrdmlahq_lane_s16(__p0_163, __p1_163, __p2_163, __p3_163) __extension__ ({ \
  36142.   int16x8_t __ret_163; \
  36143.   int16x8_t __s0_163 = __p0_163; \
  36144.   int16x8_t __s1_163 = __p1_163; \
  36145.   int16x4_t __s2_163 = __p2_163; \
  36146.   __ret_163 = vqrdmlahq_s16(__s0_163, __s1_163, splatq_lane_s16(__s2_163, __p3_163)); \
  36147.   __ret_163; \
  36148. })
  36149. #else
  36150. #define vqrdmlahq_lane_s16(__p0_164, __p1_164, __p2_164, __p3_164) __extension__ ({ \
  36151.   int16x8_t __ret_164; \
  36152.   int16x8_t __s0_164 = __p0_164; \
  36153.   int16x8_t __s1_164 = __p1_164; \
  36154.   int16x4_t __s2_164 = __p2_164; \
  36155.   int16x8_t __rev0_164;  __rev0_164 = __builtin_shufflevector(__s0_164, __s0_164, 7, 6, 5, 4, 3, 2, 1, 0); \
  36156.   int16x8_t __rev1_164;  __rev1_164 = __builtin_shufflevector(__s1_164, __s1_164, 7, 6, 5, 4, 3, 2, 1, 0); \
  36157.   int16x4_t __rev2_164;  __rev2_164 = __builtin_shufflevector(__s2_164, __s2_164, 3, 2, 1, 0); \
  36158.   __ret_164 = __noswap_vqrdmlahq_s16(__rev0_164, __rev1_164, __noswap_splatq_lane_s16(__rev2_164, __p3_164)); \
  36159.   __ret_164 = __builtin_shufflevector(__ret_164, __ret_164, 7, 6, 5, 4, 3, 2, 1, 0); \
  36160.   __ret_164; \
  36161. })
  36162. #endif
  36163.  
  36164. #ifdef __LITTLE_ENDIAN__
  36165. #define vqrdmlah_lane_s32(__p0_165, __p1_165, __p2_165, __p3_165) __extension__ ({ \
  36166.   int32x2_t __ret_165; \
  36167.   int32x2_t __s0_165 = __p0_165; \
  36168.   int32x2_t __s1_165 = __p1_165; \
  36169.   int32x2_t __s2_165 = __p2_165; \
  36170.   __ret_165 = vqrdmlah_s32(__s0_165, __s1_165, splat_lane_s32(__s2_165, __p3_165)); \
  36171.   __ret_165; \
  36172. })
  36173. #else
  36174. #define vqrdmlah_lane_s32(__p0_166, __p1_166, __p2_166, __p3_166) __extension__ ({ \
  36175.   int32x2_t __ret_166; \
  36176.   int32x2_t __s0_166 = __p0_166; \
  36177.   int32x2_t __s1_166 = __p1_166; \
  36178.   int32x2_t __s2_166 = __p2_166; \
  36179.   int32x2_t __rev0_166;  __rev0_166 = __builtin_shufflevector(__s0_166, __s0_166, 1, 0); \
  36180.   int32x2_t __rev1_166;  __rev1_166 = __builtin_shufflevector(__s1_166, __s1_166, 1, 0); \
  36181.   int32x2_t __rev2_166;  __rev2_166 = __builtin_shufflevector(__s2_166, __s2_166, 1, 0); \
  36182.   __ret_166 = __noswap_vqrdmlah_s32(__rev0_166, __rev1_166, __noswap_splat_lane_s32(__rev2_166, __p3_166)); \
  36183.   __ret_166 = __builtin_shufflevector(__ret_166, __ret_166, 1, 0); \
  36184.   __ret_166; \
  36185. })
  36186. #endif
  36187.  
  36188. #ifdef __LITTLE_ENDIAN__
  36189. #define vqrdmlah_lane_s16(__p0_167, __p1_167, __p2_167, __p3_167) __extension__ ({ \
  36190.   int16x4_t __ret_167; \
  36191.   int16x4_t __s0_167 = __p0_167; \
  36192.   int16x4_t __s1_167 = __p1_167; \
  36193.   int16x4_t __s2_167 = __p2_167; \
  36194.   __ret_167 = vqrdmlah_s16(__s0_167, __s1_167, splat_lane_s16(__s2_167, __p3_167)); \
  36195.   __ret_167; \
  36196. })
  36197. #else
  36198. #define vqrdmlah_lane_s16(__p0_168, __p1_168, __p2_168, __p3_168) __extension__ ({ \
  36199.   int16x4_t __ret_168; \
  36200.   int16x4_t __s0_168 = __p0_168; \
  36201.   int16x4_t __s1_168 = __p1_168; \
  36202.   int16x4_t __s2_168 = __p2_168; \
  36203.   int16x4_t __rev0_168;  __rev0_168 = __builtin_shufflevector(__s0_168, __s0_168, 3, 2, 1, 0); \
  36204.   int16x4_t __rev1_168;  __rev1_168 = __builtin_shufflevector(__s1_168, __s1_168, 3, 2, 1, 0); \
  36205.   int16x4_t __rev2_168;  __rev2_168 = __builtin_shufflevector(__s2_168, __s2_168, 3, 2, 1, 0); \
  36206.   __ret_168 = __noswap_vqrdmlah_s16(__rev0_168, __rev1_168, __noswap_splat_lane_s16(__rev2_168, __p3_168)); \
  36207.   __ret_168 = __builtin_shufflevector(__ret_168, __ret_168, 3, 2, 1, 0); \
  36208.   __ret_168; \
  36209. })
  36210. #endif
  36211.  
  36212. #ifdef __LITTLE_ENDIAN__
  36213. __ai __attribute__((target("v8.1a"))) int32x4_t vqrdmlshq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  36214.   int32x4_t __ret;
  36215.   __ret = (int32x4_t) __builtin_neon_vqrdmlshq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  36216.   return __ret;
  36217. }
  36218. #else
  36219. __ai __attribute__((target("v8.1a"))) int32x4_t vqrdmlshq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  36220.   int32x4_t __ret;
  36221.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36222.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36223.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  36224.   __ret = (int32x4_t) __builtin_neon_vqrdmlshq_s32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  36225.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36226.   return __ret;
  36227. }
  36228. __ai __attribute__((target("v8.1a"))) int32x4_t __noswap_vqrdmlshq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  36229.   int32x4_t __ret;
  36230.   __ret = (int32x4_t) __builtin_neon_vqrdmlshq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  36231.   return __ret;
  36232. }
  36233. #endif
  36234.  
  36235. #ifdef __LITTLE_ENDIAN__
  36236. __ai __attribute__((target("v8.1a"))) int16x8_t vqrdmlshq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  36237.   int16x8_t __ret;
  36238.   __ret = (int16x8_t) __builtin_neon_vqrdmlshq_s16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
  36239.   return __ret;
  36240. }
  36241. #else
  36242. __ai __attribute__((target("v8.1a"))) int16x8_t vqrdmlshq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  36243.   int16x8_t __ret;
  36244.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  36245.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  36246.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  36247.   __ret = (int16x8_t) __builtin_neon_vqrdmlshq_s16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 33);
  36248.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  36249.   return __ret;
  36250. }
  36251. __ai __attribute__((target("v8.1a"))) int16x8_t __noswap_vqrdmlshq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  36252.   int16x8_t __ret;
  36253.   __ret = (int16x8_t) __builtin_neon_vqrdmlshq_s16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
  36254.   return __ret;
  36255. }
  36256. #endif
  36257.  
  36258. #ifdef __LITTLE_ENDIAN__
  36259. __ai __attribute__((target("v8.1a"))) int32x2_t vqrdmlsh_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  36260.   int32x2_t __ret;
  36261.   __ret = (int32x2_t) __builtin_neon_vqrdmlsh_s32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  36262.   return __ret;
  36263. }
  36264. #else
  36265. __ai __attribute__((target("v8.1a"))) int32x2_t vqrdmlsh_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  36266.   int32x2_t __ret;
  36267.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  36268.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  36269.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  36270.   __ret = (int32x2_t) __builtin_neon_vqrdmlsh_s32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 2);
  36271.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  36272.   return __ret;
  36273. }
  36274. __ai __attribute__((target("v8.1a"))) int32x2_t __noswap_vqrdmlsh_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  36275.   int32x2_t __ret;
  36276.   __ret = (int32x2_t) __builtin_neon_vqrdmlsh_s32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
  36277.   return __ret;
  36278. }
  36279. #endif
  36280.  
  36281. #ifdef __LITTLE_ENDIAN__
  36282. __ai __attribute__((target("v8.1a"))) int16x4_t vqrdmlsh_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  36283.   int16x4_t __ret;
  36284.   __ret = (int16x4_t) __builtin_neon_vqrdmlsh_s16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 1);
  36285.   return __ret;
  36286. }
  36287. #else
  36288. __ai __attribute__((target("v8.1a"))) int16x4_t vqrdmlsh_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  36289.   int16x4_t __ret;
  36290.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36291.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36292.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  36293.   __ret = (int16x4_t) __builtin_neon_vqrdmlsh_s16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 1);
  36294.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36295.   return __ret;
  36296. }
  36297. __ai __attribute__((target("v8.1a"))) int16x4_t __noswap_vqrdmlsh_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  36298.   int16x4_t __ret;
  36299.   __ret = (int16x4_t) __builtin_neon_vqrdmlsh_s16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 1);
  36300.   return __ret;
  36301. }
  36302. #endif
  36303.  
  36304. #ifdef __LITTLE_ENDIAN__
  36305. #define vqrdmlshq_lane_s32(__p0_169, __p1_169, __p2_169, __p3_169) __extension__ ({ \
  36306.   int32x4_t __ret_169; \
  36307.   int32x4_t __s0_169 = __p0_169; \
  36308.   int32x4_t __s1_169 = __p1_169; \
  36309.   int32x2_t __s2_169 = __p2_169; \
  36310.   __ret_169 = vqrdmlshq_s32(__s0_169, __s1_169, splatq_lane_s32(__s2_169, __p3_169)); \
  36311.   __ret_169; \
  36312. })
  36313. #else
  36314. #define vqrdmlshq_lane_s32(__p0_170, __p1_170, __p2_170, __p3_170) __extension__ ({ \
  36315.   int32x4_t __ret_170; \
  36316.   int32x4_t __s0_170 = __p0_170; \
  36317.   int32x4_t __s1_170 = __p1_170; \
  36318.   int32x2_t __s2_170 = __p2_170; \
  36319.   int32x4_t __rev0_170;  __rev0_170 = __builtin_shufflevector(__s0_170, __s0_170, 3, 2, 1, 0); \
  36320.   int32x4_t __rev1_170;  __rev1_170 = __builtin_shufflevector(__s1_170, __s1_170, 3, 2, 1, 0); \
  36321.   int32x2_t __rev2_170;  __rev2_170 = __builtin_shufflevector(__s2_170, __s2_170, 1, 0); \
  36322.   __ret_170 = __noswap_vqrdmlshq_s32(__rev0_170, __rev1_170, __noswap_splatq_lane_s32(__rev2_170, __p3_170)); \
  36323.   __ret_170 = __builtin_shufflevector(__ret_170, __ret_170, 3, 2, 1, 0); \
  36324.   __ret_170; \
  36325. })
  36326. #endif
  36327.  
  36328. #ifdef __LITTLE_ENDIAN__
  36329. #define vqrdmlshq_lane_s16(__p0_171, __p1_171, __p2_171, __p3_171) __extension__ ({ \
  36330.   int16x8_t __ret_171; \
  36331.   int16x8_t __s0_171 = __p0_171; \
  36332.   int16x8_t __s1_171 = __p1_171; \
  36333.   int16x4_t __s2_171 = __p2_171; \
  36334.   __ret_171 = vqrdmlshq_s16(__s0_171, __s1_171, splatq_lane_s16(__s2_171, __p3_171)); \
  36335.   __ret_171; \
  36336. })
  36337. #else
  36338. #define vqrdmlshq_lane_s16(__p0_172, __p1_172, __p2_172, __p3_172) __extension__ ({ \
  36339.   int16x8_t __ret_172; \
  36340.   int16x8_t __s0_172 = __p0_172; \
  36341.   int16x8_t __s1_172 = __p1_172; \
  36342.   int16x4_t __s2_172 = __p2_172; \
  36343.   int16x8_t __rev0_172;  __rev0_172 = __builtin_shufflevector(__s0_172, __s0_172, 7, 6, 5, 4, 3, 2, 1, 0); \
  36344.   int16x8_t __rev1_172;  __rev1_172 = __builtin_shufflevector(__s1_172, __s1_172, 7, 6, 5, 4, 3, 2, 1, 0); \
  36345.   int16x4_t __rev2_172;  __rev2_172 = __builtin_shufflevector(__s2_172, __s2_172, 3, 2, 1, 0); \
  36346.   __ret_172 = __noswap_vqrdmlshq_s16(__rev0_172, __rev1_172, __noswap_splatq_lane_s16(__rev2_172, __p3_172)); \
  36347.   __ret_172 = __builtin_shufflevector(__ret_172, __ret_172, 7, 6, 5, 4, 3, 2, 1, 0); \
  36348.   __ret_172; \
  36349. })
  36350. #endif
  36351.  
  36352. #ifdef __LITTLE_ENDIAN__
  36353. #define vqrdmlsh_lane_s32(__p0_173, __p1_173, __p2_173, __p3_173) __extension__ ({ \
  36354.   int32x2_t __ret_173; \
  36355.   int32x2_t __s0_173 = __p0_173; \
  36356.   int32x2_t __s1_173 = __p1_173; \
  36357.   int32x2_t __s2_173 = __p2_173; \
  36358.   __ret_173 = vqrdmlsh_s32(__s0_173, __s1_173, splat_lane_s32(__s2_173, __p3_173)); \
  36359.   __ret_173; \
  36360. })
  36361. #else
  36362. #define vqrdmlsh_lane_s32(__p0_174, __p1_174, __p2_174, __p3_174) __extension__ ({ \
  36363.   int32x2_t __ret_174; \
  36364.   int32x2_t __s0_174 = __p0_174; \
  36365.   int32x2_t __s1_174 = __p1_174; \
  36366.   int32x2_t __s2_174 = __p2_174; \
  36367.   int32x2_t __rev0_174;  __rev0_174 = __builtin_shufflevector(__s0_174, __s0_174, 1, 0); \
  36368.   int32x2_t __rev1_174;  __rev1_174 = __builtin_shufflevector(__s1_174, __s1_174, 1, 0); \
  36369.   int32x2_t __rev2_174;  __rev2_174 = __builtin_shufflevector(__s2_174, __s2_174, 1, 0); \
  36370.   __ret_174 = __noswap_vqrdmlsh_s32(__rev0_174, __rev1_174, __noswap_splat_lane_s32(__rev2_174, __p3_174)); \
  36371.   __ret_174 = __builtin_shufflevector(__ret_174, __ret_174, 1, 0); \
  36372.   __ret_174; \
  36373. })
  36374. #endif
  36375.  
  36376. #ifdef __LITTLE_ENDIAN__
  36377. #define vqrdmlsh_lane_s16(__p0_175, __p1_175, __p2_175, __p3_175) __extension__ ({ \
  36378.   int16x4_t __ret_175; \
  36379.   int16x4_t __s0_175 = __p0_175; \
  36380.   int16x4_t __s1_175 = __p1_175; \
  36381.   int16x4_t __s2_175 = __p2_175; \
  36382.   __ret_175 = vqrdmlsh_s16(__s0_175, __s1_175, splat_lane_s16(__s2_175, __p3_175)); \
  36383.   __ret_175; \
  36384. })
  36385. #else
  36386. #define vqrdmlsh_lane_s16(__p0_176, __p1_176, __p2_176, __p3_176) __extension__ ({ \
  36387.   int16x4_t __ret_176; \
  36388.   int16x4_t __s0_176 = __p0_176; \
  36389.   int16x4_t __s1_176 = __p1_176; \
  36390.   int16x4_t __s2_176 = __p2_176; \
  36391.   int16x4_t __rev0_176;  __rev0_176 = __builtin_shufflevector(__s0_176, __s0_176, 3, 2, 1, 0); \
  36392.   int16x4_t __rev1_176;  __rev1_176 = __builtin_shufflevector(__s1_176, __s1_176, 3, 2, 1, 0); \
  36393.   int16x4_t __rev2_176;  __rev2_176 = __builtin_shufflevector(__s2_176, __s2_176, 3, 2, 1, 0); \
  36394.   __ret_176 = __noswap_vqrdmlsh_s16(__rev0_176, __rev1_176, __noswap_splat_lane_s16(__rev2_176, __p3_176)); \
  36395.   __ret_176 = __builtin_shufflevector(__ret_176, __ret_176, 3, 2, 1, 0); \
  36396.   __ret_176; \
  36397. })
  36398. #endif
  36399.  
  36400. #ifdef __LITTLE_ENDIAN__
  36401. __ai __attribute__((target("v8.3a"))) float32x2_t vcadd_rot270_f32(float32x2_t __p0, float32x2_t __p1) {
  36402.   float32x2_t __ret;
  36403.   __ret = (float32x2_t) __builtin_neon_vcadd_rot270_f32((int8x8_t)__p0, (int8x8_t)__p1, 9);
  36404.   return __ret;
  36405. }
  36406. #else
  36407. __ai __attribute__((target("v8.3a"))) float32x2_t vcadd_rot270_f32(float32x2_t __p0, float32x2_t __p1) {
  36408.   float32x2_t __ret;
  36409.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  36410.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  36411.   __ret = (float32x2_t) __builtin_neon_vcadd_rot270_f32((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  36412.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  36413.   return __ret;
  36414. }
  36415. #endif
  36416.  
  36417. #ifdef __LITTLE_ENDIAN__
  36418. __ai __attribute__((target("v8.3a"))) float32x2_t vcadd_rot90_f32(float32x2_t __p0, float32x2_t __p1) {
  36419.   float32x2_t __ret;
  36420.   __ret = (float32x2_t) __builtin_neon_vcadd_rot90_f32((int8x8_t)__p0, (int8x8_t)__p1, 9);
  36421.   return __ret;
  36422. }
  36423. #else
  36424. __ai __attribute__((target("v8.3a"))) float32x2_t vcadd_rot90_f32(float32x2_t __p0, float32x2_t __p1) {
  36425.   float32x2_t __ret;
  36426.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  36427.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  36428.   __ret = (float32x2_t) __builtin_neon_vcadd_rot90_f32((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  36429.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  36430.   return __ret;
  36431. }
  36432. #endif
  36433.  
  36434. #ifdef __LITTLE_ENDIAN__
  36435. __ai __attribute__((target("v8.3a"))) float32x4_t vcaddq_rot270_f32(float32x4_t __p0, float32x4_t __p1) {
  36436.   float32x4_t __ret;
  36437.   __ret = (float32x4_t) __builtin_neon_vcaddq_rot270_f32((int8x16_t)__p0, (int8x16_t)__p1, 41);
  36438.   return __ret;
  36439. }
  36440. #else
  36441. __ai __attribute__((target("v8.3a"))) float32x4_t vcaddq_rot270_f32(float32x4_t __p0, float32x4_t __p1) {
  36442.   float32x4_t __ret;
  36443.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36444.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36445.   __ret = (float32x4_t) __builtin_neon_vcaddq_rot270_f32((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  36446.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36447.   return __ret;
  36448. }
  36449. #endif
  36450.  
  36451. #ifdef __LITTLE_ENDIAN__
  36452. __ai __attribute__((target("v8.3a"))) float32x4_t vcaddq_rot90_f32(float32x4_t __p0, float32x4_t __p1) {
  36453.   float32x4_t __ret;
  36454.   __ret = (float32x4_t) __builtin_neon_vcaddq_rot90_f32((int8x16_t)__p0, (int8x16_t)__p1, 41);
  36455.   return __ret;
  36456. }
  36457. #else
  36458. __ai __attribute__((target("v8.3a"))) float32x4_t vcaddq_rot90_f32(float32x4_t __p0, float32x4_t __p1) {
  36459.   float32x4_t __ret;
  36460.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36461.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36462.   __ret = (float32x4_t) __builtin_neon_vcaddq_rot90_f32((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  36463.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36464.   return __ret;
  36465. }
  36466. #endif
  36467.  
  36468. #ifdef __LITTLE_ENDIAN__
  36469. __ai __attribute__((target("v8.3a"))) float32x4_t vcmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36470.   float32x4_t __ret;
  36471.   __ret = (float32x4_t) __builtin_neon_vcmlaq_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  36472.   return __ret;
  36473. }
  36474. #else
  36475. __ai __attribute__((target("v8.3a"))) float32x4_t vcmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36476.   float32x4_t __ret;
  36477.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36478.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36479.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  36480.   __ret = (float32x4_t) __builtin_neon_vcmlaq_f32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  36481.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36482.   return __ret;
  36483. }
  36484. __ai __attribute__((target("v8.3a"))) float32x4_t __noswap_vcmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36485.   float32x4_t __ret;
  36486.   __ret = (float32x4_t) __builtin_neon_vcmlaq_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  36487.   return __ret;
  36488. }
  36489. #endif
  36490.  
  36491. #ifdef __LITTLE_ENDIAN__
  36492. __ai __attribute__((target("v8.3a"))) float32x2_t vcmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36493.   float32x2_t __ret;
  36494.   __ret = (float32x2_t) __builtin_neon_vcmla_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  36495.   return __ret;
  36496. }
  36497. #else
  36498. __ai __attribute__((target("v8.3a"))) float32x2_t vcmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36499.   float32x2_t __ret;
  36500.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  36501.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  36502.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  36503.   __ret = (float32x2_t) __builtin_neon_vcmla_f32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  36504.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  36505.   return __ret;
  36506. }
  36507. __ai __attribute__((target("v8.3a"))) float32x2_t __noswap_vcmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36508.   float32x2_t __ret;
  36509.   __ret = (float32x2_t) __builtin_neon_vcmla_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  36510.   return __ret;
  36511. }
  36512. #endif
  36513.  
  36514. #ifdef __LITTLE_ENDIAN__
  36515. #define vcmla_lane_f32(__p0_177, __p1_177, __p2_177, __p3_177) __extension__ ({ \
  36516.   float32x2_t __ret_177; \
  36517.   float32x2_t __s0_177 = __p0_177; \
  36518.   float32x2_t __s1_177 = __p1_177; \
  36519.   float32x2_t __s2_177 = __p2_177; \
  36520. float32x2_t __reint_177 = __s2_177; \
  36521. uint64x1_t __reint1_177 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_177, __p3_177)}; \
  36522.   __ret_177 = vcmla_f32(__s0_177, __s1_177, *(float32x2_t *) &__reint1_177); \
  36523.   __ret_177; \
  36524. })
  36525. #else
  36526. #define vcmla_lane_f32(__p0_178, __p1_178, __p2_178, __p3_178) __extension__ ({ \
  36527.   float32x2_t __ret_178; \
  36528.   float32x2_t __s0_178 = __p0_178; \
  36529.   float32x2_t __s1_178 = __p1_178; \
  36530.   float32x2_t __s2_178 = __p2_178; \
  36531.   float32x2_t __rev0_178;  __rev0_178 = __builtin_shufflevector(__s0_178, __s0_178, 1, 0); \
  36532.   float32x2_t __rev1_178;  __rev1_178 = __builtin_shufflevector(__s1_178, __s1_178, 1, 0); \
  36533.   float32x2_t __rev2_178;  __rev2_178 = __builtin_shufflevector(__s2_178, __s2_178, 1, 0); \
  36534. float32x2_t __reint_178 = __rev2_178; \
  36535. uint64x1_t __reint1_178 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_178, __p3_178)}; \
  36536.   __ret_178 = __noswap_vcmla_f32(__rev0_178, __rev1_178, *(float32x2_t *) &__reint1_178); \
  36537.   __ret_178 = __builtin_shufflevector(__ret_178, __ret_178, 1, 0); \
  36538.   __ret_178; \
  36539. })
  36540. #endif
  36541.  
  36542. #ifdef __LITTLE_ENDIAN__
  36543. #define vcmlaq_lane_f32(__p0_179, __p1_179, __p2_179, __p3_179) __extension__ ({ \
  36544.   float32x4_t __ret_179; \
  36545.   float32x4_t __s0_179 = __p0_179; \
  36546.   float32x4_t __s1_179 = __p1_179; \
  36547.   float32x2_t __s2_179 = __p2_179; \
  36548. float32x2_t __reint_179 = __s2_179; \
  36549. uint64x2_t __reint1_179 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_179, __p3_179), vget_lane_u64(*(uint64x1_t *) &__reint_179, __p3_179)}; \
  36550.   __ret_179 = vcmlaq_f32(__s0_179, __s1_179, *(float32x4_t *) &__reint1_179); \
  36551.   __ret_179; \
  36552. })
  36553. #else
  36554. #define vcmlaq_lane_f32(__p0_180, __p1_180, __p2_180, __p3_180) __extension__ ({ \
  36555.   float32x4_t __ret_180; \
  36556.   float32x4_t __s0_180 = __p0_180; \
  36557.   float32x4_t __s1_180 = __p1_180; \
  36558.   float32x2_t __s2_180 = __p2_180; \
  36559.   float32x4_t __rev0_180;  __rev0_180 = __builtin_shufflevector(__s0_180, __s0_180, 3, 2, 1, 0); \
  36560.   float32x4_t __rev1_180;  __rev1_180 = __builtin_shufflevector(__s1_180, __s1_180, 3, 2, 1, 0); \
  36561.   float32x2_t __rev2_180;  __rev2_180 = __builtin_shufflevector(__s2_180, __s2_180, 1, 0); \
  36562. float32x2_t __reint_180 = __rev2_180; \
  36563. uint64x2_t __reint1_180 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_180, __p3_180), vget_lane_u64(*(uint64x1_t *) &__reint_180, __p3_180)}; \
  36564.   __ret_180 = __noswap_vcmlaq_f32(__rev0_180, __rev1_180, *(float32x4_t *) &__reint1_180); \
  36565.   __ret_180 = __builtin_shufflevector(__ret_180, __ret_180, 3, 2, 1, 0); \
  36566.   __ret_180; \
  36567. })
  36568. #endif
  36569.  
  36570. #ifdef __LITTLE_ENDIAN__
  36571. #define vcmla_laneq_f32(__p0_181, __p1_181, __p2_181, __p3_181) __extension__ ({ \
  36572.   float32x2_t __ret_181; \
  36573.   float32x2_t __s0_181 = __p0_181; \
  36574.   float32x2_t __s1_181 = __p1_181; \
  36575.   float32x4_t __s2_181 = __p2_181; \
  36576. float32x4_t __reint_181 = __s2_181; \
  36577. uint64x1_t __reint1_181 = (uint64x1_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_181, __p3_181)}; \
  36578.   __ret_181 = vcmla_f32(__s0_181, __s1_181, *(float32x2_t *) &__reint1_181); \
  36579.   __ret_181; \
  36580. })
  36581. #else
  36582. #define vcmla_laneq_f32(__p0_182, __p1_182, __p2_182, __p3_182) __extension__ ({ \
  36583.   float32x2_t __ret_182; \
  36584.   float32x2_t __s0_182 = __p0_182; \
  36585.   float32x2_t __s1_182 = __p1_182; \
  36586.   float32x4_t __s2_182 = __p2_182; \
  36587.   float32x2_t __rev0_182;  __rev0_182 = __builtin_shufflevector(__s0_182, __s0_182, 1, 0); \
  36588.   float32x2_t __rev1_182;  __rev1_182 = __builtin_shufflevector(__s1_182, __s1_182, 1, 0); \
  36589.   float32x4_t __rev2_182;  __rev2_182 = __builtin_shufflevector(__s2_182, __s2_182, 3, 2, 1, 0); \
  36590. float32x4_t __reint_182 = __rev2_182; \
  36591. uint64x1_t __reint1_182 = (uint64x1_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_182, __p3_182)}; \
  36592.   __ret_182 = __noswap_vcmla_f32(__rev0_182, __rev1_182, *(float32x2_t *) &__reint1_182); \
  36593.   __ret_182 = __builtin_shufflevector(__ret_182, __ret_182, 1, 0); \
  36594.   __ret_182; \
  36595. })
  36596. #endif
  36597.  
  36598. #ifdef __LITTLE_ENDIAN__
  36599. #define vcmlaq_laneq_f32(__p0_183, __p1_183, __p2_183, __p3_183) __extension__ ({ \
  36600.   float32x4_t __ret_183; \
  36601.   float32x4_t __s0_183 = __p0_183; \
  36602.   float32x4_t __s1_183 = __p1_183; \
  36603.   float32x4_t __s2_183 = __p2_183; \
  36604. float32x4_t __reint_183 = __s2_183; \
  36605. uint64x2_t __reint1_183 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_183, __p3_183), vgetq_lane_u64(*(uint64x2_t *) &__reint_183, __p3_183)}; \
  36606.   __ret_183 = vcmlaq_f32(__s0_183, __s1_183, *(float32x4_t *) &__reint1_183); \
  36607.   __ret_183; \
  36608. })
  36609. #else
  36610. #define vcmlaq_laneq_f32(__p0_184, __p1_184, __p2_184, __p3_184) __extension__ ({ \
  36611.   float32x4_t __ret_184; \
  36612.   float32x4_t __s0_184 = __p0_184; \
  36613.   float32x4_t __s1_184 = __p1_184; \
  36614.   float32x4_t __s2_184 = __p2_184; \
  36615.   float32x4_t __rev0_184;  __rev0_184 = __builtin_shufflevector(__s0_184, __s0_184, 3, 2, 1, 0); \
  36616.   float32x4_t __rev1_184;  __rev1_184 = __builtin_shufflevector(__s1_184, __s1_184, 3, 2, 1, 0); \
  36617.   float32x4_t __rev2_184;  __rev2_184 = __builtin_shufflevector(__s2_184, __s2_184, 3, 2, 1, 0); \
  36618. float32x4_t __reint_184 = __rev2_184; \
  36619. uint64x2_t __reint1_184 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_184, __p3_184), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_184, __p3_184)}; \
  36620.   __ret_184 = __noswap_vcmlaq_f32(__rev0_184, __rev1_184, *(float32x4_t *) &__reint1_184); \
  36621.   __ret_184 = __builtin_shufflevector(__ret_184, __ret_184, 3, 2, 1, 0); \
  36622.   __ret_184; \
  36623. })
  36624. #endif
  36625.  
  36626. #ifdef __LITTLE_ENDIAN__
  36627. __ai __attribute__((target("v8.3a"))) float32x4_t vcmlaq_rot180_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36628.   float32x4_t __ret;
  36629.   __ret = (float32x4_t) __builtin_neon_vcmlaq_rot180_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  36630.   return __ret;
  36631. }
  36632. #else
  36633. __ai __attribute__((target("v8.3a"))) float32x4_t vcmlaq_rot180_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36634.   float32x4_t __ret;
  36635.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36636.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36637.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  36638.   __ret = (float32x4_t) __builtin_neon_vcmlaq_rot180_f32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  36639.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36640.   return __ret;
  36641. }
  36642. __ai __attribute__((target("v8.3a"))) float32x4_t __noswap_vcmlaq_rot180_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36643.   float32x4_t __ret;
  36644.   __ret = (float32x4_t) __builtin_neon_vcmlaq_rot180_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  36645.   return __ret;
  36646. }
  36647. #endif
  36648.  
  36649. #ifdef __LITTLE_ENDIAN__
  36650. __ai __attribute__((target("v8.3a"))) float32x2_t vcmla_rot180_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36651.   float32x2_t __ret;
  36652.   __ret = (float32x2_t) __builtin_neon_vcmla_rot180_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  36653.   return __ret;
  36654. }
  36655. #else
  36656. __ai __attribute__((target("v8.3a"))) float32x2_t vcmla_rot180_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36657.   float32x2_t __ret;
  36658.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  36659.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  36660.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  36661.   __ret = (float32x2_t) __builtin_neon_vcmla_rot180_f32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  36662.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  36663.   return __ret;
  36664. }
  36665. __ai __attribute__((target("v8.3a"))) float32x2_t __noswap_vcmla_rot180_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36666.   float32x2_t __ret;
  36667.   __ret = (float32x2_t) __builtin_neon_vcmla_rot180_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  36668.   return __ret;
  36669. }
  36670. #endif
  36671.  
  36672. #ifdef __LITTLE_ENDIAN__
  36673. #define vcmla_rot180_lane_f32(__p0_185, __p1_185, __p2_185, __p3_185) __extension__ ({ \
  36674.   float32x2_t __ret_185; \
  36675.   float32x2_t __s0_185 = __p0_185; \
  36676.   float32x2_t __s1_185 = __p1_185; \
  36677.   float32x2_t __s2_185 = __p2_185; \
  36678. float32x2_t __reint_185 = __s2_185; \
  36679. uint64x1_t __reint1_185 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_185, __p3_185)}; \
  36680.   __ret_185 = vcmla_rot180_f32(__s0_185, __s1_185, *(float32x2_t *) &__reint1_185); \
  36681.   __ret_185; \
  36682. })
  36683. #else
  36684. #define vcmla_rot180_lane_f32(__p0_186, __p1_186, __p2_186, __p3_186) __extension__ ({ \
  36685.   float32x2_t __ret_186; \
  36686.   float32x2_t __s0_186 = __p0_186; \
  36687.   float32x2_t __s1_186 = __p1_186; \
  36688.   float32x2_t __s2_186 = __p2_186; \
  36689.   float32x2_t __rev0_186;  __rev0_186 = __builtin_shufflevector(__s0_186, __s0_186, 1, 0); \
  36690.   float32x2_t __rev1_186;  __rev1_186 = __builtin_shufflevector(__s1_186, __s1_186, 1, 0); \
  36691.   float32x2_t __rev2_186;  __rev2_186 = __builtin_shufflevector(__s2_186, __s2_186, 1, 0); \
  36692. float32x2_t __reint_186 = __rev2_186; \
  36693. uint64x1_t __reint1_186 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_186, __p3_186)}; \
  36694.   __ret_186 = __noswap_vcmla_rot180_f32(__rev0_186, __rev1_186, *(float32x2_t *) &__reint1_186); \
  36695.   __ret_186 = __builtin_shufflevector(__ret_186, __ret_186, 1, 0); \
  36696.   __ret_186; \
  36697. })
  36698. #endif
  36699.  
  36700. #ifdef __LITTLE_ENDIAN__
  36701. #define vcmlaq_rot180_lane_f32(__p0_187, __p1_187, __p2_187, __p3_187) __extension__ ({ \
  36702.   float32x4_t __ret_187; \
  36703.   float32x4_t __s0_187 = __p0_187; \
  36704.   float32x4_t __s1_187 = __p1_187; \
  36705.   float32x2_t __s2_187 = __p2_187; \
  36706. float32x2_t __reint_187 = __s2_187; \
  36707. uint64x2_t __reint1_187 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_187, __p3_187), vget_lane_u64(*(uint64x1_t *) &__reint_187, __p3_187)}; \
  36708.   __ret_187 = vcmlaq_rot180_f32(__s0_187, __s1_187, *(float32x4_t *) &__reint1_187); \
  36709.   __ret_187; \
  36710. })
  36711. #else
  36712. #define vcmlaq_rot180_lane_f32(__p0_188, __p1_188, __p2_188, __p3_188) __extension__ ({ \
  36713.   float32x4_t __ret_188; \
  36714.   float32x4_t __s0_188 = __p0_188; \
  36715.   float32x4_t __s1_188 = __p1_188; \
  36716.   float32x2_t __s2_188 = __p2_188; \
  36717.   float32x4_t __rev0_188;  __rev0_188 = __builtin_shufflevector(__s0_188, __s0_188, 3, 2, 1, 0); \
  36718.   float32x4_t __rev1_188;  __rev1_188 = __builtin_shufflevector(__s1_188, __s1_188, 3, 2, 1, 0); \
  36719.   float32x2_t __rev2_188;  __rev2_188 = __builtin_shufflevector(__s2_188, __s2_188, 1, 0); \
  36720. float32x2_t __reint_188 = __rev2_188; \
  36721. uint64x2_t __reint1_188 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_188, __p3_188), vget_lane_u64(*(uint64x1_t *) &__reint_188, __p3_188)}; \
  36722.   __ret_188 = __noswap_vcmlaq_rot180_f32(__rev0_188, __rev1_188, *(float32x4_t *) &__reint1_188); \
  36723.   __ret_188 = __builtin_shufflevector(__ret_188, __ret_188, 3, 2, 1, 0); \
  36724.   __ret_188; \
  36725. })
  36726. #endif
  36727.  
  36728. #ifdef __LITTLE_ENDIAN__
  36729. #define vcmla_rot180_laneq_f32(__p0_189, __p1_189, __p2_189, __p3_189) __extension__ ({ \
  36730.   float32x2_t __ret_189; \
  36731.   float32x2_t __s0_189 = __p0_189; \
  36732.   float32x2_t __s1_189 = __p1_189; \
  36733.   float32x4_t __s2_189 = __p2_189; \
  36734. float32x4_t __reint_189 = __s2_189; \
  36735. uint64x1_t __reint1_189 = (uint64x1_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_189, __p3_189)}; \
  36736.   __ret_189 = vcmla_rot180_f32(__s0_189, __s1_189, *(float32x2_t *) &__reint1_189); \
  36737.   __ret_189; \
  36738. })
  36739. #else
  36740. #define vcmla_rot180_laneq_f32(__p0_190, __p1_190, __p2_190, __p3_190) __extension__ ({ \
  36741.   float32x2_t __ret_190; \
  36742.   float32x2_t __s0_190 = __p0_190; \
  36743.   float32x2_t __s1_190 = __p1_190; \
  36744.   float32x4_t __s2_190 = __p2_190; \
  36745.   float32x2_t __rev0_190;  __rev0_190 = __builtin_shufflevector(__s0_190, __s0_190, 1, 0); \
  36746.   float32x2_t __rev1_190;  __rev1_190 = __builtin_shufflevector(__s1_190, __s1_190, 1, 0); \
  36747.   float32x4_t __rev2_190;  __rev2_190 = __builtin_shufflevector(__s2_190, __s2_190, 3, 2, 1, 0); \
  36748. float32x4_t __reint_190 = __rev2_190; \
  36749. uint64x1_t __reint1_190 = (uint64x1_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_190, __p3_190)}; \
  36750.   __ret_190 = __noswap_vcmla_rot180_f32(__rev0_190, __rev1_190, *(float32x2_t *) &__reint1_190); \
  36751.   __ret_190 = __builtin_shufflevector(__ret_190, __ret_190, 1, 0); \
  36752.   __ret_190; \
  36753. })
  36754. #endif
  36755.  
  36756. #ifdef __LITTLE_ENDIAN__
  36757. #define vcmlaq_rot180_laneq_f32(__p0_191, __p1_191, __p2_191, __p3_191) __extension__ ({ \
  36758.   float32x4_t __ret_191; \
  36759.   float32x4_t __s0_191 = __p0_191; \
  36760.   float32x4_t __s1_191 = __p1_191; \
  36761.   float32x4_t __s2_191 = __p2_191; \
  36762. float32x4_t __reint_191 = __s2_191; \
  36763. uint64x2_t __reint1_191 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_191, __p3_191), vgetq_lane_u64(*(uint64x2_t *) &__reint_191, __p3_191)}; \
  36764.   __ret_191 = vcmlaq_rot180_f32(__s0_191, __s1_191, *(float32x4_t *) &__reint1_191); \
  36765.   __ret_191; \
  36766. })
  36767. #else
  36768. #define vcmlaq_rot180_laneq_f32(__p0_192, __p1_192, __p2_192, __p3_192) __extension__ ({ \
  36769.   float32x4_t __ret_192; \
  36770.   float32x4_t __s0_192 = __p0_192; \
  36771.   float32x4_t __s1_192 = __p1_192; \
  36772.   float32x4_t __s2_192 = __p2_192; \
  36773.   float32x4_t __rev0_192;  __rev0_192 = __builtin_shufflevector(__s0_192, __s0_192, 3, 2, 1, 0); \
  36774.   float32x4_t __rev1_192;  __rev1_192 = __builtin_shufflevector(__s1_192, __s1_192, 3, 2, 1, 0); \
  36775.   float32x4_t __rev2_192;  __rev2_192 = __builtin_shufflevector(__s2_192, __s2_192, 3, 2, 1, 0); \
  36776. float32x4_t __reint_192 = __rev2_192; \
  36777. uint64x2_t __reint1_192 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_192, __p3_192), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_192, __p3_192)}; \
  36778.   __ret_192 = __noswap_vcmlaq_rot180_f32(__rev0_192, __rev1_192, *(float32x4_t *) &__reint1_192); \
  36779.   __ret_192 = __builtin_shufflevector(__ret_192, __ret_192, 3, 2, 1, 0); \
  36780.   __ret_192; \
  36781. })
  36782. #endif
  36783.  
  36784. #ifdef __LITTLE_ENDIAN__
  36785. __ai __attribute__((target("v8.3a"))) float32x4_t vcmlaq_rot270_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36786.   float32x4_t __ret;
  36787.   __ret = (float32x4_t) __builtin_neon_vcmlaq_rot270_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  36788.   return __ret;
  36789. }
  36790. #else
  36791. __ai __attribute__((target("v8.3a"))) float32x4_t vcmlaq_rot270_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36792.   float32x4_t __ret;
  36793.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36794.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36795.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  36796.   __ret = (float32x4_t) __builtin_neon_vcmlaq_rot270_f32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  36797.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36798.   return __ret;
  36799. }
  36800. __ai __attribute__((target("v8.3a"))) float32x4_t __noswap_vcmlaq_rot270_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36801.   float32x4_t __ret;
  36802.   __ret = (float32x4_t) __builtin_neon_vcmlaq_rot270_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  36803.   return __ret;
  36804. }
  36805. #endif
  36806.  
  36807. #ifdef __LITTLE_ENDIAN__
  36808. __ai __attribute__((target("v8.3a"))) float32x2_t vcmla_rot270_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36809.   float32x2_t __ret;
  36810.   __ret = (float32x2_t) __builtin_neon_vcmla_rot270_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  36811.   return __ret;
  36812. }
  36813. #else
  36814. __ai __attribute__((target("v8.3a"))) float32x2_t vcmla_rot270_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36815.   float32x2_t __ret;
  36816.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  36817.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  36818.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  36819.   __ret = (float32x2_t) __builtin_neon_vcmla_rot270_f32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  36820.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  36821.   return __ret;
  36822. }
  36823. __ai __attribute__((target("v8.3a"))) float32x2_t __noswap_vcmla_rot270_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36824.   float32x2_t __ret;
  36825.   __ret = (float32x2_t) __builtin_neon_vcmla_rot270_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  36826.   return __ret;
  36827. }
  36828. #endif
  36829.  
  36830. #ifdef __LITTLE_ENDIAN__
  36831. #define vcmla_rot270_lane_f32(__p0_193, __p1_193, __p2_193, __p3_193) __extension__ ({ \
  36832.   float32x2_t __ret_193; \
  36833.   float32x2_t __s0_193 = __p0_193; \
  36834.   float32x2_t __s1_193 = __p1_193; \
  36835.   float32x2_t __s2_193 = __p2_193; \
  36836. float32x2_t __reint_193 = __s2_193; \
  36837. uint64x1_t __reint1_193 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_193, __p3_193)}; \
  36838.   __ret_193 = vcmla_rot270_f32(__s0_193, __s1_193, *(float32x2_t *) &__reint1_193); \
  36839.   __ret_193; \
  36840. })
  36841. #else
  36842. #define vcmla_rot270_lane_f32(__p0_194, __p1_194, __p2_194, __p3_194) __extension__ ({ \
  36843.   float32x2_t __ret_194; \
  36844.   float32x2_t __s0_194 = __p0_194; \
  36845.   float32x2_t __s1_194 = __p1_194; \
  36846.   float32x2_t __s2_194 = __p2_194; \
  36847.   float32x2_t __rev0_194;  __rev0_194 = __builtin_shufflevector(__s0_194, __s0_194, 1, 0); \
  36848.   float32x2_t __rev1_194;  __rev1_194 = __builtin_shufflevector(__s1_194, __s1_194, 1, 0); \
  36849.   float32x2_t __rev2_194;  __rev2_194 = __builtin_shufflevector(__s2_194, __s2_194, 1, 0); \
  36850. float32x2_t __reint_194 = __rev2_194; \
  36851. uint64x1_t __reint1_194 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_194, __p3_194)}; \
  36852.   __ret_194 = __noswap_vcmla_rot270_f32(__rev0_194, __rev1_194, *(float32x2_t *) &__reint1_194); \
  36853.   __ret_194 = __builtin_shufflevector(__ret_194, __ret_194, 1, 0); \
  36854.   __ret_194; \
  36855. })
  36856. #endif
  36857.  
  36858. #ifdef __LITTLE_ENDIAN__
  36859. #define vcmlaq_rot270_lane_f32(__p0_195, __p1_195, __p2_195, __p3_195) __extension__ ({ \
  36860.   float32x4_t __ret_195; \
  36861.   float32x4_t __s0_195 = __p0_195; \
  36862.   float32x4_t __s1_195 = __p1_195; \
  36863.   float32x2_t __s2_195 = __p2_195; \
  36864. float32x2_t __reint_195 = __s2_195; \
  36865. uint64x2_t __reint1_195 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_195, __p3_195), vget_lane_u64(*(uint64x1_t *) &__reint_195, __p3_195)}; \
  36866.   __ret_195 = vcmlaq_rot270_f32(__s0_195, __s1_195, *(float32x4_t *) &__reint1_195); \
  36867.   __ret_195; \
  36868. })
  36869. #else
  36870. #define vcmlaq_rot270_lane_f32(__p0_196, __p1_196, __p2_196, __p3_196) __extension__ ({ \
  36871.   float32x4_t __ret_196; \
  36872.   float32x4_t __s0_196 = __p0_196; \
  36873.   float32x4_t __s1_196 = __p1_196; \
  36874.   float32x2_t __s2_196 = __p2_196; \
  36875.   float32x4_t __rev0_196;  __rev0_196 = __builtin_shufflevector(__s0_196, __s0_196, 3, 2, 1, 0); \
  36876.   float32x4_t __rev1_196;  __rev1_196 = __builtin_shufflevector(__s1_196, __s1_196, 3, 2, 1, 0); \
  36877.   float32x2_t __rev2_196;  __rev2_196 = __builtin_shufflevector(__s2_196, __s2_196, 1, 0); \
  36878. float32x2_t __reint_196 = __rev2_196; \
  36879. uint64x2_t __reint1_196 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_196, __p3_196), vget_lane_u64(*(uint64x1_t *) &__reint_196, __p3_196)}; \
  36880.   __ret_196 = __noswap_vcmlaq_rot270_f32(__rev0_196, __rev1_196, *(float32x4_t *) &__reint1_196); \
  36881.   __ret_196 = __builtin_shufflevector(__ret_196, __ret_196, 3, 2, 1, 0); \
  36882.   __ret_196; \
  36883. })
  36884. #endif
  36885.  
  36886. #ifdef __LITTLE_ENDIAN__
  36887. #define vcmla_rot270_laneq_f32(__p0_197, __p1_197, __p2_197, __p3_197) __extension__ ({ \
  36888.   float32x2_t __ret_197; \
  36889.   float32x2_t __s0_197 = __p0_197; \
  36890.   float32x2_t __s1_197 = __p1_197; \
  36891.   float32x4_t __s2_197 = __p2_197; \
  36892. float32x4_t __reint_197 = __s2_197; \
  36893. uint64x1_t __reint1_197 = (uint64x1_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_197, __p3_197)}; \
  36894.   __ret_197 = vcmla_rot270_f32(__s0_197, __s1_197, *(float32x2_t *) &__reint1_197); \
  36895.   __ret_197; \
  36896. })
  36897. #else
  36898. #define vcmla_rot270_laneq_f32(__p0_198, __p1_198, __p2_198, __p3_198) __extension__ ({ \
  36899.   float32x2_t __ret_198; \
  36900.   float32x2_t __s0_198 = __p0_198; \
  36901.   float32x2_t __s1_198 = __p1_198; \
  36902.   float32x4_t __s2_198 = __p2_198; \
  36903.   float32x2_t __rev0_198;  __rev0_198 = __builtin_shufflevector(__s0_198, __s0_198, 1, 0); \
  36904.   float32x2_t __rev1_198;  __rev1_198 = __builtin_shufflevector(__s1_198, __s1_198, 1, 0); \
  36905.   float32x4_t __rev2_198;  __rev2_198 = __builtin_shufflevector(__s2_198, __s2_198, 3, 2, 1, 0); \
  36906. float32x4_t __reint_198 = __rev2_198; \
  36907. uint64x1_t __reint1_198 = (uint64x1_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_198, __p3_198)}; \
  36908.   __ret_198 = __noswap_vcmla_rot270_f32(__rev0_198, __rev1_198, *(float32x2_t *) &__reint1_198); \
  36909.   __ret_198 = __builtin_shufflevector(__ret_198, __ret_198, 1, 0); \
  36910.   __ret_198; \
  36911. })
  36912. #endif
  36913.  
  36914. #ifdef __LITTLE_ENDIAN__
  36915. #define vcmlaq_rot270_laneq_f32(__p0_199, __p1_199, __p2_199, __p3_199) __extension__ ({ \
  36916.   float32x4_t __ret_199; \
  36917.   float32x4_t __s0_199 = __p0_199; \
  36918.   float32x4_t __s1_199 = __p1_199; \
  36919.   float32x4_t __s2_199 = __p2_199; \
  36920. float32x4_t __reint_199 = __s2_199; \
  36921. uint64x2_t __reint1_199 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_199, __p3_199), vgetq_lane_u64(*(uint64x2_t *) &__reint_199, __p3_199)}; \
  36922.   __ret_199 = vcmlaq_rot270_f32(__s0_199, __s1_199, *(float32x4_t *) &__reint1_199); \
  36923.   __ret_199; \
  36924. })
  36925. #else
  36926. #define vcmlaq_rot270_laneq_f32(__p0_200, __p1_200, __p2_200, __p3_200) __extension__ ({ \
  36927.   float32x4_t __ret_200; \
  36928.   float32x4_t __s0_200 = __p0_200; \
  36929.   float32x4_t __s1_200 = __p1_200; \
  36930.   float32x4_t __s2_200 = __p2_200; \
  36931.   float32x4_t __rev0_200;  __rev0_200 = __builtin_shufflevector(__s0_200, __s0_200, 3, 2, 1, 0); \
  36932.   float32x4_t __rev1_200;  __rev1_200 = __builtin_shufflevector(__s1_200, __s1_200, 3, 2, 1, 0); \
  36933.   float32x4_t __rev2_200;  __rev2_200 = __builtin_shufflevector(__s2_200, __s2_200, 3, 2, 1, 0); \
  36934. float32x4_t __reint_200 = __rev2_200; \
  36935. uint64x2_t __reint1_200 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_200, __p3_200), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_200, __p3_200)}; \
  36936.   __ret_200 = __noswap_vcmlaq_rot270_f32(__rev0_200, __rev1_200, *(float32x4_t *) &__reint1_200); \
  36937.   __ret_200 = __builtin_shufflevector(__ret_200, __ret_200, 3, 2, 1, 0); \
  36938.   __ret_200; \
  36939. })
  36940. #endif
  36941.  
  36942. #ifdef __LITTLE_ENDIAN__
  36943. __ai __attribute__((target("v8.3a"))) float32x4_t vcmlaq_rot90_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36944.   float32x4_t __ret;
  36945.   __ret = (float32x4_t) __builtin_neon_vcmlaq_rot90_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  36946.   return __ret;
  36947. }
  36948. #else
  36949. __ai __attribute__((target("v8.3a"))) float32x4_t vcmlaq_rot90_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36950.   float32x4_t __ret;
  36951.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  36952.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  36953.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  36954.   __ret = (float32x4_t) __builtin_neon_vcmlaq_rot90_f32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  36955.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  36956.   return __ret;
  36957. }
  36958. __ai __attribute__((target("v8.3a"))) float32x4_t __noswap_vcmlaq_rot90_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  36959.   float32x4_t __ret;
  36960.   __ret = (float32x4_t) __builtin_neon_vcmlaq_rot90_f32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  36961.   return __ret;
  36962. }
  36963. #endif
  36964.  
  36965. #ifdef __LITTLE_ENDIAN__
  36966. __ai __attribute__((target("v8.3a"))) float32x2_t vcmla_rot90_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36967.   float32x2_t __ret;
  36968.   __ret = (float32x2_t) __builtin_neon_vcmla_rot90_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  36969.   return __ret;
  36970. }
  36971. #else
  36972. __ai __attribute__((target("v8.3a"))) float32x2_t vcmla_rot90_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36973.   float32x2_t __ret;
  36974.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  36975.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  36976.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  36977.   __ret = (float32x2_t) __builtin_neon_vcmla_rot90_f32((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  36978.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  36979.   return __ret;
  36980. }
  36981. __ai __attribute__((target("v8.3a"))) float32x2_t __noswap_vcmla_rot90_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  36982.   float32x2_t __ret;
  36983.   __ret = (float32x2_t) __builtin_neon_vcmla_rot90_f32((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  36984.   return __ret;
  36985. }
  36986. #endif
  36987.  
  36988. #ifdef __LITTLE_ENDIAN__
  36989. #define vcmla_rot90_lane_f32(__p0_201, __p1_201, __p2_201, __p3_201) __extension__ ({ \
  36990.   float32x2_t __ret_201; \
  36991.   float32x2_t __s0_201 = __p0_201; \
  36992.   float32x2_t __s1_201 = __p1_201; \
  36993.   float32x2_t __s2_201 = __p2_201; \
  36994. float32x2_t __reint_201 = __s2_201; \
  36995. uint64x1_t __reint1_201 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_201, __p3_201)}; \
  36996.   __ret_201 = vcmla_rot90_f32(__s0_201, __s1_201, *(float32x2_t *) &__reint1_201); \
  36997.   __ret_201; \
  36998. })
  36999. #else
  37000. #define vcmla_rot90_lane_f32(__p0_202, __p1_202, __p2_202, __p3_202) __extension__ ({ \
  37001.   float32x2_t __ret_202; \
  37002.   float32x2_t __s0_202 = __p0_202; \
  37003.   float32x2_t __s1_202 = __p1_202; \
  37004.   float32x2_t __s2_202 = __p2_202; \
  37005.   float32x2_t __rev0_202;  __rev0_202 = __builtin_shufflevector(__s0_202, __s0_202, 1, 0); \
  37006.   float32x2_t __rev1_202;  __rev1_202 = __builtin_shufflevector(__s1_202, __s1_202, 1, 0); \
  37007.   float32x2_t __rev2_202;  __rev2_202 = __builtin_shufflevector(__s2_202, __s2_202, 1, 0); \
  37008. float32x2_t __reint_202 = __rev2_202; \
  37009. uint64x1_t __reint1_202 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_202, __p3_202)}; \
  37010.   __ret_202 = __noswap_vcmla_rot90_f32(__rev0_202, __rev1_202, *(float32x2_t *) &__reint1_202); \
  37011.   __ret_202 = __builtin_shufflevector(__ret_202, __ret_202, 1, 0); \
  37012.   __ret_202; \
  37013. })
  37014. #endif
  37015.  
  37016. #ifdef __LITTLE_ENDIAN__
  37017. #define vcmlaq_rot90_lane_f32(__p0_203, __p1_203, __p2_203, __p3_203) __extension__ ({ \
  37018.   float32x4_t __ret_203; \
  37019.   float32x4_t __s0_203 = __p0_203; \
  37020.   float32x4_t __s1_203 = __p1_203; \
  37021.   float32x2_t __s2_203 = __p2_203; \
  37022. float32x2_t __reint_203 = __s2_203; \
  37023. uint64x2_t __reint1_203 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_203, __p3_203), vget_lane_u64(*(uint64x1_t *) &__reint_203, __p3_203)}; \
  37024.   __ret_203 = vcmlaq_rot90_f32(__s0_203, __s1_203, *(float32x4_t *) &__reint1_203); \
  37025.   __ret_203; \
  37026. })
  37027. #else
  37028. #define vcmlaq_rot90_lane_f32(__p0_204, __p1_204, __p2_204, __p3_204) __extension__ ({ \
  37029.   float32x4_t __ret_204; \
  37030.   float32x4_t __s0_204 = __p0_204; \
  37031.   float32x4_t __s1_204 = __p1_204; \
  37032.   float32x2_t __s2_204 = __p2_204; \
  37033.   float32x4_t __rev0_204;  __rev0_204 = __builtin_shufflevector(__s0_204, __s0_204, 3, 2, 1, 0); \
  37034.   float32x4_t __rev1_204;  __rev1_204 = __builtin_shufflevector(__s1_204, __s1_204, 3, 2, 1, 0); \
  37035.   float32x2_t __rev2_204;  __rev2_204 = __builtin_shufflevector(__s2_204, __s2_204, 1, 0); \
  37036. float32x2_t __reint_204 = __rev2_204; \
  37037. uint64x2_t __reint1_204 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_204, __p3_204), vget_lane_u64(*(uint64x1_t *) &__reint_204, __p3_204)}; \
  37038.   __ret_204 = __noswap_vcmlaq_rot90_f32(__rev0_204, __rev1_204, *(float32x4_t *) &__reint1_204); \
  37039.   __ret_204 = __builtin_shufflevector(__ret_204, __ret_204, 3, 2, 1, 0); \
  37040.   __ret_204; \
  37041. })
  37042. #endif
  37043.  
  37044. #ifdef __LITTLE_ENDIAN__
  37045. #define vcmla_rot90_laneq_f32(__p0_205, __p1_205, __p2_205, __p3_205) __extension__ ({ \
  37046.   float32x2_t __ret_205; \
  37047.   float32x2_t __s0_205 = __p0_205; \
  37048.   float32x2_t __s1_205 = __p1_205; \
  37049.   float32x4_t __s2_205 = __p2_205; \
  37050. float32x4_t __reint_205 = __s2_205; \
  37051. uint64x1_t __reint1_205 = (uint64x1_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_205, __p3_205)}; \
  37052.   __ret_205 = vcmla_rot90_f32(__s0_205, __s1_205, *(float32x2_t *) &__reint1_205); \
  37053.   __ret_205; \
  37054. })
  37055. #else
  37056. #define vcmla_rot90_laneq_f32(__p0_206, __p1_206, __p2_206, __p3_206) __extension__ ({ \
  37057.   float32x2_t __ret_206; \
  37058.   float32x2_t __s0_206 = __p0_206; \
  37059.   float32x2_t __s1_206 = __p1_206; \
  37060.   float32x4_t __s2_206 = __p2_206; \
  37061.   float32x2_t __rev0_206;  __rev0_206 = __builtin_shufflevector(__s0_206, __s0_206, 1, 0); \
  37062.   float32x2_t __rev1_206;  __rev1_206 = __builtin_shufflevector(__s1_206, __s1_206, 1, 0); \
  37063.   float32x4_t __rev2_206;  __rev2_206 = __builtin_shufflevector(__s2_206, __s2_206, 3, 2, 1, 0); \
  37064. float32x4_t __reint_206 = __rev2_206; \
  37065. uint64x1_t __reint1_206 = (uint64x1_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_206, __p3_206)}; \
  37066.   __ret_206 = __noswap_vcmla_rot90_f32(__rev0_206, __rev1_206, *(float32x2_t *) &__reint1_206); \
  37067.   __ret_206 = __builtin_shufflevector(__ret_206, __ret_206, 1, 0); \
  37068.   __ret_206; \
  37069. })
  37070. #endif
  37071.  
  37072. #ifdef __LITTLE_ENDIAN__
  37073. #define vcmlaq_rot90_laneq_f32(__p0_207, __p1_207, __p2_207, __p3_207) __extension__ ({ \
  37074.   float32x4_t __ret_207; \
  37075.   float32x4_t __s0_207 = __p0_207; \
  37076.   float32x4_t __s1_207 = __p1_207; \
  37077.   float32x4_t __s2_207 = __p2_207; \
  37078. float32x4_t __reint_207 = __s2_207; \
  37079. uint64x2_t __reint1_207 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_207, __p3_207), vgetq_lane_u64(*(uint64x2_t *) &__reint_207, __p3_207)}; \
  37080.   __ret_207 = vcmlaq_rot90_f32(__s0_207, __s1_207, *(float32x4_t *) &__reint1_207); \
  37081.   __ret_207; \
  37082. })
  37083. #else
  37084. #define vcmlaq_rot90_laneq_f32(__p0_208, __p1_208, __p2_208, __p3_208) __extension__ ({ \
  37085.   float32x4_t __ret_208; \
  37086.   float32x4_t __s0_208 = __p0_208; \
  37087.   float32x4_t __s1_208 = __p1_208; \
  37088.   float32x4_t __s2_208 = __p2_208; \
  37089.   float32x4_t __rev0_208;  __rev0_208 = __builtin_shufflevector(__s0_208, __s0_208, 3, 2, 1, 0); \
  37090.   float32x4_t __rev1_208;  __rev1_208 = __builtin_shufflevector(__s1_208, __s1_208, 3, 2, 1, 0); \
  37091.   float32x4_t __rev2_208;  __rev2_208 = __builtin_shufflevector(__s2_208, __s2_208, 3, 2, 1, 0); \
  37092. float32x4_t __reint_208 = __rev2_208; \
  37093. uint64x2_t __reint1_208 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_208, __p3_208), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_208, __p3_208)}; \
  37094.   __ret_208 = __noswap_vcmlaq_rot90_f32(__rev0_208, __rev1_208, *(float32x4_t *) &__reint1_208); \
  37095.   __ret_208 = __builtin_shufflevector(__ret_208, __ret_208, 3, 2, 1, 0); \
  37096.   __ret_208; \
  37097. })
  37098. #endif
  37099.  
  37100. #ifdef __LITTLE_ENDIAN__
  37101. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcadd_rot270_f16(float16x4_t __p0, float16x4_t __p1) {
  37102.   float16x4_t __ret;
  37103.   __ret = (float16x4_t) __builtin_neon_vcadd_rot270_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  37104.   return __ret;
  37105. }
  37106. #else
  37107. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcadd_rot270_f16(float16x4_t __p0, float16x4_t __p1) {
  37108.   float16x4_t __ret;
  37109.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  37110.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  37111.   __ret = (float16x4_t) __builtin_neon_vcadd_rot270_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  37112.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  37113.   return __ret;
  37114. }
  37115. #endif
  37116.  
  37117. #ifdef __LITTLE_ENDIAN__
  37118. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcadd_rot90_f16(float16x4_t __p0, float16x4_t __p1) {
  37119.   float16x4_t __ret;
  37120.   __ret = (float16x4_t) __builtin_neon_vcadd_rot90_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  37121.   return __ret;
  37122. }
  37123. #else
  37124. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcadd_rot90_f16(float16x4_t __p0, float16x4_t __p1) {
  37125.   float16x4_t __ret;
  37126.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  37127.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  37128.   __ret = (float16x4_t) __builtin_neon_vcadd_rot90_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  37129.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  37130.   return __ret;
  37131. }
  37132. #endif
  37133.  
  37134. #ifdef __LITTLE_ENDIAN__
  37135. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcaddq_rot270_f16(float16x8_t __p0, float16x8_t __p1) {
  37136.   float16x8_t __ret;
  37137.   __ret = (float16x8_t) __builtin_neon_vcaddq_rot270_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  37138.   return __ret;
  37139. }
  37140. #else
  37141. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcaddq_rot270_f16(float16x8_t __p0, float16x8_t __p1) {
  37142.   float16x8_t __ret;
  37143.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  37144.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  37145.   __ret = (float16x8_t) __builtin_neon_vcaddq_rot270_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  37146.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  37147.   return __ret;
  37148. }
  37149. #endif
  37150.  
  37151. #ifdef __LITTLE_ENDIAN__
  37152. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcaddq_rot90_f16(float16x8_t __p0, float16x8_t __p1) {
  37153.   float16x8_t __ret;
  37154.   __ret = (float16x8_t) __builtin_neon_vcaddq_rot90_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  37155.   return __ret;
  37156. }
  37157. #else
  37158. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcaddq_rot90_f16(float16x8_t __p0, float16x8_t __p1) {
  37159.   float16x8_t __ret;
  37160.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  37161.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  37162.   __ret = (float16x8_t) __builtin_neon_vcaddq_rot90_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  37163.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  37164.   return __ret;
  37165. }
  37166. #endif
  37167.  
  37168. #ifdef __LITTLE_ENDIAN__
  37169. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcmlaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37170.   float16x8_t __ret;
  37171.   __ret = (float16x8_t) __builtin_neon_vcmlaq_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  37172.   return __ret;
  37173. }
  37174. #else
  37175. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcmlaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37176.   float16x8_t __ret;
  37177.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  37178.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  37179.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  37180.   __ret = (float16x8_t) __builtin_neon_vcmlaq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40);
  37181.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  37182.   return __ret;
  37183. }
  37184. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t __noswap_vcmlaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37185.   float16x8_t __ret;
  37186.   __ret = (float16x8_t) __builtin_neon_vcmlaq_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  37187.   return __ret;
  37188. }
  37189. #endif
  37190.  
  37191. #ifdef __LITTLE_ENDIAN__
  37192. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcmla_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37193.   float16x4_t __ret;
  37194.   __ret = (float16x4_t) __builtin_neon_vcmla_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  37195.   return __ret;
  37196. }
  37197. #else
  37198. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcmla_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37199.   float16x4_t __ret;
  37200.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  37201.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  37202.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  37203.   __ret = (float16x4_t) __builtin_neon_vcmla_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8);
  37204.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  37205.   return __ret;
  37206. }
  37207. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t __noswap_vcmla_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37208.   float16x4_t __ret;
  37209.   __ret = (float16x4_t) __builtin_neon_vcmla_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  37210.   return __ret;
  37211. }
  37212. #endif
  37213.  
  37214. #ifdef __LITTLE_ENDIAN__
  37215. #define vcmla_lane_f16(__p0_209, __p1_209, __p2_209, __p3_209) __extension__ ({ \
  37216.   float16x4_t __ret_209; \
  37217.   float16x4_t __s0_209 = __p0_209; \
  37218.   float16x4_t __s1_209 = __p1_209; \
  37219.   float16x4_t __s2_209 = __p2_209; \
  37220. float16x4_t __reint_209 = __s2_209; \
  37221. uint32x2_t __reint1_209 = (uint32x2_t) {vget_lane_u32(*(uint32x2_t *) &__reint_209, __p3_209), vget_lane_u32(*(uint32x2_t *) &__reint_209, __p3_209)}; \
  37222.   __ret_209 = vcmla_f16(__s0_209, __s1_209, *(float16x4_t *) &__reint1_209); \
  37223.   __ret_209; \
  37224. })
  37225. #else
  37226. #define vcmla_lane_f16(__p0_210, __p1_210, __p2_210, __p3_210) __extension__ ({ \
  37227.   float16x4_t __ret_210; \
  37228.   float16x4_t __s0_210 = __p0_210; \
  37229.   float16x4_t __s1_210 = __p1_210; \
  37230.   float16x4_t __s2_210 = __p2_210; \
  37231.   float16x4_t __rev0_210;  __rev0_210 = __builtin_shufflevector(__s0_210, __s0_210, 3, 2, 1, 0); \
  37232.   float16x4_t __rev1_210;  __rev1_210 = __builtin_shufflevector(__s1_210, __s1_210, 3, 2, 1, 0); \
  37233.   float16x4_t __rev2_210;  __rev2_210 = __builtin_shufflevector(__s2_210, __s2_210, 3, 2, 1, 0); \
  37234. float16x4_t __reint_210 = __rev2_210; \
  37235. uint32x2_t __reint1_210 = (uint32x2_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_210, __p3_210), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_210, __p3_210)}; \
  37236.   __ret_210 = __noswap_vcmla_f16(__rev0_210, __rev1_210, *(float16x4_t *) &__reint1_210); \
  37237.   __ret_210 = __builtin_shufflevector(__ret_210, __ret_210, 3, 2, 1, 0); \
  37238.   __ret_210; \
  37239. })
  37240. #endif
  37241.  
  37242. #ifdef __LITTLE_ENDIAN__
  37243. #define vcmlaq_lane_f16(__p0_211, __p1_211, __p2_211, __p3_211) __extension__ ({ \
  37244.   float16x8_t __ret_211; \
  37245.   float16x8_t __s0_211 = __p0_211; \
  37246.   float16x8_t __s1_211 = __p1_211; \
  37247.   float16x4_t __s2_211 = __p2_211; \
  37248. float16x4_t __reint_211 = __s2_211; \
  37249. uint32x4_t __reint1_211 = (uint32x4_t) {vget_lane_u32(*(uint32x2_t *) &__reint_211, __p3_211), vget_lane_u32(*(uint32x2_t *) &__reint_211, __p3_211), vget_lane_u32(*(uint32x2_t *) &__reint_211, __p3_211), vget_lane_u32(*(uint32x2_t *) &__reint_211, __p3_211)}; \
  37250.   __ret_211 = vcmlaq_f16(__s0_211, __s1_211, *(float16x8_t *) &__reint1_211); \
  37251.   __ret_211; \
  37252. })
  37253. #else
  37254. #define vcmlaq_lane_f16(__p0_212, __p1_212, __p2_212, __p3_212) __extension__ ({ \
  37255.   float16x8_t __ret_212; \
  37256.   float16x8_t __s0_212 = __p0_212; \
  37257.   float16x8_t __s1_212 = __p1_212; \
  37258.   float16x4_t __s2_212 = __p2_212; \
  37259.   float16x8_t __rev0_212;  __rev0_212 = __builtin_shufflevector(__s0_212, __s0_212, 7, 6, 5, 4, 3, 2, 1, 0); \
  37260.   float16x8_t __rev1_212;  __rev1_212 = __builtin_shufflevector(__s1_212, __s1_212, 7, 6, 5, 4, 3, 2, 1, 0); \
  37261.   float16x4_t __rev2_212;  __rev2_212 = __builtin_shufflevector(__s2_212, __s2_212, 3, 2, 1, 0); \
  37262. float16x4_t __reint_212 = __rev2_212; \
  37263. uint32x4_t __reint1_212 = (uint32x4_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_212, __p3_212), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_212, __p3_212), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_212, __p3_212), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_212, __p3_212)}; \
  37264.   __ret_212 = __noswap_vcmlaq_f16(__rev0_212, __rev1_212, *(float16x8_t *) &__reint1_212); \
  37265.   __ret_212 = __builtin_shufflevector(__ret_212, __ret_212, 7, 6, 5, 4, 3, 2, 1, 0); \
  37266.   __ret_212; \
  37267. })
  37268. #endif
  37269.  
  37270. #ifdef __LITTLE_ENDIAN__
  37271. #define vcmla_laneq_f16(__p0_213, __p1_213, __p2_213, __p3_213) __extension__ ({ \
  37272.   float16x4_t __ret_213; \
  37273.   float16x4_t __s0_213 = __p0_213; \
  37274.   float16x4_t __s1_213 = __p1_213; \
  37275.   float16x8_t __s2_213 = __p2_213; \
  37276. float16x8_t __reint_213 = __s2_213; \
  37277. uint32x2_t __reint1_213 = (uint32x2_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_213, __p3_213), vgetq_lane_u32(*(uint32x4_t *) &__reint_213, __p3_213)}; \
  37278.   __ret_213 = vcmla_f16(__s0_213, __s1_213, *(float16x4_t *) &__reint1_213); \
  37279.   __ret_213; \
  37280. })
  37281. #else
  37282. #define vcmla_laneq_f16(__p0_214, __p1_214, __p2_214, __p3_214) __extension__ ({ \
  37283.   float16x4_t __ret_214; \
  37284.   float16x4_t __s0_214 = __p0_214; \
  37285.   float16x4_t __s1_214 = __p1_214; \
  37286.   float16x8_t __s2_214 = __p2_214; \
  37287.   float16x4_t __rev0_214;  __rev0_214 = __builtin_shufflevector(__s0_214, __s0_214, 3, 2, 1, 0); \
  37288.   float16x4_t __rev1_214;  __rev1_214 = __builtin_shufflevector(__s1_214, __s1_214, 3, 2, 1, 0); \
  37289.   float16x8_t __rev2_214;  __rev2_214 = __builtin_shufflevector(__s2_214, __s2_214, 7, 6, 5, 4, 3, 2, 1, 0); \
  37290. float16x8_t __reint_214 = __rev2_214; \
  37291. uint32x2_t __reint1_214 = (uint32x2_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_214, __p3_214), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_214, __p3_214)}; \
  37292.   __ret_214 = __noswap_vcmla_f16(__rev0_214, __rev1_214, *(float16x4_t *) &__reint1_214); \
  37293.   __ret_214 = __builtin_shufflevector(__ret_214, __ret_214, 3, 2, 1, 0); \
  37294.   __ret_214; \
  37295. })
  37296. #endif
  37297.  
  37298. #ifdef __LITTLE_ENDIAN__
  37299. #define vcmlaq_laneq_f16(__p0_215, __p1_215, __p2_215, __p3_215) __extension__ ({ \
  37300.   float16x8_t __ret_215; \
  37301.   float16x8_t __s0_215 = __p0_215; \
  37302.   float16x8_t __s1_215 = __p1_215; \
  37303.   float16x8_t __s2_215 = __p2_215; \
  37304. float16x8_t __reint_215 = __s2_215; \
  37305. uint32x4_t __reint1_215 = (uint32x4_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_215, __p3_215), vgetq_lane_u32(*(uint32x4_t *) &__reint_215, __p3_215), vgetq_lane_u32(*(uint32x4_t *) &__reint_215, __p3_215), vgetq_lane_u32(*(uint32x4_t *) &__reint_215, __p3_215)}; \
  37306.   __ret_215 = vcmlaq_f16(__s0_215, __s1_215, *(float16x8_t *) &__reint1_215); \
  37307.   __ret_215; \
  37308. })
  37309. #else
  37310. #define vcmlaq_laneq_f16(__p0_216, __p1_216, __p2_216, __p3_216) __extension__ ({ \
  37311.   float16x8_t __ret_216; \
  37312.   float16x8_t __s0_216 = __p0_216; \
  37313.   float16x8_t __s1_216 = __p1_216; \
  37314.   float16x8_t __s2_216 = __p2_216; \
  37315.   float16x8_t __rev0_216;  __rev0_216 = __builtin_shufflevector(__s0_216, __s0_216, 7, 6, 5, 4, 3, 2, 1, 0); \
  37316.   float16x8_t __rev1_216;  __rev1_216 = __builtin_shufflevector(__s1_216, __s1_216, 7, 6, 5, 4, 3, 2, 1, 0); \
  37317.   float16x8_t __rev2_216;  __rev2_216 = __builtin_shufflevector(__s2_216, __s2_216, 7, 6, 5, 4, 3, 2, 1, 0); \
  37318. float16x8_t __reint_216 = __rev2_216; \
  37319. uint32x4_t __reint1_216 = (uint32x4_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_216, __p3_216), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_216, __p3_216), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_216, __p3_216), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_216, __p3_216)}; \
  37320.   __ret_216 = __noswap_vcmlaq_f16(__rev0_216, __rev1_216, *(float16x8_t *) &__reint1_216); \
  37321.   __ret_216 = __builtin_shufflevector(__ret_216, __ret_216, 7, 6, 5, 4, 3, 2, 1, 0); \
  37322.   __ret_216; \
  37323. })
  37324. #endif
  37325.  
  37326. #ifdef __LITTLE_ENDIAN__
  37327. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcmlaq_rot180_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37328.   float16x8_t __ret;
  37329.   __ret = (float16x8_t) __builtin_neon_vcmlaq_rot180_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  37330.   return __ret;
  37331. }
  37332. #else
  37333. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcmlaq_rot180_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37334.   float16x8_t __ret;
  37335.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  37336.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  37337.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  37338.   __ret = (float16x8_t) __builtin_neon_vcmlaq_rot180_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40);
  37339.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  37340.   return __ret;
  37341. }
  37342. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t __noswap_vcmlaq_rot180_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37343.   float16x8_t __ret;
  37344.   __ret = (float16x8_t) __builtin_neon_vcmlaq_rot180_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  37345.   return __ret;
  37346. }
  37347. #endif
  37348.  
  37349. #ifdef __LITTLE_ENDIAN__
  37350. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcmla_rot180_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37351.   float16x4_t __ret;
  37352.   __ret = (float16x4_t) __builtin_neon_vcmla_rot180_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  37353.   return __ret;
  37354. }
  37355. #else
  37356. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcmla_rot180_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37357.   float16x4_t __ret;
  37358.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  37359.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  37360.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  37361.   __ret = (float16x4_t) __builtin_neon_vcmla_rot180_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8);
  37362.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  37363.   return __ret;
  37364. }
  37365. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t __noswap_vcmla_rot180_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37366.   float16x4_t __ret;
  37367.   __ret = (float16x4_t) __builtin_neon_vcmla_rot180_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  37368.   return __ret;
  37369. }
  37370. #endif
  37371.  
  37372. #ifdef __LITTLE_ENDIAN__
  37373. #define vcmla_rot180_lane_f16(__p0_217, __p1_217, __p2_217, __p3_217) __extension__ ({ \
  37374.   float16x4_t __ret_217; \
  37375.   float16x4_t __s0_217 = __p0_217; \
  37376.   float16x4_t __s1_217 = __p1_217; \
  37377.   float16x4_t __s2_217 = __p2_217; \
  37378. float16x4_t __reint_217 = __s2_217; \
  37379. uint32x2_t __reint1_217 = (uint32x2_t) {vget_lane_u32(*(uint32x2_t *) &__reint_217, __p3_217), vget_lane_u32(*(uint32x2_t *) &__reint_217, __p3_217)}; \
  37380.   __ret_217 = vcmla_rot180_f16(__s0_217, __s1_217, *(float16x4_t *) &__reint1_217); \
  37381.   __ret_217; \
  37382. })
  37383. #else
  37384. #define vcmla_rot180_lane_f16(__p0_218, __p1_218, __p2_218, __p3_218) __extension__ ({ \
  37385.   float16x4_t __ret_218; \
  37386.   float16x4_t __s0_218 = __p0_218; \
  37387.   float16x4_t __s1_218 = __p1_218; \
  37388.   float16x4_t __s2_218 = __p2_218; \
  37389.   float16x4_t __rev0_218;  __rev0_218 = __builtin_shufflevector(__s0_218, __s0_218, 3, 2, 1, 0); \
  37390.   float16x4_t __rev1_218;  __rev1_218 = __builtin_shufflevector(__s1_218, __s1_218, 3, 2, 1, 0); \
  37391.   float16x4_t __rev2_218;  __rev2_218 = __builtin_shufflevector(__s2_218, __s2_218, 3, 2, 1, 0); \
  37392. float16x4_t __reint_218 = __rev2_218; \
  37393. uint32x2_t __reint1_218 = (uint32x2_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_218, __p3_218), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_218, __p3_218)}; \
  37394.   __ret_218 = __noswap_vcmla_rot180_f16(__rev0_218, __rev1_218, *(float16x4_t *) &__reint1_218); \
  37395.   __ret_218 = __builtin_shufflevector(__ret_218, __ret_218, 3, 2, 1, 0); \
  37396.   __ret_218; \
  37397. })
  37398. #endif
  37399.  
  37400. #ifdef __LITTLE_ENDIAN__
  37401. #define vcmlaq_rot180_lane_f16(__p0_219, __p1_219, __p2_219, __p3_219) __extension__ ({ \
  37402.   float16x8_t __ret_219; \
  37403.   float16x8_t __s0_219 = __p0_219; \
  37404.   float16x8_t __s1_219 = __p1_219; \
  37405.   float16x4_t __s2_219 = __p2_219; \
  37406. float16x4_t __reint_219 = __s2_219; \
  37407. uint32x4_t __reint1_219 = (uint32x4_t) {vget_lane_u32(*(uint32x2_t *) &__reint_219, __p3_219), vget_lane_u32(*(uint32x2_t *) &__reint_219, __p3_219), vget_lane_u32(*(uint32x2_t *) &__reint_219, __p3_219), vget_lane_u32(*(uint32x2_t *) &__reint_219, __p3_219)}; \
  37408.   __ret_219 = vcmlaq_rot180_f16(__s0_219, __s1_219, *(float16x8_t *) &__reint1_219); \
  37409.   __ret_219; \
  37410. })
  37411. #else
  37412. #define vcmlaq_rot180_lane_f16(__p0_220, __p1_220, __p2_220, __p3_220) __extension__ ({ \
  37413.   float16x8_t __ret_220; \
  37414.   float16x8_t __s0_220 = __p0_220; \
  37415.   float16x8_t __s1_220 = __p1_220; \
  37416.   float16x4_t __s2_220 = __p2_220; \
  37417.   float16x8_t __rev0_220;  __rev0_220 = __builtin_shufflevector(__s0_220, __s0_220, 7, 6, 5, 4, 3, 2, 1, 0); \
  37418.   float16x8_t __rev1_220;  __rev1_220 = __builtin_shufflevector(__s1_220, __s1_220, 7, 6, 5, 4, 3, 2, 1, 0); \
  37419.   float16x4_t __rev2_220;  __rev2_220 = __builtin_shufflevector(__s2_220, __s2_220, 3, 2, 1, 0); \
  37420. float16x4_t __reint_220 = __rev2_220; \
  37421. uint32x4_t __reint1_220 = (uint32x4_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_220, __p3_220), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_220, __p3_220), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_220, __p3_220), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_220, __p3_220)}; \
  37422.   __ret_220 = __noswap_vcmlaq_rot180_f16(__rev0_220, __rev1_220, *(float16x8_t *) &__reint1_220); \
  37423.   __ret_220 = __builtin_shufflevector(__ret_220, __ret_220, 7, 6, 5, 4, 3, 2, 1, 0); \
  37424.   __ret_220; \
  37425. })
  37426. #endif
  37427.  
  37428. #ifdef __LITTLE_ENDIAN__
  37429. #define vcmla_rot180_laneq_f16(__p0_221, __p1_221, __p2_221, __p3_221) __extension__ ({ \
  37430.   float16x4_t __ret_221; \
  37431.   float16x4_t __s0_221 = __p0_221; \
  37432.   float16x4_t __s1_221 = __p1_221; \
  37433.   float16x8_t __s2_221 = __p2_221; \
  37434. float16x8_t __reint_221 = __s2_221; \
  37435. uint32x2_t __reint1_221 = (uint32x2_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_221, __p3_221), vgetq_lane_u32(*(uint32x4_t *) &__reint_221, __p3_221)}; \
  37436.   __ret_221 = vcmla_rot180_f16(__s0_221, __s1_221, *(float16x4_t *) &__reint1_221); \
  37437.   __ret_221; \
  37438. })
  37439. #else
  37440. #define vcmla_rot180_laneq_f16(__p0_222, __p1_222, __p2_222, __p3_222) __extension__ ({ \
  37441.   float16x4_t __ret_222; \
  37442.   float16x4_t __s0_222 = __p0_222; \
  37443.   float16x4_t __s1_222 = __p1_222; \
  37444.   float16x8_t __s2_222 = __p2_222; \
  37445.   float16x4_t __rev0_222;  __rev0_222 = __builtin_shufflevector(__s0_222, __s0_222, 3, 2, 1, 0); \
  37446.   float16x4_t __rev1_222;  __rev1_222 = __builtin_shufflevector(__s1_222, __s1_222, 3, 2, 1, 0); \
  37447.   float16x8_t __rev2_222;  __rev2_222 = __builtin_shufflevector(__s2_222, __s2_222, 7, 6, 5, 4, 3, 2, 1, 0); \
  37448. float16x8_t __reint_222 = __rev2_222; \
  37449. uint32x2_t __reint1_222 = (uint32x2_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_222, __p3_222), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_222, __p3_222)}; \
  37450.   __ret_222 = __noswap_vcmla_rot180_f16(__rev0_222, __rev1_222, *(float16x4_t *) &__reint1_222); \
  37451.   __ret_222 = __builtin_shufflevector(__ret_222, __ret_222, 3, 2, 1, 0); \
  37452.   __ret_222; \
  37453. })
  37454. #endif
  37455.  
  37456. #ifdef __LITTLE_ENDIAN__
  37457. #define vcmlaq_rot180_laneq_f16(__p0_223, __p1_223, __p2_223, __p3_223) __extension__ ({ \
  37458.   float16x8_t __ret_223; \
  37459.   float16x8_t __s0_223 = __p0_223; \
  37460.   float16x8_t __s1_223 = __p1_223; \
  37461.   float16x8_t __s2_223 = __p2_223; \
  37462. float16x8_t __reint_223 = __s2_223; \
  37463. uint32x4_t __reint1_223 = (uint32x4_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_223, __p3_223), vgetq_lane_u32(*(uint32x4_t *) &__reint_223, __p3_223), vgetq_lane_u32(*(uint32x4_t *) &__reint_223, __p3_223), vgetq_lane_u32(*(uint32x4_t *) &__reint_223, __p3_223)}; \
  37464.   __ret_223 = vcmlaq_rot180_f16(__s0_223, __s1_223, *(float16x8_t *) &__reint1_223); \
  37465.   __ret_223; \
  37466. })
  37467. #else
  37468. #define vcmlaq_rot180_laneq_f16(__p0_224, __p1_224, __p2_224, __p3_224) __extension__ ({ \
  37469.   float16x8_t __ret_224; \
  37470.   float16x8_t __s0_224 = __p0_224; \
  37471.   float16x8_t __s1_224 = __p1_224; \
  37472.   float16x8_t __s2_224 = __p2_224; \
  37473.   float16x8_t __rev0_224;  __rev0_224 = __builtin_shufflevector(__s0_224, __s0_224, 7, 6, 5, 4, 3, 2, 1, 0); \
  37474.   float16x8_t __rev1_224;  __rev1_224 = __builtin_shufflevector(__s1_224, __s1_224, 7, 6, 5, 4, 3, 2, 1, 0); \
  37475.   float16x8_t __rev2_224;  __rev2_224 = __builtin_shufflevector(__s2_224, __s2_224, 7, 6, 5, 4, 3, 2, 1, 0); \
  37476. float16x8_t __reint_224 = __rev2_224; \
  37477. uint32x4_t __reint1_224 = (uint32x4_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_224, __p3_224), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_224, __p3_224), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_224, __p3_224), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_224, __p3_224)}; \
  37478.   __ret_224 = __noswap_vcmlaq_rot180_f16(__rev0_224, __rev1_224, *(float16x8_t *) &__reint1_224); \
  37479.   __ret_224 = __builtin_shufflevector(__ret_224, __ret_224, 7, 6, 5, 4, 3, 2, 1, 0); \
  37480.   __ret_224; \
  37481. })
  37482. #endif
  37483.  
  37484. #ifdef __LITTLE_ENDIAN__
  37485. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcmlaq_rot270_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37486.   float16x8_t __ret;
  37487.   __ret = (float16x8_t) __builtin_neon_vcmlaq_rot270_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  37488.   return __ret;
  37489. }
  37490. #else
  37491. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcmlaq_rot270_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37492.   float16x8_t __ret;
  37493.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  37494.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  37495.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  37496.   __ret = (float16x8_t) __builtin_neon_vcmlaq_rot270_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40);
  37497.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  37498.   return __ret;
  37499. }
  37500. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t __noswap_vcmlaq_rot270_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37501.   float16x8_t __ret;
  37502.   __ret = (float16x8_t) __builtin_neon_vcmlaq_rot270_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  37503.   return __ret;
  37504. }
  37505. #endif
  37506.  
  37507. #ifdef __LITTLE_ENDIAN__
  37508. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcmla_rot270_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37509.   float16x4_t __ret;
  37510.   __ret = (float16x4_t) __builtin_neon_vcmla_rot270_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  37511.   return __ret;
  37512. }
  37513. #else
  37514. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcmla_rot270_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37515.   float16x4_t __ret;
  37516.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  37517.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  37518.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  37519.   __ret = (float16x4_t) __builtin_neon_vcmla_rot270_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8);
  37520.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  37521.   return __ret;
  37522. }
  37523. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t __noswap_vcmla_rot270_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37524.   float16x4_t __ret;
  37525.   __ret = (float16x4_t) __builtin_neon_vcmla_rot270_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  37526.   return __ret;
  37527. }
  37528. #endif
  37529.  
  37530. #ifdef __LITTLE_ENDIAN__
  37531. #define vcmla_rot270_lane_f16(__p0_225, __p1_225, __p2_225, __p3_225) __extension__ ({ \
  37532.   float16x4_t __ret_225; \
  37533.   float16x4_t __s0_225 = __p0_225; \
  37534.   float16x4_t __s1_225 = __p1_225; \
  37535.   float16x4_t __s2_225 = __p2_225; \
  37536. float16x4_t __reint_225 = __s2_225; \
  37537. uint32x2_t __reint1_225 = (uint32x2_t) {vget_lane_u32(*(uint32x2_t *) &__reint_225, __p3_225), vget_lane_u32(*(uint32x2_t *) &__reint_225, __p3_225)}; \
  37538.   __ret_225 = vcmla_rot270_f16(__s0_225, __s1_225, *(float16x4_t *) &__reint1_225); \
  37539.   __ret_225; \
  37540. })
  37541. #else
  37542. #define vcmla_rot270_lane_f16(__p0_226, __p1_226, __p2_226, __p3_226) __extension__ ({ \
  37543.   float16x4_t __ret_226; \
  37544.   float16x4_t __s0_226 = __p0_226; \
  37545.   float16x4_t __s1_226 = __p1_226; \
  37546.   float16x4_t __s2_226 = __p2_226; \
  37547.   float16x4_t __rev0_226;  __rev0_226 = __builtin_shufflevector(__s0_226, __s0_226, 3, 2, 1, 0); \
  37548.   float16x4_t __rev1_226;  __rev1_226 = __builtin_shufflevector(__s1_226, __s1_226, 3, 2, 1, 0); \
  37549.   float16x4_t __rev2_226;  __rev2_226 = __builtin_shufflevector(__s2_226, __s2_226, 3, 2, 1, 0); \
  37550. float16x4_t __reint_226 = __rev2_226; \
  37551. uint32x2_t __reint1_226 = (uint32x2_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_226, __p3_226), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_226, __p3_226)}; \
  37552.   __ret_226 = __noswap_vcmla_rot270_f16(__rev0_226, __rev1_226, *(float16x4_t *) &__reint1_226); \
  37553.   __ret_226 = __builtin_shufflevector(__ret_226, __ret_226, 3, 2, 1, 0); \
  37554.   __ret_226; \
  37555. })
  37556. #endif
  37557.  
  37558. #ifdef __LITTLE_ENDIAN__
  37559. #define vcmlaq_rot270_lane_f16(__p0_227, __p1_227, __p2_227, __p3_227) __extension__ ({ \
  37560.   float16x8_t __ret_227; \
  37561.   float16x8_t __s0_227 = __p0_227; \
  37562.   float16x8_t __s1_227 = __p1_227; \
  37563.   float16x4_t __s2_227 = __p2_227; \
  37564. float16x4_t __reint_227 = __s2_227; \
  37565. uint32x4_t __reint1_227 = (uint32x4_t) {vget_lane_u32(*(uint32x2_t *) &__reint_227, __p3_227), vget_lane_u32(*(uint32x2_t *) &__reint_227, __p3_227), vget_lane_u32(*(uint32x2_t *) &__reint_227, __p3_227), vget_lane_u32(*(uint32x2_t *) &__reint_227, __p3_227)}; \
  37566.   __ret_227 = vcmlaq_rot270_f16(__s0_227, __s1_227, *(float16x8_t *) &__reint1_227); \
  37567.   __ret_227; \
  37568. })
  37569. #else
  37570. #define vcmlaq_rot270_lane_f16(__p0_228, __p1_228, __p2_228, __p3_228) __extension__ ({ \
  37571.   float16x8_t __ret_228; \
  37572.   float16x8_t __s0_228 = __p0_228; \
  37573.   float16x8_t __s1_228 = __p1_228; \
  37574.   float16x4_t __s2_228 = __p2_228; \
  37575.   float16x8_t __rev0_228;  __rev0_228 = __builtin_shufflevector(__s0_228, __s0_228, 7, 6, 5, 4, 3, 2, 1, 0); \
  37576.   float16x8_t __rev1_228;  __rev1_228 = __builtin_shufflevector(__s1_228, __s1_228, 7, 6, 5, 4, 3, 2, 1, 0); \
  37577.   float16x4_t __rev2_228;  __rev2_228 = __builtin_shufflevector(__s2_228, __s2_228, 3, 2, 1, 0); \
  37578. float16x4_t __reint_228 = __rev2_228; \
  37579. uint32x4_t __reint1_228 = (uint32x4_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_228, __p3_228), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_228, __p3_228), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_228, __p3_228), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_228, __p3_228)}; \
  37580.   __ret_228 = __noswap_vcmlaq_rot270_f16(__rev0_228, __rev1_228, *(float16x8_t *) &__reint1_228); \
  37581.   __ret_228 = __builtin_shufflevector(__ret_228, __ret_228, 7, 6, 5, 4, 3, 2, 1, 0); \
  37582.   __ret_228; \
  37583. })
  37584. #endif
  37585.  
  37586. #ifdef __LITTLE_ENDIAN__
  37587. #define vcmla_rot270_laneq_f16(__p0_229, __p1_229, __p2_229, __p3_229) __extension__ ({ \
  37588.   float16x4_t __ret_229; \
  37589.   float16x4_t __s0_229 = __p0_229; \
  37590.   float16x4_t __s1_229 = __p1_229; \
  37591.   float16x8_t __s2_229 = __p2_229; \
  37592. float16x8_t __reint_229 = __s2_229; \
  37593. uint32x2_t __reint1_229 = (uint32x2_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_229, __p3_229), vgetq_lane_u32(*(uint32x4_t *) &__reint_229, __p3_229)}; \
  37594.   __ret_229 = vcmla_rot270_f16(__s0_229, __s1_229, *(float16x4_t *) &__reint1_229); \
  37595.   __ret_229; \
  37596. })
  37597. #else
  37598. #define vcmla_rot270_laneq_f16(__p0_230, __p1_230, __p2_230, __p3_230) __extension__ ({ \
  37599.   float16x4_t __ret_230; \
  37600.   float16x4_t __s0_230 = __p0_230; \
  37601.   float16x4_t __s1_230 = __p1_230; \
  37602.   float16x8_t __s2_230 = __p2_230; \
  37603.   float16x4_t __rev0_230;  __rev0_230 = __builtin_shufflevector(__s0_230, __s0_230, 3, 2, 1, 0); \
  37604.   float16x4_t __rev1_230;  __rev1_230 = __builtin_shufflevector(__s1_230, __s1_230, 3, 2, 1, 0); \
  37605.   float16x8_t __rev2_230;  __rev2_230 = __builtin_shufflevector(__s2_230, __s2_230, 7, 6, 5, 4, 3, 2, 1, 0); \
  37606. float16x8_t __reint_230 = __rev2_230; \
  37607. uint32x2_t __reint1_230 = (uint32x2_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_230, __p3_230), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_230, __p3_230)}; \
  37608.   __ret_230 = __noswap_vcmla_rot270_f16(__rev0_230, __rev1_230, *(float16x4_t *) &__reint1_230); \
  37609.   __ret_230 = __builtin_shufflevector(__ret_230, __ret_230, 3, 2, 1, 0); \
  37610.   __ret_230; \
  37611. })
  37612. #endif
  37613.  
  37614. #ifdef __LITTLE_ENDIAN__
  37615. #define vcmlaq_rot270_laneq_f16(__p0_231, __p1_231, __p2_231, __p3_231) __extension__ ({ \
  37616.   float16x8_t __ret_231; \
  37617.   float16x8_t __s0_231 = __p0_231; \
  37618.   float16x8_t __s1_231 = __p1_231; \
  37619.   float16x8_t __s2_231 = __p2_231; \
  37620. float16x8_t __reint_231 = __s2_231; \
  37621. uint32x4_t __reint1_231 = (uint32x4_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_231, __p3_231), vgetq_lane_u32(*(uint32x4_t *) &__reint_231, __p3_231), vgetq_lane_u32(*(uint32x4_t *) &__reint_231, __p3_231), vgetq_lane_u32(*(uint32x4_t *) &__reint_231, __p3_231)}; \
  37622.   __ret_231 = vcmlaq_rot270_f16(__s0_231, __s1_231, *(float16x8_t *) &__reint1_231); \
  37623.   __ret_231; \
  37624. })
  37625. #else
  37626. #define vcmlaq_rot270_laneq_f16(__p0_232, __p1_232, __p2_232, __p3_232) __extension__ ({ \
  37627.   float16x8_t __ret_232; \
  37628.   float16x8_t __s0_232 = __p0_232; \
  37629.   float16x8_t __s1_232 = __p1_232; \
  37630.   float16x8_t __s2_232 = __p2_232; \
  37631.   float16x8_t __rev0_232;  __rev0_232 = __builtin_shufflevector(__s0_232, __s0_232, 7, 6, 5, 4, 3, 2, 1, 0); \
  37632.   float16x8_t __rev1_232;  __rev1_232 = __builtin_shufflevector(__s1_232, __s1_232, 7, 6, 5, 4, 3, 2, 1, 0); \
  37633.   float16x8_t __rev2_232;  __rev2_232 = __builtin_shufflevector(__s2_232, __s2_232, 7, 6, 5, 4, 3, 2, 1, 0); \
  37634. float16x8_t __reint_232 = __rev2_232; \
  37635. uint32x4_t __reint1_232 = (uint32x4_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_232, __p3_232), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_232, __p3_232), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_232, __p3_232), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_232, __p3_232)}; \
  37636.   __ret_232 = __noswap_vcmlaq_rot270_f16(__rev0_232, __rev1_232, *(float16x8_t *) &__reint1_232); \
  37637.   __ret_232 = __builtin_shufflevector(__ret_232, __ret_232, 7, 6, 5, 4, 3, 2, 1, 0); \
  37638.   __ret_232; \
  37639. })
  37640. #endif
  37641.  
  37642. #ifdef __LITTLE_ENDIAN__
  37643. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcmlaq_rot90_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37644.   float16x8_t __ret;
  37645.   __ret = (float16x8_t) __builtin_neon_vcmlaq_rot90_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  37646.   return __ret;
  37647. }
  37648. #else
  37649. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t vcmlaq_rot90_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37650.   float16x8_t __ret;
  37651.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  37652.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  37653.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  37654.   __ret = (float16x8_t) __builtin_neon_vcmlaq_rot90_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40);
  37655.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  37656.   return __ret;
  37657. }
  37658. __ai __attribute__((target("v8.3a,fullfp16"))) float16x8_t __noswap_vcmlaq_rot90_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
  37659.   float16x8_t __ret;
  37660.   __ret = (float16x8_t) __builtin_neon_vcmlaq_rot90_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
  37661.   return __ret;
  37662. }
  37663. #endif
  37664.  
  37665. #ifdef __LITTLE_ENDIAN__
  37666. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcmla_rot90_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37667.   float16x4_t __ret;
  37668.   __ret = (float16x4_t) __builtin_neon_vcmla_rot90_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  37669.   return __ret;
  37670. }
  37671. #else
  37672. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t vcmla_rot90_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37673.   float16x4_t __ret;
  37674.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  37675.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  37676.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  37677.   __ret = (float16x4_t) __builtin_neon_vcmla_rot90_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8);
  37678.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  37679.   return __ret;
  37680. }
  37681. __ai __attribute__((target("v8.3a,fullfp16"))) float16x4_t __noswap_vcmla_rot90_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
  37682.   float16x4_t __ret;
  37683.   __ret = (float16x4_t) __builtin_neon_vcmla_rot90_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
  37684.   return __ret;
  37685. }
  37686. #endif
  37687.  
  37688. #ifdef __LITTLE_ENDIAN__
  37689. #define vcmla_rot90_lane_f16(__p0_233, __p1_233, __p2_233, __p3_233) __extension__ ({ \
  37690.   float16x4_t __ret_233; \
  37691.   float16x4_t __s0_233 = __p0_233; \
  37692.   float16x4_t __s1_233 = __p1_233; \
  37693.   float16x4_t __s2_233 = __p2_233; \
  37694. float16x4_t __reint_233 = __s2_233; \
  37695. uint32x2_t __reint1_233 = (uint32x2_t) {vget_lane_u32(*(uint32x2_t *) &__reint_233, __p3_233), vget_lane_u32(*(uint32x2_t *) &__reint_233, __p3_233)}; \
  37696.   __ret_233 = vcmla_rot90_f16(__s0_233, __s1_233, *(float16x4_t *) &__reint1_233); \
  37697.   __ret_233; \
  37698. })
  37699. #else
  37700. #define vcmla_rot90_lane_f16(__p0_234, __p1_234, __p2_234, __p3_234) __extension__ ({ \
  37701.   float16x4_t __ret_234; \
  37702.   float16x4_t __s0_234 = __p0_234; \
  37703.   float16x4_t __s1_234 = __p1_234; \
  37704.   float16x4_t __s2_234 = __p2_234; \
  37705.   float16x4_t __rev0_234;  __rev0_234 = __builtin_shufflevector(__s0_234, __s0_234, 3, 2, 1, 0); \
  37706.   float16x4_t __rev1_234;  __rev1_234 = __builtin_shufflevector(__s1_234, __s1_234, 3, 2, 1, 0); \
  37707.   float16x4_t __rev2_234;  __rev2_234 = __builtin_shufflevector(__s2_234, __s2_234, 3, 2, 1, 0); \
  37708. float16x4_t __reint_234 = __rev2_234; \
  37709. uint32x2_t __reint1_234 = (uint32x2_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_234, __p3_234), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_234, __p3_234)}; \
  37710.   __ret_234 = __noswap_vcmla_rot90_f16(__rev0_234, __rev1_234, *(float16x4_t *) &__reint1_234); \
  37711.   __ret_234 = __builtin_shufflevector(__ret_234, __ret_234, 3, 2, 1, 0); \
  37712.   __ret_234; \
  37713. })
  37714. #endif
  37715.  
  37716. #ifdef __LITTLE_ENDIAN__
  37717. #define vcmlaq_rot90_lane_f16(__p0_235, __p1_235, __p2_235, __p3_235) __extension__ ({ \
  37718.   float16x8_t __ret_235; \
  37719.   float16x8_t __s0_235 = __p0_235; \
  37720.   float16x8_t __s1_235 = __p1_235; \
  37721.   float16x4_t __s2_235 = __p2_235; \
  37722. float16x4_t __reint_235 = __s2_235; \
  37723. uint32x4_t __reint1_235 = (uint32x4_t) {vget_lane_u32(*(uint32x2_t *) &__reint_235, __p3_235), vget_lane_u32(*(uint32x2_t *) &__reint_235, __p3_235), vget_lane_u32(*(uint32x2_t *) &__reint_235, __p3_235), vget_lane_u32(*(uint32x2_t *) &__reint_235, __p3_235)}; \
  37724.   __ret_235 = vcmlaq_rot90_f16(__s0_235, __s1_235, *(float16x8_t *) &__reint1_235); \
  37725.   __ret_235; \
  37726. })
  37727. #else
  37728. #define vcmlaq_rot90_lane_f16(__p0_236, __p1_236, __p2_236, __p3_236) __extension__ ({ \
  37729.   float16x8_t __ret_236; \
  37730.   float16x8_t __s0_236 = __p0_236; \
  37731.   float16x8_t __s1_236 = __p1_236; \
  37732.   float16x4_t __s2_236 = __p2_236; \
  37733.   float16x8_t __rev0_236;  __rev0_236 = __builtin_shufflevector(__s0_236, __s0_236, 7, 6, 5, 4, 3, 2, 1, 0); \
  37734.   float16x8_t __rev1_236;  __rev1_236 = __builtin_shufflevector(__s1_236, __s1_236, 7, 6, 5, 4, 3, 2, 1, 0); \
  37735.   float16x4_t __rev2_236;  __rev2_236 = __builtin_shufflevector(__s2_236, __s2_236, 3, 2, 1, 0); \
  37736. float16x4_t __reint_236 = __rev2_236; \
  37737. uint32x4_t __reint1_236 = (uint32x4_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_236, __p3_236), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_236, __p3_236), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_236, __p3_236), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_236, __p3_236)}; \
  37738.   __ret_236 = __noswap_vcmlaq_rot90_f16(__rev0_236, __rev1_236, *(float16x8_t *) &__reint1_236); \
  37739.   __ret_236 = __builtin_shufflevector(__ret_236, __ret_236, 7, 6, 5, 4, 3, 2, 1, 0); \
  37740.   __ret_236; \
  37741. })
  37742. #endif
  37743.  
  37744. #ifdef __LITTLE_ENDIAN__
  37745. #define vcmla_rot90_laneq_f16(__p0_237, __p1_237, __p2_237, __p3_237) __extension__ ({ \
  37746.   float16x4_t __ret_237; \
  37747.   float16x4_t __s0_237 = __p0_237; \
  37748.   float16x4_t __s1_237 = __p1_237; \
  37749.   float16x8_t __s2_237 = __p2_237; \
  37750. float16x8_t __reint_237 = __s2_237; \
  37751. uint32x2_t __reint1_237 = (uint32x2_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_237, __p3_237), vgetq_lane_u32(*(uint32x4_t *) &__reint_237, __p3_237)}; \
  37752.   __ret_237 = vcmla_rot90_f16(__s0_237, __s1_237, *(float16x4_t *) &__reint1_237); \
  37753.   __ret_237; \
  37754. })
  37755. #else
  37756. #define vcmla_rot90_laneq_f16(__p0_238, __p1_238, __p2_238, __p3_238) __extension__ ({ \
  37757.   float16x4_t __ret_238; \
  37758.   float16x4_t __s0_238 = __p0_238; \
  37759.   float16x4_t __s1_238 = __p1_238; \
  37760.   float16x8_t __s2_238 = __p2_238; \
  37761.   float16x4_t __rev0_238;  __rev0_238 = __builtin_shufflevector(__s0_238, __s0_238, 3, 2, 1, 0); \
  37762.   float16x4_t __rev1_238;  __rev1_238 = __builtin_shufflevector(__s1_238, __s1_238, 3, 2, 1, 0); \
  37763.   float16x8_t __rev2_238;  __rev2_238 = __builtin_shufflevector(__s2_238, __s2_238, 7, 6, 5, 4, 3, 2, 1, 0); \
  37764. float16x8_t __reint_238 = __rev2_238; \
  37765. uint32x2_t __reint1_238 = (uint32x2_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_238, __p3_238), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_238, __p3_238)}; \
  37766.   __ret_238 = __noswap_vcmla_rot90_f16(__rev0_238, __rev1_238, *(float16x4_t *) &__reint1_238); \
  37767.   __ret_238 = __builtin_shufflevector(__ret_238, __ret_238, 3, 2, 1, 0); \
  37768.   __ret_238; \
  37769. })
  37770. #endif
  37771.  
  37772. #ifdef __LITTLE_ENDIAN__
  37773. #define vcmlaq_rot90_laneq_f16(__p0_239, __p1_239, __p2_239, __p3_239) __extension__ ({ \
  37774.   float16x8_t __ret_239; \
  37775.   float16x8_t __s0_239 = __p0_239; \
  37776.   float16x8_t __s1_239 = __p1_239; \
  37777.   float16x8_t __s2_239 = __p2_239; \
  37778. float16x8_t __reint_239 = __s2_239; \
  37779. uint32x4_t __reint1_239 = (uint32x4_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_239, __p3_239), vgetq_lane_u32(*(uint32x4_t *) &__reint_239, __p3_239), vgetq_lane_u32(*(uint32x4_t *) &__reint_239, __p3_239), vgetq_lane_u32(*(uint32x4_t *) &__reint_239, __p3_239)}; \
  37780.   __ret_239 = vcmlaq_rot90_f16(__s0_239, __s1_239, *(float16x8_t *) &__reint1_239); \
  37781.   __ret_239; \
  37782. })
  37783. #else
  37784. #define vcmlaq_rot90_laneq_f16(__p0_240, __p1_240, __p2_240, __p3_240) __extension__ ({ \
  37785.   float16x8_t __ret_240; \
  37786.   float16x8_t __s0_240 = __p0_240; \
  37787.   float16x8_t __s1_240 = __p1_240; \
  37788.   float16x8_t __s2_240 = __p2_240; \
  37789.   float16x8_t __rev0_240;  __rev0_240 = __builtin_shufflevector(__s0_240, __s0_240, 7, 6, 5, 4, 3, 2, 1, 0); \
  37790.   float16x8_t __rev1_240;  __rev1_240 = __builtin_shufflevector(__s1_240, __s1_240, 7, 6, 5, 4, 3, 2, 1, 0); \
  37791.   float16x8_t __rev2_240;  __rev2_240 = __builtin_shufflevector(__s2_240, __s2_240, 7, 6, 5, 4, 3, 2, 1, 0); \
  37792. float16x8_t __reint_240 = __rev2_240; \
  37793. uint32x4_t __reint1_240 = (uint32x4_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_240, __p3_240), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_240, __p3_240), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_240, __p3_240), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_240, __p3_240)}; \
  37794.   __ret_240 = __noswap_vcmlaq_rot90_f16(__rev0_240, __rev1_240, *(float16x8_t *) &__reint1_240); \
  37795.   __ret_240 = __builtin_shufflevector(__ret_240, __ret_240, 7, 6, 5, 4, 3, 2, 1, 0); \
  37796.   __ret_240; \
  37797. })
  37798. #endif
  37799.  
  37800. #if !defined(__aarch64__)
  37801. #ifdef __LITTLE_ENDIAN__
  37802. #define vqdmulhq_lane_s32(__p0_241, __p1_241, __p2_241) __extension__ ({ \
  37803.   int32x4_t __ret_241; \
  37804.   int32x4_t __s0_241 = __p0_241; \
  37805.   int32x2_t __s1_241 = __p1_241; \
  37806.   __ret_241 = vqdmulhq_s32(__s0_241, splatq_lane_s32(__s1_241, __p2_241)); \
  37807.   __ret_241; \
  37808. })
  37809. #else
  37810. #define vqdmulhq_lane_s32(__p0_242, __p1_242, __p2_242) __extension__ ({ \
  37811.   int32x4_t __ret_242; \
  37812.   int32x4_t __s0_242 = __p0_242; \
  37813.   int32x2_t __s1_242 = __p1_242; \
  37814.   int32x4_t __rev0_242;  __rev0_242 = __builtin_shufflevector(__s0_242, __s0_242, 3, 2, 1, 0); \
  37815.   int32x2_t __rev1_242;  __rev1_242 = __builtin_shufflevector(__s1_242, __s1_242, 1, 0); \
  37816.   __ret_242 = __noswap_vqdmulhq_s32(__rev0_242, __noswap_splatq_lane_s32(__rev1_242, __p2_242)); \
  37817.   __ret_242 = __builtin_shufflevector(__ret_242, __ret_242, 3, 2, 1, 0); \
  37818.   __ret_242; \
  37819. })
  37820. #endif
  37821.  
  37822. #ifdef __LITTLE_ENDIAN__
  37823. #define vqdmulhq_lane_s16(__p0_243, __p1_243, __p2_243) __extension__ ({ \
  37824.   int16x8_t __ret_243; \
  37825.   int16x8_t __s0_243 = __p0_243; \
  37826.   int16x4_t __s1_243 = __p1_243; \
  37827.   __ret_243 = vqdmulhq_s16(__s0_243, splatq_lane_s16(__s1_243, __p2_243)); \
  37828.   __ret_243; \
  37829. })
  37830. #else
  37831. #define vqdmulhq_lane_s16(__p0_244, __p1_244, __p2_244) __extension__ ({ \
  37832.   int16x8_t __ret_244; \
  37833.   int16x8_t __s0_244 = __p0_244; \
  37834.   int16x4_t __s1_244 = __p1_244; \
  37835.   int16x8_t __rev0_244;  __rev0_244 = __builtin_shufflevector(__s0_244, __s0_244, 7, 6, 5, 4, 3, 2, 1, 0); \
  37836.   int16x4_t __rev1_244;  __rev1_244 = __builtin_shufflevector(__s1_244, __s1_244, 3, 2, 1, 0); \
  37837.   __ret_244 = __noswap_vqdmulhq_s16(__rev0_244, __noswap_splatq_lane_s16(__rev1_244, __p2_244)); \
  37838.   __ret_244 = __builtin_shufflevector(__ret_244, __ret_244, 7, 6, 5, 4, 3, 2, 1, 0); \
  37839.   __ret_244; \
  37840. })
  37841. #endif
  37842.  
  37843. #ifdef __LITTLE_ENDIAN__
  37844. #define vqdmulh_lane_s32(__p0_245, __p1_245, __p2_245) __extension__ ({ \
  37845.   int32x2_t __ret_245; \
  37846.   int32x2_t __s0_245 = __p0_245; \
  37847.   int32x2_t __s1_245 = __p1_245; \
  37848.   __ret_245 = vqdmulh_s32(__s0_245, splat_lane_s32(__s1_245, __p2_245)); \
  37849.   __ret_245; \
  37850. })
  37851. #else
  37852. #define vqdmulh_lane_s32(__p0_246, __p1_246, __p2_246) __extension__ ({ \
  37853.   int32x2_t __ret_246; \
  37854.   int32x2_t __s0_246 = __p0_246; \
  37855.   int32x2_t __s1_246 = __p1_246; \
  37856.   int32x2_t __rev0_246;  __rev0_246 = __builtin_shufflevector(__s0_246, __s0_246, 1, 0); \
  37857.   int32x2_t __rev1_246;  __rev1_246 = __builtin_shufflevector(__s1_246, __s1_246, 1, 0); \
  37858.   __ret_246 = __noswap_vqdmulh_s32(__rev0_246, __noswap_splat_lane_s32(__rev1_246, __p2_246)); \
  37859.   __ret_246 = __builtin_shufflevector(__ret_246, __ret_246, 1, 0); \
  37860.   __ret_246; \
  37861. })
  37862. #endif
  37863.  
  37864. #ifdef __LITTLE_ENDIAN__
  37865. #define vqdmulh_lane_s16(__p0_247, __p1_247, __p2_247) __extension__ ({ \
  37866.   int16x4_t __ret_247; \
  37867.   int16x4_t __s0_247 = __p0_247; \
  37868.   int16x4_t __s1_247 = __p1_247; \
  37869.   __ret_247 = vqdmulh_s16(__s0_247, splat_lane_s16(__s1_247, __p2_247)); \
  37870.   __ret_247; \
  37871. })
  37872. #else
  37873. #define vqdmulh_lane_s16(__p0_248, __p1_248, __p2_248) __extension__ ({ \
  37874.   int16x4_t __ret_248; \
  37875.   int16x4_t __s0_248 = __p0_248; \
  37876.   int16x4_t __s1_248 = __p1_248; \
  37877.   int16x4_t __rev0_248;  __rev0_248 = __builtin_shufflevector(__s0_248, __s0_248, 3, 2, 1, 0); \
  37878.   int16x4_t __rev1_248;  __rev1_248 = __builtin_shufflevector(__s1_248, __s1_248, 3, 2, 1, 0); \
  37879.   __ret_248 = __noswap_vqdmulh_s16(__rev0_248, __noswap_splat_lane_s16(__rev1_248, __p2_248)); \
  37880.   __ret_248 = __builtin_shufflevector(__ret_248, __ret_248, 3, 2, 1, 0); \
  37881.   __ret_248; \
  37882. })
  37883. #endif
  37884.  
  37885. #ifdef __LITTLE_ENDIAN__
  37886. #define vqrdmulhq_lane_s32(__p0_249, __p1_249, __p2_249) __extension__ ({ \
  37887.   int32x4_t __ret_249; \
  37888.   int32x4_t __s0_249 = __p0_249; \
  37889.   int32x2_t __s1_249 = __p1_249; \
  37890.   __ret_249 = vqrdmulhq_s32(__s0_249, splatq_lane_s32(__s1_249, __p2_249)); \
  37891.   __ret_249; \
  37892. })
  37893. #else
  37894. #define vqrdmulhq_lane_s32(__p0_250, __p1_250, __p2_250) __extension__ ({ \
  37895.   int32x4_t __ret_250; \
  37896.   int32x4_t __s0_250 = __p0_250; \
  37897.   int32x2_t __s1_250 = __p1_250; \
  37898.   int32x4_t __rev0_250;  __rev0_250 = __builtin_shufflevector(__s0_250, __s0_250, 3, 2, 1, 0); \
  37899.   int32x2_t __rev1_250;  __rev1_250 = __builtin_shufflevector(__s1_250, __s1_250, 1, 0); \
  37900.   __ret_250 = __noswap_vqrdmulhq_s32(__rev0_250, __noswap_splatq_lane_s32(__rev1_250, __p2_250)); \
  37901.   __ret_250 = __builtin_shufflevector(__ret_250, __ret_250, 3, 2, 1, 0); \
  37902.   __ret_250; \
  37903. })
  37904. #endif
  37905.  
  37906. #ifdef __LITTLE_ENDIAN__
  37907. #define vqrdmulhq_lane_s16(__p0_251, __p1_251, __p2_251) __extension__ ({ \
  37908.   int16x8_t __ret_251; \
  37909.   int16x8_t __s0_251 = __p0_251; \
  37910.   int16x4_t __s1_251 = __p1_251; \
  37911.   __ret_251 = vqrdmulhq_s16(__s0_251, splatq_lane_s16(__s1_251, __p2_251)); \
  37912.   __ret_251; \
  37913. })
  37914. #else
  37915. #define vqrdmulhq_lane_s16(__p0_252, __p1_252, __p2_252) __extension__ ({ \
  37916.   int16x8_t __ret_252; \
  37917.   int16x8_t __s0_252 = __p0_252; \
  37918.   int16x4_t __s1_252 = __p1_252; \
  37919.   int16x8_t __rev0_252;  __rev0_252 = __builtin_shufflevector(__s0_252, __s0_252, 7, 6, 5, 4, 3, 2, 1, 0); \
  37920.   int16x4_t __rev1_252;  __rev1_252 = __builtin_shufflevector(__s1_252, __s1_252, 3, 2, 1, 0); \
  37921.   __ret_252 = __noswap_vqrdmulhq_s16(__rev0_252, __noswap_splatq_lane_s16(__rev1_252, __p2_252)); \
  37922.   __ret_252 = __builtin_shufflevector(__ret_252, __ret_252, 7, 6, 5, 4, 3, 2, 1, 0); \
  37923.   __ret_252; \
  37924. })
  37925. #endif
  37926.  
  37927. #ifdef __LITTLE_ENDIAN__
  37928. #define vqrdmulh_lane_s32(__p0_253, __p1_253, __p2_253) __extension__ ({ \
  37929.   int32x2_t __ret_253; \
  37930.   int32x2_t __s0_253 = __p0_253; \
  37931.   int32x2_t __s1_253 = __p1_253; \
  37932.   __ret_253 = vqrdmulh_s32(__s0_253, splat_lane_s32(__s1_253, __p2_253)); \
  37933.   __ret_253; \
  37934. })
  37935. #else
  37936. #define vqrdmulh_lane_s32(__p0_254, __p1_254, __p2_254) __extension__ ({ \
  37937.   int32x2_t __ret_254; \
  37938.   int32x2_t __s0_254 = __p0_254; \
  37939.   int32x2_t __s1_254 = __p1_254; \
  37940.   int32x2_t __rev0_254;  __rev0_254 = __builtin_shufflevector(__s0_254, __s0_254, 1, 0); \
  37941.   int32x2_t __rev1_254;  __rev1_254 = __builtin_shufflevector(__s1_254, __s1_254, 1, 0); \
  37942.   __ret_254 = __noswap_vqrdmulh_s32(__rev0_254, __noswap_splat_lane_s32(__rev1_254, __p2_254)); \
  37943.   __ret_254 = __builtin_shufflevector(__ret_254, __ret_254, 1, 0); \
  37944.   __ret_254; \
  37945. })
  37946. #endif
  37947.  
  37948. #ifdef __LITTLE_ENDIAN__
  37949. #define vqrdmulh_lane_s16(__p0_255, __p1_255, __p2_255) __extension__ ({ \
  37950.   int16x4_t __ret_255; \
  37951.   int16x4_t __s0_255 = __p0_255; \
  37952.   int16x4_t __s1_255 = __p1_255; \
  37953.   __ret_255 = vqrdmulh_s16(__s0_255, splat_lane_s16(__s1_255, __p2_255)); \
  37954.   __ret_255; \
  37955. })
  37956. #else
  37957. #define vqrdmulh_lane_s16(__p0_256, __p1_256, __p2_256) __extension__ ({ \
  37958.   int16x4_t __ret_256; \
  37959.   int16x4_t __s0_256 = __p0_256; \
  37960.   int16x4_t __s1_256 = __p1_256; \
  37961.   int16x4_t __rev0_256;  __rev0_256 = __builtin_shufflevector(__s0_256, __s0_256, 3, 2, 1, 0); \
  37962.   int16x4_t __rev1_256;  __rev1_256 = __builtin_shufflevector(__s1_256, __s1_256, 3, 2, 1, 0); \
  37963.   __ret_256 = __noswap_vqrdmulh_s16(__rev0_256, __noswap_splat_lane_s16(__rev1_256, __p2_256)); \
  37964.   __ret_256 = __builtin_shufflevector(__ret_256, __ret_256, 3, 2, 1, 0); \
  37965.   __ret_256; \
  37966. })
  37967. #endif
  37968.  
  37969. __ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
  37970.   poly8x8_t __ret;
  37971.   __ret = (poly8x8_t)(__p0);
  37972.   return __ret;
  37973. }
  37974. __ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
  37975.   poly8x8_t __ret;
  37976.   __ret = (poly8x8_t)(__p0);
  37977.   return __ret;
  37978. }
  37979. __ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
  37980.   poly8x8_t __ret;
  37981.   __ret = (poly8x8_t)(__p0);
  37982.   return __ret;
  37983. }
  37984. __ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
  37985.   poly8x8_t __ret;
  37986.   __ret = (poly8x8_t)(__p0);
  37987.   return __ret;
  37988. }
  37989. __ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
  37990.   poly8x8_t __ret;
  37991.   __ret = (poly8x8_t)(__p0);
  37992.   return __ret;
  37993. }
  37994. __ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
  37995.   poly8x8_t __ret;
  37996.   __ret = (poly8x8_t)(__p0);
  37997.   return __ret;
  37998. }
  37999. __ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
  38000.   poly8x8_t __ret;
  38001.   __ret = (poly8x8_t)(__p0);
  38002.   return __ret;
  38003. }
  38004. __ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
  38005.   poly8x8_t __ret;
  38006.   __ret = (poly8x8_t)(__p0);
  38007.   return __ret;
  38008. }
  38009. __ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
  38010.   poly8x8_t __ret;
  38011.   __ret = (poly8x8_t)(__p0);
  38012.   return __ret;
  38013. }
  38014. __ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
  38015.   poly8x8_t __ret;
  38016.   __ret = (poly8x8_t)(__p0);
  38017.   return __ret;
  38018. }
  38019. __ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
  38020.   poly8x8_t __ret;
  38021.   __ret = (poly8x8_t)(__p0);
  38022.   return __ret;
  38023. }
  38024. __ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
  38025.   poly16x4_t __ret;
  38026.   __ret = (poly16x4_t)(__p0);
  38027.   return __ret;
  38028. }
  38029. __ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
  38030.   poly16x4_t __ret;
  38031.   __ret = (poly16x4_t)(__p0);
  38032.   return __ret;
  38033. }
  38034. __ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
  38035.   poly16x4_t __ret;
  38036.   __ret = (poly16x4_t)(__p0);
  38037.   return __ret;
  38038. }
  38039. __ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
  38040.   poly16x4_t __ret;
  38041.   __ret = (poly16x4_t)(__p0);
  38042.   return __ret;
  38043. }
  38044. __ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
  38045.   poly16x4_t __ret;
  38046.   __ret = (poly16x4_t)(__p0);
  38047.   return __ret;
  38048. }
  38049. __ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
  38050.   poly16x4_t __ret;
  38051.   __ret = (poly16x4_t)(__p0);
  38052.   return __ret;
  38053. }
  38054. __ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
  38055.   poly16x4_t __ret;
  38056.   __ret = (poly16x4_t)(__p0);
  38057.   return __ret;
  38058. }
  38059. __ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
  38060.   poly16x4_t __ret;
  38061.   __ret = (poly16x4_t)(__p0);
  38062.   return __ret;
  38063. }
  38064. __ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
  38065.   poly16x4_t __ret;
  38066.   __ret = (poly16x4_t)(__p0);
  38067.   return __ret;
  38068. }
  38069. __ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
  38070.   poly16x4_t __ret;
  38071.   __ret = (poly16x4_t)(__p0);
  38072.   return __ret;
  38073. }
  38074. __ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
  38075.   poly16x4_t __ret;
  38076.   __ret = (poly16x4_t)(__p0);
  38077.   return __ret;
  38078. }
  38079. __ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
  38080.   poly8x16_t __ret;
  38081.   __ret = (poly8x16_t)(__p0);
  38082.   return __ret;
  38083. }
  38084. __ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
  38085.   poly8x16_t __ret;
  38086.   __ret = (poly8x16_t)(__p0);
  38087.   return __ret;
  38088. }
  38089. __ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
  38090.   poly8x16_t __ret;
  38091.   __ret = (poly8x16_t)(__p0);
  38092.   return __ret;
  38093. }
  38094. __ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
  38095.   poly8x16_t __ret;
  38096.   __ret = (poly8x16_t)(__p0);
  38097.   return __ret;
  38098. }
  38099. __ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
  38100.   poly8x16_t __ret;
  38101.   __ret = (poly8x16_t)(__p0);
  38102.   return __ret;
  38103. }
  38104. __ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
  38105.   poly8x16_t __ret;
  38106.   __ret = (poly8x16_t)(__p0);
  38107.   return __ret;
  38108. }
  38109. __ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
  38110.   poly8x16_t __ret;
  38111.   __ret = (poly8x16_t)(__p0);
  38112.   return __ret;
  38113. }
  38114. __ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
  38115.   poly8x16_t __ret;
  38116.   __ret = (poly8x16_t)(__p0);
  38117.   return __ret;
  38118. }
  38119. __ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
  38120.   poly8x16_t __ret;
  38121.   __ret = (poly8x16_t)(__p0);
  38122.   return __ret;
  38123. }
  38124. __ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
  38125.   poly8x16_t __ret;
  38126.   __ret = (poly8x16_t)(__p0);
  38127.   return __ret;
  38128. }
  38129. __ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
  38130.   poly8x16_t __ret;
  38131.   __ret = (poly8x16_t)(__p0);
  38132.   return __ret;
  38133. }
  38134. __ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
  38135.   poly16x8_t __ret;
  38136.   __ret = (poly16x8_t)(__p0);
  38137.   return __ret;
  38138. }
  38139. __ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
  38140.   poly16x8_t __ret;
  38141.   __ret = (poly16x8_t)(__p0);
  38142.   return __ret;
  38143. }
  38144. __ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
  38145.   poly16x8_t __ret;
  38146.   __ret = (poly16x8_t)(__p0);
  38147.   return __ret;
  38148. }
  38149. __ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
  38150.   poly16x8_t __ret;
  38151.   __ret = (poly16x8_t)(__p0);
  38152.   return __ret;
  38153. }
  38154. __ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
  38155.   poly16x8_t __ret;
  38156.   __ret = (poly16x8_t)(__p0);
  38157.   return __ret;
  38158. }
  38159. __ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
  38160.   poly16x8_t __ret;
  38161.   __ret = (poly16x8_t)(__p0);
  38162.   return __ret;
  38163. }
  38164. __ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
  38165.   poly16x8_t __ret;
  38166.   __ret = (poly16x8_t)(__p0);
  38167.   return __ret;
  38168. }
  38169. __ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
  38170.   poly16x8_t __ret;
  38171.   __ret = (poly16x8_t)(__p0);
  38172.   return __ret;
  38173. }
  38174. __ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
  38175.   poly16x8_t __ret;
  38176.   __ret = (poly16x8_t)(__p0);
  38177.   return __ret;
  38178. }
  38179. __ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
  38180.   poly16x8_t __ret;
  38181.   __ret = (poly16x8_t)(__p0);
  38182.   return __ret;
  38183. }
  38184. __ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
  38185.   poly16x8_t __ret;
  38186.   __ret = (poly16x8_t)(__p0);
  38187.   return __ret;
  38188. }
  38189. __ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
  38190.   uint8x16_t __ret;
  38191.   __ret = (uint8x16_t)(__p0);
  38192.   return __ret;
  38193. }
  38194. __ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
  38195.   uint8x16_t __ret;
  38196.   __ret = (uint8x16_t)(__p0);
  38197.   return __ret;
  38198. }
  38199. __ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
  38200.   uint8x16_t __ret;
  38201.   __ret = (uint8x16_t)(__p0);
  38202.   return __ret;
  38203. }
  38204. __ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
  38205.   uint8x16_t __ret;
  38206.   __ret = (uint8x16_t)(__p0);
  38207.   return __ret;
  38208. }
  38209. __ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
  38210.   uint8x16_t __ret;
  38211.   __ret = (uint8x16_t)(__p0);
  38212.   return __ret;
  38213. }
  38214. __ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
  38215.   uint8x16_t __ret;
  38216.   __ret = (uint8x16_t)(__p0);
  38217.   return __ret;
  38218. }
  38219. __ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
  38220.   uint8x16_t __ret;
  38221.   __ret = (uint8x16_t)(__p0);
  38222.   return __ret;
  38223. }
  38224. __ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
  38225.   uint8x16_t __ret;
  38226.   __ret = (uint8x16_t)(__p0);
  38227.   return __ret;
  38228. }
  38229. __ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
  38230.   uint8x16_t __ret;
  38231.   __ret = (uint8x16_t)(__p0);
  38232.   return __ret;
  38233. }
  38234. __ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
  38235.   uint8x16_t __ret;
  38236.   __ret = (uint8x16_t)(__p0);
  38237.   return __ret;
  38238. }
  38239. __ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
  38240.   uint8x16_t __ret;
  38241.   __ret = (uint8x16_t)(__p0);
  38242.   return __ret;
  38243. }
  38244. __ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
  38245.   uint32x4_t __ret;
  38246.   __ret = (uint32x4_t)(__p0);
  38247.   return __ret;
  38248. }
  38249. __ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
  38250.   uint32x4_t __ret;
  38251.   __ret = (uint32x4_t)(__p0);
  38252.   return __ret;
  38253. }
  38254. __ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
  38255.   uint32x4_t __ret;
  38256.   __ret = (uint32x4_t)(__p0);
  38257.   return __ret;
  38258. }
  38259. __ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
  38260.   uint32x4_t __ret;
  38261.   __ret = (uint32x4_t)(__p0);
  38262.   return __ret;
  38263. }
  38264. __ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
  38265.   uint32x4_t __ret;
  38266.   __ret = (uint32x4_t)(__p0);
  38267.   return __ret;
  38268. }
  38269. __ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
  38270.   uint32x4_t __ret;
  38271.   __ret = (uint32x4_t)(__p0);
  38272.   return __ret;
  38273. }
  38274. __ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
  38275.   uint32x4_t __ret;
  38276.   __ret = (uint32x4_t)(__p0);
  38277.   return __ret;
  38278. }
  38279. __ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
  38280.   uint32x4_t __ret;
  38281.   __ret = (uint32x4_t)(__p0);
  38282.   return __ret;
  38283. }
  38284. __ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
  38285.   uint32x4_t __ret;
  38286.   __ret = (uint32x4_t)(__p0);
  38287.   return __ret;
  38288. }
  38289. __ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
  38290.   uint32x4_t __ret;
  38291.   __ret = (uint32x4_t)(__p0);
  38292.   return __ret;
  38293. }
  38294. __ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
  38295.   uint32x4_t __ret;
  38296.   __ret = (uint32x4_t)(__p0);
  38297.   return __ret;
  38298. }
  38299. __ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
  38300.   uint64x2_t __ret;
  38301.   __ret = (uint64x2_t)(__p0);
  38302.   return __ret;
  38303. }
  38304. __ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
  38305.   uint64x2_t __ret;
  38306.   __ret = (uint64x2_t)(__p0);
  38307.   return __ret;
  38308. }
  38309. __ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
  38310.   uint64x2_t __ret;
  38311.   __ret = (uint64x2_t)(__p0);
  38312.   return __ret;
  38313. }
  38314. __ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
  38315.   uint64x2_t __ret;
  38316.   __ret = (uint64x2_t)(__p0);
  38317.   return __ret;
  38318. }
  38319. __ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
  38320.   uint64x2_t __ret;
  38321.   __ret = (uint64x2_t)(__p0);
  38322.   return __ret;
  38323. }
  38324. __ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
  38325.   uint64x2_t __ret;
  38326.   __ret = (uint64x2_t)(__p0);
  38327.   return __ret;
  38328. }
  38329. __ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
  38330.   uint64x2_t __ret;
  38331.   __ret = (uint64x2_t)(__p0);
  38332.   return __ret;
  38333. }
  38334. __ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
  38335.   uint64x2_t __ret;
  38336.   __ret = (uint64x2_t)(__p0);
  38337.   return __ret;
  38338. }
  38339. __ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
  38340.   uint64x2_t __ret;
  38341.   __ret = (uint64x2_t)(__p0);
  38342.   return __ret;
  38343. }
  38344. __ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
  38345.   uint64x2_t __ret;
  38346.   __ret = (uint64x2_t)(__p0);
  38347.   return __ret;
  38348. }
  38349. __ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
  38350.   uint64x2_t __ret;
  38351.   __ret = (uint64x2_t)(__p0);
  38352.   return __ret;
  38353. }
  38354. __ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
  38355.   uint16x8_t __ret;
  38356.   __ret = (uint16x8_t)(__p0);
  38357.   return __ret;
  38358. }
  38359. __ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
  38360.   uint16x8_t __ret;
  38361.   __ret = (uint16x8_t)(__p0);
  38362.   return __ret;
  38363. }
  38364. __ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
  38365.   uint16x8_t __ret;
  38366.   __ret = (uint16x8_t)(__p0);
  38367.   return __ret;
  38368. }
  38369. __ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
  38370.   uint16x8_t __ret;
  38371.   __ret = (uint16x8_t)(__p0);
  38372.   return __ret;
  38373. }
  38374. __ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
  38375.   uint16x8_t __ret;
  38376.   __ret = (uint16x8_t)(__p0);
  38377.   return __ret;
  38378. }
  38379. __ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
  38380.   uint16x8_t __ret;
  38381.   __ret = (uint16x8_t)(__p0);
  38382.   return __ret;
  38383. }
  38384. __ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
  38385.   uint16x8_t __ret;
  38386.   __ret = (uint16x8_t)(__p0);
  38387.   return __ret;
  38388. }
  38389. __ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
  38390.   uint16x8_t __ret;
  38391.   __ret = (uint16x8_t)(__p0);
  38392.   return __ret;
  38393. }
  38394. __ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
  38395.   uint16x8_t __ret;
  38396.   __ret = (uint16x8_t)(__p0);
  38397.   return __ret;
  38398. }
  38399. __ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
  38400.   uint16x8_t __ret;
  38401.   __ret = (uint16x8_t)(__p0);
  38402.   return __ret;
  38403. }
  38404. __ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
  38405.   uint16x8_t __ret;
  38406.   __ret = (uint16x8_t)(__p0);
  38407.   return __ret;
  38408. }
  38409. __ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
  38410.   int8x16_t __ret;
  38411.   __ret = (int8x16_t)(__p0);
  38412.   return __ret;
  38413. }
  38414. __ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
  38415.   int8x16_t __ret;
  38416.   __ret = (int8x16_t)(__p0);
  38417.   return __ret;
  38418. }
  38419. __ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
  38420.   int8x16_t __ret;
  38421.   __ret = (int8x16_t)(__p0);
  38422.   return __ret;
  38423. }
  38424. __ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
  38425.   int8x16_t __ret;
  38426.   __ret = (int8x16_t)(__p0);
  38427.   return __ret;
  38428. }
  38429. __ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
  38430.   int8x16_t __ret;
  38431.   __ret = (int8x16_t)(__p0);
  38432.   return __ret;
  38433. }
  38434. __ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
  38435.   int8x16_t __ret;
  38436.   __ret = (int8x16_t)(__p0);
  38437.   return __ret;
  38438. }
  38439. __ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
  38440.   int8x16_t __ret;
  38441.   __ret = (int8x16_t)(__p0);
  38442.   return __ret;
  38443. }
  38444. __ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
  38445.   int8x16_t __ret;
  38446.   __ret = (int8x16_t)(__p0);
  38447.   return __ret;
  38448. }
  38449. __ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
  38450.   int8x16_t __ret;
  38451.   __ret = (int8x16_t)(__p0);
  38452.   return __ret;
  38453. }
  38454. __ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
  38455.   int8x16_t __ret;
  38456.   __ret = (int8x16_t)(__p0);
  38457.   return __ret;
  38458. }
  38459. __ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
  38460.   int8x16_t __ret;
  38461.   __ret = (int8x16_t)(__p0);
  38462.   return __ret;
  38463. }
  38464. __ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
  38465.   float32x4_t __ret;
  38466.   __ret = (float32x4_t)(__p0);
  38467.   return __ret;
  38468. }
  38469. __ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
  38470.   float32x4_t __ret;
  38471.   __ret = (float32x4_t)(__p0);
  38472.   return __ret;
  38473. }
  38474. __ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
  38475.   float32x4_t __ret;
  38476.   __ret = (float32x4_t)(__p0);
  38477.   return __ret;
  38478. }
  38479. __ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
  38480.   float32x4_t __ret;
  38481.   __ret = (float32x4_t)(__p0);
  38482.   return __ret;
  38483. }
  38484. __ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
  38485.   float32x4_t __ret;
  38486.   __ret = (float32x4_t)(__p0);
  38487.   return __ret;
  38488. }
  38489. __ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
  38490.   float32x4_t __ret;
  38491.   __ret = (float32x4_t)(__p0);
  38492.   return __ret;
  38493. }
  38494. __ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
  38495.   float32x4_t __ret;
  38496.   __ret = (float32x4_t)(__p0);
  38497.   return __ret;
  38498. }
  38499. __ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
  38500.   float32x4_t __ret;
  38501.   __ret = (float32x4_t)(__p0);
  38502.   return __ret;
  38503. }
  38504. __ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
  38505.   float32x4_t __ret;
  38506.   __ret = (float32x4_t)(__p0);
  38507.   return __ret;
  38508. }
  38509. __ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
  38510.   float32x4_t __ret;
  38511.   __ret = (float32x4_t)(__p0);
  38512.   return __ret;
  38513. }
  38514. __ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
  38515.   float32x4_t __ret;
  38516.   __ret = (float32x4_t)(__p0);
  38517.   return __ret;
  38518. }
  38519. __ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
  38520.   float16x8_t __ret;
  38521.   __ret = (float16x8_t)(__p0);
  38522.   return __ret;
  38523. }
  38524. __ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
  38525.   float16x8_t __ret;
  38526.   __ret = (float16x8_t)(__p0);
  38527.   return __ret;
  38528. }
  38529. __ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
  38530.   float16x8_t __ret;
  38531.   __ret = (float16x8_t)(__p0);
  38532.   return __ret;
  38533. }
  38534. __ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
  38535.   float16x8_t __ret;
  38536.   __ret = (float16x8_t)(__p0);
  38537.   return __ret;
  38538. }
  38539. __ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
  38540.   float16x8_t __ret;
  38541.   __ret = (float16x8_t)(__p0);
  38542.   return __ret;
  38543. }
  38544. __ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
  38545.   float16x8_t __ret;
  38546.   __ret = (float16x8_t)(__p0);
  38547.   return __ret;
  38548. }
  38549. __ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
  38550.   float16x8_t __ret;
  38551.   __ret = (float16x8_t)(__p0);
  38552.   return __ret;
  38553. }
  38554. __ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
  38555.   float16x8_t __ret;
  38556.   __ret = (float16x8_t)(__p0);
  38557.   return __ret;
  38558. }
  38559. __ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
  38560.   float16x8_t __ret;
  38561.   __ret = (float16x8_t)(__p0);
  38562.   return __ret;
  38563. }
  38564. __ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
  38565.   float16x8_t __ret;
  38566.   __ret = (float16x8_t)(__p0);
  38567.   return __ret;
  38568. }
  38569. __ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
  38570.   float16x8_t __ret;
  38571.   __ret = (float16x8_t)(__p0);
  38572.   return __ret;
  38573. }
  38574. __ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
  38575.   int32x4_t __ret;
  38576.   __ret = (int32x4_t)(__p0);
  38577.   return __ret;
  38578. }
  38579. __ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
  38580.   int32x4_t __ret;
  38581.   __ret = (int32x4_t)(__p0);
  38582.   return __ret;
  38583. }
  38584. __ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
  38585.   int32x4_t __ret;
  38586.   __ret = (int32x4_t)(__p0);
  38587.   return __ret;
  38588. }
  38589. __ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
  38590.   int32x4_t __ret;
  38591.   __ret = (int32x4_t)(__p0);
  38592.   return __ret;
  38593. }
  38594. __ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
  38595.   int32x4_t __ret;
  38596.   __ret = (int32x4_t)(__p0);
  38597.   return __ret;
  38598. }
  38599. __ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
  38600.   int32x4_t __ret;
  38601.   __ret = (int32x4_t)(__p0);
  38602.   return __ret;
  38603. }
  38604. __ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
  38605.   int32x4_t __ret;
  38606.   __ret = (int32x4_t)(__p0);
  38607.   return __ret;
  38608. }
  38609. __ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
  38610.   int32x4_t __ret;
  38611.   __ret = (int32x4_t)(__p0);
  38612.   return __ret;
  38613. }
  38614. __ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
  38615.   int32x4_t __ret;
  38616.   __ret = (int32x4_t)(__p0);
  38617.   return __ret;
  38618. }
  38619. __ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
  38620.   int32x4_t __ret;
  38621.   __ret = (int32x4_t)(__p0);
  38622.   return __ret;
  38623. }
  38624. __ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
  38625.   int32x4_t __ret;
  38626.   __ret = (int32x4_t)(__p0);
  38627.   return __ret;
  38628. }
  38629. __ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
  38630.   int64x2_t __ret;
  38631.   __ret = (int64x2_t)(__p0);
  38632.   return __ret;
  38633. }
  38634. __ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
  38635.   int64x2_t __ret;
  38636.   __ret = (int64x2_t)(__p0);
  38637.   return __ret;
  38638. }
  38639. __ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
  38640.   int64x2_t __ret;
  38641.   __ret = (int64x2_t)(__p0);
  38642.   return __ret;
  38643. }
  38644. __ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
  38645.   int64x2_t __ret;
  38646.   __ret = (int64x2_t)(__p0);
  38647.   return __ret;
  38648. }
  38649. __ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
  38650.   int64x2_t __ret;
  38651.   __ret = (int64x2_t)(__p0);
  38652.   return __ret;
  38653. }
  38654. __ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
  38655.   int64x2_t __ret;
  38656.   __ret = (int64x2_t)(__p0);
  38657.   return __ret;
  38658. }
  38659. __ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
  38660.   int64x2_t __ret;
  38661.   __ret = (int64x2_t)(__p0);
  38662.   return __ret;
  38663. }
  38664. __ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
  38665.   int64x2_t __ret;
  38666.   __ret = (int64x2_t)(__p0);
  38667.   return __ret;
  38668. }
  38669. __ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
  38670.   int64x2_t __ret;
  38671.   __ret = (int64x2_t)(__p0);
  38672.   return __ret;
  38673. }
  38674. __ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
  38675.   int64x2_t __ret;
  38676.   __ret = (int64x2_t)(__p0);
  38677.   return __ret;
  38678. }
  38679. __ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
  38680.   int64x2_t __ret;
  38681.   __ret = (int64x2_t)(__p0);
  38682.   return __ret;
  38683. }
  38684. __ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
  38685.   int16x8_t __ret;
  38686.   __ret = (int16x8_t)(__p0);
  38687.   return __ret;
  38688. }
  38689. __ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
  38690.   int16x8_t __ret;
  38691.   __ret = (int16x8_t)(__p0);
  38692.   return __ret;
  38693. }
  38694. __ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
  38695.   int16x8_t __ret;
  38696.   __ret = (int16x8_t)(__p0);
  38697.   return __ret;
  38698. }
  38699. __ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
  38700.   int16x8_t __ret;
  38701.   __ret = (int16x8_t)(__p0);
  38702.   return __ret;
  38703. }
  38704. __ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
  38705.   int16x8_t __ret;
  38706.   __ret = (int16x8_t)(__p0);
  38707.   return __ret;
  38708. }
  38709. __ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
  38710.   int16x8_t __ret;
  38711.   __ret = (int16x8_t)(__p0);
  38712.   return __ret;
  38713. }
  38714. __ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
  38715.   int16x8_t __ret;
  38716.   __ret = (int16x8_t)(__p0);
  38717.   return __ret;
  38718. }
  38719. __ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
  38720.   int16x8_t __ret;
  38721.   __ret = (int16x8_t)(__p0);
  38722.   return __ret;
  38723. }
  38724. __ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
  38725.   int16x8_t __ret;
  38726.   __ret = (int16x8_t)(__p0);
  38727.   return __ret;
  38728. }
  38729. __ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
  38730.   int16x8_t __ret;
  38731.   __ret = (int16x8_t)(__p0);
  38732.   return __ret;
  38733. }
  38734. __ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
  38735.   int16x8_t __ret;
  38736.   __ret = (int16x8_t)(__p0);
  38737.   return __ret;
  38738. }
  38739. __ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
  38740.   uint8x8_t __ret;
  38741.   __ret = (uint8x8_t)(__p0);
  38742.   return __ret;
  38743. }
  38744. __ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
  38745.   uint8x8_t __ret;
  38746.   __ret = (uint8x8_t)(__p0);
  38747.   return __ret;
  38748. }
  38749. __ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
  38750.   uint8x8_t __ret;
  38751.   __ret = (uint8x8_t)(__p0);
  38752.   return __ret;
  38753. }
  38754. __ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
  38755.   uint8x8_t __ret;
  38756.   __ret = (uint8x8_t)(__p0);
  38757.   return __ret;
  38758. }
  38759. __ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
  38760.   uint8x8_t __ret;
  38761.   __ret = (uint8x8_t)(__p0);
  38762.   return __ret;
  38763. }
  38764. __ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
  38765.   uint8x8_t __ret;
  38766.   __ret = (uint8x8_t)(__p0);
  38767.   return __ret;
  38768. }
  38769. __ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
  38770.   uint8x8_t __ret;
  38771.   __ret = (uint8x8_t)(__p0);
  38772.   return __ret;
  38773. }
  38774. __ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
  38775.   uint8x8_t __ret;
  38776.   __ret = (uint8x8_t)(__p0);
  38777.   return __ret;
  38778. }
  38779. __ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
  38780.   uint8x8_t __ret;
  38781.   __ret = (uint8x8_t)(__p0);
  38782.   return __ret;
  38783. }
  38784. __ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
  38785.   uint8x8_t __ret;
  38786.   __ret = (uint8x8_t)(__p0);
  38787.   return __ret;
  38788. }
  38789. __ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
  38790.   uint8x8_t __ret;
  38791.   __ret = (uint8x8_t)(__p0);
  38792.   return __ret;
  38793. }
  38794. __ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
  38795.   uint32x2_t __ret;
  38796.   __ret = (uint32x2_t)(__p0);
  38797.   return __ret;
  38798. }
  38799. __ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
  38800.   uint32x2_t __ret;
  38801.   __ret = (uint32x2_t)(__p0);
  38802.   return __ret;
  38803. }
  38804. __ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
  38805.   uint32x2_t __ret;
  38806.   __ret = (uint32x2_t)(__p0);
  38807.   return __ret;
  38808. }
  38809. __ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
  38810.   uint32x2_t __ret;
  38811.   __ret = (uint32x2_t)(__p0);
  38812.   return __ret;
  38813. }
  38814. __ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
  38815.   uint32x2_t __ret;
  38816.   __ret = (uint32x2_t)(__p0);
  38817.   return __ret;
  38818. }
  38819. __ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
  38820.   uint32x2_t __ret;
  38821.   __ret = (uint32x2_t)(__p0);
  38822.   return __ret;
  38823. }
  38824. __ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
  38825.   uint32x2_t __ret;
  38826.   __ret = (uint32x2_t)(__p0);
  38827.   return __ret;
  38828. }
  38829. __ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
  38830.   uint32x2_t __ret;
  38831.   __ret = (uint32x2_t)(__p0);
  38832.   return __ret;
  38833. }
  38834. __ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
  38835.   uint32x2_t __ret;
  38836.   __ret = (uint32x2_t)(__p0);
  38837.   return __ret;
  38838. }
  38839. __ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
  38840.   uint32x2_t __ret;
  38841.   __ret = (uint32x2_t)(__p0);
  38842.   return __ret;
  38843. }
  38844. __ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
  38845.   uint32x2_t __ret;
  38846.   __ret = (uint32x2_t)(__p0);
  38847.   return __ret;
  38848. }
  38849. __ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
  38850.   uint64x1_t __ret;
  38851.   __ret = (uint64x1_t)(__p0);
  38852.   return __ret;
  38853. }
  38854. __ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
  38855.   uint64x1_t __ret;
  38856.   __ret = (uint64x1_t)(__p0);
  38857.   return __ret;
  38858. }
  38859. __ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
  38860.   uint64x1_t __ret;
  38861.   __ret = (uint64x1_t)(__p0);
  38862.   return __ret;
  38863. }
  38864. __ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
  38865.   uint64x1_t __ret;
  38866.   __ret = (uint64x1_t)(__p0);
  38867.   return __ret;
  38868. }
  38869. __ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
  38870.   uint64x1_t __ret;
  38871.   __ret = (uint64x1_t)(__p0);
  38872.   return __ret;
  38873. }
  38874. __ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
  38875.   uint64x1_t __ret;
  38876.   __ret = (uint64x1_t)(__p0);
  38877.   return __ret;
  38878. }
  38879. __ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
  38880.   uint64x1_t __ret;
  38881.   __ret = (uint64x1_t)(__p0);
  38882.   return __ret;
  38883. }
  38884. __ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
  38885.   uint64x1_t __ret;
  38886.   __ret = (uint64x1_t)(__p0);
  38887.   return __ret;
  38888. }
  38889. __ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
  38890.   uint64x1_t __ret;
  38891.   __ret = (uint64x1_t)(__p0);
  38892.   return __ret;
  38893. }
  38894. __ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
  38895.   uint64x1_t __ret;
  38896.   __ret = (uint64x1_t)(__p0);
  38897.   return __ret;
  38898. }
  38899. __ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
  38900.   uint64x1_t __ret;
  38901.   __ret = (uint64x1_t)(__p0);
  38902.   return __ret;
  38903. }
  38904. __ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
  38905.   uint16x4_t __ret;
  38906.   __ret = (uint16x4_t)(__p0);
  38907.   return __ret;
  38908. }
  38909. __ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
  38910.   uint16x4_t __ret;
  38911.   __ret = (uint16x4_t)(__p0);
  38912.   return __ret;
  38913. }
  38914. __ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
  38915.   uint16x4_t __ret;
  38916.   __ret = (uint16x4_t)(__p0);
  38917.   return __ret;
  38918. }
  38919. __ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
  38920.   uint16x4_t __ret;
  38921.   __ret = (uint16x4_t)(__p0);
  38922.   return __ret;
  38923. }
  38924. __ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
  38925.   uint16x4_t __ret;
  38926.   __ret = (uint16x4_t)(__p0);
  38927.   return __ret;
  38928. }
  38929. __ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
  38930.   uint16x4_t __ret;
  38931.   __ret = (uint16x4_t)(__p0);
  38932.   return __ret;
  38933. }
  38934. __ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
  38935.   uint16x4_t __ret;
  38936.   __ret = (uint16x4_t)(__p0);
  38937.   return __ret;
  38938. }
  38939. __ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
  38940.   uint16x4_t __ret;
  38941.   __ret = (uint16x4_t)(__p0);
  38942.   return __ret;
  38943. }
  38944. __ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
  38945.   uint16x4_t __ret;
  38946.   __ret = (uint16x4_t)(__p0);
  38947.   return __ret;
  38948. }
  38949. __ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
  38950.   uint16x4_t __ret;
  38951.   __ret = (uint16x4_t)(__p0);
  38952.   return __ret;
  38953. }
  38954. __ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
  38955.   uint16x4_t __ret;
  38956.   __ret = (uint16x4_t)(__p0);
  38957.   return __ret;
  38958. }
  38959. __ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
  38960.   int8x8_t __ret;
  38961.   __ret = (int8x8_t)(__p0);
  38962.   return __ret;
  38963. }
  38964. __ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
  38965.   int8x8_t __ret;
  38966.   __ret = (int8x8_t)(__p0);
  38967.   return __ret;
  38968. }
  38969. __ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
  38970.   int8x8_t __ret;
  38971.   __ret = (int8x8_t)(__p0);
  38972.   return __ret;
  38973. }
  38974. __ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
  38975.   int8x8_t __ret;
  38976.   __ret = (int8x8_t)(__p0);
  38977.   return __ret;
  38978. }
  38979. __ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
  38980.   int8x8_t __ret;
  38981.   __ret = (int8x8_t)(__p0);
  38982.   return __ret;
  38983. }
  38984. __ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
  38985.   int8x8_t __ret;
  38986.   __ret = (int8x8_t)(__p0);
  38987.   return __ret;
  38988. }
  38989. __ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
  38990.   int8x8_t __ret;
  38991.   __ret = (int8x8_t)(__p0);
  38992.   return __ret;
  38993. }
  38994. __ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
  38995.   int8x8_t __ret;
  38996.   __ret = (int8x8_t)(__p0);
  38997.   return __ret;
  38998. }
  38999. __ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
  39000.   int8x8_t __ret;
  39001.   __ret = (int8x8_t)(__p0);
  39002.   return __ret;
  39003. }
  39004. __ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
  39005.   int8x8_t __ret;
  39006.   __ret = (int8x8_t)(__p0);
  39007.   return __ret;
  39008. }
  39009. __ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
  39010.   int8x8_t __ret;
  39011.   __ret = (int8x8_t)(__p0);
  39012.   return __ret;
  39013. }
  39014. __ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
  39015.   float32x2_t __ret;
  39016.   __ret = (float32x2_t)(__p0);
  39017.   return __ret;
  39018. }
  39019. __ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
  39020.   float32x2_t __ret;
  39021.   __ret = (float32x2_t)(__p0);
  39022.   return __ret;
  39023. }
  39024. __ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
  39025.   float32x2_t __ret;
  39026.   __ret = (float32x2_t)(__p0);
  39027.   return __ret;
  39028. }
  39029. __ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
  39030.   float32x2_t __ret;
  39031.   __ret = (float32x2_t)(__p0);
  39032.   return __ret;
  39033. }
  39034. __ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
  39035.   float32x2_t __ret;
  39036.   __ret = (float32x2_t)(__p0);
  39037.   return __ret;
  39038. }
  39039. __ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
  39040.   float32x2_t __ret;
  39041.   __ret = (float32x2_t)(__p0);
  39042.   return __ret;
  39043. }
  39044. __ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
  39045.   float32x2_t __ret;
  39046.   __ret = (float32x2_t)(__p0);
  39047.   return __ret;
  39048. }
  39049. __ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
  39050.   float32x2_t __ret;
  39051.   __ret = (float32x2_t)(__p0);
  39052.   return __ret;
  39053. }
  39054. __ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
  39055.   float32x2_t __ret;
  39056.   __ret = (float32x2_t)(__p0);
  39057.   return __ret;
  39058. }
  39059. __ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
  39060.   float32x2_t __ret;
  39061.   __ret = (float32x2_t)(__p0);
  39062.   return __ret;
  39063. }
  39064. __ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
  39065.   float32x2_t __ret;
  39066.   __ret = (float32x2_t)(__p0);
  39067.   return __ret;
  39068. }
  39069. __ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
  39070.   float16x4_t __ret;
  39071.   __ret = (float16x4_t)(__p0);
  39072.   return __ret;
  39073. }
  39074. __ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
  39075.   float16x4_t __ret;
  39076.   __ret = (float16x4_t)(__p0);
  39077.   return __ret;
  39078. }
  39079. __ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
  39080.   float16x4_t __ret;
  39081.   __ret = (float16x4_t)(__p0);
  39082.   return __ret;
  39083. }
  39084. __ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
  39085.   float16x4_t __ret;
  39086.   __ret = (float16x4_t)(__p0);
  39087.   return __ret;
  39088. }
  39089. __ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
  39090.   float16x4_t __ret;
  39091.   __ret = (float16x4_t)(__p0);
  39092.   return __ret;
  39093. }
  39094. __ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
  39095.   float16x4_t __ret;
  39096.   __ret = (float16x4_t)(__p0);
  39097.   return __ret;
  39098. }
  39099. __ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
  39100.   float16x4_t __ret;
  39101.   __ret = (float16x4_t)(__p0);
  39102.   return __ret;
  39103. }
  39104. __ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
  39105.   float16x4_t __ret;
  39106.   __ret = (float16x4_t)(__p0);
  39107.   return __ret;
  39108. }
  39109. __ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
  39110.   float16x4_t __ret;
  39111.   __ret = (float16x4_t)(__p0);
  39112.   return __ret;
  39113. }
  39114. __ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
  39115.   float16x4_t __ret;
  39116.   __ret = (float16x4_t)(__p0);
  39117.   return __ret;
  39118. }
  39119. __ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
  39120.   float16x4_t __ret;
  39121.   __ret = (float16x4_t)(__p0);
  39122.   return __ret;
  39123. }
  39124. __ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
  39125.   int32x2_t __ret;
  39126.   __ret = (int32x2_t)(__p0);
  39127.   return __ret;
  39128. }
  39129. __ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
  39130.   int32x2_t __ret;
  39131.   __ret = (int32x2_t)(__p0);
  39132.   return __ret;
  39133. }
  39134. __ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
  39135.   int32x2_t __ret;
  39136.   __ret = (int32x2_t)(__p0);
  39137.   return __ret;
  39138. }
  39139. __ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
  39140.   int32x2_t __ret;
  39141.   __ret = (int32x2_t)(__p0);
  39142.   return __ret;
  39143. }
  39144. __ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
  39145.   int32x2_t __ret;
  39146.   __ret = (int32x2_t)(__p0);
  39147.   return __ret;
  39148. }
  39149. __ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
  39150.   int32x2_t __ret;
  39151.   __ret = (int32x2_t)(__p0);
  39152.   return __ret;
  39153. }
  39154. __ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
  39155.   int32x2_t __ret;
  39156.   __ret = (int32x2_t)(__p0);
  39157.   return __ret;
  39158. }
  39159. __ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
  39160.   int32x2_t __ret;
  39161.   __ret = (int32x2_t)(__p0);
  39162.   return __ret;
  39163. }
  39164. __ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
  39165.   int32x2_t __ret;
  39166.   __ret = (int32x2_t)(__p0);
  39167.   return __ret;
  39168. }
  39169. __ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
  39170.   int32x2_t __ret;
  39171.   __ret = (int32x2_t)(__p0);
  39172.   return __ret;
  39173. }
  39174. __ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
  39175.   int32x2_t __ret;
  39176.   __ret = (int32x2_t)(__p0);
  39177.   return __ret;
  39178. }
  39179. __ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
  39180.   int64x1_t __ret;
  39181.   __ret = (int64x1_t)(__p0);
  39182.   return __ret;
  39183. }
  39184. __ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
  39185.   int64x1_t __ret;
  39186.   __ret = (int64x1_t)(__p0);
  39187.   return __ret;
  39188. }
  39189. __ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
  39190.   int64x1_t __ret;
  39191.   __ret = (int64x1_t)(__p0);
  39192.   return __ret;
  39193. }
  39194. __ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
  39195.   int64x1_t __ret;
  39196.   __ret = (int64x1_t)(__p0);
  39197.   return __ret;
  39198. }
  39199. __ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
  39200.   int64x1_t __ret;
  39201.   __ret = (int64x1_t)(__p0);
  39202.   return __ret;
  39203. }
  39204. __ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
  39205.   int64x1_t __ret;
  39206.   __ret = (int64x1_t)(__p0);
  39207.   return __ret;
  39208. }
  39209. __ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
  39210.   int64x1_t __ret;
  39211.   __ret = (int64x1_t)(__p0);
  39212.   return __ret;
  39213. }
  39214. __ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
  39215.   int64x1_t __ret;
  39216.   __ret = (int64x1_t)(__p0);
  39217.   return __ret;
  39218. }
  39219. __ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
  39220.   int64x1_t __ret;
  39221.   __ret = (int64x1_t)(__p0);
  39222.   return __ret;
  39223. }
  39224. __ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
  39225.   int64x1_t __ret;
  39226.   __ret = (int64x1_t)(__p0);
  39227.   return __ret;
  39228. }
  39229. __ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
  39230.   int64x1_t __ret;
  39231.   __ret = (int64x1_t)(__p0);
  39232.   return __ret;
  39233. }
  39234. __ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
  39235.   int16x4_t __ret;
  39236.   __ret = (int16x4_t)(__p0);
  39237.   return __ret;
  39238. }
  39239. __ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
  39240.   int16x4_t __ret;
  39241.   __ret = (int16x4_t)(__p0);
  39242.   return __ret;
  39243. }
  39244. __ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
  39245.   int16x4_t __ret;
  39246.   __ret = (int16x4_t)(__p0);
  39247.   return __ret;
  39248. }
  39249. __ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
  39250.   int16x4_t __ret;
  39251.   __ret = (int16x4_t)(__p0);
  39252.   return __ret;
  39253. }
  39254. __ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
  39255.   int16x4_t __ret;
  39256.   __ret = (int16x4_t)(__p0);
  39257.   return __ret;
  39258. }
  39259. __ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
  39260.   int16x4_t __ret;
  39261.   __ret = (int16x4_t)(__p0);
  39262.   return __ret;
  39263. }
  39264. __ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
  39265.   int16x4_t __ret;
  39266.   __ret = (int16x4_t)(__p0);
  39267.   return __ret;
  39268. }
  39269. __ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
  39270.   int16x4_t __ret;
  39271.   __ret = (int16x4_t)(__p0);
  39272.   return __ret;
  39273. }
  39274. __ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
  39275.   int16x4_t __ret;
  39276.   __ret = (int16x4_t)(__p0);
  39277.   return __ret;
  39278. }
  39279. __ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
  39280.   int16x4_t __ret;
  39281.   __ret = (int16x4_t)(__p0);
  39282.   return __ret;
  39283. }
  39284. __ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
  39285.   int16x4_t __ret;
  39286.   __ret = (int16x4_t)(__p0);
  39287.   return __ret;
  39288. }
  39289. #ifdef __LITTLE_ENDIAN__
  39290. __ai __attribute__((target("bf16"))) bfloat16x4_t __a32_vcvt_bf16_f32(float32x4_t __p0) {
  39291.   bfloat16x4_t __ret;
  39292.   __ret = (bfloat16x4_t) __builtin_neon___a32_vcvt_bf16_f32((int8x16_t)__p0, 11);
  39293.   return __ret;
  39294. }
  39295. #else
  39296. __ai __attribute__((target("bf16"))) bfloat16x4_t __a32_vcvt_bf16_f32(float32x4_t __p0) {
  39297.   bfloat16x4_t __ret;
  39298.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  39299.   __ret = (bfloat16x4_t) __builtin_neon___a32_vcvt_bf16_f32((int8x16_t)__rev0, 11);
  39300.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  39301.   return __ret;
  39302. }
  39303. __ai __attribute__((target("bf16"))) bfloat16x4_t __noswap___a32_vcvt_bf16_f32(float32x4_t __p0) {
  39304.   bfloat16x4_t __ret;
  39305.   __ret = (bfloat16x4_t) __builtin_neon___a32_vcvt_bf16_f32((int8x16_t)__p0, 11);
  39306.   return __ret;
  39307. }
  39308. #endif
  39309.  
  39310. #ifdef __LITTLE_ENDIAN__
  39311. __ai __attribute__((target("bf16"))) bfloat16x4_t vcvt_bf16_f32(float32x4_t __p0) {
  39312.   bfloat16x4_t __ret;
  39313.   __ret = __a32_vcvt_bf16_f32(__p0);
  39314.   return __ret;
  39315. }
  39316. #else
  39317. __ai __attribute__((target("bf16"))) bfloat16x4_t vcvt_bf16_f32(float32x4_t __p0) {
  39318.   bfloat16x4_t __ret;
  39319.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  39320.   __ret = __noswap___a32_vcvt_bf16_f32(__rev0);
  39321.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  39322.   return __ret;
  39323. }
  39324. #endif
  39325.  
  39326. #ifdef __LITTLE_ENDIAN__
  39327. __ai __attribute__((target("bf16"))) bfloat16x8_t vcvtq_high_bf16_f32(bfloat16x8_t __p0, float32x4_t __p1) {
  39328.   bfloat16x8_t __ret;
  39329.   __ret = vcombine_bf16(__a32_vcvt_bf16_f32(__p1), vget_low_bf16(__p0));
  39330.   return __ret;
  39331. }
  39332. #else
  39333. __ai __attribute__((target("bf16"))) bfloat16x8_t vcvtq_high_bf16_f32(bfloat16x8_t __p0, float32x4_t __p1) {
  39334.   bfloat16x8_t __ret;
  39335.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  39336.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  39337.   __ret = __noswap_vcombine_bf16(__noswap___a32_vcvt_bf16_f32(__rev1), __noswap_vget_low_bf16(__rev0));
  39338.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  39339.   return __ret;
  39340. }
  39341. #endif
  39342.  
  39343. #ifdef __LITTLE_ENDIAN__
  39344. __ai __attribute__((target("bf16"))) bfloat16x8_t vcvtq_low_bf16_f32(float32x4_t __p0) {
  39345.   bfloat16x8_t __ret;
  39346.   __ret = vcombine_bf16((bfloat16x4_t)(0ULL), __a32_vcvt_bf16_f32(__p0));
  39347.   return __ret;
  39348. }
  39349. #else
  39350. __ai __attribute__((target("bf16"))) bfloat16x8_t vcvtq_low_bf16_f32(float32x4_t __p0) {
  39351.   bfloat16x8_t __ret;
  39352.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  39353.   __ret = __noswap_vcombine_bf16((bfloat16x4_t)(0ULL), __noswap___a32_vcvt_bf16_f32(__rev0));
  39354.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  39355.   return __ret;
  39356. }
  39357. #endif
  39358.  
  39359. __ai __attribute__((target("bf16"))) poly8x8_t vreinterpret_p8_bf16(bfloat16x4_t __p0) {
  39360.   poly8x8_t __ret;
  39361.   __ret = (poly8x8_t)(__p0);
  39362.   return __ret;
  39363. }
  39364. __ai __attribute__((target("bf16"))) poly64x1_t vreinterpret_p64_bf16(bfloat16x4_t __p0) {
  39365.   poly64x1_t __ret;
  39366.   __ret = (poly64x1_t)(__p0);
  39367.   return __ret;
  39368. }
  39369. __ai __attribute__((target("bf16"))) poly16x4_t vreinterpret_p16_bf16(bfloat16x4_t __p0) {
  39370.   poly16x4_t __ret;
  39371.   __ret = (poly16x4_t)(__p0);
  39372.   return __ret;
  39373. }
  39374. __ai __attribute__((target("bf16"))) poly8x16_t vreinterpretq_p8_bf16(bfloat16x8_t __p0) {
  39375.   poly8x16_t __ret;
  39376.   __ret = (poly8x16_t)(__p0);
  39377.   return __ret;
  39378. }
  39379. __ai __attribute__((target("bf16"))) poly64x2_t vreinterpretq_p64_bf16(bfloat16x8_t __p0) {
  39380.   poly64x2_t __ret;
  39381.   __ret = (poly64x2_t)(__p0);
  39382.   return __ret;
  39383. }
  39384. __ai __attribute__((target("bf16"))) poly16x8_t vreinterpretq_p16_bf16(bfloat16x8_t __p0) {
  39385.   poly16x8_t __ret;
  39386.   __ret = (poly16x8_t)(__p0);
  39387.   return __ret;
  39388. }
  39389. __ai __attribute__((target("bf16"))) uint8x16_t vreinterpretq_u8_bf16(bfloat16x8_t __p0) {
  39390.   uint8x16_t __ret;
  39391.   __ret = (uint8x16_t)(__p0);
  39392.   return __ret;
  39393. }
  39394. __ai __attribute__((target("bf16"))) uint32x4_t vreinterpretq_u32_bf16(bfloat16x8_t __p0) {
  39395.   uint32x4_t __ret;
  39396.   __ret = (uint32x4_t)(__p0);
  39397.   return __ret;
  39398. }
  39399. __ai __attribute__((target("bf16"))) uint64x2_t vreinterpretq_u64_bf16(bfloat16x8_t __p0) {
  39400.   uint64x2_t __ret;
  39401.   __ret = (uint64x2_t)(__p0);
  39402.   return __ret;
  39403. }
  39404. __ai __attribute__((target("bf16"))) uint16x8_t vreinterpretq_u16_bf16(bfloat16x8_t __p0) {
  39405.   uint16x8_t __ret;
  39406.   __ret = (uint16x8_t)(__p0);
  39407.   return __ret;
  39408. }
  39409. __ai __attribute__((target("bf16"))) int8x16_t vreinterpretq_s8_bf16(bfloat16x8_t __p0) {
  39410.   int8x16_t __ret;
  39411.   __ret = (int8x16_t)(__p0);
  39412.   return __ret;
  39413. }
  39414. __ai __attribute__((target("bf16"))) float32x4_t vreinterpretq_f32_bf16(bfloat16x8_t __p0) {
  39415.   float32x4_t __ret;
  39416.   __ret = (float32x4_t)(__p0);
  39417.   return __ret;
  39418. }
  39419. __ai __attribute__((target("bf16"))) float16x8_t vreinterpretq_f16_bf16(bfloat16x8_t __p0) {
  39420.   float16x8_t __ret;
  39421.   __ret = (float16x8_t)(__p0);
  39422.   return __ret;
  39423. }
  39424. __ai __attribute__((target("bf16"))) int32x4_t vreinterpretq_s32_bf16(bfloat16x8_t __p0) {
  39425.   int32x4_t __ret;
  39426.   __ret = (int32x4_t)(__p0);
  39427.   return __ret;
  39428. }
  39429. __ai __attribute__((target("bf16"))) int64x2_t vreinterpretq_s64_bf16(bfloat16x8_t __p0) {
  39430.   int64x2_t __ret;
  39431.   __ret = (int64x2_t)(__p0);
  39432.   return __ret;
  39433. }
  39434. __ai __attribute__((target("bf16"))) int16x8_t vreinterpretq_s16_bf16(bfloat16x8_t __p0) {
  39435.   int16x8_t __ret;
  39436.   __ret = (int16x8_t)(__p0);
  39437.   return __ret;
  39438. }
  39439. __ai __attribute__((target("bf16"))) uint8x8_t vreinterpret_u8_bf16(bfloat16x4_t __p0) {
  39440.   uint8x8_t __ret;
  39441.   __ret = (uint8x8_t)(__p0);
  39442.   return __ret;
  39443. }
  39444. __ai __attribute__((target("bf16"))) uint32x2_t vreinterpret_u32_bf16(bfloat16x4_t __p0) {
  39445.   uint32x2_t __ret;
  39446.   __ret = (uint32x2_t)(__p0);
  39447.   return __ret;
  39448. }
  39449. __ai __attribute__((target("bf16"))) uint64x1_t vreinterpret_u64_bf16(bfloat16x4_t __p0) {
  39450.   uint64x1_t __ret;
  39451.   __ret = (uint64x1_t)(__p0);
  39452.   return __ret;
  39453. }
  39454. __ai __attribute__((target("bf16"))) uint16x4_t vreinterpret_u16_bf16(bfloat16x4_t __p0) {
  39455.   uint16x4_t __ret;
  39456.   __ret = (uint16x4_t)(__p0);
  39457.   return __ret;
  39458. }
  39459. __ai __attribute__((target("bf16"))) int8x8_t vreinterpret_s8_bf16(bfloat16x4_t __p0) {
  39460.   int8x8_t __ret;
  39461.   __ret = (int8x8_t)(__p0);
  39462.   return __ret;
  39463. }
  39464. __ai __attribute__((target("bf16"))) float32x2_t vreinterpret_f32_bf16(bfloat16x4_t __p0) {
  39465.   float32x2_t __ret;
  39466.   __ret = (float32x2_t)(__p0);
  39467.   return __ret;
  39468. }
  39469. __ai __attribute__((target("bf16"))) float16x4_t vreinterpret_f16_bf16(bfloat16x4_t __p0) {
  39470.   float16x4_t __ret;
  39471.   __ret = (float16x4_t)(__p0);
  39472.   return __ret;
  39473. }
  39474. __ai __attribute__((target("bf16"))) int32x2_t vreinterpret_s32_bf16(bfloat16x4_t __p0) {
  39475.   int32x2_t __ret;
  39476.   __ret = (int32x2_t)(__p0);
  39477.   return __ret;
  39478. }
  39479. __ai __attribute__((target("bf16"))) int64x1_t vreinterpret_s64_bf16(bfloat16x4_t __p0) {
  39480.   int64x1_t __ret;
  39481.   __ret = (int64x1_t)(__p0);
  39482.   return __ret;
  39483. }
  39484. __ai __attribute__((target("bf16"))) int16x4_t vreinterpret_s16_bf16(bfloat16x4_t __p0) {
  39485.   int16x4_t __ret;
  39486.   __ret = (int16x4_t)(__p0);
  39487.   return __ret;
  39488. }
  39489. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_p8(poly8x16_t __p0) {
  39490.   bfloat16x8_t __ret;
  39491.   __ret = (bfloat16x8_t)(__p0);
  39492.   return __ret;
  39493. }
  39494. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_p64(poly64x2_t __p0) {
  39495.   bfloat16x8_t __ret;
  39496.   __ret = (bfloat16x8_t)(__p0);
  39497.   return __ret;
  39498. }
  39499. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_p16(poly16x8_t __p0) {
  39500.   bfloat16x8_t __ret;
  39501.   __ret = (bfloat16x8_t)(__p0);
  39502.   return __ret;
  39503. }
  39504. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_u8(uint8x16_t __p0) {
  39505.   bfloat16x8_t __ret;
  39506.   __ret = (bfloat16x8_t)(__p0);
  39507.   return __ret;
  39508. }
  39509. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_u32(uint32x4_t __p0) {
  39510.   bfloat16x8_t __ret;
  39511.   __ret = (bfloat16x8_t)(__p0);
  39512.   return __ret;
  39513. }
  39514. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_u64(uint64x2_t __p0) {
  39515.   bfloat16x8_t __ret;
  39516.   __ret = (bfloat16x8_t)(__p0);
  39517.   return __ret;
  39518. }
  39519. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_u16(uint16x8_t __p0) {
  39520.   bfloat16x8_t __ret;
  39521.   __ret = (bfloat16x8_t)(__p0);
  39522.   return __ret;
  39523. }
  39524. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_s8(int8x16_t __p0) {
  39525.   bfloat16x8_t __ret;
  39526.   __ret = (bfloat16x8_t)(__p0);
  39527.   return __ret;
  39528. }
  39529. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_f32(float32x4_t __p0) {
  39530.   bfloat16x8_t __ret;
  39531.   __ret = (bfloat16x8_t)(__p0);
  39532.   return __ret;
  39533. }
  39534. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_f16(float16x8_t __p0) {
  39535.   bfloat16x8_t __ret;
  39536.   __ret = (bfloat16x8_t)(__p0);
  39537.   return __ret;
  39538. }
  39539. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_s32(int32x4_t __p0) {
  39540.   bfloat16x8_t __ret;
  39541.   __ret = (bfloat16x8_t)(__p0);
  39542.   return __ret;
  39543. }
  39544. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_s64(int64x2_t __p0) {
  39545.   bfloat16x8_t __ret;
  39546.   __ret = (bfloat16x8_t)(__p0);
  39547.   return __ret;
  39548. }
  39549. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_s16(int16x8_t __p0) {
  39550.   bfloat16x8_t __ret;
  39551.   __ret = (bfloat16x8_t)(__p0);
  39552.   return __ret;
  39553. }
  39554. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_p8(poly8x8_t __p0) {
  39555.   bfloat16x4_t __ret;
  39556.   __ret = (bfloat16x4_t)(__p0);
  39557.   return __ret;
  39558. }
  39559. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_p64(poly64x1_t __p0) {
  39560.   bfloat16x4_t __ret;
  39561.   __ret = (bfloat16x4_t)(__p0);
  39562.   return __ret;
  39563. }
  39564. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_p16(poly16x4_t __p0) {
  39565.   bfloat16x4_t __ret;
  39566.   __ret = (bfloat16x4_t)(__p0);
  39567.   return __ret;
  39568. }
  39569. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_u8(uint8x8_t __p0) {
  39570.   bfloat16x4_t __ret;
  39571.   __ret = (bfloat16x4_t)(__p0);
  39572.   return __ret;
  39573. }
  39574. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_u32(uint32x2_t __p0) {
  39575.   bfloat16x4_t __ret;
  39576.   __ret = (bfloat16x4_t)(__p0);
  39577.   return __ret;
  39578. }
  39579. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_u64(uint64x1_t __p0) {
  39580.   bfloat16x4_t __ret;
  39581.   __ret = (bfloat16x4_t)(__p0);
  39582.   return __ret;
  39583. }
  39584. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_u16(uint16x4_t __p0) {
  39585.   bfloat16x4_t __ret;
  39586.   __ret = (bfloat16x4_t)(__p0);
  39587.   return __ret;
  39588. }
  39589. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_s8(int8x8_t __p0) {
  39590.   bfloat16x4_t __ret;
  39591.   __ret = (bfloat16x4_t)(__p0);
  39592.   return __ret;
  39593. }
  39594. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_f32(float32x2_t __p0) {
  39595.   bfloat16x4_t __ret;
  39596.   __ret = (bfloat16x4_t)(__p0);
  39597.   return __ret;
  39598. }
  39599. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_f16(float16x4_t __p0) {
  39600.   bfloat16x4_t __ret;
  39601.   __ret = (bfloat16x4_t)(__p0);
  39602.   return __ret;
  39603. }
  39604. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_s32(int32x2_t __p0) {
  39605.   bfloat16x4_t __ret;
  39606.   __ret = (bfloat16x4_t)(__p0);
  39607.   return __ret;
  39608. }
  39609. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_s64(int64x1_t __p0) {
  39610.   bfloat16x4_t __ret;
  39611.   __ret = (bfloat16x4_t)(__p0);
  39612.   return __ret;
  39613. }
  39614. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_s16(int16x4_t __p0) {
  39615.   bfloat16x4_t __ret;
  39616.   __ret = (bfloat16x4_t)(__p0);
  39617.   return __ret;
  39618. }
  39619. #endif
  39620. #if (__ARM_FP & 2)
  39621. #ifdef __LITTLE_ENDIAN__
  39622. __ai float16x4_t vcvt_f16_f32(float32x4_t __p0) {
  39623.   float16x4_t __ret;
  39624.   __ret = (float16x4_t) __builtin_neon_vcvt_f16_f32((int8x16_t)__p0, 41);
  39625.   return __ret;
  39626. }
  39627. #else
  39628. __ai float16x4_t vcvt_f16_f32(float32x4_t __p0) {
  39629.   float16x4_t __ret;
  39630.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  39631.   __ret = (float16x4_t) __builtin_neon_vcvt_f16_f32((int8x16_t)__rev0, 41);
  39632.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  39633.   return __ret;
  39634. }
  39635. __ai float16x4_t __noswap_vcvt_f16_f32(float32x4_t __p0) {
  39636.   float16x4_t __ret;
  39637.   __ret = (float16x4_t) __builtin_neon_vcvt_f16_f32((int8x16_t)__p0, 41);
  39638.   return __ret;
  39639. }
  39640. #endif
  39641.  
  39642. #ifdef __LITTLE_ENDIAN__
  39643. __ai float32x4_t vcvt_f32_f16(float16x4_t __p0) {
  39644.   float32x4_t __ret;
  39645.   __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__p0, 8);
  39646.   return __ret;
  39647. }
  39648. #else
  39649. __ai float32x4_t vcvt_f32_f16(float16x4_t __p0) {
  39650.   float32x4_t __ret;
  39651.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  39652.   __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__rev0, 8);
  39653.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  39654.   return __ret;
  39655. }
  39656. __ai float32x4_t __noswap_vcvt_f32_f16(float16x4_t __p0) {
  39657.   float32x4_t __ret;
  39658.   __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__p0, 8);
  39659.   return __ret;
  39660. }
  39661. #endif
  39662.  
  39663. #ifdef __LITTLE_ENDIAN__
  39664. #define vld1q_f16(__p0) __extension__ ({ \
  39665.   float16x8_t __ret; \
  39666.   __ret = (float16x8_t) __builtin_neon_vld1q_v(__p0, 40); \
  39667.   __ret; \
  39668. })
  39669. #else
  39670. #define vld1q_f16(__p0) __extension__ ({ \
  39671.   float16x8_t __ret; \
  39672.   __ret = (float16x8_t) __builtin_neon_vld1q_v(__p0, 40); \
  39673.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  39674.   __ret; \
  39675. })
  39676. #endif
  39677.  
  39678. #ifdef __LITTLE_ENDIAN__
  39679. #define vld1_f16(__p0) __extension__ ({ \
  39680.   float16x4_t __ret; \
  39681.   __ret = (float16x4_t) __builtin_neon_vld1_v(__p0, 8); \
  39682.   __ret; \
  39683. })
  39684. #else
  39685. #define vld1_f16(__p0) __extension__ ({ \
  39686.   float16x4_t __ret; \
  39687.   __ret = (float16x4_t) __builtin_neon_vld1_v(__p0, 8); \
  39688.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  39689.   __ret; \
  39690. })
  39691. #endif
  39692.  
  39693. #ifdef __LITTLE_ENDIAN__
  39694. #define vld1q_dup_f16(__p0) __extension__ ({ \
  39695.   float16x8_t __ret; \
  39696.   __ret = (float16x8_t) __builtin_neon_vld1q_dup_v(__p0, 40); \
  39697.   __ret; \
  39698. })
  39699. #else
  39700. #define vld1q_dup_f16(__p0) __extension__ ({ \
  39701.   float16x8_t __ret; \
  39702.   __ret = (float16x8_t) __builtin_neon_vld1q_dup_v(__p0, 40); \
  39703.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  39704.   __ret; \
  39705. })
  39706. #endif
  39707.  
  39708. #ifdef __LITTLE_ENDIAN__
  39709. #define vld1_dup_f16(__p0) __extension__ ({ \
  39710.   float16x4_t __ret; \
  39711.   __ret = (float16x4_t) __builtin_neon_vld1_dup_v(__p0, 8); \
  39712.   __ret; \
  39713. })
  39714. #else
  39715. #define vld1_dup_f16(__p0) __extension__ ({ \
  39716.   float16x4_t __ret; \
  39717.   __ret = (float16x4_t) __builtin_neon_vld1_dup_v(__p0, 8); \
  39718.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  39719.   __ret; \
  39720. })
  39721. #endif
  39722.  
  39723. #ifdef __LITTLE_ENDIAN__
  39724. #define vld1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  39725.   float16x8_t __ret; \
  39726.   float16x8_t __s1 = __p1; \
  39727.   __ret = (float16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 40); \
  39728.   __ret; \
  39729. })
  39730. #else
  39731. #define vld1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  39732.   float16x8_t __ret; \
  39733.   float16x8_t __s1 = __p1; \
  39734.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  39735.   __ret = (float16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 40); \
  39736.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  39737.   __ret; \
  39738. })
  39739. #endif
  39740.  
  39741. #ifdef __LITTLE_ENDIAN__
  39742. #define vld1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  39743.   float16x4_t __ret; \
  39744.   float16x4_t __s1 = __p1; \
  39745.   __ret = (float16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 8); \
  39746.   __ret; \
  39747. })
  39748. #else
  39749. #define vld1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  39750.   float16x4_t __ret; \
  39751.   float16x4_t __s1 = __p1; \
  39752.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  39753.   __ret = (float16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 8); \
  39754.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  39755.   __ret; \
  39756. })
  39757. #endif
  39758.  
  39759. #ifdef __LITTLE_ENDIAN__
  39760. #define vld1q_f16_x2(__p0) __extension__ ({ \
  39761.   float16x8x2_t __ret; \
  39762.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 40); \
  39763.   __ret; \
  39764. })
  39765. #else
  39766. #define vld1q_f16_x2(__p0) __extension__ ({ \
  39767.   float16x8x2_t __ret; \
  39768.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 40); \
  39769.  \
  39770.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  39771.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  39772.   __ret; \
  39773. })
  39774. #endif
  39775.  
  39776. #ifdef __LITTLE_ENDIAN__
  39777. #define vld1_f16_x2(__p0) __extension__ ({ \
  39778.   float16x4x2_t __ret; \
  39779.   __builtin_neon_vld1_x2_v(&__ret, __p0, 8); \
  39780.   __ret; \
  39781. })
  39782. #else
  39783. #define vld1_f16_x2(__p0) __extension__ ({ \
  39784.   float16x4x2_t __ret; \
  39785.   __builtin_neon_vld1_x2_v(&__ret, __p0, 8); \
  39786.  \
  39787.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  39788.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  39789.   __ret; \
  39790. })
  39791. #endif
  39792.  
  39793. #ifdef __LITTLE_ENDIAN__
  39794. #define vld1q_f16_x3(__p0) __extension__ ({ \
  39795.   float16x8x3_t __ret; \
  39796.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 40); \
  39797.   __ret; \
  39798. })
  39799. #else
  39800. #define vld1q_f16_x3(__p0) __extension__ ({ \
  39801.   float16x8x3_t __ret; \
  39802.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 40); \
  39803.  \
  39804.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  39805.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  39806.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  39807.   __ret; \
  39808. })
  39809. #endif
  39810.  
  39811. #ifdef __LITTLE_ENDIAN__
  39812. #define vld1_f16_x3(__p0) __extension__ ({ \
  39813.   float16x4x3_t __ret; \
  39814.   __builtin_neon_vld1_x3_v(&__ret, __p0, 8); \
  39815.   __ret; \
  39816. })
  39817. #else
  39818. #define vld1_f16_x3(__p0) __extension__ ({ \
  39819.   float16x4x3_t __ret; \
  39820.   __builtin_neon_vld1_x3_v(&__ret, __p0, 8); \
  39821.  \
  39822.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  39823.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  39824.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  39825.   __ret; \
  39826. })
  39827. #endif
  39828.  
  39829. #ifdef __LITTLE_ENDIAN__
  39830. #define vld1q_f16_x4(__p0) __extension__ ({ \
  39831.   float16x8x4_t __ret; \
  39832.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 40); \
  39833.   __ret; \
  39834. })
  39835. #else
  39836. #define vld1q_f16_x4(__p0) __extension__ ({ \
  39837.   float16x8x4_t __ret; \
  39838.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 40); \
  39839.  \
  39840.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  39841.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  39842.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  39843.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  39844.   __ret; \
  39845. })
  39846. #endif
  39847.  
  39848. #ifdef __LITTLE_ENDIAN__
  39849. #define vld1_f16_x4(__p0) __extension__ ({ \
  39850.   float16x4x4_t __ret; \
  39851.   __builtin_neon_vld1_x4_v(&__ret, __p0, 8); \
  39852.   __ret; \
  39853. })
  39854. #else
  39855. #define vld1_f16_x4(__p0) __extension__ ({ \
  39856.   float16x4x4_t __ret; \
  39857.   __builtin_neon_vld1_x4_v(&__ret, __p0, 8); \
  39858.  \
  39859.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  39860.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  39861.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  39862.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  39863.   __ret; \
  39864. })
  39865. #endif
  39866.  
  39867. #ifdef __LITTLE_ENDIAN__
  39868. #define vld2q_f16(__p0) __extension__ ({ \
  39869.   float16x8x2_t __ret; \
  39870.   __builtin_neon_vld2q_v(&__ret, __p0, 40); \
  39871.   __ret; \
  39872. })
  39873. #else
  39874. #define vld2q_f16(__p0) __extension__ ({ \
  39875.   float16x8x2_t __ret; \
  39876.   __builtin_neon_vld2q_v(&__ret, __p0, 40); \
  39877.  \
  39878.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  39879.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  39880.   __ret; \
  39881. })
  39882. #endif
  39883.  
  39884. #ifdef __LITTLE_ENDIAN__
  39885. #define vld2_f16(__p0) __extension__ ({ \
  39886.   float16x4x2_t __ret; \
  39887.   __builtin_neon_vld2_v(&__ret, __p0, 8); \
  39888.   __ret; \
  39889. })
  39890. #else
  39891. #define vld2_f16(__p0) __extension__ ({ \
  39892.   float16x4x2_t __ret; \
  39893.   __builtin_neon_vld2_v(&__ret, __p0, 8); \
  39894.  \
  39895.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  39896.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  39897.   __ret; \
  39898. })
  39899. #endif
  39900.  
  39901. #ifdef __LITTLE_ENDIAN__
  39902. #define vld2q_dup_f16(__p0) __extension__ ({ \
  39903.   float16x8x2_t __ret; \
  39904.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 40); \
  39905.   __ret; \
  39906. })
  39907. #else
  39908. #define vld2q_dup_f16(__p0) __extension__ ({ \
  39909.   float16x8x2_t __ret; \
  39910.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 40); \
  39911.  \
  39912.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  39913.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  39914.   __ret; \
  39915. })
  39916. #endif
  39917.  
  39918. #ifdef __LITTLE_ENDIAN__
  39919. #define vld2_dup_f16(__p0) __extension__ ({ \
  39920.   float16x4x2_t __ret; \
  39921.   __builtin_neon_vld2_dup_v(&__ret, __p0, 8); \
  39922.   __ret; \
  39923. })
  39924. #else
  39925. #define vld2_dup_f16(__p0) __extension__ ({ \
  39926.   float16x4x2_t __ret; \
  39927.   __builtin_neon_vld2_dup_v(&__ret, __p0, 8); \
  39928.  \
  39929.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  39930.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  39931.   __ret; \
  39932. })
  39933. #endif
  39934.  
  39935. #ifdef __LITTLE_ENDIAN__
  39936. #define vld2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  39937.   float16x8x2_t __ret; \
  39938.   float16x8x2_t __s1 = __p1; \
  39939.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 40); \
  39940.   __ret; \
  39941. })
  39942. #else
  39943. #define vld2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  39944.   float16x8x2_t __ret; \
  39945.   float16x8x2_t __s1 = __p1; \
  39946.   float16x8x2_t __rev1; \
  39947.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  39948.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  39949.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 40); \
  39950.  \
  39951.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  39952.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  39953.   __ret; \
  39954. })
  39955. #endif
  39956.  
  39957. #ifdef __LITTLE_ENDIAN__
  39958. #define vld2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  39959.   float16x4x2_t __ret; \
  39960.   float16x4x2_t __s1 = __p1; \
  39961.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 8); \
  39962.   __ret; \
  39963. })
  39964. #else
  39965. #define vld2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  39966.   float16x4x2_t __ret; \
  39967.   float16x4x2_t __s1 = __p1; \
  39968.   float16x4x2_t __rev1; \
  39969.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  39970.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  39971.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 8); \
  39972.  \
  39973.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  39974.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  39975.   __ret; \
  39976. })
  39977. #endif
  39978.  
  39979. #ifdef __LITTLE_ENDIAN__
  39980. #define vld3q_f16(__p0) __extension__ ({ \
  39981.   float16x8x3_t __ret; \
  39982.   __builtin_neon_vld3q_v(&__ret, __p0, 40); \
  39983.   __ret; \
  39984. })
  39985. #else
  39986. #define vld3q_f16(__p0) __extension__ ({ \
  39987.   float16x8x3_t __ret; \
  39988.   __builtin_neon_vld3q_v(&__ret, __p0, 40); \
  39989.  \
  39990.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  39991.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  39992.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  39993.   __ret; \
  39994. })
  39995. #endif
  39996.  
  39997. #ifdef __LITTLE_ENDIAN__
  39998. #define vld3_f16(__p0) __extension__ ({ \
  39999.   float16x4x3_t __ret; \
  40000.   __builtin_neon_vld3_v(&__ret, __p0, 8); \
  40001.   __ret; \
  40002. })
  40003. #else
  40004. #define vld3_f16(__p0) __extension__ ({ \
  40005.   float16x4x3_t __ret; \
  40006.   __builtin_neon_vld3_v(&__ret, __p0, 8); \
  40007.  \
  40008.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  40009.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  40010.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  40011.   __ret; \
  40012. })
  40013. #endif
  40014.  
  40015. #ifdef __LITTLE_ENDIAN__
  40016. #define vld3q_dup_f16(__p0) __extension__ ({ \
  40017.   float16x8x3_t __ret; \
  40018.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 40); \
  40019.   __ret; \
  40020. })
  40021. #else
  40022. #define vld3q_dup_f16(__p0) __extension__ ({ \
  40023.   float16x8x3_t __ret; \
  40024.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 40); \
  40025.  \
  40026.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40027.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40028.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40029.   __ret; \
  40030. })
  40031. #endif
  40032.  
  40033. #ifdef __LITTLE_ENDIAN__
  40034. #define vld3_dup_f16(__p0) __extension__ ({ \
  40035.   float16x4x3_t __ret; \
  40036.   __builtin_neon_vld3_dup_v(&__ret, __p0, 8); \
  40037.   __ret; \
  40038. })
  40039. #else
  40040. #define vld3_dup_f16(__p0) __extension__ ({ \
  40041.   float16x4x3_t __ret; \
  40042.   __builtin_neon_vld3_dup_v(&__ret, __p0, 8); \
  40043.  \
  40044.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  40045.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  40046.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  40047.   __ret; \
  40048. })
  40049. #endif
  40050.  
  40051. #ifdef __LITTLE_ENDIAN__
  40052. #define vld3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40053.   float16x8x3_t __ret; \
  40054.   float16x8x3_t __s1 = __p1; \
  40055.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 40); \
  40056.   __ret; \
  40057. })
  40058. #else
  40059. #define vld3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40060.   float16x8x3_t __ret; \
  40061.   float16x8x3_t __s1 = __p1; \
  40062.   float16x8x3_t __rev1; \
  40063.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40064.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40065.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40066.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 40); \
  40067.  \
  40068.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40069.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40070.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40071.   __ret; \
  40072. })
  40073. #endif
  40074.  
  40075. #ifdef __LITTLE_ENDIAN__
  40076. #define vld3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40077.   float16x4x3_t __ret; \
  40078.   float16x4x3_t __s1 = __p1; \
  40079.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 8); \
  40080.   __ret; \
  40081. })
  40082. #else
  40083. #define vld3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40084.   float16x4x3_t __ret; \
  40085.   float16x4x3_t __s1 = __p1; \
  40086.   float16x4x3_t __rev1; \
  40087.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40088.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40089.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  40090.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 8); \
  40091.  \
  40092.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  40093.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  40094.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  40095.   __ret; \
  40096. })
  40097. #endif
  40098.  
  40099. #ifdef __LITTLE_ENDIAN__
  40100. #define vld4q_f16(__p0) __extension__ ({ \
  40101.   float16x8x4_t __ret; \
  40102.   __builtin_neon_vld4q_v(&__ret, __p0, 40); \
  40103.   __ret; \
  40104. })
  40105. #else
  40106. #define vld4q_f16(__p0) __extension__ ({ \
  40107.   float16x8x4_t __ret; \
  40108.   __builtin_neon_vld4q_v(&__ret, __p0, 40); \
  40109.  \
  40110.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40111.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40112.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40113.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  40114.   __ret; \
  40115. })
  40116. #endif
  40117.  
  40118. #ifdef __LITTLE_ENDIAN__
  40119. #define vld4_f16(__p0) __extension__ ({ \
  40120.   float16x4x4_t __ret; \
  40121.   __builtin_neon_vld4_v(&__ret, __p0, 8); \
  40122.   __ret; \
  40123. })
  40124. #else
  40125. #define vld4_f16(__p0) __extension__ ({ \
  40126.   float16x4x4_t __ret; \
  40127.   __builtin_neon_vld4_v(&__ret, __p0, 8); \
  40128.  \
  40129.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  40130.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  40131.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  40132.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  40133.   __ret; \
  40134. })
  40135. #endif
  40136.  
  40137. #ifdef __LITTLE_ENDIAN__
  40138. #define vld4q_dup_f16(__p0) __extension__ ({ \
  40139.   float16x8x4_t __ret; \
  40140.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 40); \
  40141.   __ret; \
  40142. })
  40143. #else
  40144. #define vld4q_dup_f16(__p0) __extension__ ({ \
  40145.   float16x8x4_t __ret; \
  40146.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 40); \
  40147.  \
  40148.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40149.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40150.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40151.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  40152.   __ret; \
  40153. })
  40154. #endif
  40155.  
  40156. #ifdef __LITTLE_ENDIAN__
  40157. #define vld4_dup_f16(__p0) __extension__ ({ \
  40158.   float16x4x4_t __ret; \
  40159.   __builtin_neon_vld4_dup_v(&__ret, __p0, 8); \
  40160.   __ret; \
  40161. })
  40162. #else
  40163. #define vld4_dup_f16(__p0) __extension__ ({ \
  40164.   float16x4x4_t __ret; \
  40165.   __builtin_neon_vld4_dup_v(&__ret, __p0, 8); \
  40166.  \
  40167.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  40168.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  40169.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  40170.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  40171.   __ret; \
  40172. })
  40173. #endif
  40174.  
  40175. #ifdef __LITTLE_ENDIAN__
  40176. #define vld4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40177.   float16x8x4_t __ret; \
  40178.   float16x8x4_t __s1 = __p1; \
  40179.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 40); \
  40180.   __ret; \
  40181. })
  40182. #else
  40183. #define vld4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40184.   float16x8x4_t __ret; \
  40185.   float16x8x4_t __s1 = __p1; \
  40186.   float16x8x4_t __rev1; \
  40187.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40188.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40189.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40190.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  40191.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 40); \
  40192.  \
  40193.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40194.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40195.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40196.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  40197.   __ret; \
  40198. })
  40199. #endif
  40200.  
  40201. #ifdef __LITTLE_ENDIAN__
  40202. #define vld4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40203.   float16x4x4_t __ret; \
  40204.   float16x4x4_t __s1 = __p1; \
  40205.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 8); \
  40206.   __ret; \
  40207. })
  40208. #else
  40209. #define vld4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40210.   float16x4x4_t __ret; \
  40211.   float16x4x4_t __s1 = __p1; \
  40212.   float16x4x4_t __rev1; \
  40213.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40214.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40215.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  40216.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  40217.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 8); \
  40218.  \
  40219.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
  40220.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
  40221.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
  40222.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
  40223.   __ret; \
  40224. })
  40225. #endif
  40226.  
  40227. #ifdef __LITTLE_ENDIAN__
  40228. #define vst1q_f16(__p0, __p1) __extension__ ({ \
  40229.   float16x8_t __s1 = __p1; \
  40230.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 40); \
  40231. })
  40232. #else
  40233. #define vst1q_f16(__p0, __p1) __extension__ ({ \
  40234.   float16x8_t __s1 = __p1; \
  40235.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  40236.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 40); \
  40237. })
  40238. #endif
  40239.  
  40240. #ifdef __LITTLE_ENDIAN__
  40241. #define vst1_f16(__p0, __p1) __extension__ ({ \
  40242.   float16x4_t __s1 = __p1; \
  40243.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 8); \
  40244. })
  40245. #else
  40246. #define vst1_f16(__p0, __p1) __extension__ ({ \
  40247.   float16x4_t __s1 = __p1; \
  40248.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  40249.   __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 8); \
  40250. })
  40251. #endif
  40252.  
  40253. #ifdef __LITTLE_ENDIAN__
  40254. #define vst1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40255.   float16x8_t __s1 = __p1; \
  40256.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 40); \
  40257. })
  40258. #else
  40259. #define vst1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40260.   float16x8_t __s1 = __p1; \
  40261.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  40262.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 40); \
  40263. })
  40264. #endif
  40265.  
  40266. #ifdef __LITTLE_ENDIAN__
  40267. #define vst1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40268.   float16x4_t __s1 = __p1; \
  40269.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 8); \
  40270. })
  40271. #else
  40272. #define vst1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40273.   float16x4_t __s1 = __p1; \
  40274.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  40275.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 8); \
  40276. })
  40277. #endif
  40278.  
  40279. #ifdef __LITTLE_ENDIAN__
  40280. #define vst1q_f16_x2(__p0, __p1) __extension__ ({ \
  40281.   float16x8x2_t __s1 = __p1; \
  40282.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 40); \
  40283. })
  40284. #else
  40285. #define vst1q_f16_x2(__p0, __p1) __extension__ ({ \
  40286.   float16x8x2_t __s1 = __p1; \
  40287.   float16x8x2_t __rev1; \
  40288.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40289.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40290.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 40); \
  40291. })
  40292. #endif
  40293.  
  40294. #ifdef __LITTLE_ENDIAN__
  40295. #define vst1_f16_x2(__p0, __p1) __extension__ ({ \
  40296.   float16x4x2_t __s1 = __p1; \
  40297.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 8); \
  40298. })
  40299. #else
  40300. #define vst1_f16_x2(__p0, __p1) __extension__ ({ \
  40301.   float16x4x2_t __s1 = __p1; \
  40302.   float16x4x2_t __rev1; \
  40303.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40304.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40305.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 8); \
  40306. })
  40307. #endif
  40308.  
  40309. #ifdef __LITTLE_ENDIAN__
  40310. #define vst1q_f16_x3(__p0, __p1) __extension__ ({ \
  40311.   float16x8x3_t __s1 = __p1; \
  40312.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 40); \
  40313. })
  40314. #else
  40315. #define vst1q_f16_x3(__p0, __p1) __extension__ ({ \
  40316.   float16x8x3_t __s1 = __p1; \
  40317.   float16x8x3_t __rev1; \
  40318.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40319.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40320.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40321.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 40); \
  40322. })
  40323. #endif
  40324.  
  40325. #ifdef __LITTLE_ENDIAN__
  40326. #define vst1_f16_x3(__p0, __p1) __extension__ ({ \
  40327.   float16x4x3_t __s1 = __p1; \
  40328.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 8); \
  40329. })
  40330. #else
  40331. #define vst1_f16_x3(__p0, __p1) __extension__ ({ \
  40332.   float16x4x3_t __s1 = __p1; \
  40333.   float16x4x3_t __rev1; \
  40334.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40335.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40336.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  40337.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 8); \
  40338. })
  40339. #endif
  40340.  
  40341. #ifdef __LITTLE_ENDIAN__
  40342. #define vst1q_f16_x4(__p0, __p1) __extension__ ({ \
  40343.   float16x8x4_t __s1 = __p1; \
  40344.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 40); \
  40345. })
  40346. #else
  40347. #define vst1q_f16_x4(__p0, __p1) __extension__ ({ \
  40348.   float16x8x4_t __s1 = __p1; \
  40349.   float16x8x4_t __rev1; \
  40350.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40351.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40352.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40353.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  40354.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 40); \
  40355. })
  40356. #endif
  40357.  
  40358. #ifdef __LITTLE_ENDIAN__
  40359. #define vst1_f16_x4(__p0, __p1) __extension__ ({ \
  40360.   float16x4x4_t __s1 = __p1; \
  40361.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 8); \
  40362. })
  40363. #else
  40364. #define vst1_f16_x4(__p0, __p1) __extension__ ({ \
  40365.   float16x4x4_t __s1 = __p1; \
  40366.   float16x4x4_t __rev1; \
  40367.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40368.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40369.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  40370.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  40371.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 8); \
  40372. })
  40373. #endif
  40374.  
  40375. #ifdef __LITTLE_ENDIAN__
  40376. #define vst2q_f16(__p0, __p1) __extension__ ({ \
  40377.   float16x8x2_t __s1 = __p1; \
  40378.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 40); \
  40379. })
  40380. #else
  40381. #define vst2q_f16(__p0, __p1) __extension__ ({ \
  40382.   float16x8x2_t __s1 = __p1; \
  40383.   float16x8x2_t __rev1; \
  40384.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40385.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40386.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 40); \
  40387. })
  40388. #endif
  40389.  
  40390. #ifdef __LITTLE_ENDIAN__
  40391. #define vst2_f16(__p0, __p1) __extension__ ({ \
  40392.   float16x4x2_t __s1 = __p1; \
  40393.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 8); \
  40394. })
  40395. #else
  40396. #define vst2_f16(__p0, __p1) __extension__ ({ \
  40397.   float16x4x2_t __s1 = __p1; \
  40398.   float16x4x2_t __rev1; \
  40399.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40400.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40401.   __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 8); \
  40402. })
  40403. #endif
  40404.  
  40405. #ifdef __LITTLE_ENDIAN__
  40406. #define vst2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40407.   float16x8x2_t __s1 = __p1; \
  40408.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 40); \
  40409. })
  40410. #else
  40411. #define vst2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40412.   float16x8x2_t __s1 = __p1; \
  40413.   float16x8x2_t __rev1; \
  40414.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40415.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40416.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 40); \
  40417. })
  40418. #endif
  40419.  
  40420. #ifdef __LITTLE_ENDIAN__
  40421. #define vst2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40422.   float16x4x2_t __s1 = __p1; \
  40423.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 8); \
  40424. })
  40425. #else
  40426. #define vst2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40427.   float16x4x2_t __s1 = __p1; \
  40428.   float16x4x2_t __rev1; \
  40429.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40430.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40431.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 8); \
  40432. })
  40433. #endif
  40434.  
  40435. #ifdef __LITTLE_ENDIAN__
  40436. #define vst3q_f16(__p0, __p1) __extension__ ({ \
  40437.   float16x8x3_t __s1 = __p1; \
  40438.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 40); \
  40439. })
  40440. #else
  40441. #define vst3q_f16(__p0, __p1) __extension__ ({ \
  40442.   float16x8x3_t __s1 = __p1; \
  40443.   float16x8x3_t __rev1; \
  40444.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40445.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40446.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40447.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 40); \
  40448. })
  40449. #endif
  40450.  
  40451. #ifdef __LITTLE_ENDIAN__
  40452. #define vst3_f16(__p0, __p1) __extension__ ({ \
  40453.   float16x4x3_t __s1 = __p1; \
  40454.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 8); \
  40455. })
  40456. #else
  40457. #define vst3_f16(__p0, __p1) __extension__ ({ \
  40458.   float16x4x3_t __s1 = __p1; \
  40459.   float16x4x3_t __rev1; \
  40460.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40461.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40462.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  40463.   __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 8); \
  40464. })
  40465. #endif
  40466.  
  40467. #ifdef __LITTLE_ENDIAN__
  40468. #define vst3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40469.   float16x8x3_t __s1 = __p1; \
  40470.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 40); \
  40471. })
  40472. #else
  40473. #define vst3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40474.   float16x8x3_t __s1 = __p1; \
  40475.   float16x8x3_t __rev1; \
  40476.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40477.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40478.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40479.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 40); \
  40480. })
  40481. #endif
  40482.  
  40483. #ifdef __LITTLE_ENDIAN__
  40484. #define vst3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40485.   float16x4x3_t __s1 = __p1; \
  40486.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 8); \
  40487. })
  40488. #else
  40489. #define vst3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40490.   float16x4x3_t __s1 = __p1; \
  40491.   float16x4x3_t __rev1; \
  40492.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40493.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40494.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  40495.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 8); \
  40496. })
  40497. #endif
  40498.  
  40499. #ifdef __LITTLE_ENDIAN__
  40500. #define vst4q_f16(__p0, __p1) __extension__ ({ \
  40501.   float16x8x4_t __s1 = __p1; \
  40502.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 40); \
  40503. })
  40504. #else
  40505. #define vst4q_f16(__p0, __p1) __extension__ ({ \
  40506.   float16x8x4_t __s1 = __p1; \
  40507.   float16x8x4_t __rev1; \
  40508.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40509.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40510.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40511.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  40512.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 40); \
  40513. })
  40514. #endif
  40515.  
  40516. #ifdef __LITTLE_ENDIAN__
  40517. #define vst4_f16(__p0, __p1) __extension__ ({ \
  40518.   float16x4x4_t __s1 = __p1; \
  40519.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 8); \
  40520. })
  40521. #else
  40522. #define vst4_f16(__p0, __p1) __extension__ ({ \
  40523.   float16x4x4_t __s1 = __p1; \
  40524.   float16x4x4_t __rev1; \
  40525.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40526.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40527.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  40528.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  40529.   __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 8); \
  40530. })
  40531. #endif
  40532.  
  40533. #ifdef __LITTLE_ENDIAN__
  40534. #define vst4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40535.   float16x8x4_t __s1 = __p1; \
  40536.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 40); \
  40537. })
  40538. #else
  40539. #define vst4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40540.   float16x8x4_t __s1 = __p1; \
  40541.   float16x8x4_t __rev1; \
  40542.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
  40543.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
  40544.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
  40545.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
  40546.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 40); \
  40547. })
  40548. #endif
  40549.  
  40550. #ifdef __LITTLE_ENDIAN__
  40551. #define vst4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40552.   float16x4x4_t __s1 = __p1; \
  40553.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 8); \
  40554. })
  40555. #else
  40556. #define vst4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  40557.   float16x4x4_t __s1 = __p1; \
  40558.   float16x4x4_t __rev1; \
  40559.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
  40560.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
  40561.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
  40562.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
  40563.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 8); \
  40564. })
  40565. #endif
  40566.  
  40567. #endif
  40568. #if __ARM_ARCH >= 8
  40569. #ifdef __LITTLE_ENDIAN__
  40570. __ai int32x4_t vcvtaq_s32_f32(float32x4_t __p0) {
  40571.   int32x4_t __ret;
  40572.   __ret = (int32x4_t) __builtin_neon_vcvtaq_s32_v((int8x16_t)__p0, 34);
  40573.   return __ret;
  40574. }
  40575. #else
  40576. __ai int32x4_t vcvtaq_s32_f32(float32x4_t __p0) {
  40577.   int32x4_t __ret;
  40578.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40579.   __ret = (int32x4_t) __builtin_neon_vcvtaq_s32_v((int8x16_t)__rev0, 34);
  40580.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40581.   return __ret;
  40582. }
  40583. #endif
  40584.  
  40585. #ifdef __LITTLE_ENDIAN__
  40586. __ai int32x2_t vcvta_s32_f32(float32x2_t __p0) {
  40587.   int32x2_t __ret;
  40588.   __ret = (int32x2_t) __builtin_neon_vcvta_s32_v((int8x8_t)__p0, 2);
  40589.   return __ret;
  40590. }
  40591. #else
  40592. __ai int32x2_t vcvta_s32_f32(float32x2_t __p0) {
  40593.   int32x2_t __ret;
  40594.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  40595.   __ret = (int32x2_t) __builtin_neon_vcvta_s32_v((int8x8_t)__rev0, 2);
  40596.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  40597.   return __ret;
  40598. }
  40599. #endif
  40600.  
  40601. #ifdef __LITTLE_ENDIAN__
  40602. __ai uint32x4_t vcvtaq_u32_f32(float32x4_t __p0) {
  40603.   uint32x4_t __ret;
  40604.   __ret = (uint32x4_t) __builtin_neon_vcvtaq_u32_v((int8x16_t)__p0, 50);
  40605.   return __ret;
  40606. }
  40607. #else
  40608. __ai uint32x4_t vcvtaq_u32_f32(float32x4_t __p0) {
  40609.   uint32x4_t __ret;
  40610.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40611.   __ret = (uint32x4_t) __builtin_neon_vcvtaq_u32_v((int8x16_t)__rev0, 50);
  40612.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40613.   return __ret;
  40614. }
  40615. #endif
  40616.  
  40617. #ifdef __LITTLE_ENDIAN__
  40618. __ai uint32x2_t vcvta_u32_f32(float32x2_t __p0) {
  40619.   uint32x2_t __ret;
  40620.   __ret = (uint32x2_t) __builtin_neon_vcvta_u32_v((int8x8_t)__p0, 18);
  40621.   return __ret;
  40622. }
  40623. #else
  40624. __ai uint32x2_t vcvta_u32_f32(float32x2_t __p0) {
  40625.   uint32x2_t __ret;
  40626.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  40627.   __ret = (uint32x2_t) __builtin_neon_vcvta_u32_v((int8x8_t)__rev0, 18);
  40628.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  40629.   return __ret;
  40630. }
  40631. #endif
  40632.  
  40633. #ifdef __LITTLE_ENDIAN__
  40634. __ai int32x4_t vcvtmq_s32_f32(float32x4_t __p0) {
  40635.   int32x4_t __ret;
  40636.   __ret = (int32x4_t) __builtin_neon_vcvtmq_s32_v((int8x16_t)__p0, 34);
  40637.   return __ret;
  40638. }
  40639. #else
  40640. __ai int32x4_t vcvtmq_s32_f32(float32x4_t __p0) {
  40641.   int32x4_t __ret;
  40642.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40643.   __ret = (int32x4_t) __builtin_neon_vcvtmq_s32_v((int8x16_t)__rev0, 34);
  40644.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40645.   return __ret;
  40646. }
  40647. #endif
  40648.  
  40649. #ifdef __LITTLE_ENDIAN__
  40650. __ai int32x2_t vcvtm_s32_f32(float32x2_t __p0) {
  40651.   int32x2_t __ret;
  40652.   __ret = (int32x2_t) __builtin_neon_vcvtm_s32_v((int8x8_t)__p0, 2);
  40653.   return __ret;
  40654. }
  40655. #else
  40656. __ai int32x2_t vcvtm_s32_f32(float32x2_t __p0) {
  40657.   int32x2_t __ret;
  40658.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  40659.   __ret = (int32x2_t) __builtin_neon_vcvtm_s32_v((int8x8_t)__rev0, 2);
  40660.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  40661.   return __ret;
  40662. }
  40663. #endif
  40664.  
  40665. #ifdef __LITTLE_ENDIAN__
  40666. __ai uint32x4_t vcvtmq_u32_f32(float32x4_t __p0) {
  40667.   uint32x4_t __ret;
  40668.   __ret = (uint32x4_t) __builtin_neon_vcvtmq_u32_v((int8x16_t)__p0, 50);
  40669.   return __ret;
  40670. }
  40671. #else
  40672. __ai uint32x4_t vcvtmq_u32_f32(float32x4_t __p0) {
  40673.   uint32x4_t __ret;
  40674.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40675.   __ret = (uint32x4_t) __builtin_neon_vcvtmq_u32_v((int8x16_t)__rev0, 50);
  40676.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40677.   return __ret;
  40678. }
  40679. #endif
  40680.  
  40681. #ifdef __LITTLE_ENDIAN__
  40682. __ai uint32x2_t vcvtm_u32_f32(float32x2_t __p0) {
  40683.   uint32x2_t __ret;
  40684.   __ret = (uint32x2_t) __builtin_neon_vcvtm_u32_v((int8x8_t)__p0, 18);
  40685.   return __ret;
  40686. }
  40687. #else
  40688. __ai uint32x2_t vcvtm_u32_f32(float32x2_t __p0) {
  40689.   uint32x2_t __ret;
  40690.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  40691.   __ret = (uint32x2_t) __builtin_neon_vcvtm_u32_v((int8x8_t)__rev0, 18);
  40692.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  40693.   return __ret;
  40694. }
  40695. #endif
  40696.  
  40697. #ifdef __LITTLE_ENDIAN__
  40698. __ai int32x4_t vcvtnq_s32_f32(float32x4_t __p0) {
  40699.   int32x4_t __ret;
  40700.   __ret = (int32x4_t) __builtin_neon_vcvtnq_s32_v((int8x16_t)__p0, 34);
  40701.   return __ret;
  40702. }
  40703. #else
  40704. __ai int32x4_t vcvtnq_s32_f32(float32x4_t __p0) {
  40705.   int32x4_t __ret;
  40706.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40707.   __ret = (int32x4_t) __builtin_neon_vcvtnq_s32_v((int8x16_t)__rev0, 34);
  40708.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40709.   return __ret;
  40710. }
  40711. #endif
  40712.  
  40713. #ifdef __LITTLE_ENDIAN__
  40714. __ai int32x2_t vcvtn_s32_f32(float32x2_t __p0) {
  40715.   int32x2_t __ret;
  40716.   __ret = (int32x2_t) __builtin_neon_vcvtn_s32_v((int8x8_t)__p0, 2);
  40717.   return __ret;
  40718. }
  40719. #else
  40720. __ai int32x2_t vcvtn_s32_f32(float32x2_t __p0) {
  40721.   int32x2_t __ret;
  40722.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  40723.   __ret = (int32x2_t) __builtin_neon_vcvtn_s32_v((int8x8_t)__rev0, 2);
  40724.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  40725.   return __ret;
  40726. }
  40727. #endif
  40728.  
  40729. #ifdef __LITTLE_ENDIAN__
  40730. __ai uint32x4_t vcvtnq_u32_f32(float32x4_t __p0) {
  40731.   uint32x4_t __ret;
  40732.   __ret = (uint32x4_t) __builtin_neon_vcvtnq_u32_v((int8x16_t)__p0, 50);
  40733.   return __ret;
  40734. }
  40735. #else
  40736. __ai uint32x4_t vcvtnq_u32_f32(float32x4_t __p0) {
  40737.   uint32x4_t __ret;
  40738.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40739.   __ret = (uint32x4_t) __builtin_neon_vcvtnq_u32_v((int8x16_t)__rev0, 50);
  40740.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40741.   return __ret;
  40742. }
  40743. #endif
  40744.  
  40745. #ifdef __LITTLE_ENDIAN__
  40746. __ai uint32x2_t vcvtn_u32_f32(float32x2_t __p0) {
  40747.   uint32x2_t __ret;
  40748.   __ret = (uint32x2_t) __builtin_neon_vcvtn_u32_v((int8x8_t)__p0, 18);
  40749.   return __ret;
  40750. }
  40751. #else
  40752. __ai uint32x2_t vcvtn_u32_f32(float32x2_t __p0) {
  40753.   uint32x2_t __ret;
  40754.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  40755.   __ret = (uint32x2_t) __builtin_neon_vcvtn_u32_v((int8x8_t)__rev0, 18);
  40756.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  40757.   return __ret;
  40758. }
  40759. #endif
  40760.  
  40761. #ifdef __LITTLE_ENDIAN__
  40762. __ai int32x4_t vcvtpq_s32_f32(float32x4_t __p0) {
  40763.   int32x4_t __ret;
  40764.   __ret = (int32x4_t) __builtin_neon_vcvtpq_s32_v((int8x16_t)__p0, 34);
  40765.   return __ret;
  40766. }
  40767. #else
  40768. __ai int32x4_t vcvtpq_s32_f32(float32x4_t __p0) {
  40769.   int32x4_t __ret;
  40770.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40771.   __ret = (int32x4_t) __builtin_neon_vcvtpq_s32_v((int8x16_t)__rev0, 34);
  40772.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40773.   return __ret;
  40774. }
  40775. #endif
  40776.  
  40777. #ifdef __LITTLE_ENDIAN__
  40778. __ai int32x2_t vcvtp_s32_f32(float32x2_t __p0) {
  40779.   int32x2_t __ret;
  40780.   __ret = (int32x2_t) __builtin_neon_vcvtp_s32_v((int8x8_t)__p0, 2);
  40781.   return __ret;
  40782. }
  40783. #else
  40784. __ai int32x2_t vcvtp_s32_f32(float32x2_t __p0) {
  40785.   int32x2_t __ret;
  40786.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  40787.   __ret = (int32x2_t) __builtin_neon_vcvtp_s32_v((int8x8_t)__rev0, 2);
  40788.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  40789.   return __ret;
  40790. }
  40791. #endif
  40792.  
  40793. #ifdef __LITTLE_ENDIAN__
  40794. __ai uint32x4_t vcvtpq_u32_f32(float32x4_t __p0) {
  40795.   uint32x4_t __ret;
  40796.   __ret = (uint32x4_t) __builtin_neon_vcvtpq_u32_v((int8x16_t)__p0, 50);
  40797.   return __ret;
  40798. }
  40799. #else
  40800. __ai uint32x4_t vcvtpq_u32_f32(float32x4_t __p0) {
  40801.   uint32x4_t __ret;
  40802.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40803.   __ret = (uint32x4_t) __builtin_neon_vcvtpq_u32_v((int8x16_t)__rev0, 50);
  40804.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40805.   return __ret;
  40806. }
  40807. #endif
  40808.  
  40809. #ifdef __LITTLE_ENDIAN__
  40810. __ai uint32x2_t vcvtp_u32_f32(float32x2_t __p0) {
  40811.   uint32x2_t __ret;
  40812.   __ret = (uint32x2_t) __builtin_neon_vcvtp_u32_v((int8x8_t)__p0, 18);
  40813.   return __ret;
  40814. }
  40815. #else
  40816. __ai uint32x2_t vcvtp_u32_f32(float32x2_t __p0) {
  40817.   uint32x2_t __ret;
  40818.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  40819.   __ret = (uint32x2_t) __builtin_neon_vcvtp_u32_v((int8x8_t)__rev0, 18);
  40820.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  40821.   return __ret;
  40822. }
  40823. #endif
  40824.  
  40825. #ifdef __LITTLE_ENDIAN__
  40826. __ai __attribute__((target("aes"))) uint8x16_t vaesdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  40827.   uint8x16_t __ret;
  40828.   __ret = (uint8x16_t) __builtin_neon_vaesdq_u8((int8x16_t)__p0, (int8x16_t)__p1, 48);
  40829.   return __ret;
  40830. }
  40831. #else
  40832. __ai __attribute__((target("aes"))) uint8x16_t vaesdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  40833.   uint8x16_t __ret;
  40834.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40835.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40836.   __ret = (uint8x16_t) __builtin_neon_vaesdq_u8((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  40837.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40838.   return __ret;
  40839. }
  40840. #endif
  40841.  
  40842. #ifdef __LITTLE_ENDIAN__
  40843. __ai __attribute__((target("aes"))) uint8x16_t vaeseq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  40844.   uint8x16_t __ret;
  40845.   __ret = (uint8x16_t) __builtin_neon_vaeseq_u8((int8x16_t)__p0, (int8x16_t)__p1, 48);
  40846.   return __ret;
  40847. }
  40848. #else
  40849. __ai __attribute__((target("aes"))) uint8x16_t vaeseq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  40850.   uint8x16_t __ret;
  40851.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40852.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40853.   __ret = (uint8x16_t) __builtin_neon_vaeseq_u8((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  40854.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40855.   return __ret;
  40856. }
  40857. #endif
  40858.  
  40859. #ifdef __LITTLE_ENDIAN__
  40860. __ai __attribute__((target("aes"))) uint8x16_t vaesimcq_u8(uint8x16_t __p0) {
  40861.   uint8x16_t __ret;
  40862.   __ret = (uint8x16_t) __builtin_neon_vaesimcq_u8((int8x16_t)__p0, 48);
  40863.   return __ret;
  40864. }
  40865. #else
  40866. __ai __attribute__((target("aes"))) uint8x16_t vaesimcq_u8(uint8x16_t __p0) {
  40867.   uint8x16_t __ret;
  40868.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40869.   __ret = (uint8x16_t) __builtin_neon_vaesimcq_u8((int8x16_t)__rev0, 48);
  40870.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40871.   return __ret;
  40872. }
  40873. #endif
  40874.  
  40875. #ifdef __LITTLE_ENDIAN__
  40876. __ai __attribute__((target("aes"))) uint8x16_t vaesmcq_u8(uint8x16_t __p0) {
  40877.   uint8x16_t __ret;
  40878.   __ret = (uint8x16_t) __builtin_neon_vaesmcq_u8((int8x16_t)__p0, 48);
  40879.   return __ret;
  40880. }
  40881. #else
  40882. __ai __attribute__((target("aes"))) uint8x16_t vaesmcq_u8(uint8x16_t __p0) {
  40883.   uint8x16_t __ret;
  40884.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40885.   __ret = (uint8x16_t) __builtin_neon_vaesmcq_u8((int8x16_t)__rev0, 48);
  40886.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  40887.   return __ret;
  40888. }
  40889. #endif
  40890.  
  40891. #ifdef __LITTLE_ENDIAN__
  40892. __ai __attribute__((target("sha2"))) uint32x4_t vsha1cq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
  40893.   uint32x4_t __ret;
  40894.   __ret = (uint32x4_t) __builtin_neon_vsha1cq_u32(__p0, __p1, __p2);
  40895.   return __ret;
  40896. }
  40897. #else
  40898. __ai __attribute__((target("sha2"))) uint32x4_t vsha1cq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
  40899.   uint32x4_t __ret;
  40900.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40901.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  40902.   __ret = (uint32x4_t) __builtin_neon_vsha1cq_u32(__rev0, __p1, __rev2);
  40903.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40904.   return __ret;
  40905. }
  40906. #endif
  40907.  
  40908. __ai __attribute__((target("sha2"))) uint32_t vsha1h_u32(uint32_t __p0) {
  40909.   uint32_t __ret;
  40910.   __ret = (uint32_t) __builtin_neon_vsha1h_u32(__p0);
  40911.   return __ret;
  40912. }
  40913. #ifdef __LITTLE_ENDIAN__
  40914. __ai __attribute__((target("sha2"))) uint32x4_t vsha1mq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
  40915.   uint32x4_t __ret;
  40916.   __ret = (uint32x4_t) __builtin_neon_vsha1mq_u32(__p0, __p1, __p2);
  40917.   return __ret;
  40918. }
  40919. #else
  40920. __ai __attribute__((target("sha2"))) uint32x4_t vsha1mq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
  40921.   uint32x4_t __ret;
  40922.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40923.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  40924.   __ret = (uint32x4_t) __builtin_neon_vsha1mq_u32(__rev0, __p1, __rev2);
  40925.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40926.   return __ret;
  40927. }
  40928. #endif
  40929.  
  40930. #ifdef __LITTLE_ENDIAN__
  40931. __ai __attribute__((target("sha2"))) uint32x4_t vsha1pq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
  40932.   uint32x4_t __ret;
  40933.   __ret = (uint32x4_t) __builtin_neon_vsha1pq_u32(__p0, __p1, __p2);
  40934.   return __ret;
  40935. }
  40936. #else
  40937. __ai __attribute__((target("sha2"))) uint32x4_t vsha1pq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
  40938.   uint32x4_t __ret;
  40939.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40940.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  40941.   __ret = (uint32x4_t) __builtin_neon_vsha1pq_u32(__rev0, __p1, __rev2);
  40942.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40943.   return __ret;
  40944. }
  40945. #endif
  40946.  
  40947. #ifdef __LITTLE_ENDIAN__
  40948. __ai __attribute__((target("sha2"))) uint32x4_t vsha1su0q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  40949.   uint32x4_t __ret;
  40950.   __ret = (uint32x4_t) __builtin_neon_vsha1su0q_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  40951.   return __ret;
  40952. }
  40953. #else
  40954. __ai __attribute__((target("sha2"))) uint32x4_t vsha1su0q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  40955.   uint32x4_t __ret;
  40956.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40957.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  40958.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  40959.   __ret = (uint32x4_t) __builtin_neon_vsha1su0q_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  40960.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40961.   return __ret;
  40962. }
  40963. #endif
  40964.  
  40965. #ifdef __LITTLE_ENDIAN__
  40966. __ai __attribute__((target("sha2"))) uint32x4_t vsha1su1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  40967.   uint32x4_t __ret;
  40968.   __ret = (uint32x4_t) __builtin_neon_vsha1su1q_u32((int8x16_t)__p0, (int8x16_t)__p1, 50);
  40969.   return __ret;
  40970. }
  40971. #else
  40972. __ai __attribute__((target("sha2"))) uint32x4_t vsha1su1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  40973.   uint32x4_t __ret;
  40974.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40975.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  40976.   __ret = (uint32x4_t) __builtin_neon_vsha1su1q_u32((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  40977.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40978.   return __ret;
  40979. }
  40980. #endif
  40981.  
  40982. #ifdef __LITTLE_ENDIAN__
  40983. __ai __attribute__((target("sha2"))) uint32x4_t vsha256hq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  40984.   uint32x4_t __ret;
  40985.   __ret = (uint32x4_t) __builtin_neon_vsha256hq_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  40986.   return __ret;
  40987. }
  40988. #else
  40989. __ai __attribute__((target("sha2"))) uint32x4_t vsha256hq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  40990.   uint32x4_t __ret;
  40991.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  40992.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  40993.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  40994.   __ret = (uint32x4_t) __builtin_neon_vsha256hq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  40995.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  40996.   return __ret;
  40997. }
  40998. #endif
  40999.  
  41000. #ifdef __LITTLE_ENDIAN__
  41001. __ai __attribute__((target("sha2"))) uint32x4_t vsha256h2q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  41002.   uint32x4_t __ret;
  41003.   __ret = (uint32x4_t) __builtin_neon_vsha256h2q_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  41004.   return __ret;
  41005. }
  41006. #else
  41007. __ai __attribute__((target("sha2"))) uint32x4_t vsha256h2q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  41008.   uint32x4_t __ret;
  41009.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41010.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41011.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  41012.   __ret = (uint32x4_t) __builtin_neon_vsha256h2q_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  41013.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41014.   return __ret;
  41015. }
  41016. #endif
  41017.  
  41018. #ifdef __LITTLE_ENDIAN__
  41019. __ai __attribute__((target("sha2"))) uint32x4_t vsha256su0q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  41020.   uint32x4_t __ret;
  41021.   __ret = (uint32x4_t) __builtin_neon_vsha256su0q_u32((int8x16_t)__p0, (int8x16_t)__p1, 50);
  41022.   return __ret;
  41023. }
  41024. #else
  41025. __ai __attribute__((target("sha2"))) uint32x4_t vsha256su0q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  41026.   uint32x4_t __ret;
  41027.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41028.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41029.   __ret = (uint32x4_t) __builtin_neon_vsha256su0q_u32((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  41030.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41031.   return __ret;
  41032. }
  41033. #endif
  41034.  
  41035. #ifdef __LITTLE_ENDIAN__
  41036. __ai __attribute__((target("sha2"))) uint32x4_t vsha256su1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  41037.   uint32x4_t __ret;
  41038.   __ret = (uint32x4_t) __builtin_neon_vsha256su1q_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  41039.   return __ret;
  41040. }
  41041. #else
  41042. __ai __attribute__((target("sha2"))) uint32x4_t vsha256su1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  41043.   uint32x4_t __ret;
  41044.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41045.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41046.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  41047.   __ret = (uint32x4_t) __builtin_neon_vsha256su1q_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  41048.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41049.   return __ret;
  41050. }
  41051. #endif
  41052.  
  41053. #endif
  41054. #if __ARM_ARCH >= 8 && defined(__ARM_FEATURE_DIRECTED_ROUNDING)
  41055. #ifdef __LITTLE_ENDIAN__
  41056. __ai float32x4_t vrndq_f32(float32x4_t __p0) {
  41057.   float32x4_t __ret;
  41058.   __ret = (float32x4_t) __builtin_neon_vrndq_v((int8x16_t)__p0, 41);
  41059.   return __ret;
  41060. }
  41061. #else
  41062. __ai float32x4_t vrndq_f32(float32x4_t __p0) {
  41063.   float32x4_t __ret;
  41064.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41065.   __ret = (float32x4_t) __builtin_neon_vrndq_v((int8x16_t)__rev0, 41);
  41066.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41067.   return __ret;
  41068. }
  41069. #endif
  41070.  
  41071. #ifdef __LITTLE_ENDIAN__
  41072. __ai float32x2_t vrnd_f32(float32x2_t __p0) {
  41073.   float32x2_t __ret;
  41074.   __ret = (float32x2_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 9);
  41075.   return __ret;
  41076. }
  41077. #else
  41078. __ai float32x2_t vrnd_f32(float32x2_t __p0) {
  41079.   float32x2_t __ret;
  41080.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41081.   __ret = (float32x2_t) __builtin_neon_vrnd_v((int8x8_t)__rev0, 9);
  41082.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41083.   return __ret;
  41084. }
  41085. #endif
  41086.  
  41087. #ifdef __LITTLE_ENDIAN__
  41088. __ai float32x4_t vrndaq_f32(float32x4_t __p0) {
  41089.   float32x4_t __ret;
  41090.   __ret = (float32x4_t) __builtin_neon_vrndaq_v((int8x16_t)__p0, 41);
  41091.   return __ret;
  41092. }
  41093. #else
  41094. __ai float32x4_t vrndaq_f32(float32x4_t __p0) {
  41095.   float32x4_t __ret;
  41096.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41097.   __ret = (float32x4_t) __builtin_neon_vrndaq_v((int8x16_t)__rev0, 41);
  41098.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41099.   return __ret;
  41100. }
  41101. #endif
  41102.  
  41103. #ifdef __LITTLE_ENDIAN__
  41104. __ai float32x2_t vrnda_f32(float32x2_t __p0) {
  41105.   float32x2_t __ret;
  41106.   __ret = (float32x2_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 9);
  41107.   return __ret;
  41108. }
  41109. #else
  41110. __ai float32x2_t vrnda_f32(float32x2_t __p0) {
  41111.   float32x2_t __ret;
  41112.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41113.   __ret = (float32x2_t) __builtin_neon_vrnda_v((int8x8_t)__rev0, 9);
  41114.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41115.   return __ret;
  41116. }
  41117. #endif
  41118.  
  41119. #ifdef __LITTLE_ENDIAN__
  41120. __ai float32x4_t vrndiq_f32(float32x4_t __p0) {
  41121.   float32x4_t __ret;
  41122.   __ret = (float32x4_t) __builtin_neon_vrndiq_v((int8x16_t)__p0, 41);
  41123.   return __ret;
  41124. }
  41125. #else
  41126. __ai float32x4_t vrndiq_f32(float32x4_t __p0) {
  41127.   float32x4_t __ret;
  41128.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41129.   __ret = (float32x4_t) __builtin_neon_vrndiq_v((int8x16_t)__rev0, 41);
  41130.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41131.   return __ret;
  41132. }
  41133. #endif
  41134.  
  41135. #ifdef __LITTLE_ENDIAN__
  41136. __ai float32x2_t vrndi_f32(float32x2_t __p0) {
  41137.   float32x2_t __ret;
  41138.   __ret = (float32x2_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 9);
  41139.   return __ret;
  41140. }
  41141. #else
  41142. __ai float32x2_t vrndi_f32(float32x2_t __p0) {
  41143.   float32x2_t __ret;
  41144.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41145.   __ret = (float32x2_t) __builtin_neon_vrndi_v((int8x8_t)__rev0, 9);
  41146.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41147.   return __ret;
  41148. }
  41149. #endif
  41150.  
  41151. #ifdef __LITTLE_ENDIAN__
  41152. __ai float32x4_t vrndmq_f32(float32x4_t __p0) {
  41153.   float32x4_t __ret;
  41154.   __ret = (float32x4_t) __builtin_neon_vrndmq_v((int8x16_t)__p0, 41);
  41155.   return __ret;
  41156. }
  41157. #else
  41158. __ai float32x4_t vrndmq_f32(float32x4_t __p0) {
  41159.   float32x4_t __ret;
  41160.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41161.   __ret = (float32x4_t) __builtin_neon_vrndmq_v((int8x16_t)__rev0, 41);
  41162.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41163.   return __ret;
  41164. }
  41165. #endif
  41166.  
  41167. #ifdef __LITTLE_ENDIAN__
  41168. __ai float32x2_t vrndm_f32(float32x2_t __p0) {
  41169.   float32x2_t __ret;
  41170.   __ret = (float32x2_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 9);
  41171.   return __ret;
  41172. }
  41173. #else
  41174. __ai float32x2_t vrndm_f32(float32x2_t __p0) {
  41175.   float32x2_t __ret;
  41176.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41177.   __ret = (float32x2_t) __builtin_neon_vrndm_v((int8x8_t)__rev0, 9);
  41178.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41179.   return __ret;
  41180. }
  41181. #endif
  41182.  
  41183. #ifdef __LITTLE_ENDIAN__
  41184. __ai float32x4_t vrndnq_f32(float32x4_t __p0) {
  41185.   float32x4_t __ret;
  41186.   __ret = (float32x4_t) __builtin_neon_vrndnq_v((int8x16_t)__p0, 41);
  41187.   return __ret;
  41188. }
  41189. #else
  41190. __ai float32x4_t vrndnq_f32(float32x4_t __p0) {
  41191.   float32x4_t __ret;
  41192.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41193.   __ret = (float32x4_t) __builtin_neon_vrndnq_v((int8x16_t)__rev0, 41);
  41194.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41195.   return __ret;
  41196. }
  41197. #endif
  41198.  
  41199. #ifdef __LITTLE_ENDIAN__
  41200. __ai float32x2_t vrndn_f32(float32x2_t __p0) {
  41201.   float32x2_t __ret;
  41202.   __ret = (float32x2_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 9);
  41203.   return __ret;
  41204. }
  41205. #else
  41206. __ai float32x2_t vrndn_f32(float32x2_t __p0) {
  41207.   float32x2_t __ret;
  41208.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41209.   __ret = (float32x2_t) __builtin_neon_vrndn_v((int8x8_t)__rev0, 9);
  41210.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41211.   return __ret;
  41212. }
  41213. #endif
  41214.  
  41215. __ai float32_t vrndns_f32(float32_t __p0) {
  41216.   float32_t __ret;
  41217.   __ret = (float32_t) __builtin_neon_vrndns_f32(__p0);
  41218.   return __ret;
  41219. }
  41220. #ifdef __LITTLE_ENDIAN__
  41221. __ai float32x4_t vrndpq_f32(float32x4_t __p0) {
  41222.   float32x4_t __ret;
  41223.   __ret = (float32x4_t) __builtin_neon_vrndpq_v((int8x16_t)__p0, 41);
  41224.   return __ret;
  41225. }
  41226. #else
  41227. __ai float32x4_t vrndpq_f32(float32x4_t __p0) {
  41228.   float32x4_t __ret;
  41229.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41230.   __ret = (float32x4_t) __builtin_neon_vrndpq_v((int8x16_t)__rev0, 41);
  41231.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41232.   return __ret;
  41233. }
  41234. #endif
  41235.  
  41236. #ifdef __LITTLE_ENDIAN__
  41237. __ai float32x2_t vrndp_f32(float32x2_t __p0) {
  41238.   float32x2_t __ret;
  41239.   __ret = (float32x2_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 9);
  41240.   return __ret;
  41241. }
  41242. #else
  41243. __ai float32x2_t vrndp_f32(float32x2_t __p0) {
  41244.   float32x2_t __ret;
  41245.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41246.   __ret = (float32x2_t) __builtin_neon_vrndp_v((int8x8_t)__rev0, 9);
  41247.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41248.   return __ret;
  41249. }
  41250. #endif
  41251.  
  41252. #ifdef __LITTLE_ENDIAN__
  41253. __ai float32x4_t vrndxq_f32(float32x4_t __p0) {
  41254.   float32x4_t __ret;
  41255.   __ret = (float32x4_t) __builtin_neon_vrndxq_v((int8x16_t)__p0, 41);
  41256.   return __ret;
  41257. }
  41258. #else
  41259. __ai float32x4_t vrndxq_f32(float32x4_t __p0) {
  41260.   float32x4_t __ret;
  41261.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41262.   __ret = (float32x4_t) __builtin_neon_vrndxq_v((int8x16_t)__rev0, 41);
  41263.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41264.   return __ret;
  41265. }
  41266. #endif
  41267.  
  41268. #ifdef __LITTLE_ENDIAN__
  41269. __ai float32x2_t vrndx_f32(float32x2_t __p0) {
  41270.   float32x2_t __ret;
  41271.   __ret = (float32x2_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 9);
  41272.   return __ret;
  41273. }
  41274. #else
  41275. __ai float32x2_t vrndx_f32(float32x2_t __p0) {
  41276.   float32x2_t __ret;
  41277.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41278.   __ret = (float32x2_t) __builtin_neon_vrndx_v((int8x8_t)__rev0, 9);
  41279.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41280.   return __ret;
  41281. }
  41282. #endif
  41283.  
  41284. #ifdef __LITTLE_ENDIAN__
  41285. __ai __attribute__((target("fullfp16"))) float16x8_t vrndq_f16(float16x8_t __p0) {
  41286.   float16x8_t __ret;
  41287.   __ret = (float16x8_t) __builtin_neon_vrndq_f16((int8x16_t)__p0, 40);
  41288.   return __ret;
  41289. }
  41290. #else
  41291. __ai __attribute__((target("fullfp16"))) float16x8_t vrndq_f16(float16x8_t __p0) {
  41292.   float16x8_t __ret;
  41293.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41294.   __ret = (float16x8_t) __builtin_neon_vrndq_f16((int8x16_t)__rev0, 40);
  41295.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41296.   return __ret;
  41297. }
  41298. #endif
  41299.  
  41300. #ifdef __LITTLE_ENDIAN__
  41301. __ai __attribute__((target("fullfp16"))) float16x4_t vrnd_f16(float16x4_t __p0) {
  41302.   float16x4_t __ret;
  41303.   __ret = (float16x4_t) __builtin_neon_vrnd_f16((int8x8_t)__p0, 8);
  41304.   return __ret;
  41305. }
  41306. #else
  41307. __ai __attribute__((target("fullfp16"))) float16x4_t vrnd_f16(float16x4_t __p0) {
  41308.   float16x4_t __ret;
  41309.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41310.   __ret = (float16x4_t) __builtin_neon_vrnd_f16((int8x8_t)__rev0, 8);
  41311.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41312.   return __ret;
  41313. }
  41314. #endif
  41315.  
  41316. #ifdef __LITTLE_ENDIAN__
  41317. __ai __attribute__((target("fullfp16"))) float16x8_t vrndaq_f16(float16x8_t __p0) {
  41318.   float16x8_t __ret;
  41319.   __ret = (float16x8_t) __builtin_neon_vrndaq_f16((int8x16_t)__p0, 40);
  41320.   return __ret;
  41321. }
  41322. #else
  41323. __ai __attribute__((target("fullfp16"))) float16x8_t vrndaq_f16(float16x8_t __p0) {
  41324.   float16x8_t __ret;
  41325.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41326.   __ret = (float16x8_t) __builtin_neon_vrndaq_f16((int8x16_t)__rev0, 40);
  41327.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41328.   return __ret;
  41329. }
  41330. #endif
  41331.  
  41332. #ifdef __LITTLE_ENDIAN__
  41333. __ai __attribute__((target("fullfp16"))) float16x4_t vrnda_f16(float16x4_t __p0) {
  41334.   float16x4_t __ret;
  41335.   __ret = (float16x4_t) __builtin_neon_vrnda_f16((int8x8_t)__p0, 8);
  41336.   return __ret;
  41337. }
  41338. #else
  41339. __ai __attribute__((target("fullfp16"))) float16x4_t vrnda_f16(float16x4_t __p0) {
  41340.   float16x4_t __ret;
  41341.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41342.   __ret = (float16x4_t) __builtin_neon_vrnda_f16((int8x8_t)__rev0, 8);
  41343.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41344.   return __ret;
  41345. }
  41346. #endif
  41347.  
  41348. #ifdef __LITTLE_ENDIAN__
  41349. __ai __attribute__((target("fullfp16"))) float16x8_t vrndmq_f16(float16x8_t __p0) {
  41350.   float16x8_t __ret;
  41351.   __ret = (float16x8_t) __builtin_neon_vrndmq_f16((int8x16_t)__p0, 40);
  41352.   return __ret;
  41353. }
  41354. #else
  41355. __ai __attribute__((target("fullfp16"))) float16x8_t vrndmq_f16(float16x8_t __p0) {
  41356.   float16x8_t __ret;
  41357.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41358.   __ret = (float16x8_t) __builtin_neon_vrndmq_f16((int8x16_t)__rev0, 40);
  41359.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41360.   return __ret;
  41361. }
  41362. #endif
  41363.  
  41364. #ifdef __LITTLE_ENDIAN__
  41365. __ai __attribute__((target("fullfp16"))) float16x4_t vrndm_f16(float16x4_t __p0) {
  41366.   float16x4_t __ret;
  41367.   __ret = (float16x4_t) __builtin_neon_vrndm_f16((int8x8_t)__p0, 8);
  41368.   return __ret;
  41369. }
  41370. #else
  41371. __ai __attribute__((target("fullfp16"))) float16x4_t vrndm_f16(float16x4_t __p0) {
  41372.   float16x4_t __ret;
  41373.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41374.   __ret = (float16x4_t) __builtin_neon_vrndm_f16((int8x8_t)__rev0, 8);
  41375.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41376.   return __ret;
  41377. }
  41378. #endif
  41379.  
  41380. #ifdef __LITTLE_ENDIAN__
  41381. __ai __attribute__((target("fullfp16"))) float16x8_t vrndnq_f16(float16x8_t __p0) {
  41382.   float16x8_t __ret;
  41383.   __ret = (float16x8_t) __builtin_neon_vrndnq_f16((int8x16_t)__p0, 40);
  41384.   return __ret;
  41385. }
  41386. #else
  41387. __ai __attribute__((target("fullfp16"))) float16x8_t vrndnq_f16(float16x8_t __p0) {
  41388.   float16x8_t __ret;
  41389.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41390.   __ret = (float16x8_t) __builtin_neon_vrndnq_f16((int8x16_t)__rev0, 40);
  41391.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41392.   return __ret;
  41393. }
  41394. #endif
  41395.  
  41396. #ifdef __LITTLE_ENDIAN__
  41397. __ai __attribute__((target("fullfp16"))) float16x4_t vrndn_f16(float16x4_t __p0) {
  41398.   float16x4_t __ret;
  41399.   __ret = (float16x4_t) __builtin_neon_vrndn_f16((int8x8_t)__p0, 8);
  41400.   return __ret;
  41401. }
  41402. #else
  41403. __ai __attribute__((target("fullfp16"))) float16x4_t vrndn_f16(float16x4_t __p0) {
  41404.   float16x4_t __ret;
  41405.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41406.   __ret = (float16x4_t) __builtin_neon_vrndn_f16((int8x8_t)__rev0, 8);
  41407.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41408.   return __ret;
  41409. }
  41410. #endif
  41411.  
  41412. #ifdef __LITTLE_ENDIAN__
  41413. __ai __attribute__((target("fullfp16"))) float16x8_t vrndpq_f16(float16x8_t __p0) {
  41414.   float16x8_t __ret;
  41415.   __ret = (float16x8_t) __builtin_neon_vrndpq_f16((int8x16_t)__p0, 40);
  41416.   return __ret;
  41417. }
  41418. #else
  41419. __ai __attribute__((target("fullfp16"))) float16x8_t vrndpq_f16(float16x8_t __p0) {
  41420.   float16x8_t __ret;
  41421.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41422.   __ret = (float16x8_t) __builtin_neon_vrndpq_f16((int8x16_t)__rev0, 40);
  41423.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41424.   return __ret;
  41425. }
  41426. #endif
  41427.  
  41428. #ifdef __LITTLE_ENDIAN__
  41429. __ai __attribute__((target("fullfp16"))) float16x4_t vrndp_f16(float16x4_t __p0) {
  41430.   float16x4_t __ret;
  41431.   __ret = (float16x4_t) __builtin_neon_vrndp_f16((int8x8_t)__p0, 8);
  41432.   return __ret;
  41433. }
  41434. #else
  41435. __ai __attribute__((target("fullfp16"))) float16x4_t vrndp_f16(float16x4_t __p0) {
  41436.   float16x4_t __ret;
  41437.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41438.   __ret = (float16x4_t) __builtin_neon_vrndp_f16((int8x8_t)__rev0, 8);
  41439.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41440.   return __ret;
  41441. }
  41442. #endif
  41443.  
  41444. #ifdef __LITTLE_ENDIAN__
  41445. __ai __attribute__((target("fullfp16"))) float16x8_t vrndxq_f16(float16x8_t __p0) {
  41446.   float16x8_t __ret;
  41447.   __ret = (float16x8_t) __builtin_neon_vrndxq_f16((int8x16_t)__p0, 40);
  41448.   return __ret;
  41449. }
  41450. #else
  41451. __ai __attribute__((target("fullfp16"))) float16x8_t vrndxq_f16(float16x8_t __p0) {
  41452.   float16x8_t __ret;
  41453.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41454.   __ret = (float16x8_t) __builtin_neon_vrndxq_f16((int8x16_t)__rev0, 40);
  41455.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41456.   return __ret;
  41457. }
  41458. #endif
  41459.  
  41460. #ifdef __LITTLE_ENDIAN__
  41461. __ai __attribute__((target("fullfp16"))) float16x4_t vrndx_f16(float16x4_t __p0) {
  41462.   float16x4_t __ret;
  41463.   __ret = (float16x4_t) __builtin_neon_vrndx_f16((int8x8_t)__p0, 8);
  41464.   return __ret;
  41465. }
  41466. #else
  41467. __ai __attribute__((target("fullfp16"))) float16x4_t vrndx_f16(float16x4_t __p0) {
  41468.   float16x4_t __ret;
  41469.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41470.   __ret = (float16x4_t) __builtin_neon_vrndx_f16((int8x8_t)__rev0, 8);
  41471.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41472.   return __ret;
  41473. }
  41474. #endif
  41475.  
  41476. #endif
  41477. #if __ARM_ARCH >= 8 && defined(__ARM_FEATURE_NUMERIC_MAXMIN)
  41478. #ifdef __LITTLE_ENDIAN__
  41479. __ai float32x4_t vmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
  41480.   float32x4_t __ret;
  41481.   __ret = (float32x4_t) __builtin_neon_vmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  41482.   return __ret;
  41483. }
  41484. #else
  41485. __ai float32x4_t vmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
  41486.   float32x4_t __ret;
  41487.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41488.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41489.   __ret = (float32x4_t) __builtin_neon_vmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  41490.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41491.   return __ret;
  41492. }
  41493. #endif
  41494.  
  41495. #ifdef __LITTLE_ENDIAN__
  41496. __ai float32x2_t vmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
  41497.   float32x2_t __ret;
  41498.   __ret = (float32x2_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  41499.   return __ret;
  41500. }
  41501. #else
  41502. __ai float32x2_t vmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
  41503.   float32x2_t __ret;
  41504.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41505.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  41506.   __ret = (float32x2_t) __builtin_neon_vmaxnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  41507.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41508.   return __ret;
  41509. }
  41510. #endif
  41511.  
  41512. #ifdef __LITTLE_ENDIAN__
  41513. __ai float32x4_t vminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
  41514.   float32x4_t __ret;
  41515.   __ret = (float32x4_t) __builtin_neon_vminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  41516.   return __ret;
  41517. }
  41518. #else
  41519. __ai float32x4_t vminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
  41520.   float32x4_t __ret;
  41521.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41522.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41523.   __ret = (float32x4_t) __builtin_neon_vminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  41524.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41525.   return __ret;
  41526. }
  41527. #endif
  41528.  
  41529. #ifdef __LITTLE_ENDIAN__
  41530. __ai float32x2_t vminnm_f32(float32x2_t __p0, float32x2_t __p1) {
  41531.   float32x2_t __ret;
  41532.   __ret = (float32x2_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  41533.   return __ret;
  41534. }
  41535. #else
  41536. __ai float32x2_t vminnm_f32(float32x2_t __p0, float32x2_t __p1) {
  41537.   float32x2_t __ret;
  41538.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41539.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  41540.   __ret = (float32x2_t) __builtin_neon_vminnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  41541.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41542.   return __ret;
  41543. }
  41544. #endif
  41545.  
  41546. #ifdef __LITTLE_ENDIAN__
  41547. __ai __attribute__((target("fullfp16"))) float16x8_t vmaxnmq_f16(float16x8_t __p0, float16x8_t __p1) {
  41548.   float16x8_t __ret;
  41549.   __ret = (float16x8_t) __builtin_neon_vmaxnmq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  41550.   return __ret;
  41551. }
  41552. #else
  41553. __ai __attribute__((target("fullfp16"))) float16x8_t vmaxnmq_f16(float16x8_t __p0, float16x8_t __p1) {
  41554.   float16x8_t __ret;
  41555.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41556.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  41557.   __ret = (float16x8_t) __builtin_neon_vmaxnmq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  41558.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41559.   return __ret;
  41560. }
  41561. #endif
  41562.  
  41563. #ifdef __LITTLE_ENDIAN__
  41564. __ai __attribute__((target("fullfp16"))) float16x4_t vmaxnm_f16(float16x4_t __p0, float16x4_t __p1) {
  41565.   float16x4_t __ret;
  41566.   __ret = (float16x4_t) __builtin_neon_vmaxnm_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  41567.   return __ret;
  41568. }
  41569. #else
  41570. __ai __attribute__((target("fullfp16"))) float16x4_t vmaxnm_f16(float16x4_t __p0, float16x4_t __p1) {
  41571.   float16x4_t __ret;
  41572.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41573.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41574.   __ret = (float16x4_t) __builtin_neon_vmaxnm_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  41575.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41576.   return __ret;
  41577. }
  41578. #endif
  41579.  
  41580. #ifdef __LITTLE_ENDIAN__
  41581. __ai __attribute__((target("fullfp16"))) float16x8_t vminnmq_f16(float16x8_t __p0, float16x8_t __p1) {
  41582.   float16x8_t __ret;
  41583.   __ret = (float16x8_t) __builtin_neon_vminnmq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  41584.   return __ret;
  41585. }
  41586. #else
  41587. __ai __attribute__((target("fullfp16"))) float16x8_t vminnmq_f16(float16x8_t __p0, float16x8_t __p1) {
  41588.   float16x8_t __ret;
  41589.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41590.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  41591.   __ret = (float16x8_t) __builtin_neon_vminnmq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  41592.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41593.   return __ret;
  41594. }
  41595. #endif
  41596.  
  41597. #ifdef __LITTLE_ENDIAN__
  41598. __ai __attribute__((target("fullfp16"))) float16x4_t vminnm_f16(float16x4_t __p0, float16x4_t __p1) {
  41599.   float16x4_t __ret;
  41600.   __ret = (float16x4_t) __builtin_neon_vminnm_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  41601.   return __ret;
  41602. }
  41603. #else
  41604. __ai __attribute__((target("fullfp16"))) float16x4_t vminnm_f16(float16x4_t __p0, float16x4_t __p1) {
  41605.   float16x4_t __ret;
  41606.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41607.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41608.   __ret = (float16x4_t) __builtin_neon_vminnm_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  41609.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41610.   return __ret;
  41611. }
  41612. #endif
  41613.  
  41614. #endif
  41615. #if defined(__ARM_FEATURE_FMA)
  41616. #ifdef __LITTLE_ENDIAN__
  41617. __ai float32x4_t vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  41618.   float32x4_t __ret;
  41619.   __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  41620.   return __ret;
  41621. }
  41622. #else
  41623. __ai float32x4_t vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  41624.   float32x4_t __ret;
  41625.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41626.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41627.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  41628.   __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  41629.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41630.   return __ret;
  41631. }
  41632. __ai float32x4_t __noswap_vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  41633.   float32x4_t __ret;
  41634.   __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  41635.   return __ret;
  41636. }
  41637. #endif
  41638.  
  41639. #ifdef __LITTLE_ENDIAN__
  41640. __ai float32x2_t vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  41641.   float32x2_t __ret;
  41642.   __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  41643.   return __ret;
  41644. }
  41645. #else
  41646. __ai float32x2_t vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  41647.   float32x2_t __ret;
  41648.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41649.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  41650.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  41651.   __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  41652.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41653.   return __ret;
  41654. }
  41655. __ai float32x2_t __noswap_vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  41656.   float32x2_t __ret;
  41657.   __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  41658.   return __ret;
  41659. }
  41660. #endif
  41661.  
  41662. #ifdef __LITTLE_ENDIAN__
  41663. __ai float32x4_t vfmaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
  41664.   float32x4_t __ret;
  41665.   __ret = vfmaq_f32(__p0, __p1, (float32x4_t) {__p2, __p2, __p2, __p2});
  41666.   return __ret;
  41667. }
  41668. #else
  41669. __ai float32x4_t vfmaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
  41670.   float32x4_t __ret;
  41671.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41672.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41673.   __ret = __noswap_vfmaq_f32(__rev0, __rev1, (float32x4_t) {__p2, __p2, __p2, __p2});
  41674.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41675.   return __ret;
  41676. }
  41677. #endif
  41678.  
  41679. #ifdef __LITTLE_ENDIAN__
  41680. __ai float32x2_t vfma_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
  41681.   float32x2_t __ret;
  41682.   __ret = vfma_f32(__p0, __p1, (float32x2_t) {__p2, __p2});
  41683.   return __ret;
  41684. }
  41685. #else
  41686. __ai float32x2_t vfma_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
  41687.   float32x2_t __ret;
  41688.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41689.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  41690.   __ret = __noswap_vfma_f32(__rev0, __rev1, (float32x2_t) {__p2, __p2});
  41691.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41692.   return __ret;
  41693. }
  41694. #endif
  41695.  
  41696. #ifdef __LITTLE_ENDIAN__
  41697. __ai float32x4_t vfmsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  41698.   float32x4_t __ret;
  41699.   __ret = vfmaq_f32(__p0, -__p1, __p2);
  41700.   return __ret;
  41701. }
  41702. #else
  41703. __ai float32x4_t vfmsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
  41704.   float32x4_t __ret;
  41705.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41706.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41707.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  41708.   __ret = __noswap_vfmaq_f32(__rev0, -__rev1, __rev2);
  41709.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41710.   return __ret;
  41711. }
  41712. #endif
  41713.  
  41714. #ifdef __LITTLE_ENDIAN__
  41715. __ai float32x2_t vfms_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  41716.   float32x2_t __ret;
  41717.   __ret = vfma_f32(__p0, -__p1, __p2);
  41718.   return __ret;
  41719. }
  41720. #else
  41721. __ai float32x2_t vfms_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
  41722.   float32x2_t __ret;
  41723.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41724.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  41725.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  41726.   __ret = __noswap_vfma_f32(__rev0, -__rev1, __rev2);
  41727.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41728.   return __ret;
  41729. }
  41730. #endif
  41731.  
  41732. #endif
  41733. #if defined(__aarch64__)
  41734. #ifdef __LITTLE_ENDIAN__
  41735. __ai float64x2_t vabdq_f64(float64x2_t __p0, float64x2_t __p1) {
  41736.   float64x2_t __ret;
  41737.   __ret = (float64x2_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  41738.   return __ret;
  41739. }
  41740. #else
  41741. __ai float64x2_t vabdq_f64(float64x2_t __p0, float64x2_t __p1) {
  41742.   float64x2_t __ret;
  41743.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41744.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  41745.   __ret = (float64x2_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  41746.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41747.   return __ret;
  41748. }
  41749. #endif
  41750.  
  41751. __ai float64x1_t vabd_f64(float64x1_t __p0, float64x1_t __p1) {
  41752.   float64x1_t __ret;
  41753.   __ret = (float64x1_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
  41754.   return __ret;
  41755. }
  41756. __ai float64_t vabdd_f64(float64_t __p0, float64_t __p1) {
  41757.   float64_t __ret;
  41758.   __ret = (float64_t) __builtin_neon_vabdd_f64(__p0, __p1);
  41759.   return __ret;
  41760. }
  41761. __ai float32_t vabds_f32(float32_t __p0, float32_t __p1) {
  41762.   float32_t __ret;
  41763.   __ret = (float32_t) __builtin_neon_vabds_f32(__p0, __p1);
  41764.   return __ret;
  41765. }
  41766. #ifdef __LITTLE_ENDIAN__
  41767. __ai float64x2_t vabsq_f64(float64x2_t __p0) {
  41768.   float64x2_t __ret;
  41769.   __ret = (float64x2_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 42);
  41770.   return __ret;
  41771. }
  41772. #else
  41773. __ai float64x2_t vabsq_f64(float64x2_t __p0) {
  41774.   float64x2_t __ret;
  41775.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41776.   __ret = (float64x2_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 42);
  41777.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41778.   return __ret;
  41779. }
  41780. #endif
  41781.  
  41782. #ifdef __LITTLE_ENDIAN__
  41783. __ai int64x2_t vabsq_s64(int64x2_t __p0) {
  41784.   int64x2_t __ret;
  41785.   __ret = (int64x2_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 35);
  41786.   return __ret;
  41787. }
  41788. #else
  41789. __ai int64x2_t vabsq_s64(int64x2_t __p0) {
  41790.   int64x2_t __ret;
  41791.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41792.   __ret = (int64x2_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 35);
  41793.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41794.   return __ret;
  41795. }
  41796. #endif
  41797.  
  41798. __ai float64x1_t vabs_f64(float64x1_t __p0) {
  41799.   float64x1_t __ret;
  41800.   __ret = (float64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 10);
  41801.   return __ret;
  41802. }
  41803. __ai int64x1_t vabs_s64(int64x1_t __p0) {
  41804.   int64x1_t __ret;
  41805.   __ret = (int64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 3);
  41806.   return __ret;
  41807. }
  41808. __ai int64_t vabsd_s64(int64_t __p0) {
  41809.   int64_t __ret;
  41810.   __ret = (int64_t) __builtin_neon_vabsd_s64(__p0);
  41811.   return __ret;
  41812. }
  41813. #ifdef __LITTLE_ENDIAN__
  41814. __ai float64x2_t vaddq_f64(float64x2_t __p0, float64x2_t __p1) {
  41815.   float64x2_t __ret;
  41816.   __ret = __p0 + __p1;
  41817.   return __ret;
  41818. }
  41819. #else
  41820. __ai float64x2_t vaddq_f64(float64x2_t __p0, float64x2_t __p1) {
  41821.   float64x2_t __ret;
  41822.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41823.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  41824.   __ret = __rev0 + __rev1;
  41825.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  41826.   return __ret;
  41827. }
  41828. #endif
  41829.  
  41830. __ai float64x1_t vadd_f64(float64x1_t __p0, float64x1_t __p1) {
  41831.   float64x1_t __ret;
  41832.   __ret = __p0 + __p1;
  41833.   return __ret;
  41834. }
  41835. __ai uint64_t vaddd_u64(uint64_t __p0, uint64_t __p1) {
  41836.   uint64_t __ret;
  41837.   __ret = (uint64_t) __builtin_neon_vaddd_u64(__p0, __p1);
  41838.   return __ret;
  41839. }
  41840. __ai int64_t vaddd_s64(int64_t __p0, int64_t __p1) {
  41841.   int64_t __ret;
  41842.   __ret = (int64_t) __builtin_neon_vaddd_s64(__p0, __p1);
  41843.   return __ret;
  41844. }
  41845. __ai poly128_t vaddq_p128(poly128_t __p0, poly128_t __p1) {
  41846.   poly128_t __ret;
  41847.   __ret = (poly128_t) __builtin_neon_vaddq_p128(__p0, __p1);
  41848.   return __ret;
  41849. }
  41850. #ifdef __LITTLE_ENDIAN__
  41851. __ai uint16x8_t vaddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  41852.   uint16x8_t __ret;
  41853.   __ret = vcombine_u16(__p0, vaddhn_u32(__p1, __p2));
  41854.   return __ret;
  41855. }
  41856. #else
  41857. __ai uint16x8_t vaddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  41858.   uint16x8_t __ret;
  41859.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41860.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41861.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  41862.   __ret = __noswap_vcombine_u16(__rev0, __noswap_vaddhn_u32(__rev1, __rev2));
  41863.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41864.   return __ret;
  41865. }
  41866. #endif
  41867.  
  41868. #ifdef __LITTLE_ENDIAN__
  41869. __ai uint32x4_t vaddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  41870.   uint32x4_t __ret;
  41871.   __ret = vcombine_u32(__p0, vaddhn_u64(__p1, __p2));
  41872.   return __ret;
  41873. }
  41874. #else
  41875. __ai uint32x4_t vaddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  41876.   uint32x4_t __ret;
  41877.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41878.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  41879.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  41880.   __ret = __noswap_vcombine_u32(__rev0, __noswap_vaddhn_u64(__rev1, __rev2));
  41881.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41882.   return __ret;
  41883. }
  41884. #endif
  41885.  
  41886. #ifdef __LITTLE_ENDIAN__
  41887. __ai uint8x16_t vaddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  41888.   uint8x16_t __ret;
  41889.   __ret = vcombine_u8(__p0, vaddhn_u16(__p1, __p2));
  41890.   return __ret;
  41891. }
  41892. #else
  41893. __ai uint8x16_t vaddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  41894.   uint8x16_t __ret;
  41895.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41896.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  41897.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  41898.   __ret = __noswap_vcombine_u8(__rev0, __noswap_vaddhn_u16(__rev1, __rev2));
  41899.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  41900.   return __ret;
  41901. }
  41902. #endif
  41903.  
  41904. #ifdef __LITTLE_ENDIAN__
  41905. __ai int16x8_t vaddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  41906.   int16x8_t __ret;
  41907.   __ret = vcombine_s16(__p0, vaddhn_s32(__p1, __p2));
  41908.   return __ret;
  41909. }
  41910. #else
  41911. __ai int16x8_t vaddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  41912.   int16x8_t __ret;
  41913.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41914.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  41915.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  41916.   __ret = __noswap_vcombine_s16(__rev0, __noswap_vaddhn_s32(__rev1, __rev2));
  41917.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  41918.   return __ret;
  41919. }
  41920. #endif
  41921.  
  41922. #ifdef __LITTLE_ENDIAN__
  41923. __ai int32x4_t vaddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  41924.   int32x4_t __ret;
  41925.   __ret = vcombine_s32(__p0, vaddhn_s64(__p1, __p2));
  41926.   return __ret;
  41927. }
  41928. #else
  41929. __ai int32x4_t vaddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  41930.   int32x4_t __ret;
  41931.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  41932.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  41933.   int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  41934.   __ret = __noswap_vcombine_s32(__rev0, __noswap_vaddhn_s64(__rev1, __rev2));
  41935.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  41936.   return __ret;
  41937. }
  41938. #endif
  41939.  
  41940. #ifdef __LITTLE_ENDIAN__
  41941. __ai int8x16_t vaddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  41942.   int8x16_t __ret;
  41943.   __ret = vcombine_s8(__p0, vaddhn_s16(__p1, __p2));
  41944.   return __ret;
  41945. }
  41946. #else
  41947. __ai int8x16_t vaddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  41948.   int8x16_t __ret;
  41949.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41950.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  41951.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  41952.   __ret = __noswap_vcombine_s8(__rev0, __noswap_vaddhn_s16(__rev1, __rev2));
  41953.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  41954.   return __ret;
  41955. }
  41956. #endif
  41957.  
  41958. #ifdef __LITTLE_ENDIAN__
  41959. __ai uint16_t vaddlvq_u8(uint8x16_t __p0) {
  41960.   uint16_t __ret;
  41961.   __ret = (uint16_t) __builtin_neon_vaddlvq_u8(__p0);
  41962.   return __ret;
  41963. }
  41964. #else
  41965. __ai uint16_t vaddlvq_u8(uint8x16_t __p0) {
  41966.   uint16_t __ret;
  41967.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  41968.   __ret = (uint16_t) __builtin_neon_vaddlvq_u8(__rev0);
  41969.   return __ret;
  41970. }
  41971. #endif
  41972.  
  41973. #ifdef __LITTLE_ENDIAN__
  41974. __ai uint64_t vaddlvq_u32(uint32x4_t __p0) {
  41975.   uint64_t __ret;
  41976.   __ret = (uint64_t) __builtin_neon_vaddlvq_u32(__p0);
  41977.   return __ret;
  41978. }
  41979. #else
  41980. __ai uint64_t vaddlvq_u32(uint32x4_t __p0) {
  41981.   uint64_t __ret;
  41982.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  41983.   __ret = (uint64_t) __builtin_neon_vaddlvq_u32(__rev0);
  41984.   return __ret;
  41985. }
  41986. #endif
  41987.  
  41988. #ifdef __LITTLE_ENDIAN__
  41989. __ai uint32_t vaddlvq_u16(uint16x8_t __p0) {
  41990.   uint32_t __ret;
  41991.   __ret = (uint32_t) __builtin_neon_vaddlvq_u16(__p0);
  41992.   return __ret;
  41993. }
  41994. #else
  41995. __ai uint32_t vaddlvq_u16(uint16x8_t __p0) {
  41996.   uint32_t __ret;
  41997.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  41998.   __ret = (uint32_t) __builtin_neon_vaddlvq_u16(__rev0);
  41999.   return __ret;
  42000. }
  42001. #endif
  42002.  
  42003. #ifdef __LITTLE_ENDIAN__
  42004. __ai int16_t vaddlvq_s8(int8x16_t __p0) {
  42005.   int16_t __ret;
  42006.   __ret = (int16_t) __builtin_neon_vaddlvq_s8(__p0);
  42007.   return __ret;
  42008. }
  42009. #else
  42010. __ai int16_t vaddlvq_s8(int8x16_t __p0) {
  42011.   int16_t __ret;
  42012.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  42013.   __ret = (int16_t) __builtin_neon_vaddlvq_s8(__rev0);
  42014.   return __ret;
  42015. }
  42016. #endif
  42017.  
  42018. #ifdef __LITTLE_ENDIAN__
  42019. __ai int64_t vaddlvq_s32(int32x4_t __p0) {
  42020.   int64_t __ret;
  42021.   __ret = (int64_t) __builtin_neon_vaddlvq_s32(__p0);
  42022.   return __ret;
  42023. }
  42024. #else
  42025. __ai int64_t vaddlvq_s32(int32x4_t __p0) {
  42026.   int64_t __ret;
  42027.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42028.   __ret = (int64_t) __builtin_neon_vaddlvq_s32(__rev0);
  42029.   return __ret;
  42030. }
  42031. #endif
  42032.  
  42033. #ifdef __LITTLE_ENDIAN__
  42034. __ai int32_t vaddlvq_s16(int16x8_t __p0) {
  42035.   int32_t __ret;
  42036.   __ret = (int32_t) __builtin_neon_vaddlvq_s16(__p0);
  42037.   return __ret;
  42038. }
  42039. #else
  42040. __ai int32_t vaddlvq_s16(int16x8_t __p0) {
  42041.   int32_t __ret;
  42042.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42043.   __ret = (int32_t) __builtin_neon_vaddlvq_s16(__rev0);
  42044.   return __ret;
  42045. }
  42046. #endif
  42047.  
  42048. #ifdef __LITTLE_ENDIAN__
  42049. __ai uint16_t vaddlv_u8(uint8x8_t __p0) {
  42050.   uint16_t __ret;
  42051.   __ret = (uint16_t) __builtin_neon_vaddlv_u8(__p0);
  42052.   return __ret;
  42053. }
  42054. #else
  42055. __ai uint16_t vaddlv_u8(uint8x8_t __p0) {
  42056.   uint16_t __ret;
  42057.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42058.   __ret = (uint16_t) __builtin_neon_vaddlv_u8(__rev0);
  42059.   return __ret;
  42060. }
  42061. #endif
  42062.  
  42063. #ifdef __LITTLE_ENDIAN__
  42064. __ai uint64_t vaddlv_u32(uint32x2_t __p0) {
  42065.   uint64_t __ret;
  42066.   __ret = (uint64_t) __builtin_neon_vaddlv_u32(__p0);
  42067.   return __ret;
  42068. }
  42069. #else
  42070. __ai uint64_t vaddlv_u32(uint32x2_t __p0) {
  42071.   uint64_t __ret;
  42072.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42073.   __ret = (uint64_t) __builtin_neon_vaddlv_u32(__rev0);
  42074.   return __ret;
  42075. }
  42076. #endif
  42077.  
  42078. #ifdef __LITTLE_ENDIAN__
  42079. __ai uint32_t vaddlv_u16(uint16x4_t __p0) {
  42080.   uint32_t __ret;
  42081.   __ret = (uint32_t) __builtin_neon_vaddlv_u16(__p0);
  42082.   return __ret;
  42083. }
  42084. #else
  42085. __ai uint32_t vaddlv_u16(uint16x4_t __p0) {
  42086.   uint32_t __ret;
  42087.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42088.   __ret = (uint32_t) __builtin_neon_vaddlv_u16(__rev0);
  42089.   return __ret;
  42090. }
  42091. #endif
  42092.  
  42093. #ifdef __LITTLE_ENDIAN__
  42094. __ai int16_t vaddlv_s8(int8x8_t __p0) {
  42095.   int16_t __ret;
  42096.   __ret = (int16_t) __builtin_neon_vaddlv_s8(__p0);
  42097.   return __ret;
  42098. }
  42099. #else
  42100. __ai int16_t vaddlv_s8(int8x8_t __p0) {
  42101.   int16_t __ret;
  42102.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42103.   __ret = (int16_t) __builtin_neon_vaddlv_s8(__rev0);
  42104.   return __ret;
  42105. }
  42106. #endif
  42107.  
  42108. #ifdef __LITTLE_ENDIAN__
  42109. __ai int64_t vaddlv_s32(int32x2_t __p0) {
  42110.   int64_t __ret;
  42111.   __ret = (int64_t) __builtin_neon_vaddlv_s32(__p0);
  42112.   return __ret;
  42113. }
  42114. #else
  42115. __ai int64_t vaddlv_s32(int32x2_t __p0) {
  42116.   int64_t __ret;
  42117.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42118.   __ret = (int64_t) __builtin_neon_vaddlv_s32(__rev0);
  42119.   return __ret;
  42120. }
  42121. #endif
  42122.  
  42123. #ifdef __LITTLE_ENDIAN__
  42124. __ai int32_t vaddlv_s16(int16x4_t __p0) {
  42125.   int32_t __ret;
  42126.   __ret = (int32_t) __builtin_neon_vaddlv_s16(__p0);
  42127.   return __ret;
  42128. }
  42129. #else
  42130. __ai int32_t vaddlv_s16(int16x4_t __p0) {
  42131.   int32_t __ret;
  42132.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42133.   __ret = (int32_t) __builtin_neon_vaddlv_s16(__rev0);
  42134.   return __ret;
  42135. }
  42136. #endif
  42137.  
  42138. #ifdef __LITTLE_ENDIAN__
  42139. __ai uint8_t vaddvq_u8(uint8x16_t __p0) {
  42140.   uint8_t __ret;
  42141.   __ret = (uint8_t) __builtin_neon_vaddvq_u8(__p0);
  42142.   return __ret;
  42143. }
  42144. #else
  42145. __ai uint8_t vaddvq_u8(uint8x16_t __p0) {
  42146.   uint8_t __ret;
  42147.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  42148.   __ret = (uint8_t) __builtin_neon_vaddvq_u8(__rev0);
  42149.   return __ret;
  42150. }
  42151. #endif
  42152.  
  42153. #ifdef __LITTLE_ENDIAN__
  42154. __ai uint32_t vaddvq_u32(uint32x4_t __p0) {
  42155.   uint32_t __ret;
  42156.   __ret = (uint32_t) __builtin_neon_vaddvq_u32(__p0);
  42157.   return __ret;
  42158. }
  42159. #else
  42160. __ai uint32_t vaddvq_u32(uint32x4_t __p0) {
  42161.   uint32_t __ret;
  42162.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42163.   __ret = (uint32_t) __builtin_neon_vaddvq_u32(__rev0);
  42164.   return __ret;
  42165. }
  42166. #endif
  42167.  
  42168. #ifdef __LITTLE_ENDIAN__
  42169. __ai uint64_t vaddvq_u64(uint64x2_t __p0) {
  42170.   uint64_t __ret;
  42171.   __ret = (uint64_t) __builtin_neon_vaddvq_u64(__p0);
  42172.   return __ret;
  42173. }
  42174. #else
  42175. __ai uint64_t vaddvq_u64(uint64x2_t __p0) {
  42176.   uint64_t __ret;
  42177.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42178.   __ret = (uint64_t) __builtin_neon_vaddvq_u64(__rev0);
  42179.   return __ret;
  42180. }
  42181. #endif
  42182.  
  42183. #ifdef __LITTLE_ENDIAN__
  42184. __ai uint16_t vaddvq_u16(uint16x8_t __p0) {
  42185.   uint16_t __ret;
  42186.   __ret = (uint16_t) __builtin_neon_vaddvq_u16(__p0);
  42187.   return __ret;
  42188. }
  42189. #else
  42190. __ai uint16_t vaddvq_u16(uint16x8_t __p0) {
  42191.   uint16_t __ret;
  42192.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42193.   __ret = (uint16_t) __builtin_neon_vaddvq_u16(__rev0);
  42194.   return __ret;
  42195. }
  42196. #endif
  42197.  
  42198. #ifdef __LITTLE_ENDIAN__
  42199. __ai int8_t vaddvq_s8(int8x16_t __p0) {
  42200.   int8_t __ret;
  42201.   __ret = (int8_t) __builtin_neon_vaddvq_s8(__p0);
  42202.   return __ret;
  42203. }
  42204. #else
  42205. __ai int8_t vaddvq_s8(int8x16_t __p0) {
  42206.   int8_t __ret;
  42207.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  42208.   __ret = (int8_t) __builtin_neon_vaddvq_s8(__rev0);
  42209.   return __ret;
  42210. }
  42211. #endif
  42212.  
  42213. #ifdef __LITTLE_ENDIAN__
  42214. __ai float64_t vaddvq_f64(float64x2_t __p0) {
  42215.   float64_t __ret;
  42216.   __ret = (float64_t) __builtin_neon_vaddvq_f64(__p0);
  42217.   return __ret;
  42218. }
  42219. #else
  42220. __ai float64_t vaddvq_f64(float64x2_t __p0) {
  42221.   float64_t __ret;
  42222.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42223.   __ret = (float64_t) __builtin_neon_vaddvq_f64(__rev0);
  42224.   return __ret;
  42225. }
  42226. #endif
  42227.  
  42228. #ifdef __LITTLE_ENDIAN__
  42229. __ai float32_t vaddvq_f32(float32x4_t __p0) {
  42230.   float32_t __ret;
  42231.   __ret = (float32_t) __builtin_neon_vaddvq_f32(__p0);
  42232.   return __ret;
  42233. }
  42234. #else
  42235. __ai float32_t vaddvq_f32(float32x4_t __p0) {
  42236.   float32_t __ret;
  42237.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42238.   __ret = (float32_t) __builtin_neon_vaddvq_f32(__rev0);
  42239.   return __ret;
  42240. }
  42241. #endif
  42242.  
  42243. #ifdef __LITTLE_ENDIAN__
  42244. __ai int32_t vaddvq_s32(int32x4_t __p0) {
  42245.   int32_t __ret;
  42246.   __ret = (int32_t) __builtin_neon_vaddvq_s32(__p0);
  42247.   return __ret;
  42248. }
  42249. #else
  42250. __ai int32_t vaddvq_s32(int32x4_t __p0) {
  42251.   int32_t __ret;
  42252.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42253.   __ret = (int32_t) __builtin_neon_vaddvq_s32(__rev0);
  42254.   return __ret;
  42255. }
  42256. #endif
  42257.  
  42258. #ifdef __LITTLE_ENDIAN__
  42259. __ai int64_t vaddvq_s64(int64x2_t __p0) {
  42260.   int64_t __ret;
  42261.   __ret = (int64_t) __builtin_neon_vaddvq_s64(__p0);
  42262.   return __ret;
  42263. }
  42264. #else
  42265. __ai int64_t vaddvq_s64(int64x2_t __p0) {
  42266.   int64_t __ret;
  42267.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42268.   __ret = (int64_t) __builtin_neon_vaddvq_s64(__rev0);
  42269.   return __ret;
  42270. }
  42271. #endif
  42272.  
  42273. #ifdef __LITTLE_ENDIAN__
  42274. __ai int16_t vaddvq_s16(int16x8_t __p0) {
  42275.   int16_t __ret;
  42276.   __ret = (int16_t) __builtin_neon_vaddvq_s16(__p0);
  42277.   return __ret;
  42278. }
  42279. #else
  42280. __ai int16_t vaddvq_s16(int16x8_t __p0) {
  42281.   int16_t __ret;
  42282.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42283.   __ret = (int16_t) __builtin_neon_vaddvq_s16(__rev0);
  42284.   return __ret;
  42285. }
  42286. #endif
  42287.  
  42288. #ifdef __LITTLE_ENDIAN__
  42289. __ai uint8_t vaddv_u8(uint8x8_t __p0) {
  42290.   uint8_t __ret;
  42291.   __ret = (uint8_t) __builtin_neon_vaddv_u8(__p0);
  42292.   return __ret;
  42293. }
  42294. #else
  42295. __ai uint8_t vaddv_u8(uint8x8_t __p0) {
  42296.   uint8_t __ret;
  42297.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42298.   __ret = (uint8_t) __builtin_neon_vaddv_u8(__rev0);
  42299.   return __ret;
  42300. }
  42301. #endif
  42302.  
  42303. #ifdef __LITTLE_ENDIAN__
  42304. __ai uint32_t vaddv_u32(uint32x2_t __p0) {
  42305.   uint32_t __ret;
  42306.   __ret = (uint32_t) __builtin_neon_vaddv_u32(__p0);
  42307.   return __ret;
  42308. }
  42309. #else
  42310. __ai uint32_t vaddv_u32(uint32x2_t __p0) {
  42311.   uint32_t __ret;
  42312.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42313.   __ret = (uint32_t) __builtin_neon_vaddv_u32(__rev0);
  42314.   return __ret;
  42315. }
  42316. #endif
  42317.  
  42318. #ifdef __LITTLE_ENDIAN__
  42319. __ai uint16_t vaddv_u16(uint16x4_t __p0) {
  42320.   uint16_t __ret;
  42321.   __ret = (uint16_t) __builtin_neon_vaddv_u16(__p0);
  42322.   return __ret;
  42323. }
  42324. #else
  42325. __ai uint16_t vaddv_u16(uint16x4_t __p0) {
  42326.   uint16_t __ret;
  42327.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42328.   __ret = (uint16_t) __builtin_neon_vaddv_u16(__rev0);
  42329.   return __ret;
  42330. }
  42331. #endif
  42332.  
  42333. #ifdef __LITTLE_ENDIAN__
  42334. __ai int8_t vaddv_s8(int8x8_t __p0) {
  42335.   int8_t __ret;
  42336.   __ret = (int8_t) __builtin_neon_vaddv_s8(__p0);
  42337.   return __ret;
  42338. }
  42339. #else
  42340. __ai int8_t vaddv_s8(int8x8_t __p0) {
  42341.   int8_t __ret;
  42342.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42343.   __ret = (int8_t) __builtin_neon_vaddv_s8(__rev0);
  42344.   return __ret;
  42345. }
  42346. #endif
  42347.  
  42348. #ifdef __LITTLE_ENDIAN__
  42349. __ai float32_t vaddv_f32(float32x2_t __p0) {
  42350.   float32_t __ret;
  42351.   __ret = (float32_t) __builtin_neon_vaddv_f32(__p0);
  42352.   return __ret;
  42353. }
  42354. #else
  42355. __ai float32_t vaddv_f32(float32x2_t __p0) {
  42356.   float32_t __ret;
  42357.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42358.   __ret = (float32_t) __builtin_neon_vaddv_f32(__rev0);
  42359.   return __ret;
  42360. }
  42361. #endif
  42362.  
  42363. #ifdef __LITTLE_ENDIAN__
  42364. __ai int32_t vaddv_s32(int32x2_t __p0) {
  42365.   int32_t __ret;
  42366.   __ret = (int32_t) __builtin_neon_vaddv_s32(__p0);
  42367.   return __ret;
  42368. }
  42369. #else
  42370. __ai int32_t vaddv_s32(int32x2_t __p0) {
  42371.   int32_t __ret;
  42372.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42373.   __ret = (int32_t) __builtin_neon_vaddv_s32(__rev0);
  42374.   return __ret;
  42375. }
  42376. #endif
  42377.  
  42378. #ifdef __LITTLE_ENDIAN__
  42379. __ai int16_t vaddv_s16(int16x4_t __p0) {
  42380.   int16_t __ret;
  42381.   __ret = (int16_t) __builtin_neon_vaddv_s16(__p0);
  42382.   return __ret;
  42383. }
  42384. #else
  42385. __ai int16_t vaddv_s16(int16x4_t __p0) {
  42386.   int16_t __ret;
  42387.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42388.   __ret = (int16_t) __builtin_neon_vaddv_s16(__rev0);
  42389.   return __ret;
  42390. }
  42391. #endif
  42392.  
  42393. __ai poly64x1_t vbsl_p64(uint64x1_t __p0, poly64x1_t __p1, poly64x1_t __p2) {
  42394.   poly64x1_t __ret;
  42395.   __ret = (poly64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 6);
  42396.   return __ret;
  42397. }
  42398. #ifdef __LITTLE_ENDIAN__
  42399. __ai poly64x2_t vbslq_p64(uint64x2_t __p0, poly64x2_t __p1, poly64x2_t __p2) {
  42400.   poly64x2_t __ret;
  42401.   __ret = (poly64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 38);
  42402.   return __ret;
  42403. }
  42404. #else
  42405. __ai poly64x2_t vbslq_p64(uint64x2_t __p0, poly64x2_t __p1, poly64x2_t __p2) {
  42406.   poly64x2_t __ret;
  42407.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42408.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42409.   poly64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  42410.   __ret = (poly64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 38);
  42411.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42412.   return __ret;
  42413. }
  42414. #endif
  42415.  
  42416. #ifdef __LITTLE_ENDIAN__
  42417. __ai float64x2_t vbslq_f64(uint64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  42418.   float64x2_t __ret;
  42419.   __ret = (float64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  42420.   return __ret;
  42421. }
  42422. #else
  42423. __ai float64x2_t vbslq_f64(uint64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  42424.   float64x2_t __ret;
  42425.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42426.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42427.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  42428.   __ret = (float64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
  42429.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42430.   return __ret;
  42431. }
  42432. #endif
  42433.  
  42434. __ai float64x1_t vbsl_f64(uint64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
  42435.   float64x1_t __ret;
  42436.   __ret = (float64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
  42437.   return __ret;
  42438. }
  42439. #ifdef __LITTLE_ENDIAN__
  42440. __ai uint64x2_t vcageq_f64(float64x2_t __p0, float64x2_t __p1) {
  42441.   uint64x2_t __ret;
  42442.   __ret = (uint64x2_t) __builtin_neon_vcageq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  42443.   return __ret;
  42444. }
  42445. #else
  42446. __ai uint64x2_t vcageq_f64(float64x2_t __p0, float64x2_t __p1) {
  42447.   uint64x2_t __ret;
  42448.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42449.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42450.   __ret = (uint64x2_t) __builtin_neon_vcageq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  42451.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42452.   return __ret;
  42453. }
  42454. #endif
  42455.  
  42456. __ai uint64x1_t vcage_f64(float64x1_t __p0, float64x1_t __p1) {
  42457.   uint64x1_t __ret;
  42458.   __ret = (uint64x1_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  42459.   return __ret;
  42460. }
  42461. __ai uint64_t vcaged_f64(float64_t __p0, float64_t __p1) {
  42462.   uint64_t __ret;
  42463.   __ret = (uint64_t) __builtin_neon_vcaged_f64(__p0, __p1);
  42464.   return __ret;
  42465. }
  42466. __ai uint32_t vcages_f32(float32_t __p0, float32_t __p1) {
  42467.   uint32_t __ret;
  42468.   __ret = (uint32_t) __builtin_neon_vcages_f32(__p0, __p1);
  42469.   return __ret;
  42470. }
  42471. #ifdef __LITTLE_ENDIAN__
  42472. __ai uint64x2_t vcagtq_f64(float64x2_t __p0, float64x2_t __p1) {
  42473.   uint64x2_t __ret;
  42474.   __ret = (uint64x2_t) __builtin_neon_vcagtq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  42475.   return __ret;
  42476. }
  42477. #else
  42478. __ai uint64x2_t vcagtq_f64(float64x2_t __p0, float64x2_t __p1) {
  42479.   uint64x2_t __ret;
  42480.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42481.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42482.   __ret = (uint64x2_t) __builtin_neon_vcagtq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  42483.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42484.   return __ret;
  42485. }
  42486. #endif
  42487.  
  42488. __ai uint64x1_t vcagt_f64(float64x1_t __p0, float64x1_t __p1) {
  42489.   uint64x1_t __ret;
  42490.   __ret = (uint64x1_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  42491.   return __ret;
  42492. }
  42493. __ai uint64_t vcagtd_f64(float64_t __p0, float64_t __p1) {
  42494.   uint64_t __ret;
  42495.   __ret = (uint64_t) __builtin_neon_vcagtd_f64(__p0, __p1);
  42496.   return __ret;
  42497. }
  42498. __ai uint32_t vcagts_f32(float32_t __p0, float32_t __p1) {
  42499.   uint32_t __ret;
  42500.   __ret = (uint32_t) __builtin_neon_vcagts_f32(__p0, __p1);
  42501.   return __ret;
  42502. }
  42503. #ifdef __LITTLE_ENDIAN__
  42504. __ai uint64x2_t vcaleq_f64(float64x2_t __p0, float64x2_t __p1) {
  42505.   uint64x2_t __ret;
  42506.   __ret = (uint64x2_t) __builtin_neon_vcaleq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  42507.   return __ret;
  42508. }
  42509. #else
  42510. __ai uint64x2_t vcaleq_f64(float64x2_t __p0, float64x2_t __p1) {
  42511.   uint64x2_t __ret;
  42512.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42513.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42514.   __ret = (uint64x2_t) __builtin_neon_vcaleq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  42515.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42516.   return __ret;
  42517. }
  42518. #endif
  42519.  
  42520. __ai uint64x1_t vcale_f64(float64x1_t __p0, float64x1_t __p1) {
  42521.   uint64x1_t __ret;
  42522.   __ret = (uint64x1_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  42523.   return __ret;
  42524. }
  42525. __ai uint64_t vcaled_f64(float64_t __p0, float64_t __p1) {
  42526.   uint64_t __ret;
  42527.   __ret = (uint64_t) __builtin_neon_vcaled_f64(__p0, __p1);
  42528.   return __ret;
  42529. }
  42530. __ai uint32_t vcales_f32(float32_t __p0, float32_t __p1) {
  42531.   uint32_t __ret;
  42532.   __ret = (uint32_t) __builtin_neon_vcales_f32(__p0, __p1);
  42533.   return __ret;
  42534. }
  42535. #ifdef __LITTLE_ENDIAN__
  42536. __ai uint64x2_t vcaltq_f64(float64x2_t __p0, float64x2_t __p1) {
  42537.   uint64x2_t __ret;
  42538.   __ret = (uint64x2_t) __builtin_neon_vcaltq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  42539.   return __ret;
  42540. }
  42541. #else
  42542. __ai uint64x2_t vcaltq_f64(float64x2_t __p0, float64x2_t __p1) {
  42543.   uint64x2_t __ret;
  42544.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42545.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42546.   __ret = (uint64x2_t) __builtin_neon_vcaltq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  42547.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42548.   return __ret;
  42549. }
  42550. #endif
  42551.  
  42552. __ai uint64x1_t vcalt_f64(float64x1_t __p0, float64x1_t __p1) {
  42553.   uint64x1_t __ret;
  42554.   __ret = (uint64x1_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  42555.   return __ret;
  42556. }
  42557. __ai uint64_t vcaltd_f64(float64_t __p0, float64_t __p1) {
  42558.   uint64_t __ret;
  42559.   __ret = (uint64_t) __builtin_neon_vcaltd_f64(__p0, __p1);
  42560.   return __ret;
  42561. }
  42562. __ai uint32_t vcalts_f32(float32_t __p0, float32_t __p1) {
  42563.   uint32_t __ret;
  42564.   __ret = (uint32_t) __builtin_neon_vcalts_f32(__p0, __p1);
  42565.   return __ret;
  42566. }
  42567. __ai uint64x1_t vceq_p64(poly64x1_t __p0, poly64x1_t __p1) {
  42568.   uint64x1_t __ret;
  42569.   __ret = (uint64x1_t)(__p0 == __p1);
  42570.   return __ret;
  42571. }
  42572. #ifdef __LITTLE_ENDIAN__
  42573. __ai uint64x2_t vceqq_p64(poly64x2_t __p0, poly64x2_t __p1) {
  42574.   uint64x2_t __ret;
  42575.   __ret = (uint64x2_t)(__p0 == __p1);
  42576.   return __ret;
  42577. }
  42578. #else
  42579. __ai uint64x2_t vceqq_p64(poly64x2_t __p0, poly64x2_t __p1) {
  42580.   uint64x2_t __ret;
  42581.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42582.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42583.   __ret = (uint64x2_t)(__rev0 == __rev1);
  42584.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42585.   return __ret;
  42586. }
  42587. #endif
  42588.  
  42589. #ifdef __LITTLE_ENDIAN__
  42590. __ai uint64x2_t vceqq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  42591.   uint64x2_t __ret;
  42592.   __ret = (uint64x2_t)(__p0 == __p1);
  42593.   return __ret;
  42594. }
  42595. #else
  42596. __ai uint64x2_t vceqq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  42597.   uint64x2_t __ret;
  42598.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42599.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42600.   __ret = (uint64x2_t)(__rev0 == __rev1);
  42601.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42602.   return __ret;
  42603. }
  42604. #endif
  42605.  
  42606. #ifdef __LITTLE_ENDIAN__
  42607. __ai uint64x2_t vceqq_f64(float64x2_t __p0, float64x2_t __p1) {
  42608.   uint64x2_t __ret;
  42609.   __ret = (uint64x2_t)(__p0 == __p1);
  42610.   return __ret;
  42611. }
  42612. #else
  42613. __ai uint64x2_t vceqq_f64(float64x2_t __p0, float64x2_t __p1) {
  42614.   uint64x2_t __ret;
  42615.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42616.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42617.   __ret = (uint64x2_t)(__rev0 == __rev1);
  42618.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42619.   return __ret;
  42620. }
  42621. #endif
  42622.  
  42623. #ifdef __LITTLE_ENDIAN__
  42624. __ai uint64x2_t vceqq_s64(int64x2_t __p0, int64x2_t __p1) {
  42625.   uint64x2_t __ret;
  42626.   __ret = (uint64x2_t)(__p0 == __p1);
  42627.   return __ret;
  42628. }
  42629. #else
  42630. __ai uint64x2_t vceqq_s64(int64x2_t __p0, int64x2_t __p1) {
  42631.   uint64x2_t __ret;
  42632.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42633.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  42634.   __ret = (uint64x2_t)(__rev0 == __rev1);
  42635.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42636.   return __ret;
  42637. }
  42638. #endif
  42639.  
  42640. __ai uint64x1_t vceq_u64(uint64x1_t __p0, uint64x1_t __p1) {
  42641.   uint64x1_t __ret;
  42642.   __ret = (uint64x1_t)(__p0 == __p1);
  42643.   return __ret;
  42644. }
  42645. __ai uint64x1_t vceq_f64(float64x1_t __p0, float64x1_t __p1) {
  42646.   uint64x1_t __ret;
  42647.   __ret = (uint64x1_t)(__p0 == __p1);
  42648.   return __ret;
  42649. }
  42650. __ai uint64x1_t vceq_s64(int64x1_t __p0, int64x1_t __p1) {
  42651.   uint64x1_t __ret;
  42652.   __ret = (uint64x1_t)(__p0 == __p1);
  42653.   return __ret;
  42654. }
  42655. __ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) {
  42656.   uint64_t __ret;
  42657.   __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1);
  42658.   return __ret;
  42659. }
  42660. __ai uint64_t vceqd_s64(int64_t __p0, int64_t __p1) {
  42661.   uint64_t __ret;
  42662.   __ret = (uint64_t) __builtin_neon_vceqd_s64(__p0, __p1);
  42663.   return __ret;
  42664. }
  42665. __ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) {
  42666.   uint64_t __ret;
  42667.   __ret = (uint64_t) __builtin_neon_vceqd_f64(__p0, __p1);
  42668.   return __ret;
  42669. }
  42670. __ai uint32_t vceqs_f32(float32_t __p0, float32_t __p1) {
  42671.   uint32_t __ret;
  42672.   __ret = (uint32_t) __builtin_neon_vceqs_f32(__p0, __p1);
  42673.   return __ret;
  42674. }
  42675. #ifdef __LITTLE_ENDIAN__
  42676. __ai uint8x8_t vceqz_p8(poly8x8_t __p0) {
  42677.   uint8x8_t __ret;
  42678.   __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 16);
  42679.   return __ret;
  42680. }
  42681. #else
  42682. __ai uint8x8_t vceqz_p8(poly8x8_t __p0) {
  42683.   uint8x8_t __ret;
  42684.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42685.   __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 16);
  42686.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  42687.   return __ret;
  42688. }
  42689. #endif
  42690.  
  42691. __ai uint64x1_t vceqz_p64(poly64x1_t __p0) {
  42692.   uint64x1_t __ret;
  42693.   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
  42694.   return __ret;
  42695. }
  42696. #ifdef __LITTLE_ENDIAN__
  42697. __ai uint8x16_t vceqzq_p8(poly8x16_t __p0) {
  42698.   uint8x16_t __ret;
  42699.   __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48);
  42700.   return __ret;
  42701. }
  42702. #else
  42703. __ai uint8x16_t vceqzq_p8(poly8x16_t __p0) {
  42704.   uint8x16_t __ret;
  42705.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  42706.   __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 48);
  42707.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  42708.   return __ret;
  42709. }
  42710. #endif
  42711.  
  42712. #ifdef __LITTLE_ENDIAN__
  42713. __ai uint64x2_t vceqzq_p64(poly64x2_t __p0) {
  42714.   uint64x2_t __ret;
  42715.   __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
  42716.   return __ret;
  42717. }
  42718. #else
  42719. __ai uint64x2_t vceqzq_p64(poly64x2_t __p0) {
  42720.   uint64x2_t __ret;
  42721.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42722.   __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
  42723.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42724.   return __ret;
  42725. }
  42726. #endif
  42727.  
  42728. #ifdef __LITTLE_ENDIAN__
  42729. __ai uint8x16_t vceqzq_u8(uint8x16_t __p0) {
  42730.   uint8x16_t __ret;
  42731.   __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48);
  42732.   return __ret;
  42733. }
  42734. #else
  42735. __ai uint8x16_t vceqzq_u8(uint8x16_t __p0) {
  42736.   uint8x16_t __ret;
  42737.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  42738.   __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 48);
  42739.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  42740.   return __ret;
  42741. }
  42742. #endif
  42743.  
  42744. #ifdef __LITTLE_ENDIAN__
  42745. __ai uint32x4_t vceqzq_u32(uint32x4_t __p0) {
  42746.   uint32x4_t __ret;
  42747.   __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 50);
  42748.   return __ret;
  42749. }
  42750. #else
  42751. __ai uint32x4_t vceqzq_u32(uint32x4_t __p0) {
  42752.   uint32x4_t __ret;
  42753.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42754.   __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 50);
  42755.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  42756.   return __ret;
  42757. }
  42758. #endif
  42759.  
  42760. #ifdef __LITTLE_ENDIAN__
  42761. __ai uint64x2_t vceqzq_u64(uint64x2_t __p0) {
  42762.   uint64x2_t __ret;
  42763.   __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
  42764.   return __ret;
  42765. }
  42766. #else
  42767. __ai uint64x2_t vceqzq_u64(uint64x2_t __p0) {
  42768.   uint64x2_t __ret;
  42769.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42770.   __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
  42771.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42772.   return __ret;
  42773. }
  42774. #endif
  42775.  
  42776. #ifdef __LITTLE_ENDIAN__
  42777. __ai uint16x8_t vceqzq_u16(uint16x8_t __p0) {
  42778.   uint16x8_t __ret;
  42779.   __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49);
  42780.   return __ret;
  42781. }
  42782. #else
  42783. __ai uint16x8_t vceqzq_u16(uint16x8_t __p0) {
  42784.   uint16x8_t __ret;
  42785.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42786.   __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49);
  42787.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  42788.   return __ret;
  42789. }
  42790. #endif
  42791.  
  42792. #ifdef __LITTLE_ENDIAN__
  42793. __ai uint8x16_t vceqzq_s8(int8x16_t __p0) {
  42794.   uint8x16_t __ret;
  42795.   __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48);
  42796.   return __ret;
  42797. }
  42798. #else
  42799. __ai uint8x16_t vceqzq_s8(int8x16_t __p0) {
  42800.   uint8x16_t __ret;
  42801.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  42802.   __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 48);
  42803.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  42804.   return __ret;
  42805. }
  42806. #endif
  42807.  
  42808. #ifdef __LITTLE_ENDIAN__
  42809. __ai uint64x2_t vceqzq_f64(float64x2_t __p0) {
  42810.   uint64x2_t __ret;
  42811.   __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
  42812.   return __ret;
  42813. }
  42814. #else
  42815. __ai uint64x2_t vceqzq_f64(float64x2_t __p0) {
  42816.   uint64x2_t __ret;
  42817.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42818.   __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
  42819.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42820.   return __ret;
  42821. }
  42822. #endif
  42823.  
  42824. #ifdef __LITTLE_ENDIAN__
  42825. __ai uint32x4_t vceqzq_f32(float32x4_t __p0) {
  42826.   uint32x4_t __ret;
  42827.   __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 50);
  42828.   return __ret;
  42829. }
  42830. #else
  42831. __ai uint32x4_t vceqzq_f32(float32x4_t __p0) {
  42832.   uint32x4_t __ret;
  42833.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42834.   __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 50);
  42835.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  42836.   return __ret;
  42837. }
  42838. #endif
  42839.  
  42840. #ifdef __LITTLE_ENDIAN__
  42841. __ai uint32x4_t vceqzq_s32(int32x4_t __p0) {
  42842.   uint32x4_t __ret;
  42843.   __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 50);
  42844.   return __ret;
  42845. }
  42846. #else
  42847. __ai uint32x4_t vceqzq_s32(int32x4_t __p0) {
  42848.   uint32x4_t __ret;
  42849.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42850.   __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 50);
  42851.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  42852.   return __ret;
  42853. }
  42854. #endif
  42855.  
  42856. #ifdef __LITTLE_ENDIAN__
  42857. __ai uint64x2_t vceqzq_s64(int64x2_t __p0) {
  42858.   uint64x2_t __ret;
  42859.   __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
  42860.   return __ret;
  42861. }
  42862. #else
  42863. __ai uint64x2_t vceqzq_s64(int64x2_t __p0) {
  42864.   uint64x2_t __ret;
  42865.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42866.   __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
  42867.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42868.   return __ret;
  42869. }
  42870. #endif
  42871.  
  42872. #ifdef __LITTLE_ENDIAN__
  42873. __ai uint16x8_t vceqzq_s16(int16x8_t __p0) {
  42874.   uint16x8_t __ret;
  42875.   __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49);
  42876.   return __ret;
  42877. }
  42878. #else
  42879. __ai uint16x8_t vceqzq_s16(int16x8_t __p0) {
  42880.   uint16x8_t __ret;
  42881.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42882.   __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49);
  42883.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  42884.   return __ret;
  42885. }
  42886. #endif
  42887.  
  42888. #ifdef __LITTLE_ENDIAN__
  42889. __ai uint8x8_t vceqz_u8(uint8x8_t __p0) {
  42890.   uint8x8_t __ret;
  42891.   __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 16);
  42892.   return __ret;
  42893. }
  42894. #else
  42895. __ai uint8x8_t vceqz_u8(uint8x8_t __p0) {
  42896.   uint8x8_t __ret;
  42897.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42898.   __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 16);
  42899.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  42900.   return __ret;
  42901. }
  42902. #endif
  42903.  
  42904. #ifdef __LITTLE_ENDIAN__
  42905. __ai uint32x2_t vceqz_u32(uint32x2_t __p0) {
  42906.   uint32x2_t __ret;
  42907.   __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 18);
  42908.   return __ret;
  42909. }
  42910. #else
  42911. __ai uint32x2_t vceqz_u32(uint32x2_t __p0) {
  42912.   uint32x2_t __ret;
  42913.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42914.   __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 18);
  42915.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42916.   return __ret;
  42917. }
  42918. #endif
  42919.  
  42920. __ai uint64x1_t vceqz_u64(uint64x1_t __p0) {
  42921.   uint64x1_t __ret;
  42922.   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
  42923.   return __ret;
  42924. }
  42925. #ifdef __LITTLE_ENDIAN__
  42926. __ai uint16x4_t vceqz_u16(uint16x4_t __p0) {
  42927.   uint16x4_t __ret;
  42928.   __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17);
  42929.   return __ret;
  42930. }
  42931. #else
  42932. __ai uint16x4_t vceqz_u16(uint16x4_t __p0) {
  42933.   uint16x4_t __ret;
  42934.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  42935.   __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17);
  42936.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  42937.   return __ret;
  42938. }
  42939. #endif
  42940.  
  42941. #ifdef __LITTLE_ENDIAN__
  42942. __ai uint8x8_t vceqz_s8(int8x8_t __p0) {
  42943.   uint8x8_t __ret;
  42944.   __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 16);
  42945.   return __ret;
  42946. }
  42947. #else
  42948. __ai uint8x8_t vceqz_s8(int8x8_t __p0) {
  42949.   uint8x8_t __ret;
  42950.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  42951.   __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 16);
  42952.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  42953.   return __ret;
  42954. }
  42955. #endif
  42956.  
  42957. __ai uint64x1_t vceqz_f64(float64x1_t __p0) {
  42958.   uint64x1_t __ret;
  42959.   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
  42960.   return __ret;
  42961. }
  42962. #ifdef __LITTLE_ENDIAN__
  42963. __ai uint32x2_t vceqz_f32(float32x2_t __p0) {
  42964.   uint32x2_t __ret;
  42965.   __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 18);
  42966.   return __ret;
  42967. }
  42968. #else
  42969. __ai uint32x2_t vceqz_f32(float32x2_t __p0) {
  42970.   uint32x2_t __ret;
  42971.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42972.   __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 18);
  42973.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42974.   return __ret;
  42975. }
  42976. #endif
  42977.  
  42978. #ifdef __LITTLE_ENDIAN__
  42979. __ai uint32x2_t vceqz_s32(int32x2_t __p0) {
  42980.   uint32x2_t __ret;
  42981.   __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 18);
  42982.   return __ret;
  42983. }
  42984. #else
  42985. __ai uint32x2_t vceqz_s32(int32x2_t __p0) {
  42986.   uint32x2_t __ret;
  42987.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  42988.   __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 18);
  42989.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  42990.   return __ret;
  42991. }
  42992. #endif
  42993.  
  42994. __ai uint64x1_t vceqz_s64(int64x1_t __p0) {
  42995.   uint64x1_t __ret;
  42996.   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
  42997.   return __ret;
  42998. }
  42999. #ifdef __LITTLE_ENDIAN__
  43000. __ai uint16x4_t vceqz_s16(int16x4_t __p0) {
  43001.   uint16x4_t __ret;
  43002.   __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17);
  43003.   return __ret;
  43004. }
  43005. #else
  43006. __ai uint16x4_t vceqz_s16(int16x4_t __p0) {
  43007.   uint16x4_t __ret;
  43008.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43009.   __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17);
  43010.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43011.   return __ret;
  43012. }
  43013. #endif
  43014.  
  43015. __ai uint64_t vceqzd_u64(uint64_t __p0) {
  43016.   uint64_t __ret;
  43017.   __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0);
  43018.   return __ret;
  43019. }
  43020. __ai uint64_t vceqzd_s64(int64_t __p0) {
  43021.   uint64_t __ret;
  43022.   __ret = (uint64_t) __builtin_neon_vceqzd_s64(__p0);
  43023.   return __ret;
  43024. }
  43025. __ai uint64_t vceqzd_f64(float64_t __p0) {
  43026.   uint64_t __ret;
  43027.   __ret = (uint64_t) __builtin_neon_vceqzd_f64(__p0);
  43028.   return __ret;
  43029. }
  43030. __ai uint32_t vceqzs_f32(float32_t __p0) {
  43031.   uint32_t __ret;
  43032.   __ret = (uint32_t) __builtin_neon_vceqzs_f32(__p0);
  43033.   return __ret;
  43034. }
  43035. #ifdef __LITTLE_ENDIAN__
  43036. __ai uint64x2_t vcgeq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  43037.   uint64x2_t __ret;
  43038.   __ret = (uint64x2_t)(__p0 >= __p1);
  43039.   return __ret;
  43040. }
  43041. #else
  43042. __ai uint64x2_t vcgeq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  43043.   uint64x2_t __ret;
  43044.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43045.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43046.   __ret = (uint64x2_t)(__rev0 >= __rev1);
  43047.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43048.   return __ret;
  43049. }
  43050. #endif
  43051.  
  43052. #ifdef __LITTLE_ENDIAN__
  43053. __ai uint64x2_t vcgeq_f64(float64x2_t __p0, float64x2_t __p1) {
  43054.   uint64x2_t __ret;
  43055.   __ret = (uint64x2_t)(__p0 >= __p1);
  43056.   return __ret;
  43057. }
  43058. #else
  43059. __ai uint64x2_t vcgeq_f64(float64x2_t __p0, float64x2_t __p1) {
  43060.   uint64x2_t __ret;
  43061.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43062.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43063.   __ret = (uint64x2_t)(__rev0 >= __rev1);
  43064.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43065.   return __ret;
  43066. }
  43067. #endif
  43068.  
  43069. #ifdef __LITTLE_ENDIAN__
  43070. __ai uint64x2_t vcgeq_s64(int64x2_t __p0, int64x2_t __p1) {
  43071.   uint64x2_t __ret;
  43072.   __ret = (uint64x2_t)(__p0 >= __p1);
  43073.   return __ret;
  43074. }
  43075. #else
  43076. __ai uint64x2_t vcgeq_s64(int64x2_t __p0, int64x2_t __p1) {
  43077.   uint64x2_t __ret;
  43078.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43079.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43080.   __ret = (uint64x2_t)(__rev0 >= __rev1);
  43081.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43082.   return __ret;
  43083. }
  43084. #endif
  43085.  
  43086. __ai uint64x1_t vcge_u64(uint64x1_t __p0, uint64x1_t __p1) {
  43087.   uint64x1_t __ret;
  43088.   __ret = (uint64x1_t)(__p0 >= __p1);
  43089.   return __ret;
  43090. }
  43091. __ai uint64x1_t vcge_f64(float64x1_t __p0, float64x1_t __p1) {
  43092.   uint64x1_t __ret;
  43093.   __ret = (uint64x1_t)(__p0 >= __p1);
  43094.   return __ret;
  43095. }
  43096. __ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) {
  43097.   uint64x1_t __ret;
  43098.   __ret = (uint64x1_t)(__p0 >= __p1);
  43099.   return __ret;
  43100. }
  43101. __ai uint64_t vcged_s64(int64_t __p0, int64_t __p1) {
  43102.   uint64_t __ret;
  43103.   __ret = (uint64_t) __builtin_neon_vcged_s64(__p0, __p1);
  43104.   return __ret;
  43105. }
  43106. __ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) {
  43107.   uint64_t __ret;
  43108.   __ret = (uint64_t) __builtin_neon_vcged_u64(__p0, __p1);
  43109.   return __ret;
  43110. }
  43111. __ai uint64_t vcged_f64(float64_t __p0, float64_t __p1) {
  43112.   uint64_t __ret;
  43113.   __ret = (uint64_t) __builtin_neon_vcged_f64(__p0, __p1);
  43114.   return __ret;
  43115. }
  43116. __ai uint32_t vcges_f32(float32_t __p0, float32_t __p1) {
  43117.   uint32_t __ret;
  43118.   __ret = (uint32_t) __builtin_neon_vcges_f32(__p0, __p1);
  43119.   return __ret;
  43120. }
  43121. #ifdef __LITTLE_ENDIAN__
  43122. __ai uint8x16_t vcgezq_s8(int8x16_t __p0) {
  43123.   uint8x16_t __ret;
  43124.   __ret = (uint8x16_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 48);
  43125.   return __ret;
  43126. }
  43127. #else
  43128. __ai uint8x16_t vcgezq_s8(int8x16_t __p0) {
  43129.   uint8x16_t __ret;
  43130.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  43131.   __ret = (uint8x16_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 48);
  43132.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  43133.   return __ret;
  43134. }
  43135. #endif
  43136.  
  43137. #ifdef __LITTLE_ENDIAN__
  43138. __ai uint64x2_t vcgezq_f64(float64x2_t __p0) {
  43139.   uint64x2_t __ret;
  43140.   __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 51);
  43141.   return __ret;
  43142. }
  43143. #else
  43144. __ai uint64x2_t vcgezq_f64(float64x2_t __p0) {
  43145.   uint64x2_t __ret;
  43146.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43147.   __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 51);
  43148.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43149.   return __ret;
  43150. }
  43151. #endif
  43152.  
  43153. #ifdef __LITTLE_ENDIAN__
  43154. __ai uint32x4_t vcgezq_f32(float32x4_t __p0) {
  43155.   uint32x4_t __ret;
  43156.   __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 50);
  43157.   return __ret;
  43158. }
  43159. #else
  43160. __ai uint32x4_t vcgezq_f32(float32x4_t __p0) {
  43161.   uint32x4_t __ret;
  43162.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43163.   __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 50);
  43164.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43165.   return __ret;
  43166. }
  43167. #endif
  43168.  
  43169. #ifdef __LITTLE_ENDIAN__
  43170. __ai uint32x4_t vcgezq_s32(int32x4_t __p0) {
  43171.   uint32x4_t __ret;
  43172.   __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 50);
  43173.   return __ret;
  43174. }
  43175. #else
  43176. __ai uint32x4_t vcgezq_s32(int32x4_t __p0) {
  43177.   uint32x4_t __ret;
  43178.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43179.   __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 50);
  43180.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43181.   return __ret;
  43182. }
  43183. #endif
  43184.  
  43185. #ifdef __LITTLE_ENDIAN__
  43186. __ai uint64x2_t vcgezq_s64(int64x2_t __p0) {
  43187.   uint64x2_t __ret;
  43188.   __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 51);
  43189.   return __ret;
  43190. }
  43191. #else
  43192. __ai uint64x2_t vcgezq_s64(int64x2_t __p0) {
  43193.   uint64x2_t __ret;
  43194.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43195.   __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 51);
  43196.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43197.   return __ret;
  43198. }
  43199. #endif
  43200.  
  43201. #ifdef __LITTLE_ENDIAN__
  43202. __ai uint16x8_t vcgezq_s16(int16x8_t __p0) {
  43203.   uint16x8_t __ret;
  43204.   __ret = (uint16x8_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 49);
  43205.   return __ret;
  43206. }
  43207. #else
  43208. __ai uint16x8_t vcgezq_s16(int16x8_t __p0) {
  43209.   uint16x8_t __ret;
  43210.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  43211.   __ret = (uint16x8_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 49);
  43212.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  43213.   return __ret;
  43214. }
  43215. #endif
  43216.  
  43217. #ifdef __LITTLE_ENDIAN__
  43218. __ai uint8x8_t vcgez_s8(int8x8_t __p0) {
  43219.   uint8x8_t __ret;
  43220.   __ret = (uint8x8_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 16);
  43221.   return __ret;
  43222. }
  43223. #else
  43224. __ai uint8x8_t vcgez_s8(int8x8_t __p0) {
  43225.   uint8x8_t __ret;
  43226.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  43227.   __ret = (uint8x8_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 16);
  43228.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  43229.   return __ret;
  43230. }
  43231. #endif
  43232.  
  43233. __ai uint64x1_t vcgez_f64(float64x1_t __p0) {
  43234.   uint64x1_t __ret;
  43235.   __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
  43236.   return __ret;
  43237. }
  43238. #ifdef __LITTLE_ENDIAN__
  43239. __ai uint32x2_t vcgez_f32(float32x2_t __p0) {
  43240.   uint32x2_t __ret;
  43241.   __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 18);
  43242.   return __ret;
  43243. }
  43244. #else
  43245. __ai uint32x2_t vcgez_f32(float32x2_t __p0) {
  43246.   uint32x2_t __ret;
  43247.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43248.   __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 18);
  43249.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43250.   return __ret;
  43251. }
  43252. #endif
  43253.  
  43254. #ifdef __LITTLE_ENDIAN__
  43255. __ai uint32x2_t vcgez_s32(int32x2_t __p0) {
  43256.   uint32x2_t __ret;
  43257.   __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 18);
  43258.   return __ret;
  43259. }
  43260. #else
  43261. __ai uint32x2_t vcgez_s32(int32x2_t __p0) {
  43262.   uint32x2_t __ret;
  43263.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43264.   __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 18);
  43265.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43266.   return __ret;
  43267. }
  43268. #endif
  43269.  
  43270. __ai uint64x1_t vcgez_s64(int64x1_t __p0) {
  43271.   uint64x1_t __ret;
  43272.   __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
  43273.   return __ret;
  43274. }
  43275. #ifdef __LITTLE_ENDIAN__
  43276. __ai uint16x4_t vcgez_s16(int16x4_t __p0) {
  43277.   uint16x4_t __ret;
  43278.   __ret = (uint16x4_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 17);
  43279.   return __ret;
  43280. }
  43281. #else
  43282. __ai uint16x4_t vcgez_s16(int16x4_t __p0) {
  43283.   uint16x4_t __ret;
  43284.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43285.   __ret = (uint16x4_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 17);
  43286.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43287.   return __ret;
  43288. }
  43289. #endif
  43290.  
  43291. __ai uint64_t vcgezd_s64(int64_t __p0) {
  43292.   uint64_t __ret;
  43293.   __ret = (uint64_t) __builtin_neon_vcgezd_s64(__p0);
  43294.   return __ret;
  43295. }
  43296. __ai uint64_t vcgezd_f64(float64_t __p0) {
  43297.   uint64_t __ret;
  43298.   __ret = (uint64_t) __builtin_neon_vcgezd_f64(__p0);
  43299.   return __ret;
  43300. }
  43301. __ai uint32_t vcgezs_f32(float32_t __p0) {
  43302.   uint32_t __ret;
  43303.   __ret = (uint32_t) __builtin_neon_vcgezs_f32(__p0);
  43304.   return __ret;
  43305. }
  43306. #ifdef __LITTLE_ENDIAN__
  43307. __ai uint64x2_t vcgtq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  43308.   uint64x2_t __ret;
  43309.   __ret = (uint64x2_t)(__p0 > __p1);
  43310.   return __ret;
  43311. }
  43312. #else
  43313. __ai uint64x2_t vcgtq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  43314.   uint64x2_t __ret;
  43315.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43316.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43317.   __ret = (uint64x2_t)(__rev0 > __rev1);
  43318.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43319.   return __ret;
  43320. }
  43321. #endif
  43322.  
  43323. #ifdef __LITTLE_ENDIAN__
  43324. __ai uint64x2_t vcgtq_f64(float64x2_t __p0, float64x2_t __p1) {
  43325.   uint64x2_t __ret;
  43326.   __ret = (uint64x2_t)(__p0 > __p1);
  43327.   return __ret;
  43328. }
  43329. #else
  43330. __ai uint64x2_t vcgtq_f64(float64x2_t __p0, float64x2_t __p1) {
  43331.   uint64x2_t __ret;
  43332.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43333.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43334.   __ret = (uint64x2_t)(__rev0 > __rev1);
  43335.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43336.   return __ret;
  43337. }
  43338. #endif
  43339.  
  43340. #ifdef __LITTLE_ENDIAN__
  43341. __ai uint64x2_t vcgtq_s64(int64x2_t __p0, int64x2_t __p1) {
  43342.   uint64x2_t __ret;
  43343.   __ret = (uint64x2_t)(__p0 > __p1);
  43344.   return __ret;
  43345. }
  43346. #else
  43347. __ai uint64x2_t vcgtq_s64(int64x2_t __p0, int64x2_t __p1) {
  43348.   uint64x2_t __ret;
  43349.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43350.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43351.   __ret = (uint64x2_t)(__rev0 > __rev1);
  43352.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43353.   return __ret;
  43354. }
  43355. #endif
  43356.  
  43357. __ai uint64x1_t vcgt_u64(uint64x1_t __p0, uint64x1_t __p1) {
  43358.   uint64x1_t __ret;
  43359.   __ret = (uint64x1_t)(__p0 > __p1);
  43360.   return __ret;
  43361. }
  43362. __ai uint64x1_t vcgt_f64(float64x1_t __p0, float64x1_t __p1) {
  43363.   uint64x1_t __ret;
  43364.   __ret = (uint64x1_t)(__p0 > __p1);
  43365.   return __ret;
  43366. }
  43367. __ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) {
  43368.   uint64x1_t __ret;
  43369.   __ret = (uint64x1_t)(__p0 > __p1);
  43370.   return __ret;
  43371. }
  43372. __ai uint64_t vcgtd_s64(int64_t __p0, int64_t __p1) {
  43373.   uint64_t __ret;
  43374.   __ret = (uint64_t) __builtin_neon_vcgtd_s64(__p0, __p1);
  43375.   return __ret;
  43376. }
  43377. __ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) {
  43378.   uint64_t __ret;
  43379.   __ret = (uint64_t) __builtin_neon_vcgtd_u64(__p0, __p1);
  43380.   return __ret;
  43381. }
  43382. __ai uint64_t vcgtd_f64(float64_t __p0, float64_t __p1) {
  43383.   uint64_t __ret;
  43384.   __ret = (uint64_t) __builtin_neon_vcgtd_f64(__p0, __p1);
  43385.   return __ret;
  43386. }
  43387. __ai uint32_t vcgts_f32(float32_t __p0, float32_t __p1) {
  43388.   uint32_t __ret;
  43389.   __ret = (uint32_t) __builtin_neon_vcgts_f32(__p0, __p1);
  43390.   return __ret;
  43391. }
  43392. #ifdef __LITTLE_ENDIAN__
  43393. __ai uint8x16_t vcgtzq_s8(int8x16_t __p0) {
  43394.   uint8x16_t __ret;
  43395.   __ret = (uint8x16_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 48);
  43396.   return __ret;
  43397. }
  43398. #else
  43399. __ai uint8x16_t vcgtzq_s8(int8x16_t __p0) {
  43400.   uint8x16_t __ret;
  43401.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  43402.   __ret = (uint8x16_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 48);
  43403.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  43404.   return __ret;
  43405. }
  43406. #endif
  43407.  
  43408. #ifdef __LITTLE_ENDIAN__
  43409. __ai uint64x2_t vcgtzq_f64(float64x2_t __p0) {
  43410.   uint64x2_t __ret;
  43411.   __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 51);
  43412.   return __ret;
  43413. }
  43414. #else
  43415. __ai uint64x2_t vcgtzq_f64(float64x2_t __p0) {
  43416.   uint64x2_t __ret;
  43417.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43418.   __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 51);
  43419.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43420.   return __ret;
  43421. }
  43422. #endif
  43423.  
  43424. #ifdef __LITTLE_ENDIAN__
  43425. __ai uint32x4_t vcgtzq_f32(float32x4_t __p0) {
  43426.   uint32x4_t __ret;
  43427.   __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 50);
  43428.   return __ret;
  43429. }
  43430. #else
  43431. __ai uint32x4_t vcgtzq_f32(float32x4_t __p0) {
  43432.   uint32x4_t __ret;
  43433.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43434.   __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 50);
  43435.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43436.   return __ret;
  43437. }
  43438. #endif
  43439.  
  43440. #ifdef __LITTLE_ENDIAN__
  43441. __ai uint32x4_t vcgtzq_s32(int32x4_t __p0) {
  43442.   uint32x4_t __ret;
  43443.   __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 50);
  43444.   return __ret;
  43445. }
  43446. #else
  43447. __ai uint32x4_t vcgtzq_s32(int32x4_t __p0) {
  43448.   uint32x4_t __ret;
  43449.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43450.   __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 50);
  43451.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43452.   return __ret;
  43453. }
  43454. #endif
  43455.  
  43456. #ifdef __LITTLE_ENDIAN__
  43457. __ai uint64x2_t vcgtzq_s64(int64x2_t __p0) {
  43458.   uint64x2_t __ret;
  43459.   __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 51);
  43460.   return __ret;
  43461. }
  43462. #else
  43463. __ai uint64x2_t vcgtzq_s64(int64x2_t __p0) {
  43464.   uint64x2_t __ret;
  43465.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43466.   __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 51);
  43467.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43468.   return __ret;
  43469. }
  43470. #endif
  43471.  
  43472. #ifdef __LITTLE_ENDIAN__
  43473. __ai uint16x8_t vcgtzq_s16(int16x8_t __p0) {
  43474.   uint16x8_t __ret;
  43475.   __ret = (uint16x8_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 49);
  43476.   return __ret;
  43477. }
  43478. #else
  43479. __ai uint16x8_t vcgtzq_s16(int16x8_t __p0) {
  43480.   uint16x8_t __ret;
  43481.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  43482.   __ret = (uint16x8_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 49);
  43483.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  43484.   return __ret;
  43485. }
  43486. #endif
  43487.  
  43488. #ifdef __LITTLE_ENDIAN__
  43489. __ai uint8x8_t vcgtz_s8(int8x8_t __p0) {
  43490.   uint8x8_t __ret;
  43491.   __ret = (uint8x8_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 16);
  43492.   return __ret;
  43493. }
  43494. #else
  43495. __ai uint8x8_t vcgtz_s8(int8x8_t __p0) {
  43496.   uint8x8_t __ret;
  43497.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  43498.   __ret = (uint8x8_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 16);
  43499.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  43500.   return __ret;
  43501. }
  43502. #endif
  43503.  
  43504. __ai uint64x1_t vcgtz_f64(float64x1_t __p0) {
  43505.   uint64x1_t __ret;
  43506.   __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
  43507.   return __ret;
  43508. }
  43509. #ifdef __LITTLE_ENDIAN__
  43510. __ai uint32x2_t vcgtz_f32(float32x2_t __p0) {
  43511.   uint32x2_t __ret;
  43512.   __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 18);
  43513.   return __ret;
  43514. }
  43515. #else
  43516. __ai uint32x2_t vcgtz_f32(float32x2_t __p0) {
  43517.   uint32x2_t __ret;
  43518.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43519.   __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 18);
  43520.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43521.   return __ret;
  43522. }
  43523. #endif
  43524.  
  43525. #ifdef __LITTLE_ENDIAN__
  43526. __ai uint32x2_t vcgtz_s32(int32x2_t __p0) {
  43527.   uint32x2_t __ret;
  43528.   __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 18);
  43529.   return __ret;
  43530. }
  43531. #else
  43532. __ai uint32x2_t vcgtz_s32(int32x2_t __p0) {
  43533.   uint32x2_t __ret;
  43534.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43535.   __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 18);
  43536.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43537.   return __ret;
  43538. }
  43539. #endif
  43540.  
  43541. __ai uint64x1_t vcgtz_s64(int64x1_t __p0) {
  43542.   uint64x1_t __ret;
  43543.   __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
  43544.   return __ret;
  43545. }
  43546. #ifdef __LITTLE_ENDIAN__
  43547. __ai uint16x4_t vcgtz_s16(int16x4_t __p0) {
  43548.   uint16x4_t __ret;
  43549.   __ret = (uint16x4_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 17);
  43550.   return __ret;
  43551. }
  43552. #else
  43553. __ai uint16x4_t vcgtz_s16(int16x4_t __p0) {
  43554.   uint16x4_t __ret;
  43555.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43556.   __ret = (uint16x4_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 17);
  43557.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43558.   return __ret;
  43559. }
  43560. #endif
  43561.  
  43562. __ai uint64_t vcgtzd_s64(int64_t __p0) {
  43563.   uint64_t __ret;
  43564.   __ret = (uint64_t) __builtin_neon_vcgtzd_s64(__p0);
  43565.   return __ret;
  43566. }
  43567. __ai uint64_t vcgtzd_f64(float64_t __p0) {
  43568.   uint64_t __ret;
  43569.   __ret = (uint64_t) __builtin_neon_vcgtzd_f64(__p0);
  43570.   return __ret;
  43571. }
  43572. __ai uint32_t vcgtzs_f32(float32_t __p0) {
  43573.   uint32_t __ret;
  43574.   __ret = (uint32_t) __builtin_neon_vcgtzs_f32(__p0);
  43575.   return __ret;
  43576. }
  43577. #ifdef __LITTLE_ENDIAN__
  43578. __ai uint64x2_t vcleq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  43579.   uint64x2_t __ret;
  43580.   __ret = (uint64x2_t)(__p0 <= __p1);
  43581.   return __ret;
  43582. }
  43583. #else
  43584. __ai uint64x2_t vcleq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  43585.   uint64x2_t __ret;
  43586.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43587.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43588.   __ret = (uint64x2_t)(__rev0 <= __rev1);
  43589.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43590.   return __ret;
  43591. }
  43592. #endif
  43593.  
  43594. #ifdef __LITTLE_ENDIAN__
  43595. __ai uint64x2_t vcleq_f64(float64x2_t __p0, float64x2_t __p1) {
  43596.   uint64x2_t __ret;
  43597.   __ret = (uint64x2_t)(__p0 <= __p1);
  43598.   return __ret;
  43599. }
  43600. #else
  43601. __ai uint64x2_t vcleq_f64(float64x2_t __p0, float64x2_t __p1) {
  43602.   uint64x2_t __ret;
  43603.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43604.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43605.   __ret = (uint64x2_t)(__rev0 <= __rev1);
  43606.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43607.   return __ret;
  43608. }
  43609. #endif
  43610.  
  43611. #ifdef __LITTLE_ENDIAN__
  43612. __ai uint64x2_t vcleq_s64(int64x2_t __p0, int64x2_t __p1) {
  43613.   uint64x2_t __ret;
  43614.   __ret = (uint64x2_t)(__p0 <= __p1);
  43615.   return __ret;
  43616. }
  43617. #else
  43618. __ai uint64x2_t vcleq_s64(int64x2_t __p0, int64x2_t __p1) {
  43619.   uint64x2_t __ret;
  43620.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43621.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43622.   __ret = (uint64x2_t)(__rev0 <= __rev1);
  43623.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43624.   return __ret;
  43625. }
  43626. #endif
  43627.  
  43628. __ai uint64x1_t vcle_u64(uint64x1_t __p0, uint64x1_t __p1) {
  43629.   uint64x1_t __ret;
  43630.   __ret = (uint64x1_t)(__p0 <= __p1);
  43631.   return __ret;
  43632. }
  43633. __ai uint64x1_t vcle_f64(float64x1_t __p0, float64x1_t __p1) {
  43634.   uint64x1_t __ret;
  43635.   __ret = (uint64x1_t)(__p0 <= __p1);
  43636.   return __ret;
  43637. }
  43638. __ai uint64x1_t vcle_s64(int64x1_t __p0, int64x1_t __p1) {
  43639.   uint64x1_t __ret;
  43640.   __ret = (uint64x1_t)(__p0 <= __p1);
  43641.   return __ret;
  43642. }
  43643. __ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) {
  43644.   uint64_t __ret;
  43645.   __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1);
  43646.   return __ret;
  43647. }
  43648. __ai uint64_t vcled_s64(int64_t __p0, int64_t __p1) {
  43649.   uint64_t __ret;
  43650.   __ret = (uint64_t) __builtin_neon_vcled_s64(__p0, __p1);
  43651.   return __ret;
  43652. }
  43653. __ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) {
  43654.   uint64_t __ret;
  43655.   __ret = (uint64_t) __builtin_neon_vcled_f64(__p0, __p1);
  43656.   return __ret;
  43657. }
  43658. __ai uint32_t vcles_f32(float32_t __p0, float32_t __p1) {
  43659.   uint32_t __ret;
  43660.   __ret = (uint32_t) __builtin_neon_vcles_f32(__p0, __p1);
  43661.   return __ret;
  43662. }
  43663. #ifdef __LITTLE_ENDIAN__
  43664. __ai uint8x16_t vclezq_s8(int8x16_t __p0) {
  43665.   uint8x16_t __ret;
  43666.   __ret = (uint8x16_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 48);
  43667.   return __ret;
  43668. }
  43669. #else
  43670. __ai uint8x16_t vclezq_s8(int8x16_t __p0) {
  43671.   uint8x16_t __ret;
  43672.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  43673.   __ret = (uint8x16_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 48);
  43674.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  43675.   return __ret;
  43676. }
  43677. #endif
  43678.  
  43679. #ifdef __LITTLE_ENDIAN__
  43680. __ai uint64x2_t vclezq_f64(float64x2_t __p0) {
  43681.   uint64x2_t __ret;
  43682.   __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 51);
  43683.   return __ret;
  43684. }
  43685. #else
  43686. __ai uint64x2_t vclezq_f64(float64x2_t __p0) {
  43687.   uint64x2_t __ret;
  43688.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43689.   __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 51);
  43690.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43691.   return __ret;
  43692. }
  43693. #endif
  43694.  
  43695. #ifdef __LITTLE_ENDIAN__
  43696. __ai uint32x4_t vclezq_f32(float32x4_t __p0) {
  43697.   uint32x4_t __ret;
  43698.   __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 50);
  43699.   return __ret;
  43700. }
  43701. #else
  43702. __ai uint32x4_t vclezq_f32(float32x4_t __p0) {
  43703.   uint32x4_t __ret;
  43704.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43705.   __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 50);
  43706.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43707.   return __ret;
  43708. }
  43709. #endif
  43710.  
  43711. #ifdef __LITTLE_ENDIAN__
  43712. __ai uint32x4_t vclezq_s32(int32x4_t __p0) {
  43713.   uint32x4_t __ret;
  43714.   __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 50);
  43715.   return __ret;
  43716. }
  43717. #else
  43718. __ai uint32x4_t vclezq_s32(int32x4_t __p0) {
  43719.   uint32x4_t __ret;
  43720.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43721.   __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 50);
  43722.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43723.   return __ret;
  43724. }
  43725. #endif
  43726.  
  43727. #ifdef __LITTLE_ENDIAN__
  43728. __ai uint64x2_t vclezq_s64(int64x2_t __p0) {
  43729.   uint64x2_t __ret;
  43730.   __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 51);
  43731.   return __ret;
  43732. }
  43733. #else
  43734. __ai uint64x2_t vclezq_s64(int64x2_t __p0) {
  43735.   uint64x2_t __ret;
  43736.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43737.   __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 51);
  43738.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43739.   return __ret;
  43740. }
  43741. #endif
  43742.  
  43743. #ifdef __LITTLE_ENDIAN__
  43744. __ai uint16x8_t vclezq_s16(int16x8_t __p0) {
  43745.   uint16x8_t __ret;
  43746.   __ret = (uint16x8_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 49);
  43747.   return __ret;
  43748. }
  43749. #else
  43750. __ai uint16x8_t vclezq_s16(int16x8_t __p0) {
  43751.   uint16x8_t __ret;
  43752.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  43753.   __ret = (uint16x8_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 49);
  43754.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  43755.   return __ret;
  43756. }
  43757. #endif
  43758.  
  43759. #ifdef __LITTLE_ENDIAN__
  43760. __ai uint8x8_t vclez_s8(int8x8_t __p0) {
  43761.   uint8x8_t __ret;
  43762.   __ret = (uint8x8_t) __builtin_neon_vclez_v((int8x8_t)__p0, 16);
  43763.   return __ret;
  43764. }
  43765. #else
  43766. __ai uint8x8_t vclez_s8(int8x8_t __p0) {
  43767.   uint8x8_t __ret;
  43768.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  43769.   __ret = (uint8x8_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 16);
  43770.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  43771.   return __ret;
  43772. }
  43773. #endif
  43774.  
  43775. __ai uint64x1_t vclez_f64(float64x1_t __p0) {
  43776.   uint64x1_t __ret;
  43777.   __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
  43778.   return __ret;
  43779. }
  43780. #ifdef __LITTLE_ENDIAN__
  43781. __ai uint32x2_t vclez_f32(float32x2_t __p0) {
  43782.   uint32x2_t __ret;
  43783.   __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__p0, 18);
  43784.   return __ret;
  43785. }
  43786. #else
  43787. __ai uint32x2_t vclez_f32(float32x2_t __p0) {
  43788.   uint32x2_t __ret;
  43789.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43790.   __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 18);
  43791.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43792.   return __ret;
  43793. }
  43794. #endif
  43795.  
  43796. #ifdef __LITTLE_ENDIAN__
  43797. __ai uint32x2_t vclez_s32(int32x2_t __p0) {
  43798.   uint32x2_t __ret;
  43799.   __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__p0, 18);
  43800.   return __ret;
  43801. }
  43802. #else
  43803. __ai uint32x2_t vclez_s32(int32x2_t __p0) {
  43804.   uint32x2_t __ret;
  43805.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43806.   __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 18);
  43807.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43808.   return __ret;
  43809. }
  43810. #endif
  43811.  
  43812. __ai uint64x1_t vclez_s64(int64x1_t __p0) {
  43813.   uint64x1_t __ret;
  43814.   __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
  43815.   return __ret;
  43816. }
  43817. #ifdef __LITTLE_ENDIAN__
  43818. __ai uint16x4_t vclez_s16(int16x4_t __p0) {
  43819.   uint16x4_t __ret;
  43820.   __ret = (uint16x4_t) __builtin_neon_vclez_v((int8x8_t)__p0, 17);
  43821.   return __ret;
  43822. }
  43823. #else
  43824. __ai uint16x4_t vclez_s16(int16x4_t __p0) {
  43825.   uint16x4_t __ret;
  43826.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43827.   __ret = (uint16x4_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 17);
  43828.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43829.   return __ret;
  43830. }
  43831. #endif
  43832.  
  43833. __ai uint64_t vclezd_s64(int64_t __p0) {
  43834.   uint64_t __ret;
  43835.   __ret = (uint64_t) __builtin_neon_vclezd_s64(__p0);
  43836.   return __ret;
  43837. }
  43838. __ai uint64_t vclezd_f64(float64_t __p0) {
  43839.   uint64_t __ret;
  43840.   __ret = (uint64_t) __builtin_neon_vclezd_f64(__p0);
  43841.   return __ret;
  43842. }
  43843. __ai uint32_t vclezs_f32(float32_t __p0) {
  43844.   uint32_t __ret;
  43845.   __ret = (uint32_t) __builtin_neon_vclezs_f32(__p0);
  43846.   return __ret;
  43847. }
  43848. #ifdef __LITTLE_ENDIAN__
  43849. __ai uint64x2_t vcltq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  43850.   uint64x2_t __ret;
  43851.   __ret = (uint64x2_t)(__p0 < __p1);
  43852.   return __ret;
  43853. }
  43854. #else
  43855. __ai uint64x2_t vcltq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  43856.   uint64x2_t __ret;
  43857.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43858.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43859.   __ret = (uint64x2_t)(__rev0 < __rev1);
  43860.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43861.   return __ret;
  43862. }
  43863. #endif
  43864.  
  43865. #ifdef __LITTLE_ENDIAN__
  43866. __ai uint64x2_t vcltq_f64(float64x2_t __p0, float64x2_t __p1) {
  43867.   uint64x2_t __ret;
  43868.   __ret = (uint64x2_t)(__p0 < __p1);
  43869.   return __ret;
  43870. }
  43871. #else
  43872. __ai uint64x2_t vcltq_f64(float64x2_t __p0, float64x2_t __p1) {
  43873.   uint64x2_t __ret;
  43874.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43875.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43876.   __ret = (uint64x2_t)(__rev0 < __rev1);
  43877.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43878.   return __ret;
  43879. }
  43880. #endif
  43881.  
  43882. #ifdef __LITTLE_ENDIAN__
  43883. __ai uint64x2_t vcltq_s64(int64x2_t __p0, int64x2_t __p1) {
  43884.   uint64x2_t __ret;
  43885.   __ret = (uint64x2_t)(__p0 < __p1);
  43886.   return __ret;
  43887. }
  43888. #else
  43889. __ai uint64x2_t vcltq_s64(int64x2_t __p0, int64x2_t __p1) {
  43890.   uint64x2_t __ret;
  43891.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43892.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  43893.   __ret = (uint64x2_t)(__rev0 < __rev1);
  43894.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43895.   return __ret;
  43896. }
  43897. #endif
  43898.  
  43899. __ai uint64x1_t vclt_u64(uint64x1_t __p0, uint64x1_t __p1) {
  43900.   uint64x1_t __ret;
  43901.   __ret = (uint64x1_t)(__p0 < __p1);
  43902.   return __ret;
  43903. }
  43904. __ai uint64x1_t vclt_f64(float64x1_t __p0, float64x1_t __p1) {
  43905.   uint64x1_t __ret;
  43906.   __ret = (uint64x1_t)(__p0 < __p1);
  43907.   return __ret;
  43908. }
  43909. __ai uint64x1_t vclt_s64(int64x1_t __p0, int64x1_t __p1) {
  43910.   uint64x1_t __ret;
  43911.   __ret = (uint64x1_t)(__p0 < __p1);
  43912.   return __ret;
  43913. }
  43914. __ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) {
  43915.   uint64_t __ret;
  43916.   __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1);
  43917.   return __ret;
  43918. }
  43919. __ai uint64_t vcltd_s64(int64_t __p0, int64_t __p1) {
  43920.   uint64_t __ret;
  43921.   __ret = (uint64_t) __builtin_neon_vcltd_s64(__p0, __p1);
  43922.   return __ret;
  43923. }
  43924. __ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) {
  43925.   uint64_t __ret;
  43926.   __ret = (uint64_t) __builtin_neon_vcltd_f64(__p0, __p1);
  43927.   return __ret;
  43928. }
  43929. __ai uint32_t vclts_f32(float32_t __p0, float32_t __p1) {
  43930.   uint32_t __ret;
  43931.   __ret = (uint32_t) __builtin_neon_vclts_f32(__p0, __p1);
  43932.   return __ret;
  43933. }
  43934. #ifdef __LITTLE_ENDIAN__
  43935. __ai uint8x16_t vcltzq_s8(int8x16_t __p0) {
  43936.   uint8x16_t __ret;
  43937.   __ret = (uint8x16_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 48);
  43938.   return __ret;
  43939. }
  43940. #else
  43941. __ai uint8x16_t vcltzq_s8(int8x16_t __p0) {
  43942.   uint8x16_t __ret;
  43943.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  43944.   __ret = (uint8x16_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 48);
  43945.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  43946.   return __ret;
  43947. }
  43948. #endif
  43949.  
  43950. #ifdef __LITTLE_ENDIAN__
  43951. __ai uint64x2_t vcltzq_f64(float64x2_t __p0) {
  43952.   uint64x2_t __ret;
  43953.   __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 51);
  43954.   return __ret;
  43955. }
  43956. #else
  43957. __ai uint64x2_t vcltzq_f64(float64x2_t __p0) {
  43958.   uint64x2_t __ret;
  43959.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  43960.   __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 51);
  43961.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  43962.   return __ret;
  43963. }
  43964. #endif
  43965.  
  43966. #ifdef __LITTLE_ENDIAN__
  43967. __ai uint32x4_t vcltzq_f32(float32x4_t __p0) {
  43968.   uint32x4_t __ret;
  43969.   __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 50);
  43970.   return __ret;
  43971. }
  43972. #else
  43973. __ai uint32x4_t vcltzq_f32(float32x4_t __p0) {
  43974.   uint32x4_t __ret;
  43975.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43976.   __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 50);
  43977.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43978.   return __ret;
  43979. }
  43980. #endif
  43981.  
  43982. #ifdef __LITTLE_ENDIAN__
  43983. __ai uint32x4_t vcltzq_s32(int32x4_t __p0) {
  43984.   uint32x4_t __ret;
  43985.   __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 50);
  43986.   return __ret;
  43987. }
  43988. #else
  43989. __ai uint32x4_t vcltzq_s32(int32x4_t __p0) {
  43990.   uint32x4_t __ret;
  43991.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  43992.   __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 50);
  43993.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  43994.   return __ret;
  43995. }
  43996. #endif
  43997.  
  43998. #ifdef __LITTLE_ENDIAN__
  43999. __ai uint64x2_t vcltzq_s64(int64x2_t __p0) {
  44000.   uint64x2_t __ret;
  44001.   __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 51);
  44002.   return __ret;
  44003. }
  44004. #else
  44005. __ai uint64x2_t vcltzq_s64(int64x2_t __p0) {
  44006.   uint64x2_t __ret;
  44007.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  44008.   __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 51);
  44009.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  44010.   return __ret;
  44011. }
  44012. #endif
  44013.  
  44014. #ifdef __LITTLE_ENDIAN__
  44015. __ai uint16x8_t vcltzq_s16(int16x8_t __p0) {
  44016.   uint16x8_t __ret;
  44017.   __ret = (uint16x8_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 49);
  44018.   return __ret;
  44019. }
  44020. #else
  44021. __ai uint16x8_t vcltzq_s16(int16x8_t __p0) {
  44022.   uint16x8_t __ret;
  44023.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  44024.   __ret = (uint16x8_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 49);
  44025.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  44026.   return __ret;
  44027. }
  44028. #endif
  44029.  
  44030. #ifdef __LITTLE_ENDIAN__
  44031. __ai uint8x8_t vcltz_s8(int8x8_t __p0) {
  44032.   uint8x8_t __ret;
  44033.   __ret = (uint8x8_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 16);
  44034.   return __ret;
  44035. }
  44036. #else
  44037. __ai uint8x8_t vcltz_s8(int8x8_t __p0) {
  44038.   uint8x8_t __ret;
  44039.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  44040.   __ret = (uint8x8_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 16);
  44041.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  44042.   return __ret;
  44043. }
  44044. #endif
  44045.  
  44046. __ai uint64x1_t vcltz_f64(float64x1_t __p0) {
  44047.   uint64x1_t __ret;
  44048.   __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
  44049.   return __ret;
  44050. }
  44051. #ifdef __LITTLE_ENDIAN__
  44052. __ai uint32x2_t vcltz_f32(float32x2_t __p0) {
  44053.   uint32x2_t __ret;
  44054.   __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 18);
  44055.   return __ret;
  44056. }
  44057. #else
  44058. __ai uint32x2_t vcltz_f32(float32x2_t __p0) {
  44059.   uint32x2_t __ret;
  44060.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  44061.   __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 18);
  44062.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  44063.   return __ret;
  44064. }
  44065. #endif
  44066.  
  44067. #ifdef __LITTLE_ENDIAN__
  44068. __ai uint32x2_t vcltz_s32(int32x2_t __p0) {
  44069.   uint32x2_t __ret;
  44070.   __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 18);
  44071.   return __ret;
  44072. }
  44073. #else
  44074. __ai uint32x2_t vcltz_s32(int32x2_t __p0) {
  44075.   uint32x2_t __ret;
  44076.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  44077.   __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 18);
  44078.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  44079.   return __ret;
  44080. }
  44081. #endif
  44082.  
  44083. __ai uint64x1_t vcltz_s64(int64x1_t __p0) {
  44084.   uint64x1_t __ret;
  44085.   __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
  44086.   return __ret;
  44087. }
  44088. #ifdef __LITTLE_ENDIAN__
  44089. __ai uint16x4_t vcltz_s16(int16x4_t __p0) {
  44090.   uint16x4_t __ret;
  44091.   __ret = (uint16x4_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 17);
  44092.   return __ret;
  44093. }
  44094. #else
  44095. __ai uint16x4_t vcltz_s16(int16x4_t __p0) {
  44096.   uint16x4_t __ret;
  44097.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  44098.   __ret = (uint16x4_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 17);
  44099.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  44100.   return __ret;
  44101. }
  44102. #endif
  44103.  
  44104. __ai uint64_t vcltzd_s64(int64_t __p0) {
  44105.   uint64_t __ret;
  44106.   __ret = (uint64_t) __builtin_neon_vcltzd_s64(__p0);
  44107.   return __ret;
  44108. }
  44109. __ai uint64_t vcltzd_f64(float64_t __p0) {
  44110.   uint64_t __ret;
  44111.   __ret = (uint64_t) __builtin_neon_vcltzd_f64(__p0);
  44112.   return __ret;
  44113. }
  44114. __ai uint32_t vcltzs_f32(float32_t __p0) {
  44115.   uint32_t __ret;
  44116.   __ret = (uint32_t) __builtin_neon_vcltzs_f32(__p0);
  44117.   return __ret;
  44118. }
  44119. #ifdef __LITTLE_ENDIAN__
  44120. __ai poly64x2_t vcombine_p64(poly64x1_t __p0, poly64x1_t __p1) {
  44121.   poly64x2_t __ret;
  44122.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  44123.   return __ret;
  44124. }
  44125. #else
  44126. __ai poly64x2_t vcombine_p64(poly64x1_t __p0, poly64x1_t __p1) {
  44127.   poly64x2_t __ret;
  44128.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  44129.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  44130.   return __ret;
  44131. }
  44132. #endif
  44133.  
  44134. #ifdef __LITTLE_ENDIAN__
  44135. __ai float64x2_t vcombine_f64(float64x1_t __p0, float64x1_t __p1) {
  44136.   float64x2_t __ret;
  44137.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  44138.   return __ret;
  44139. }
  44140. #else
  44141. __ai float64x2_t vcombine_f64(float64x1_t __p0, float64x1_t __p1) {
  44142.   float64x2_t __ret;
  44143.   __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
  44144.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  44145.   return __ret;
  44146. }
  44147. #endif
  44148.  
  44149. #ifdef __LITTLE_ENDIAN__
  44150. #define vcopyq_lane_p8(__p0_257, __p1_257, __p2_257, __p3_257) __extension__ ({ \
  44151.   poly8x16_t __ret_257; \
  44152.   poly8x16_t __s0_257 = __p0_257; \
  44153.   poly8x8_t __s2_257 = __p2_257; \
  44154.   __ret_257 = vsetq_lane_p8(vget_lane_p8(__s2_257, __p3_257), __s0_257, __p1_257); \
  44155.   __ret_257; \
  44156. })
  44157. #else
  44158. #define vcopyq_lane_p8(__p0_258, __p1_258, __p2_258, __p3_258) __extension__ ({ \
  44159.   poly8x16_t __ret_258; \
  44160.   poly8x16_t __s0_258 = __p0_258; \
  44161.   poly8x8_t __s2_258 = __p2_258; \
  44162.   poly8x16_t __rev0_258;  __rev0_258 = __builtin_shufflevector(__s0_258, __s0_258, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44163.   poly8x8_t __rev2_258;  __rev2_258 = __builtin_shufflevector(__s2_258, __s2_258, 7, 6, 5, 4, 3, 2, 1, 0); \
  44164.   __ret_258 = __noswap_vsetq_lane_p8(__noswap_vget_lane_p8(__rev2_258, __p3_258), __rev0_258, __p1_258); \
  44165.   __ret_258 = __builtin_shufflevector(__ret_258, __ret_258, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44166.   __ret_258; \
  44167. })
  44168. #endif
  44169.  
  44170. #ifdef __LITTLE_ENDIAN__
  44171. #define vcopyq_lane_p16(__p0_259, __p1_259, __p2_259, __p3_259) __extension__ ({ \
  44172.   poly16x8_t __ret_259; \
  44173.   poly16x8_t __s0_259 = __p0_259; \
  44174.   poly16x4_t __s2_259 = __p2_259; \
  44175.   __ret_259 = vsetq_lane_p16(vget_lane_p16(__s2_259, __p3_259), __s0_259, __p1_259); \
  44176.   __ret_259; \
  44177. })
  44178. #else
  44179. #define vcopyq_lane_p16(__p0_260, __p1_260, __p2_260, __p3_260) __extension__ ({ \
  44180.   poly16x8_t __ret_260; \
  44181.   poly16x8_t __s0_260 = __p0_260; \
  44182.   poly16x4_t __s2_260 = __p2_260; \
  44183.   poly16x8_t __rev0_260;  __rev0_260 = __builtin_shufflevector(__s0_260, __s0_260, 7, 6, 5, 4, 3, 2, 1, 0); \
  44184.   poly16x4_t __rev2_260;  __rev2_260 = __builtin_shufflevector(__s2_260, __s2_260, 3, 2, 1, 0); \
  44185.   __ret_260 = __noswap_vsetq_lane_p16(__noswap_vget_lane_p16(__rev2_260, __p3_260), __rev0_260, __p1_260); \
  44186.   __ret_260 = __builtin_shufflevector(__ret_260, __ret_260, 7, 6, 5, 4, 3, 2, 1, 0); \
  44187.   __ret_260; \
  44188. })
  44189. #endif
  44190.  
  44191. #ifdef __LITTLE_ENDIAN__
  44192. #define vcopyq_lane_u8(__p0_261, __p1_261, __p2_261, __p3_261) __extension__ ({ \
  44193.   uint8x16_t __ret_261; \
  44194.   uint8x16_t __s0_261 = __p0_261; \
  44195.   uint8x8_t __s2_261 = __p2_261; \
  44196.   __ret_261 = vsetq_lane_u8(vget_lane_u8(__s2_261, __p3_261), __s0_261, __p1_261); \
  44197.   __ret_261; \
  44198. })
  44199. #else
  44200. #define vcopyq_lane_u8(__p0_262, __p1_262, __p2_262, __p3_262) __extension__ ({ \
  44201.   uint8x16_t __ret_262; \
  44202.   uint8x16_t __s0_262 = __p0_262; \
  44203.   uint8x8_t __s2_262 = __p2_262; \
  44204.   uint8x16_t __rev0_262;  __rev0_262 = __builtin_shufflevector(__s0_262, __s0_262, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44205.   uint8x8_t __rev2_262;  __rev2_262 = __builtin_shufflevector(__s2_262, __s2_262, 7, 6, 5, 4, 3, 2, 1, 0); \
  44206.   __ret_262 = __noswap_vsetq_lane_u8(__noswap_vget_lane_u8(__rev2_262, __p3_262), __rev0_262, __p1_262); \
  44207.   __ret_262 = __builtin_shufflevector(__ret_262, __ret_262, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44208.   __ret_262; \
  44209. })
  44210. #endif
  44211.  
  44212. #ifdef __LITTLE_ENDIAN__
  44213. #define vcopyq_lane_u32(__p0_263, __p1_263, __p2_263, __p3_263) __extension__ ({ \
  44214.   uint32x4_t __ret_263; \
  44215.   uint32x4_t __s0_263 = __p0_263; \
  44216.   uint32x2_t __s2_263 = __p2_263; \
  44217.   __ret_263 = vsetq_lane_u32(vget_lane_u32(__s2_263, __p3_263), __s0_263, __p1_263); \
  44218.   __ret_263; \
  44219. })
  44220. #else
  44221. #define vcopyq_lane_u32(__p0_264, __p1_264, __p2_264, __p3_264) __extension__ ({ \
  44222.   uint32x4_t __ret_264; \
  44223.   uint32x4_t __s0_264 = __p0_264; \
  44224.   uint32x2_t __s2_264 = __p2_264; \
  44225.   uint32x4_t __rev0_264;  __rev0_264 = __builtin_shufflevector(__s0_264, __s0_264, 3, 2, 1, 0); \
  44226.   uint32x2_t __rev2_264;  __rev2_264 = __builtin_shufflevector(__s2_264, __s2_264, 1, 0); \
  44227.   __ret_264 = __noswap_vsetq_lane_u32(__noswap_vget_lane_u32(__rev2_264, __p3_264), __rev0_264, __p1_264); \
  44228.   __ret_264 = __builtin_shufflevector(__ret_264, __ret_264, 3, 2, 1, 0); \
  44229.   __ret_264; \
  44230. })
  44231. #endif
  44232.  
  44233. #ifdef __LITTLE_ENDIAN__
  44234. #define vcopyq_lane_u64(__p0_265, __p1_265, __p2_265, __p3_265) __extension__ ({ \
  44235.   uint64x2_t __ret_265; \
  44236.   uint64x2_t __s0_265 = __p0_265; \
  44237.   uint64x1_t __s2_265 = __p2_265; \
  44238.   __ret_265 = vsetq_lane_u64(vget_lane_u64(__s2_265, __p3_265), __s0_265, __p1_265); \
  44239.   __ret_265; \
  44240. })
  44241. #else
  44242. #define vcopyq_lane_u64(__p0_266, __p1_266, __p2_266, __p3_266) __extension__ ({ \
  44243.   uint64x2_t __ret_266; \
  44244.   uint64x2_t __s0_266 = __p0_266; \
  44245.   uint64x1_t __s2_266 = __p2_266; \
  44246.   uint64x2_t __rev0_266;  __rev0_266 = __builtin_shufflevector(__s0_266, __s0_266, 1, 0); \
  44247.   __ret_266 = __noswap_vsetq_lane_u64(vget_lane_u64(__s2_266, __p3_266), __rev0_266, __p1_266); \
  44248.   __ret_266 = __builtin_shufflevector(__ret_266, __ret_266, 1, 0); \
  44249.   __ret_266; \
  44250. })
  44251. #endif
  44252.  
  44253. #ifdef __LITTLE_ENDIAN__
  44254. #define vcopyq_lane_u16(__p0_267, __p1_267, __p2_267, __p3_267) __extension__ ({ \
  44255.   uint16x8_t __ret_267; \
  44256.   uint16x8_t __s0_267 = __p0_267; \
  44257.   uint16x4_t __s2_267 = __p2_267; \
  44258.   __ret_267 = vsetq_lane_u16(vget_lane_u16(__s2_267, __p3_267), __s0_267, __p1_267); \
  44259.   __ret_267; \
  44260. })
  44261. #else
  44262. #define vcopyq_lane_u16(__p0_268, __p1_268, __p2_268, __p3_268) __extension__ ({ \
  44263.   uint16x8_t __ret_268; \
  44264.   uint16x8_t __s0_268 = __p0_268; \
  44265.   uint16x4_t __s2_268 = __p2_268; \
  44266.   uint16x8_t __rev0_268;  __rev0_268 = __builtin_shufflevector(__s0_268, __s0_268, 7, 6, 5, 4, 3, 2, 1, 0); \
  44267.   uint16x4_t __rev2_268;  __rev2_268 = __builtin_shufflevector(__s2_268, __s2_268, 3, 2, 1, 0); \
  44268.   __ret_268 = __noswap_vsetq_lane_u16(__noswap_vget_lane_u16(__rev2_268, __p3_268), __rev0_268, __p1_268); \
  44269.   __ret_268 = __builtin_shufflevector(__ret_268, __ret_268, 7, 6, 5, 4, 3, 2, 1, 0); \
  44270.   __ret_268; \
  44271. })
  44272. #endif
  44273.  
  44274. #ifdef __LITTLE_ENDIAN__
  44275. #define vcopyq_lane_s8(__p0_269, __p1_269, __p2_269, __p3_269) __extension__ ({ \
  44276.   int8x16_t __ret_269; \
  44277.   int8x16_t __s0_269 = __p0_269; \
  44278.   int8x8_t __s2_269 = __p2_269; \
  44279.   __ret_269 = vsetq_lane_s8(vget_lane_s8(__s2_269, __p3_269), __s0_269, __p1_269); \
  44280.   __ret_269; \
  44281. })
  44282. #else
  44283. #define vcopyq_lane_s8(__p0_270, __p1_270, __p2_270, __p3_270) __extension__ ({ \
  44284.   int8x16_t __ret_270; \
  44285.   int8x16_t __s0_270 = __p0_270; \
  44286.   int8x8_t __s2_270 = __p2_270; \
  44287.   int8x16_t __rev0_270;  __rev0_270 = __builtin_shufflevector(__s0_270, __s0_270, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44288.   int8x8_t __rev2_270;  __rev2_270 = __builtin_shufflevector(__s2_270, __s2_270, 7, 6, 5, 4, 3, 2, 1, 0); \
  44289.   __ret_270 = __noswap_vsetq_lane_s8(__noswap_vget_lane_s8(__rev2_270, __p3_270), __rev0_270, __p1_270); \
  44290.   __ret_270 = __builtin_shufflevector(__ret_270, __ret_270, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44291.   __ret_270; \
  44292. })
  44293. #endif
  44294.  
  44295. #ifdef __LITTLE_ENDIAN__
  44296. #define vcopyq_lane_f32(__p0_271, __p1_271, __p2_271, __p3_271) __extension__ ({ \
  44297.   float32x4_t __ret_271; \
  44298.   float32x4_t __s0_271 = __p0_271; \
  44299.   float32x2_t __s2_271 = __p2_271; \
  44300.   __ret_271 = vsetq_lane_f32(vget_lane_f32(__s2_271, __p3_271), __s0_271, __p1_271); \
  44301.   __ret_271; \
  44302. })
  44303. #else
  44304. #define vcopyq_lane_f32(__p0_272, __p1_272, __p2_272, __p3_272) __extension__ ({ \
  44305.   float32x4_t __ret_272; \
  44306.   float32x4_t __s0_272 = __p0_272; \
  44307.   float32x2_t __s2_272 = __p2_272; \
  44308.   float32x4_t __rev0_272;  __rev0_272 = __builtin_shufflevector(__s0_272, __s0_272, 3, 2, 1, 0); \
  44309.   float32x2_t __rev2_272;  __rev2_272 = __builtin_shufflevector(__s2_272, __s2_272, 1, 0); \
  44310.   __ret_272 = __noswap_vsetq_lane_f32(__noswap_vget_lane_f32(__rev2_272, __p3_272), __rev0_272, __p1_272); \
  44311.   __ret_272 = __builtin_shufflevector(__ret_272, __ret_272, 3, 2, 1, 0); \
  44312.   __ret_272; \
  44313. })
  44314. #endif
  44315.  
  44316. #ifdef __LITTLE_ENDIAN__
  44317. #define vcopyq_lane_s32(__p0_273, __p1_273, __p2_273, __p3_273) __extension__ ({ \
  44318.   int32x4_t __ret_273; \
  44319.   int32x4_t __s0_273 = __p0_273; \
  44320.   int32x2_t __s2_273 = __p2_273; \
  44321.   __ret_273 = vsetq_lane_s32(vget_lane_s32(__s2_273, __p3_273), __s0_273, __p1_273); \
  44322.   __ret_273; \
  44323. })
  44324. #else
  44325. #define vcopyq_lane_s32(__p0_274, __p1_274, __p2_274, __p3_274) __extension__ ({ \
  44326.   int32x4_t __ret_274; \
  44327.   int32x4_t __s0_274 = __p0_274; \
  44328.   int32x2_t __s2_274 = __p2_274; \
  44329.   int32x4_t __rev0_274;  __rev0_274 = __builtin_shufflevector(__s0_274, __s0_274, 3, 2, 1, 0); \
  44330.   int32x2_t __rev2_274;  __rev2_274 = __builtin_shufflevector(__s2_274, __s2_274, 1, 0); \
  44331.   __ret_274 = __noswap_vsetq_lane_s32(__noswap_vget_lane_s32(__rev2_274, __p3_274), __rev0_274, __p1_274); \
  44332.   __ret_274 = __builtin_shufflevector(__ret_274, __ret_274, 3, 2, 1, 0); \
  44333.   __ret_274; \
  44334. })
  44335. #endif
  44336.  
  44337. #ifdef __LITTLE_ENDIAN__
  44338. #define vcopyq_lane_s64(__p0_275, __p1_275, __p2_275, __p3_275) __extension__ ({ \
  44339.   int64x2_t __ret_275; \
  44340.   int64x2_t __s0_275 = __p0_275; \
  44341.   int64x1_t __s2_275 = __p2_275; \
  44342.   __ret_275 = vsetq_lane_s64(vget_lane_s64(__s2_275, __p3_275), __s0_275, __p1_275); \
  44343.   __ret_275; \
  44344. })
  44345. #else
  44346. #define vcopyq_lane_s64(__p0_276, __p1_276, __p2_276, __p3_276) __extension__ ({ \
  44347.   int64x2_t __ret_276; \
  44348.   int64x2_t __s0_276 = __p0_276; \
  44349.   int64x1_t __s2_276 = __p2_276; \
  44350.   int64x2_t __rev0_276;  __rev0_276 = __builtin_shufflevector(__s0_276, __s0_276, 1, 0); \
  44351.   __ret_276 = __noswap_vsetq_lane_s64(vget_lane_s64(__s2_276, __p3_276), __rev0_276, __p1_276); \
  44352.   __ret_276 = __builtin_shufflevector(__ret_276, __ret_276, 1, 0); \
  44353.   __ret_276; \
  44354. })
  44355. #endif
  44356.  
  44357. #ifdef __LITTLE_ENDIAN__
  44358. #define vcopyq_lane_s16(__p0_277, __p1_277, __p2_277, __p3_277) __extension__ ({ \
  44359.   int16x8_t __ret_277; \
  44360.   int16x8_t __s0_277 = __p0_277; \
  44361.   int16x4_t __s2_277 = __p2_277; \
  44362.   __ret_277 = vsetq_lane_s16(vget_lane_s16(__s2_277, __p3_277), __s0_277, __p1_277); \
  44363.   __ret_277; \
  44364. })
  44365. #else
  44366. #define vcopyq_lane_s16(__p0_278, __p1_278, __p2_278, __p3_278) __extension__ ({ \
  44367.   int16x8_t __ret_278; \
  44368.   int16x8_t __s0_278 = __p0_278; \
  44369.   int16x4_t __s2_278 = __p2_278; \
  44370.   int16x8_t __rev0_278;  __rev0_278 = __builtin_shufflevector(__s0_278, __s0_278, 7, 6, 5, 4, 3, 2, 1, 0); \
  44371.   int16x4_t __rev2_278;  __rev2_278 = __builtin_shufflevector(__s2_278, __s2_278, 3, 2, 1, 0); \
  44372.   __ret_278 = __noswap_vsetq_lane_s16(__noswap_vget_lane_s16(__rev2_278, __p3_278), __rev0_278, __p1_278); \
  44373.   __ret_278 = __builtin_shufflevector(__ret_278, __ret_278, 7, 6, 5, 4, 3, 2, 1, 0); \
  44374.   __ret_278; \
  44375. })
  44376. #endif
  44377.  
  44378. #ifdef __LITTLE_ENDIAN__
  44379. #define vcopy_lane_p8(__p0_279, __p1_279, __p2_279, __p3_279) __extension__ ({ \
  44380.   poly8x8_t __ret_279; \
  44381.   poly8x8_t __s0_279 = __p0_279; \
  44382.   poly8x8_t __s2_279 = __p2_279; \
  44383.   __ret_279 = vset_lane_p8(vget_lane_p8(__s2_279, __p3_279), __s0_279, __p1_279); \
  44384.   __ret_279; \
  44385. })
  44386. #else
  44387. #define vcopy_lane_p8(__p0_280, __p1_280, __p2_280, __p3_280) __extension__ ({ \
  44388.   poly8x8_t __ret_280; \
  44389.   poly8x8_t __s0_280 = __p0_280; \
  44390.   poly8x8_t __s2_280 = __p2_280; \
  44391.   poly8x8_t __rev0_280;  __rev0_280 = __builtin_shufflevector(__s0_280, __s0_280, 7, 6, 5, 4, 3, 2, 1, 0); \
  44392.   poly8x8_t __rev2_280;  __rev2_280 = __builtin_shufflevector(__s2_280, __s2_280, 7, 6, 5, 4, 3, 2, 1, 0); \
  44393.   __ret_280 = __noswap_vset_lane_p8(__noswap_vget_lane_p8(__rev2_280, __p3_280), __rev0_280, __p1_280); \
  44394.   __ret_280 = __builtin_shufflevector(__ret_280, __ret_280, 7, 6, 5, 4, 3, 2, 1, 0); \
  44395.   __ret_280; \
  44396. })
  44397. #endif
  44398.  
  44399. #ifdef __LITTLE_ENDIAN__
  44400. #define vcopy_lane_p16(__p0_281, __p1_281, __p2_281, __p3_281) __extension__ ({ \
  44401.   poly16x4_t __ret_281; \
  44402.   poly16x4_t __s0_281 = __p0_281; \
  44403.   poly16x4_t __s2_281 = __p2_281; \
  44404.   __ret_281 = vset_lane_p16(vget_lane_p16(__s2_281, __p3_281), __s0_281, __p1_281); \
  44405.   __ret_281; \
  44406. })
  44407. #else
  44408. #define vcopy_lane_p16(__p0_282, __p1_282, __p2_282, __p3_282) __extension__ ({ \
  44409.   poly16x4_t __ret_282; \
  44410.   poly16x4_t __s0_282 = __p0_282; \
  44411.   poly16x4_t __s2_282 = __p2_282; \
  44412.   poly16x4_t __rev0_282;  __rev0_282 = __builtin_shufflevector(__s0_282, __s0_282, 3, 2, 1, 0); \
  44413.   poly16x4_t __rev2_282;  __rev2_282 = __builtin_shufflevector(__s2_282, __s2_282, 3, 2, 1, 0); \
  44414.   __ret_282 = __noswap_vset_lane_p16(__noswap_vget_lane_p16(__rev2_282, __p3_282), __rev0_282, __p1_282); \
  44415.   __ret_282 = __builtin_shufflevector(__ret_282, __ret_282, 3, 2, 1, 0); \
  44416.   __ret_282; \
  44417. })
  44418. #endif
  44419.  
  44420. #ifdef __LITTLE_ENDIAN__
  44421. #define vcopy_lane_u8(__p0_283, __p1_283, __p2_283, __p3_283) __extension__ ({ \
  44422.   uint8x8_t __ret_283; \
  44423.   uint8x8_t __s0_283 = __p0_283; \
  44424.   uint8x8_t __s2_283 = __p2_283; \
  44425.   __ret_283 = vset_lane_u8(vget_lane_u8(__s2_283, __p3_283), __s0_283, __p1_283); \
  44426.   __ret_283; \
  44427. })
  44428. #else
  44429. #define vcopy_lane_u8(__p0_284, __p1_284, __p2_284, __p3_284) __extension__ ({ \
  44430.   uint8x8_t __ret_284; \
  44431.   uint8x8_t __s0_284 = __p0_284; \
  44432.   uint8x8_t __s2_284 = __p2_284; \
  44433.   uint8x8_t __rev0_284;  __rev0_284 = __builtin_shufflevector(__s0_284, __s0_284, 7, 6, 5, 4, 3, 2, 1, 0); \
  44434.   uint8x8_t __rev2_284;  __rev2_284 = __builtin_shufflevector(__s2_284, __s2_284, 7, 6, 5, 4, 3, 2, 1, 0); \
  44435.   __ret_284 = __noswap_vset_lane_u8(__noswap_vget_lane_u8(__rev2_284, __p3_284), __rev0_284, __p1_284); \
  44436.   __ret_284 = __builtin_shufflevector(__ret_284, __ret_284, 7, 6, 5, 4, 3, 2, 1, 0); \
  44437.   __ret_284; \
  44438. })
  44439. #endif
  44440.  
  44441. #ifdef __LITTLE_ENDIAN__
  44442. #define vcopy_lane_u32(__p0_285, __p1_285, __p2_285, __p3_285) __extension__ ({ \
  44443.   uint32x2_t __ret_285; \
  44444.   uint32x2_t __s0_285 = __p0_285; \
  44445.   uint32x2_t __s2_285 = __p2_285; \
  44446.   __ret_285 = vset_lane_u32(vget_lane_u32(__s2_285, __p3_285), __s0_285, __p1_285); \
  44447.   __ret_285; \
  44448. })
  44449. #else
  44450. #define vcopy_lane_u32(__p0_286, __p1_286, __p2_286, __p3_286) __extension__ ({ \
  44451.   uint32x2_t __ret_286; \
  44452.   uint32x2_t __s0_286 = __p0_286; \
  44453.   uint32x2_t __s2_286 = __p2_286; \
  44454.   uint32x2_t __rev0_286;  __rev0_286 = __builtin_shufflevector(__s0_286, __s0_286, 1, 0); \
  44455.   uint32x2_t __rev2_286;  __rev2_286 = __builtin_shufflevector(__s2_286, __s2_286, 1, 0); \
  44456.   __ret_286 = __noswap_vset_lane_u32(__noswap_vget_lane_u32(__rev2_286, __p3_286), __rev0_286, __p1_286); \
  44457.   __ret_286 = __builtin_shufflevector(__ret_286, __ret_286, 1, 0); \
  44458.   __ret_286; \
  44459. })
  44460. #endif
  44461.  
  44462. #define vcopy_lane_u64(__p0_287, __p1_287, __p2_287, __p3_287) __extension__ ({ \
  44463.   uint64x1_t __ret_287; \
  44464.   uint64x1_t __s0_287 = __p0_287; \
  44465.   uint64x1_t __s2_287 = __p2_287; \
  44466.   __ret_287 = vset_lane_u64(vget_lane_u64(__s2_287, __p3_287), __s0_287, __p1_287); \
  44467.   __ret_287; \
  44468. })
  44469. #ifdef __LITTLE_ENDIAN__
  44470. #define vcopy_lane_u16(__p0_288, __p1_288, __p2_288, __p3_288) __extension__ ({ \
  44471.   uint16x4_t __ret_288; \
  44472.   uint16x4_t __s0_288 = __p0_288; \
  44473.   uint16x4_t __s2_288 = __p2_288; \
  44474.   __ret_288 = vset_lane_u16(vget_lane_u16(__s2_288, __p3_288), __s0_288, __p1_288); \
  44475.   __ret_288; \
  44476. })
  44477. #else
  44478. #define vcopy_lane_u16(__p0_289, __p1_289, __p2_289, __p3_289) __extension__ ({ \
  44479.   uint16x4_t __ret_289; \
  44480.   uint16x4_t __s0_289 = __p0_289; \
  44481.   uint16x4_t __s2_289 = __p2_289; \
  44482.   uint16x4_t __rev0_289;  __rev0_289 = __builtin_shufflevector(__s0_289, __s0_289, 3, 2, 1, 0); \
  44483.   uint16x4_t __rev2_289;  __rev2_289 = __builtin_shufflevector(__s2_289, __s2_289, 3, 2, 1, 0); \
  44484.   __ret_289 = __noswap_vset_lane_u16(__noswap_vget_lane_u16(__rev2_289, __p3_289), __rev0_289, __p1_289); \
  44485.   __ret_289 = __builtin_shufflevector(__ret_289, __ret_289, 3, 2, 1, 0); \
  44486.   __ret_289; \
  44487. })
  44488. #endif
  44489.  
  44490. #ifdef __LITTLE_ENDIAN__
  44491. #define vcopy_lane_s8(__p0_290, __p1_290, __p2_290, __p3_290) __extension__ ({ \
  44492.   int8x8_t __ret_290; \
  44493.   int8x8_t __s0_290 = __p0_290; \
  44494.   int8x8_t __s2_290 = __p2_290; \
  44495.   __ret_290 = vset_lane_s8(vget_lane_s8(__s2_290, __p3_290), __s0_290, __p1_290); \
  44496.   __ret_290; \
  44497. })
  44498. #else
  44499. #define vcopy_lane_s8(__p0_291, __p1_291, __p2_291, __p3_291) __extension__ ({ \
  44500.   int8x8_t __ret_291; \
  44501.   int8x8_t __s0_291 = __p0_291; \
  44502.   int8x8_t __s2_291 = __p2_291; \
  44503.   int8x8_t __rev0_291;  __rev0_291 = __builtin_shufflevector(__s0_291, __s0_291, 7, 6, 5, 4, 3, 2, 1, 0); \
  44504.   int8x8_t __rev2_291;  __rev2_291 = __builtin_shufflevector(__s2_291, __s2_291, 7, 6, 5, 4, 3, 2, 1, 0); \
  44505.   __ret_291 = __noswap_vset_lane_s8(__noswap_vget_lane_s8(__rev2_291, __p3_291), __rev0_291, __p1_291); \
  44506.   __ret_291 = __builtin_shufflevector(__ret_291, __ret_291, 7, 6, 5, 4, 3, 2, 1, 0); \
  44507.   __ret_291; \
  44508. })
  44509. #endif
  44510.  
  44511. #ifdef __LITTLE_ENDIAN__
  44512. #define vcopy_lane_f32(__p0_292, __p1_292, __p2_292, __p3_292) __extension__ ({ \
  44513.   float32x2_t __ret_292; \
  44514.   float32x2_t __s0_292 = __p0_292; \
  44515.   float32x2_t __s2_292 = __p2_292; \
  44516.   __ret_292 = vset_lane_f32(vget_lane_f32(__s2_292, __p3_292), __s0_292, __p1_292); \
  44517.   __ret_292; \
  44518. })
  44519. #else
  44520. #define vcopy_lane_f32(__p0_293, __p1_293, __p2_293, __p3_293) __extension__ ({ \
  44521.   float32x2_t __ret_293; \
  44522.   float32x2_t __s0_293 = __p0_293; \
  44523.   float32x2_t __s2_293 = __p2_293; \
  44524.   float32x2_t __rev0_293;  __rev0_293 = __builtin_shufflevector(__s0_293, __s0_293, 1, 0); \
  44525.   float32x2_t __rev2_293;  __rev2_293 = __builtin_shufflevector(__s2_293, __s2_293, 1, 0); \
  44526.   __ret_293 = __noswap_vset_lane_f32(__noswap_vget_lane_f32(__rev2_293, __p3_293), __rev0_293, __p1_293); \
  44527.   __ret_293 = __builtin_shufflevector(__ret_293, __ret_293, 1, 0); \
  44528.   __ret_293; \
  44529. })
  44530. #endif
  44531.  
  44532. #ifdef __LITTLE_ENDIAN__
  44533. #define vcopy_lane_s32(__p0_294, __p1_294, __p2_294, __p3_294) __extension__ ({ \
  44534.   int32x2_t __ret_294; \
  44535.   int32x2_t __s0_294 = __p0_294; \
  44536.   int32x2_t __s2_294 = __p2_294; \
  44537.   __ret_294 = vset_lane_s32(vget_lane_s32(__s2_294, __p3_294), __s0_294, __p1_294); \
  44538.   __ret_294; \
  44539. })
  44540. #else
  44541. #define vcopy_lane_s32(__p0_295, __p1_295, __p2_295, __p3_295) __extension__ ({ \
  44542.   int32x2_t __ret_295; \
  44543.   int32x2_t __s0_295 = __p0_295; \
  44544.   int32x2_t __s2_295 = __p2_295; \
  44545.   int32x2_t __rev0_295;  __rev0_295 = __builtin_shufflevector(__s0_295, __s0_295, 1, 0); \
  44546.   int32x2_t __rev2_295;  __rev2_295 = __builtin_shufflevector(__s2_295, __s2_295, 1, 0); \
  44547.   __ret_295 = __noswap_vset_lane_s32(__noswap_vget_lane_s32(__rev2_295, __p3_295), __rev0_295, __p1_295); \
  44548.   __ret_295 = __builtin_shufflevector(__ret_295, __ret_295, 1, 0); \
  44549.   __ret_295; \
  44550. })
  44551. #endif
  44552.  
  44553. #define vcopy_lane_s64(__p0_296, __p1_296, __p2_296, __p3_296) __extension__ ({ \
  44554.   int64x1_t __ret_296; \
  44555.   int64x1_t __s0_296 = __p0_296; \
  44556.   int64x1_t __s2_296 = __p2_296; \
  44557.   __ret_296 = vset_lane_s64(vget_lane_s64(__s2_296, __p3_296), __s0_296, __p1_296); \
  44558.   __ret_296; \
  44559. })
  44560. #ifdef __LITTLE_ENDIAN__
  44561. #define vcopy_lane_s16(__p0_297, __p1_297, __p2_297, __p3_297) __extension__ ({ \
  44562.   int16x4_t __ret_297; \
  44563.   int16x4_t __s0_297 = __p0_297; \
  44564.   int16x4_t __s2_297 = __p2_297; \
  44565.   __ret_297 = vset_lane_s16(vget_lane_s16(__s2_297, __p3_297), __s0_297, __p1_297); \
  44566.   __ret_297; \
  44567. })
  44568. #else
  44569. #define vcopy_lane_s16(__p0_298, __p1_298, __p2_298, __p3_298) __extension__ ({ \
  44570.   int16x4_t __ret_298; \
  44571.   int16x4_t __s0_298 = __p0_298; \
  44572.   int16x4_t __s2_298 = __p2_298; \
  44573.   int16x4_t __rev0_298;  __rev0_298 = __builtin_shufflevector(__s0_298, __s0_298, 3, 2, 1, 0); \
  44574.   int16x4_t __rev2_298;  __rev2_298 = __builtin_shufflevector(__s2_298, __s2_298, 3, 2, 1, 0); \
  44575.   __ret_298 = __noswap_vset_lane_s16(__noswap_vget_lane_s16(__rev2_298, __p3_298), __rev0_298, __p1_298); \
  44576.   __ret_298 = __builtin_shufflevector(__ret_298, __ret_298, 3, 2, 1, 0); \
  44577.   __ret_298; \
  44578. })
  44579. #endif
  44580.  
  44581. #ifdef __LITTLE_ENDIAN__
  44582. #define vcopyq_laneq_p8(__p0_299, __p1_299, __p2_299, __p3_299) __extension__ ({ \
  44583.   poly8x16_t __ret_299; \
  44584.   poly8x16_t __s0_299 = __p0_299; \
  44585.   poly8x16_t __s2_299 = __p2_299; \
  44586.   __ret_299 = vsetq_lane_p8(vgetq_lane_p8(__s2_299, __p3_299), __s0_299, __p1_299); \
  44587.   __ret_299; \
  44588. })
  44589. #else
  44590. #define vcopyq_laneq_p8(__p0_300, __p1_300, __p2_300, __p3_300) __extension__ ({ \
  44591.   poly8x16_t __ret_300; \
  44592.   poly8x16_t __s0_300 = __p0_300; \
  44593.   poly8x16_t __s2_300 = __p2_300; \
  44594.   poly8x16_t __rev0_300;  __rev0_300 = __builtin_shufflevector(__s0_300, __s0_300, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44595.   poly8x16_t __rev2_300;  __rev2_300 = __builtin_shufflevector(__s2_300, __s2_300, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44596.   __ret_300 = __noswap_vsetq_lane_p8(__noswap_vgetq_lane_p8(__rev2_300, __p3_300), __rev0_300, __p1_300); \
  44597.   __ret_300 = __builtin_shufflevector(__ret_300, __ret_300, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44598.   __ret_300; \
  44599. })
  44600. #endif
  44601.  
  44602. #ifdef __LITTLE_ENDIAN__
  44603. #define vcopyq_laneq_p16(__p0_301, __p1_301, __p2_301, __p3_301) __extension__ ({ \
  44604.   poly16x8_t __ret_301; \
  44605.   poly16x8_t __s0_301 = __p0_301; \
  44606.   poly16x8_t __s2_301 = __p2_301; \
  44607.   __ret_301 = vsetq_lane_p16(vgetq_lane_p16(__s2_301, __p3_301), __s0_301, __p1_301); \
  44608.   __ret_301; \
  44609. })
  44610. #else
  44611. #define vcopyq_laneq_p16(__p0_302, __p1_302, __p2_302, __p3_302) __extension__ ({ \
  44612.   poly16x8_t __ret_302; \
  44613.   poly16x8_t __s0_302 = __p0_302; \
  44614.   poly16x8_t __s2_302 = __p2_302; \
  44615.   poly16x8_t __rev0_302;  __rev0_302 = __builtin_shufflevector(__s0_302, __s0_302, 7, 6, 5, 4, 3, 2, 1, 0); \
  44616.   poly16x8_t __rev2_302;  __rev2_302 = __builtin_shufflevector(__s2_302, __s2_302, 7, 6, 5, 4, 3, 2, 1, 0); \
  44617.   __ret_302 = __noswap_vsetq_lane_p16(__noswap_vgetq_lane_p16(__rev2_302, __p3_302), __rev0_302, __p1_302); \
  44618.   __ret_302 = __builtin_shufflevector(__ret_302, __ret_302, 7, 6, 5, 4, 3, 2, 1, 0); \
  44619.   __ret_302; \
  44620. })
  44621. #endif
  44622.  
  44623. #ifdef __LITTLE_ENDIAN__
  44624. #define vcopyq_laneq_u8(__p0_303, __p1_303, __p2_303, __p3_303) __extension__ ({ \
  44625.   uint8x16_t __ret_303; \
  44626.   uint8x16_t __s0_303 = __p0_303; \
  44627.   uint8x16_t __s2_303 = __p2_303; \
  44628.   __ret_303 = vsetq_lane_u8(vgetq_lane_u8(__s2_303, __p3_303), __s0_303, __p1_303); \
  44629.   __ret_303; \
  44630. })
  44631. #else
  44632. #define vcopyq_laneq_u8(__p0_304, __p1_304, __p2_304, __p3_304) __extension__ ({ \
  44633.   uint8x16_t __ret_304; \
  44634.   uint8x16_t __s0_304 = __p0_304; \
  44635.   uint8x16_t __s2_304 = __p2_304; \
  44636.   uint8x16_t __rev0_304;  __rev0_304 = __builtin_shufflevector(__s0_304, __s0_304, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44637.   uint8x16_t __rev2_304;  __rev2_304 = __builtin_shufflevector(__s2_304, __s2_304, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44638.   __ret_304 = __noswap_vsetq_lane_u8(__noswap_vgetq_lane_u8(__rev2_304, __p3_304), __rev0_304, __p1_304); \
  44639.   __ret_304 = __builtin_shufflevector(__ret_304, __ret_304, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44640.   __ret_304; \
  44641. })
  44642. #endif
  44643.  
  44644. #ifdef __LITTLE_ENDIAN__
  44645. #define vcopyq_laneq_u32(__p0_305, __p1_305, __p2_305, __p3_305) __extension__ ({ \
  44646.   uint32x4_t __ret_305; \
  44647.   uint32x4_t __s0_305 = __p0_305; \
  44648.   uint32x4_t __s2_305 = __p2_305; \
  44649.   __ret_305 = vsetq_lane_u32(vgetq_lane_u32(__s2_305, __p3_305), __s0_305, __p1_305); \
  44650.   __ret_305; \
  44651. })
  44652. #else
  44653. #define vcopyq_laneq_u32(__p0_306, __p1_306, __p2_306, __p3_306) __extension__ ({ \
  44654.   uint32x4_t __ret_306; \
  44655.   uint32x4_t __s0_306 = __p0_306; \
  44656.   uint32x4_t __s2_306 = __p2_306; \
  44657.   uint32x4_t __rev0_306;  __rev0_306 = __builtin_shufflevector(__s0_306, __s0_306, 3, 2, 1, 0); \
  44658.   uint32x4_t __rev2_306;  __rev2_306 = __builtin_shufflevector(__s2_306, __s2_306, 3, 2, 1, 0); \
  44659.   __ret_306 = __noswap_vsetq_lane_u32(__noswap_vgetq_lane_u32(__rev2_306, __p3_306), __rev0_306, __p1_306); \
  44660.   __ret_306 = __builtin_shufflevector(__ret_306, __ret_306, 3, 2, 1, 0); \
  44661.   __ret_306; \
  44662. })
  44663. #endif
  44664.  
  44665. #ifdef __LITTLE_ENDIAN__
  44666. #define vcopyq_laneq_u64(__p0_307, __p1_307, __p2_307, __p3_307) __extension__ ({ \
  44667.   uint64x2_t __ret_307; \
  44668.   uint64x2_t __s0_307 = __p0_307; \
  44669.   uint64x2_t __s2_307 = __p2_307; \
  44670.   __ret_307 = vsetq_lane_u64(vgetq_lane_u64(__s2_307, __p3_307), __s0_307, __p1_307); \
  44671.   __ret_307; \
  44672. })
  44673. #else
  44674. #define vcopyq_laneq_u64(__p0_308, __p1_308, __p2_308, __p3_308) __extension__ ({ \
  44675.   uint64x2_t __ret_308; \
  44676.   uint64x2_t __s0_308 = __p0_308; \
  44677.   uint64x2_t __s2_308 = __p2_308; \
  44678.   uint64x2_t __rev0_308;  __rev0_308 = __builtin_shufflevector(__s0_308, __s0_308, 1, 0); \
  44679.   uint64x2_t __rev2_308;  __rev2_308 = __builtin_shufflevector(__s2_308, __s2_308, 1, 0); \
  44680.   __ret_308 = __noswap_vsetq_lane_u64(__noswap_vgetq_lane_u64(__rev2_308, __p3_308), __rev0_308, __p1_308); \
  44681.   __ret_308 = __builtin_shufflevector(__ret_308, __ret_308, 1, 0); \
  44682.   __ret_308; \
  44683. })
  44684. #endif
  44685.  
  44686. #ifdef __LITTLE_ENDIAN__
  44687. #define vcopyq_laneq_u16(__p0_309, __p1_309, __p2_309, __p3_309) __extension__ ({ \
  44688.   uint16x8_t __ret_309; \
  44689.   uint16x8_t __s0_309 = __p0_309; \
  44690.   uint16x8_t __s2_309 = __p2_309; \
  44691.   __ret_309 = vsetq_lane_u16(vgetq_lane_u16(__s2_309, __p3_309), __s0_309, __p1_309); \
  44692.   __ret_309; \
  44693. })
  44694. #else
  44695. #define vcopyq_laneq_u16(__p0_310, __p1_310, __p2_310, __p3_310) __extension__ ({ \
  44696.   uint16x8_t __ret_310; \
  44697.   uint16x8_t __s0_310 = __p0_310; \
  44698.   uint16x8_t __s2_310 = __p2_310; \
  44699.   uint16x8_t __rev0_310;  __rev0_310 = __builtin_shufflevector(__s0_310, __s0_310, 7, 6, 5, 4, 3, 2, 1, 0); \
  44700.   uint16x8_t __rev2_310;  __rev2_310 = __builtin_shufflevector(__s2_310, __s2_310, 7, 6, 5, 4, 3, 2, 1, 0); \
  44701.   __ret_310 = __noswap_vsetq_lane_u16(__noswap_vgetq_lane_u16(__rev2_310, __p3_310), __rev0_310, __p1_310); \
  44702.   __ret_310 = __builtin_shufflevector(__ret_310, __ret_310, 7, 6, 5, 4, 3, 2, 1, 0); \
  44703.   __ret_310; \
  44704. })
  44705. #endif
  44706.  
  44707. #ifdef __LITTLE_ENDIAN__
  44708. #define vcopyq_laneq_s8(__p0_311, __p1_311, __p2_311, __p3_311) __extension__ ({ \
  44709.   int8x16_t __ret_311; \
  44710.   int8x16_t __s0_311 = __p0_311; \
  44711.   int8x16_t __s2_311 = __p2_311; \
  44712.   __ret_311 = vsetq_lane_s8(vgetq_lane_s8(__s2_311, __p3_311), __s0_311, __p1_311); \
  44713.   __ret_311; \
  44714. })
  44715. #else
  44716. #define vcopyq_laneq_s8(__p0_312, __p1_312, __p2_312, __p3_312) __extension__ ({ \
  44717.   int8x16_t __ret_312; \
  44718.   int8x16_t __s0_312 = __p0_312; \
  44719.   int8x16_t __s2_312 = __p2_312; \
  44720.   int8x16_t __rev0_312;  __rev0_312 = __builtin_shufflevector(__s0_312, __s0_312, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44721.   int8x16_t __rev2_312;  __rev2_312 = __builtin_shufflevector(__s2_312, __s2_312, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44722.   __ret_312 = __noswap_vsetq_lane_s8(__noswap_vgetq_lane_s8(__rev2_312, __p3_312), __rev0_312, __p1_312); \
  44723.   __ret_312 = __builtin_shufflevector(__ret_312, __ret_312, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44724.   __ret_312; \
  44725. })
  44726. #endif
  44727.  
  44728. #ifdef __LITTLE_ENDIAN__
  44729. #define vcopyq_laneq_f32(__p0_313, __p1_313, __p2_313, __p3_313) __extension__ ({ \
  44730.   float32x4_t __ret_313; \
  44731.   float32x4_t __s0_313 = __p0_313; \
  44732.   float32x4_t __s2_313 = __p2_313; \
  44733.   __ret_313 = vsetq_lane_f32(vgetq_lane_f32(__s2_313, __p3_313), __s0_313, __p1_313); \
  44734.   __ret_313; \
  44735. })
  44736. #else
  44737. #define vcopyq_laneq_f32(__p0_314, __p1_314, __p2_314, __p3_314) __extension__ ({ \
  44738.   float32x4_t __ret_314; \
  44739.   float32x4_t __s0_314 = __p0_314; \
  44740.   float32x4_t __s2_314 = __p2_314; \
  44741.   float32x4_t __rev0_314;  __rev0_314 = __builtin_shufflevector(__s0_314, __s0_314, 3, 2, 1, 0); \
  44742.   float32x4_t __rev2_314;  __rev2_314 = __builtin_shufflevector(__s2_314, __s2_314, 3, 2, 1, 0); \
  44743.   __ret_314 = __noswap_vsetq_lane_f32(__noswap_vgetq_lane_f32(__rev2_314, __p3_314), __rev0_314, __p1_314); \
  44744.   __ret_314 = __builtin_shufflevector(__ret_314, __ret_314, 3, 2, 1, 0); \
  44745.   __ret_314; \
  44746. })
  44747. #endif
  44748.  
  44749. #ifdef __LITTLE_ENDIAN__
  44750. #define vcopyq_laneq_s32(__p0_315, __p1_315, __p2_315, __p3_315) __extension__ ({ \
  44751.   int32x4_t __ret_315; \
  44752.   int32x4_t __s0_315 = __p0_315; \
  44753.   int32x4_t __s2_315 = __p2_315; \
  44754.   __ret_315 = vsetq_lane_s32(vgetq_lane_s32(__s2_315, __p3_315), __s0_315, __p1_315); \
  44755.   __ret_315; \
  44756. })
  44757. #else
  44758. #define vcopyq_laneq_s32(__p0_316, __p1_316, __p2_316, __p3_316) __extension__ ({ \
  44759.   int32x4_t __ret_316; \
  44760.   int32x4_t __s0_316 = __p0_316; \
  44761.   int32x4_t __s2_316 = __p2_316; \
  44762.   int32x4_t __rev0_316;  __rev0_316 = __builtin_shufflevector(__s0_316, __s0_316, 3, 2, 1, 0); \
  44763.   int32x4_t __rev2_316;  __rev2_316 = __builtin_shufflevector(__s2_316, __s2_316, 3, 2, 1, 0); \
  44764.   __ret_316 = __noswap_vsetq_lane_s32(__noswap_vgetq_lane_s32(__rev2_316, __p3_316), __rev0_316, __p1_316); \
  44765.   __ret_316 = __builtin_shufflevector(__ret_316, __ret_316, 3, 2, 1, 0); \
  44766.   __ret_316; \
  44767. })
  44768. #endif
  44769.  
  44770. #ifdef __LITTLE_ENDIAN__
  44771. #define vcopyq_laneq_s64(__p0_317, __p1_317, __p2_317, __p3_317) __extension__ ({ \
  44772.   int64x2_t __ret_317; \
  44773.   int64x2_t __s0_317 = __p0_317; \
  44774.   int64x2_t __s2_317 = __p2_317; \
  44775.   __ret_317 = vsetq_lane_s64(vgetq_lane_s64(__s2_317, __p3_317), __s0_317, __p1_317); \
  44776.   __ret_317; \
  44777. })
  44778. #else
  44779. #define vcopyq_laneq_s64(__p0_318, __p1_318, __p2_318, __p3_318) __extension__ ({ \
  44780.   int64x2_t __ret_318; \
  44781.   int64x2_t __s0_318 = __p0_318; \
  44782.   int64x2_t __s2_318 = __p2_318; \
  44783.   int64x2_t __rev0_318;  __rev0_318 = __builtin_shufflevector(__s0_318, __s0_318, 1, 0); \
  44784.   int64x2_t __rev2_318;  __rev2_318 = __builtin_shufflevector(__s2_318, __s2_318, 1, 0); \
  44785.   __ret_318 = __noswap_vsetq_lane_s64(__noswap_vgetq_lane_s64(__rev2_318, __p3_318), __rev0_318, __p1_318); \
  44786.   __ret_318 = __builtin_shufflevector(__ret_318, __ret_318, 1, 0); \
  44787.   __ret_318; \
  44788. })
  44789. #endif
  44790.  
  44791. #ifdef __LITTLE_ENDIAN__
  44792. #define vcopyq_laneq_s16(__p0_319, __p1_319, __p2_319, __p3_319) __extension__ ({ \
  44793.   int16x8_t __ret_319; \
  44794.   int16x8_t __s0_319 = __p0_319; \
  44795.   int16x8_t __s2_319 = __p2_319; \
  44796.   __ret_319 = vsetq_lane_s16(vgetq_lane_s16(__s2_319, __p3_319), __s0_319, __p1_319); \
  44797.   __ret_319; \
  44798. })
  44799. #else
  44800. #define vcopyq_laneq_s16(__p0_320, __p1_320, __p2_320, __p3_320) __extension__ ({ \
  44801.   int16x8_t __ret_320; \
  44802.   int16x8_t __s0_320 = __p0_320; \
  44803.   int16x8_t __s2_320 = __p2_320; \
  44804.   int16x8_t __rev0_320;  __rev0_320 = __builtin_shufflevector(__s0_320, __s0_320, 7, 6, 5, 4, 3, 2, 1, 0); \
  44805.   int16x8_t __rev2_320;  __rev2_320 = __builtin_shufflevector(__s2_320, __s2_320, 7, 6, 5, 4, 3, 2, 1, 0); \
  44806.   __ret_320 = __noswap_vsetq_lane_s16(__noswap_vgetq_lane_s16(__rev2_320, __p3_320), __rev0_320, __p1_320); \
  44807.   __ret_320 = __builtin_shufflevector(__ret_320, __ret_320, 7, 6, 5, 4, 3, 2, 1, 0); \
  44808.   __ret_320; \
  44809. })
  44810. #endif
  44811.  
  44812. #ifdef __LITTLE_ENDIAN__
  44813. #define vcopy_laneq_p8(__p0_321, __p1_321, __p2_321, __p3_321) __extension__ ({ \
  44814.   poly8x8_t __ret_321; \
  44815.   poly8x8_t __s0_321 = __p0_321; \
  44816.   poly8x16_t __s2_321 = __p2_321; \
  44817.   __ret_321 = vset_lane_p8(vgetq_lane_p8(__s2_321, __p3_321), __s0_321, __p1_321); \
  44818.   __ret_321; \
  44819. })
  44820. #else
  44821. #define vcopy_laneq_p8(__p0_322, __p1_322, __p2_322, __p3_322) __extension__ ({ \
  44822.   poly8x8_t __ret_322; \
  44823.   poly8x8_t __s0_322 = __p0_322; \
  44824.   poly8x16_t __s2_322 = __p2_322; \
  44825.   poly8x8_t __rev0_322;  __rev0_322 = __builtin_shufflevector(__s0_322, __s0_322, 7, 6, 5, 4, 3, 2, 1, 0); \
  44826.   poly8x16_t __rev2_322;  __rev2_322 = __builtin_shufflevector(__s2_322, __s2_322, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44827.   __ret_322 = __noswap_vset_lane_p8(__noswap_vgetq_lane_p8(__rev2_322, __p3_322), __rev0_322, __p1_322); \
  44828.   __ret_322 = __builtin_shufflevector(__ret_322, __ret_322, 7, 6, 5, 4, 3, 2, 1, 0); \
  44829.   __ret_322; \
  44830. })
  44831. #endif
  44832.  
  44833. #ifdef __LITTLE_ENDIAN__
  44834. #define vcopy_laneq_p16(__p0_323, __p1_323, __p2_323, __p3_323) __extension__ ({ \
  44835.   poly16x4_t __ret_323; \
  44836.   poly16x4_t __s0_323 = __p0_323; \
  44837.   poly16x8_t __s2_323 = __p2_323; \
  44838.   __ret_323 = vset_lane_p16(vgetq_lane_p16(__s2_323, __p3_323), __s0_323, __p1_323); \
  44839.   __ret_323; \
  44840. })
  44841. #else
  44842. #define vcopy_laneq_p16(__p0_324, __p1_324, __p2_324, __p3_324) __extension__ ({ \
  44843.   poly16x4_t __ret_324; \
  44844.   poly16x4_t __s0_324 = __p0_324; \
  44845.   poly16x8_t __s2_324 = __p2_324; \
  44846.   poly16x4_t __rev0_324;  __rev0_324 = __builtin_shufflevector(__s0_324, __s0_324, 3, 2, 1, 0); \
  44847.   poly16x8_t __rev2_324;  __rev2_324 = __builtin_shufflevector(__s2_324, __s2_324, 7, 6, 5, 4, 3, 2, 1, 0); \
  44848.   __ret_324 = __noswap_vset_lane_p16(__noswap_vgetq_lane_p16(__rev2_324, __p3_324), __rev0_324, __p1_324); \
  44849.   __ret_324 = __builtin_shufflevector(__ret_324, __ret_324, 3, 2, 1, 0); \
  44850.   __ret_324; \
  44851. })
  44852. #endif
  44853.  
  44854. #ifdef __LITTLE_ENDIAN__
  44855. #define vcopy_laneq_u8(__p0_325, __p1_325, __p2_325, __p3_325) __extension__ ({ \
  44856.   uint8x8_t __ret_325; \
  44857.   uint8x8_t __s0_325 = __p0_325; \
  44858.   uint8x16_t __s2_325 = __p2_325; \
  44859.   __ret_325 = vset_lane_u8(vgetq_lane_u8(__s2_325, __p3_325), __s0_325, __p1_325); \
  44860.   __ret_325; \
  44861. })
  44862. #else
  44863. #define vcopy_laneq_u8(__p0_326, __p1_326, __p2_326, __p3_326) __extension__ ({ \
  44864.   uint8x8_t __ret_326; \
  44865.   uint8x8_t __s0_326 = __p0_326; \
  44866.   uint8x16_t __s2_326 = __p2_326; \
  44867.   uint8x8_t __rev0_326;  __rev0_326 = __builtin_shufflevector(__s0_326, __s0_326, 7, 6, 5, 4, 3, 2, 1, 0); \
  44868.   uint8x16_t __rev2_326;  __rev2_326 = __builtin_shufflevector(__s2_326, __s2_326, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44869.   __ret_326 = __noswap_vset_lane_u8(__noswap_vgetq_lane_u8(__rev2_326, __p3_326), __rev0_326, __p1_326); \
  44870.   __ret_326 = __builtin_shufflevector(__ret_326, __ret_326, 7, 6, 5, 4, 3, 2, 1, 0); \
  44871.   __ret_326; \
  44872. })
  44873. #endif
  44874.  
  44875. #ifdef __LITTLE_ENDIAN__
  44876. #define vcopy_laneq_u32(__p0_327, __p1_327, __p2_327, __p3_327) __extension__ ({ \
  44877.   uint32x2_t __ret_327; \
  44878.   uint32x2_t __s0_327 = __p0_327; \
  44879.   uint32x4_t __s2_327 = __p2_327; \
  44880.   __ret_327 = vset_lane_u32(vgetq_lane_u32(__s2_327, __p3_327), __s0_327, __p1_327); \
  44881.   __ret_327; \
  44882. })
  44883. #else
  44884. #define vcopy_laneq_u32(__p0_328, __p1_328, __p2_328, __p3_328) __extension__ ({ \
  44885.   uint32x2_t __ret_328; \
  44886.   uint32x2_t __s0_328 = __p0_328; \
  44887.   uint32x4_t __s2_328 = __p2_328; \
  44888.   uint32x2_t __rev0_328;  __rev0_328 = __builtin_shufflevector(__s0_328, __s0_328, 1, 0); \
  44889.   uint32x4_t __rev2_328;  __rev2_328 = __builtin_shufflevector(__s2_328, __s2_328, 3, 2, 1, 0); \
  44890.   __ret_328 = __noswap_vset_lane_u32(__noswap_vgetq_lane_u32(__rev2_328, __p3_328), __rev0_328, __p1_328); \
  44891.   __ret_328 = __builtin_shufflevector(__ret_328, __ret_328, 1, 0); \
  44892.   __ret_328; \
  44893. })
  44894. #endif
  44895.  
  44896. #ifdef __LITTLE_ENDIAN__
  44897. #define vcopy_laneq_u64(__p0_329, __p1_329, __p2_329, __p3_329) __extension__ ({ \
  44898.   uint64x1_t __ret_329; \
  44899.   uint64x1_t __s0_329 = __p0_329; \
  44900.   uint64x2_t __s2_329 = __p2_329; \
  44901.   __ret_329 = vset_lane_u64(vgetq_lane_u64(__s2_329, __p3_329), __s0_329, __p1_329); \
  44902.   __ret_329; \
  44903. })
  44904. #else
  44905. #define vcopy_laneq_u64(__p0_330, __p1_330, __p2_330, __p3_330) __extension__ ({ \
  44906.   uint64x1_t __ret_330; \
  44907.   uint64x1_t __s0_330 = __p0_330; \
  44908.   uint64x2_t __s2_330 = __p2_330; \
  44909.   uint64x2_t __rev2_330;  __rev2_330 = __builtin_shufflevector(__s2_330, __s2_330, 1, 0); \
  44910.   __ret_330 = vset_lane_u64(__noswap_vgetq_lane_u64(__rev2_330, __p3_330), __s0_330, __p1_330); \
  44911.   __ret_330; \
  44912. })
  44913. #endif
  44914.  
  44915. #ifdef __LITTLE_ENDIAN__
  44916. #define vcopy_laneq_u16(__p0_331, __p1_331, __p2_331, __p3_331) __extension__ ({ \
  44917.   uint16x4_t __ret_331; \
  44918.   uint16x4_t __s0_331 = __p0_331; \
  44919.   uint16x8_t __s2_331 = __p2_331; \
  44920.   __ret_331 = vset_lane_u16(vgetq_lane_u16(__s2_331, __p3_331), __s0_331, __p1_331); \
  44921.   __ret_331; \
  44922. })
  44923. #else
  44924. #define vcopy_laneq_u16(__p0_332, __p1_332, __p2_332, __p3_332) __extension__ ({ \
  44925.   uint16x4_t __ret_332; \
  44926.   uint16x4_t __s0_332 = __p0_332; \
  44927.   uint16x8_t __s2_332 = __p2_332; \
  44928.   uint16x4_t __rev0_332;  __rev0_332 = __builtin_shufflevector(__s0_332, __s0_332, 3, 2, 1, 0); \
  44929.   uint16x8_t __rev2_332;  __rev2_332 = __builtin_shufflevector(__s2_332, __s2_332, 7, 6, 5, 4, 3, 2, 1, 0); \
  44930.   __ret_332 = __noswap_vset_lane_u16(__noswap_vgetq_lane_u16(__rev2_332, __p3_332), __rev0_332, __p1_332); \
  44931.   __ret_332 = __builtin_shufflevector(__ret_332, __ret_332, 3, 2, 1, 0); \
  44932.   __ret_332; \
  44933. })
  44934. #endif
  44935.  
  44936. #ifdef __LITTLE_ENDIAN__
  44937. #define vcopy_laneq_s8(__p0_333, __p1_333, __p2_333, __p3_333) __extension__ ({ \
  44938.   int8x8_t __ret_333; \
  44939.   int8x8_t __s0_333 = __p0_333; \
  44940.   int8x16_t __s2_333 = __p2_333; \
  44941.   __ret_333 = vset_lane_s8(vgetq_lane_s8(__s2_333, __p3_333), __s0_333, __p1_333); \
  44942.   __ret_333; \
  44943. })
  44944. #else
  44945. #define vcopy_laneq_s8(__p0_334, __p1_334, __p2_334, __p3_334) __extension__ ({ \
  44946.   int8x8_t __ret_334; \
  44947.   int8x8_t __s0_334 = __p0_334; \
  44948.   int8x16_t __s2_334 = __p2_334; \
  44949.   int8x8_t __rev0_334;  __rev0_334 = __builtin_shufflevector(__s0_334, __s0_334, 7, 6, 5, 4, 3, 2, 1, 0); \
  44950.   int8x16_t __rev2_334;  __rev2_334 = __builtin_shufflevector(__s2_334, __s2_334, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  44951.   __ret_334 = __noswap_vset_lane_s8(__noswap_vgetq_lane_s8(__rev2_334, __p3_334), __rev0_334, __p1_334); \
  44952.   __ret_334 = __builtin_shufflevector(__ret_334, __ret_334, 7, 6, 5, 4, 3, 2, 1, 0); \
  44953.   __ret_334; \
  44954. })
  44955. #endif
  44956.  
  44957. #ifdef __LITTLE_ENDIAN__
  44958. #define vcopy_laneq_f32(__p0_335, __p1_335, __p2_335, __p3_335) __extension__ ({ \
  44959.   float32x2_t __ret_335; \
  44960.   float32x2_t __s0_335 = __p0_335; \
  44961.   float32x4_t __s2_335 = __p2_335; \
  44962.   __ret_335 = vset_lane_f32(vgetq_lane_f32(__s2_335, __p3_335), __s0_335, __p1_335); \
  44963.   __ret_335; \
  44964. })
  44965. #else
  44966. #define vcopy_laneq_f32(__p0_336, __p1_336, __p2_336, __p3_336) __extension__ ({ \
  44967.   float32x2_t __ret_336; \
  44968.   float32x2_t __s0_336 = __p0_336; \
  44969.   float32x4_t __s2_336 = __p2_336; \
  44970.   float32x2_t __rev0_336;  __rev0_336 = __builtin_shufflevector(__s0_336, __s0_336, 1, 0); \
  44971.   float32x4_t __rev2_336;  __rev2_336 = __builtin_shufflevector(__s2_336, __s2_336, 3, 2, 1, 0); \
  44972.   __ret_336 = __noswap_vset_lane_f32(__noswap_vgetq_lane_f32(__rev2_336, __p3_336), __rev0_336, __p1_336); \
  44973.   __ret_336 = __builtin_shufflevector(__ret_336, __ret_336, 1, 0); \
  44974.   __ret_336; \
  44975. })
  44976. #endif
  44977.  
  44978. #ifdef __LITTLE_ENDIAN__
  44979. #define vcopy_laneq_s32(__p0_337, __p1_337, __p2_337, __p3_337) __extension__ ({ \
  44980.   int32x2_t __ret_337; \
  44981.   int32x2_t __s0_337 = __p0_337; \
  44982.   int32x4_t __s2_337 = __p2_337; \
  44983.   __ret_337 = vset_lane_s32(vgetq_lane_s32(__s2_337, __p3_337), __s0_337, __p1_337); \
  44984.   __ret_337; \
  44985. })
  44986. #else
  44987. #define vcopy_laneq_s32(__p0_338, __p1_338, __p2_338, __p3_338) __extension__ ({ \
  44988.   int32x2_t __ret_338; \
  44989.   int32x2_t __s0_338 = __p0_338; \
  44990.   int32x4_t __s2_338 = __p2_338; \
  44991.   int32x2_t __rev0_338;  __rev0_338 = __builtin_shufflevector(__s0_338, __s0_338, 1, 0); \
  44992.   int32x4_t __rev2_338;  __rev2_338 = __builtin_shufflevector(__s2_338, __s2_338, 3, 2, 1, 0); \
  44993.   __ret_338 = __noswap_vset_lane_s32(__noswap_vgetq_lane_s32(__rev2_338, __p3_338), __rev0_338, __p1_338); \
  44994.   __ret_338 = __builtin_shufflevector(__ret_338, __ret_338, 1, 0); \
  44995.   __ret_338; \
  44996. })
  44997. #endif
  44998.  
  44999. #ifdef __LITTLE_ENDIAN__
  45000. #define vcopy_laneq_s64(__p0_339, __p1_339, __p2_339, __p3_339) __extension__ ({ \
  45001.   int64x1_t __ret_339; \
  45002.   int64x1_t __s0_339 = __p0_339; \
  45003.   int64x2_t __s2_339 = __p2_339; \
  45004.   __ret_339 = vset_lane_s64(vgetq_lane_s64(__s2_339, __p3_339), __s0_339, __p1_339); \
  45005.   __ret_339; \
  45006. })
  45007. #else
  45008. #define vcopy_laneq_s64(__p0_340, __p1_340, __p2_340, __p3_340) __extension__ ({ \
  45009.   int64x1_t __ret_340; \
  45010.   int64x1_t __s0_340 = __p0_340; \
  45011.   int64x2_t __s2_340 = __p2_340; \
  45012.   int64x2_t __rev2_340;  __rev2_340 = __builtin_shufflevector(__s2_340, __s2_340, 1, 0); \
  45013.   __ret_340 = vset_lane_s64(__noswap_vgetq_lane_s64(__rev2_340, __p3_340), __s0_340, __p1_340); \
  45014.   __ret_340; \
  45015. })
  45016. #endif
  45017.  
  45018. #ifdef __LITTLE_ENDIAN__
  45019. #define vcopy_laneq_s16(__p0_341, __p1_341, __p2_341, __p3_341) __extension__ ({ \
  45020.   int16x4_t __ret_341; \
  45021.   int16x4_t __s0_341 = __p0_341; \
  45022.   int16x8_t __s2_341 = __p2_341; \
  45023.   __ret_341 = vset_lane_s16(vgetq_lane_s16(__s2_341, __p3_341), __s0_341, __p1_341); \
  45024.   __ret_341; \
  45025. })
  45026. #else
  45027. #define vcopy_laneq_s16(__p0_342, __p1_342, __p2_342, __p3_342) __extension__ ({ \
  45028.   int16x4_t __ret_342; \
  45029.   int16x4_t __s0_342 = __p0_342; \
  45030.   int16x8_t __s2_342 = __p2_342; \
  45031.   int16x4_t __rev0_342;  __rev0_342 = __builtin_shufflevector(__s0_342, __s0_342, 3, 2, 1, 0); \
  45032.   int16x8_t __rev2_342;  __rev2_342 = __builtin_shufflevector(__s2_342, __s2_342, 7, 6, 5, 4, 3, 2, 1, 0); \
  45033.   __ret_342 = __noswap_vset_lane_s16(__noswap_vgetq_lane_s16(__rev2_342, __p3_342), __rev0_342, __p1_342); \
  45034.   __ret_342 = __builtin_shufflevector(__ret_342, __ret_342, 3, 2, 1, 0); \
  45035.   __ret_342; \
  45036. })
  45037. #endif
  45038.  
  45039. #define vcreate_p64(__p0) __extension__ ({ \
  45040.   poly64x1_t __ret; \
  45041.   uint64_t __promote = __p0; \
  45042.   __ret = (poly64x1_t)(__promote); \
  45043.   __ret; \
  45044. })
  45045. #define vcreate_f64(__p0) __extension__ ({ \
  45046.   float64x1_t __ret; \
  45047.   uint64_t __promote = __p0; \
  45048.   __ret = (float64x1_t)(__promote); \
  45049.   __ret; \
  45050. })
  45051. __ai float32_t vcvts_f32_s32(int32_t __p0) {
  45052.   float32_t __ret;
  45053.   __ret = (float32_t) __builtin_neon_vcvts_f32_s32(__p0);
  45054.   return __ret;
  45055. }
  45056. __ai float32_t vcvts_f32_u32(uint32_t __p0) {
  45057.   float32_t __ret;
  45058.   __ret = (float32_t) __builtin_neon_vcvts_f32_u32(__p0);
  45059.   return __ret;
  45060. }
  45061. #ifdef __LITTLE_ENDIAN__
  45062. __ai float32x2_t vcvt_f32_f64(float64x2_t __p0) {
  45063.   float32x2_t __ret;
  45064.   __ret = (float32x2_t) __builtin_neon_vcvt_f32_f64((int8x16_t)__p0, 9);
  45065.   return __ret;
  45066. }
  45067. #else
  45068. __ai float32x2_t vcvt_f32_f64(float64x2_t __p0) {
  45069.   float32x2_t __ret;
  45070.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45071.   __ret = (float32x2_t) __builtin_neon_vcvt_f32_f64((int8x16_t)__rev0, 9);
  45072.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45073.   return __ret;
  45074. }
  45075. __ai float32x2_t __noswap_vcvt_f32_f64(float64x2_t __p0) {
  45076.   float32x2_t __ret;
  45077.   __ret = (float32x2_t) __builtin_neon_vcvt_f32_f64((int8x16_t)__p0, 9);
  45078.   return __ret;
  45079. }
  45080. #endif
  45081.  
  45082. __ai float64_t vcvtd_f64_s64(int64_t __p0) {
  45083.   float64_t __ret;
  45084.   __ret = (float64_t) __builtin_neon_vcvtd_f64_s64(__p0);
  45085.   return __ret;
  45086. }
  45087. __ai float64_t vcvtd_f64_u64(uint64_t __p0) {
  45088.   float64_t __ret;
  45089.   __ret = (float64_t) __builtin_neon_vcvtd_f64_u64(__p0);
  45090.   return __ret;
  45091. }
  45092. #ifdef __LITTLE_ENDIAN__
  45093. __ai float64x2_t vcvtq_f64_u64(uint64x2_t __p0) {
  45094.   float64x2_t __ret;
  45095.   __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__p0, 51);
  45096.   return __ret;
  45097. }
  45098. #else
  45099. __ai float64x2_t vcvtq_f64_u64(uint64x2_t __p0) {
  45100.   float64x2_t __ret;
  45101.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45102.   __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__rev0, 51);
  45103.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45104.   return __ret;
  45105. }
  45106. #endif
  45107.  
  45108. #ifdef __LITTLE_ENDIAN__
  45109. __ai float64x2_t vcvtq_f64_s64(int64x2_t __p0) {
  45110.   float64x2_t __ret;
  45111.   __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__p0, 35);
  45112.   return __ret;
  45113. }
  45114. #else
  45115. __ai float64x2_t vcvtq_f64_s64(int64x2_t __p0) {
  45116.   float64x2_t __ret;
  45117.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45118.   __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__rev0, 35);
  45119.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45120.   return __ret;
  45121. }
  45122. #endif
  45123.  
  45124. __ai float64x1_t vcvt_f64_u64(uint64x1_t __p0) {
  45125.   float64x1_t __ret;
  45126.   __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 19);
  45127.   return __ret;
  45128. }
  45129. __ai float64x1_t vcvt_f64_s64(int64x1_t __p0) {
  45130.   float64x1_t __ret;
  45131.   __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 3);
  45132.   return __ret;
  45133. }
  45134. #ifdef __LITTLE_ENDIAN__
  45135. __ai float64x2_t vcvt_f64_f32(float32x2_t __p0) {
  45136.   float64x2_t __ret;
  45137.   __ret = (float64x2_t) __builtin_neon_vcvt_f64_f32((int8x8_t)__p0, 42);
  45138.   return __ret;
  45139. }
  45140. #else
  45141. __ai float64x2_t vcvt_f64_f32(float32x2_t __p0) {
  45142.   float64x2_t __ret;
  45143.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45144.   __ret = (float64x2_t) __builtin_neon_vcvt_f64_f32((int8x8_t)__rev0, 42);
  45145.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45146.   return __ret;
  45147. }
  45148. __ai float64x2_t __noswap_vcvt_f64_f32(float32x2_t __p0) {
  45149.   float64x2_t __ret;
  45150.   __ret = (float64x2_t) __builtin_neon_vcvt_f64_f32((int8x8_t)__p0, 42);
  45151.   return __ret;
  45152. }
  45153. #endif
  45154.  
  45155. #ifdef __LITTLE_ENDIAN__
  45156. __ai float16x8_t vcvt_high_f16_f32(float16x4_t __p0, float32x4_t __p1) {
  45157.   float16x8_t __ret;
  45158.   __ret = vcombine_f16(__p0, vcvt_f16_f32(__p1));
  45159.   return __ret;
  45160. }
  45161. #else
  45162. __ai float16x8_t vcvt_high_f16_f32(float16x4_t __p0, float32x4_t __p1) {
  45163.   float16x8_t __ret;
  45164.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  45165.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  45166.   __ret = __noswap_vcombine_f16(__rev0, __noswap_vcvt_f16_f32(__rev1));
  45167.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  45168.   return __ret;
  45169. }
  45170. #endif
  45171.  
  45172. #ifdef __LITTLE_ENDIAN__
  45173. __ai float32x4_t vcvt_high_f32_f16(float16x8_t __p0) {
  45174.   float32x4_t __ret;
  45175.   __ret = vcvt_f32_f16(vget_high_f16(__p0));
  45176.   return __ret;
  45177. }
  45178. #else
  45179. __ai float32x4_t vcvt_high_f32_f16(float16x8_t __p0) {
  45180.   float32x4_t __ret;
  45181.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  45182.   __ret = __noswap_vcvt_f32_f16(__noswap_vget_high_f16(__rev0));
  45183.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  45184.   return __ret;
  45185. }
  45186. #endif
  45187.  
  45188. #ifdef __LITTLE_ENDIAN__
  45189. __ai float32x4_t vcvt_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
  45190.   float32x4_t __ret;
  45191.   __ret = vcombine_f32(__p0, vcvt_f32_f64(__p1));
  45192.   return __ret;
  45193. }
  45194. #else
  45195. __ai float32x4_t vcvt_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
  45196.   float32x4_t __ret;
  45197.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45198.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  45199.   __ret = __noswap_vcombine_f32(__rev0, __noswap_vcvt_f32_f64(__rev1));
  45200.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  45201.   return __ret;
  45202. }
  45203. #endif
  45204.  
  45205. #ifdef __LITTLE_ENDIAN__
  45206. __ai float64x2_t vcvt_high_f64_f32(float32x4_t __p0) {
  45207.   float64x2_t __ret;
  45208.   __ret = vcvt_f64_f32(vget_high_f32(__p0));
  45209.   return __ret;
  45210. }
  45211. #else
  45212. __ai float64x2_t vcvt_high_f64_f32(float32x4_t __p0) {
  45213.   float64x2_t __ret;
  45214.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  45215.   __ret = __noswap_vcvt_f64_f32(__noswap_vget_high_f32(__rev0));
  45216.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45217.   return __ret;
  45218. }
  45219. #endif
  45220.  
  45221. #define vcvts_n_f32_u32(__p0, __p1) __extension__ ({ \
  45222.   float32_t __ret; \
  45223.   uint32_t __s0 = __p0; \
  45224.   __ret = (float32_t) __builtin_neon_vcvts_n_f32_u32(__s0, __p1); \
  45225.   __ret; \
  45226. })
  45227. #define vcvts_n_f32_s32(__p0, __p1) __extension__ ({ \
  45228.   float32_t __ret; \
  45229.   int32_t __s0 = __p0; \
  45230.   __ret = (float32_t) __builtin_neon_vcvts_n_f32_s32(__s0, __p1); \
  45231.   __ret; \
  45232. })
  45233. #ifdef __LITTLE_ENDIAN__
  45234. #define vcvtq_n_f64_u64(__p0, __p1) __extension__ ({ \
  45235.   float64x2_t __ret; \
  45236.   uint64x2_t __s0 = __p0; \
  45237.   __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__s0, __p1, 51); \
  45238.   __ret; \
  45239. })
  45240. #else
  45241. #define vcvtq_n_f64_u64(__p0, __p1) __extension__ ({ \
  45242.   float64x2_t __ret; \
  45243.   uint64x2_t __s0 = __p0; \
  45244.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  45245.   __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__rev0, __p1, 51); \
  45246.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  45247.   __ret; \
  45248. })
  45249. #endif
  45250.  
  45251. #ifdef __LITTLE_ENDIAN__
  45252. #define vcvtq_n_f64_s64(__p0, __p1) __extension__ ({ \
  45253.   float64x2_t __ret; \
  45254.   int64x2_t __s0 = __p0; \
  45255.   __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__s0, __p1, 35); \
  45256.   __ret; \
  45257. })
  45258. #else
  45259. #define vcvtq_n_f64_s64(__p0, __p1) __extension__ ({ \
  45260.   float64x2_t __ret; \
  45261.   int64x2_t __s0 = __p0; \
  45262.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  45263.   __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__rev0, __p1, 35); \
  45264.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  45265.   __ret; \
  45266. })
  45267. #endif
  45268.  
  45269. #define vcvt_n_f64_u64(__p0, __p1) __extension__ ({ \
  45270.   float64x1_t __ret; \
  45271.   uint64x1_t __s0 = __p0; \
  45272.   __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 19); \
  45273.   __ret; \
  45274. })
  45275. #define vcvt_n_f64_s64(__p0, __p1) __extension__ ({ \
  45276.   float64x1_t __ret; \
  45277.   int64x1_t __s0 = __p0; \
  45278.   __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 3); \
  45279.   __ret; \
  45280. })
  45281. #define vcvtd_n_f64_u64(__p0, __p1) __extension__ ({ \
  45282.   float64_t __ret; \
  45283.   uint64_t __s0 = __p0; \
  45284.   __ret = (float64_t) __builtin_neon_vcvtd_n_f64_u64(__s0, __p1); \
  45285.   __ret; \
  45286. })
  45287. #define vcvtd_n_f64_s64(__p0, __p1) __extension__ ({ \
  45288.   float64_t __ret; \
  45289.   int64_t __s0 = __p0; \
  45290.   __ret = (float64_t) __builtin_neon_vcvtd_n_f64_s64(__s0, __p1); \
  45291.   __ret; \
  45292. })
  45293. #define vcvts_n_s32_f32(__p0, __p1) __extension__ ({ \
  45294.   int32_t __ret; \
  45295.   float32_t __s0 = __p0; \
  45296.   __ret = (int32_t) __builtin_neon_vcvts_n_s32_f32(__s0, __p1); \
  45297.   __ret; \
  45298. })
  45299. #ifdef __LITTLE_ENDIAN__
  45300. #define vcvtq_n_s64_f64(__p0, __p1) __extension__ ({ \
  45301.   int64x2_t __ret; \
  45302.   float64x2_t __s0 = __p0; \
  45303.   __ret = (int64x2_t) __builtin_neon_vcvtq_n_s64_v((int8x16_t)__s0, __p1, 35); \
  45304.   __ret; \
  45305. })
  45306. #else
  45307. #define vcvtq_n_s64_f64(__p0, __p1) __extension__ ({ \
  45308.   int64x2_t __ret; \
  45309.   float64x2_t __s0 = __p0; \
  45310.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  45311.   __ret = (int64x2_t) __builtin_neon_vcvtq_n_s64_v((int8x16_t)__rev0, __p1, 35); \
  45312.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  45313.   __ret; \
  45314. })
  45315. #endif
  45316.  
  45317. #define vcvt_n_s64_f64(__p0, __p1) __extension__ ({ \
  45318.   int64x1_t __ret; \
  45319.   float64x1_t __s0 = __p0; \
  45320.   __ret = (int64x1_t) __builtin_neon_vcvt_n_s64_v((int8x8_t)__s0, __p1, 3); \
  45321.   __ret; \
  45322. })
  45323. #define vcvtd_n_s64_f64(__p0, __p1) __extension__ ({ \
  45324.   int64_t __ret; \
  45325.   float64_t __s0 = __p0; \
  45326.   __ret = (int64_t) __builtin_neon_vcvtd_n_s64_f64(__s0, __p1); \
  45327.   __ret; \
  45328. })
  45329. #define vcvts_n_u32_f32(__p0, __p1) __extension__ ({ \
  45330.   uint32_t __ret; \
  45331.   float32_t __s0 = __p0; \
  45332.   __ret = (uint32_t) __builtin_neon_vcvts_n_u32_f32(__s0, __p1); \
  45333.   __ret; \
  45334. })
  45335. #ifdef __LITTLE_ENDIAN__
  45336. #define vcvtq_n_u64_f64(__p0, __p1) __extension__ ({ \
  45337.   uint64x2_t __ret; \
  45338.   float64x2_t __s0 = __p0; \
  45339.   __ret = (uint64x2_t) __builtin_neon_vcvtq_n_u64_v((int8x16_t)__s0, __p1, 51); \
  45340.   __ret; \
  45341. })
  45342. #else
  45343. #define vcvtq_n_u64_f64(__p0, __p1) __extension__ ({ \
  45344.   uint64x2_t __ret; \
  45345.   float64x2_t __s0 = __p0; \
  45346.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  45347.   __ret = (uint64x2_t) __builtin_neon_vcvtq_n_u64_v((int8x16_t)__rev0, __p1, 51); \
  45348.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  45349.   __ret; \
  45350. })
  45351. #endif
  45352.  
  45353. #define vcvt_n_u64_f64(__p0, __p1) __extension__ ({ \
  45354.   uint64x1_t __ret; \
  45355.   float64x1_t __s0 = __p0; \
  45356.   __ret = (uint64x1_t) __builtin_neon_vcvt_n_u64_v((int8x8_t)__s0, __p1, 19); \
  45357.   __ret; \
  45358. })
  45359. #define vcvtd_n_u64_f64(__p0, __p1) __extension__ ({ \
  45360.   uint64_t __ret; \
  45361.   float64_t __s0 = __p0; \
  45362.   __ret = (uint64_t) __builtin_neon_vcvtd_n_u64_f64(__s0, __p1); \
  45363.   __ret; \
  45364. })
  45365. __ai int32_t vcvts_s32_f32(float32_t __p0) {
  45366.   int32_t __ret;
  45367.   __ret = (int32_t) __builtin_neon_vcvts_s32_f32(__p0);
  45368.   return __ret;
  45369. }
  45370. __ai int64_t vcvtd_s64_f64(float64_t __p0) {
  45371.   int64_t __ret;
  45372.   __ret = (int64_t) __builtin_neon_vcvtd_s64_f64(__p0);
  45373.   return __ret;
  45374. }
  45375. #ifdef __LITTLE_ENDIAN__
  45376. __ai int64x2_t vcvtq_s64_f64(float64x2_t __p0) {
  45377.   int64x2_t __ret;
  45378.   __ret = (int64x2_t) __builtin_neon_vcvtq_s64_v((int8x16_t)__p0, 35);
  45379.   return __ret;
  45380. }
  45381. #else
  45382. __ai int64x2_t vcvtq_s64_f64(float64x2_t __p0) {
  45383.   int64x2_t __ret;
  45384.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45385.   __ret = (int64x2_t) __builtin_neon_vcvtq_s64_v((int8x16_t)__rev0, 35);
  45386.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45387.   return __ret;
  45388. }
  45389. #endif
  45390.  
  45391. __ai int64x1_t vcvt_s64_f64(float64x1_t __p0) {
  45392.   int64x1_t __ret;
  45393.   __ret = (int64x1_t) __builtin_neon_vcvt_s64_v((int8x8_t)__p0, 3);
  45394.   return __ret;
  45395. }
  45396. __ai uint32_t vcvts_u32_f32(float32_t __p0) {
  45397.   uint32_t __ret;
  45398.   __ret = (uint32_t) __builtin_neon_vcvts_u32_f32(__p0);
  45399.   return __ret;
  45400. }
  45401. __ai uint64_t vcvtd_u64_f64(float64_t __p0) {
  45402.   uint64_t __ret;
  45403.   __ret = (uint64_t) __builtin_neon_vcvtd_u64_f64(__p0);
  45404.   return __ret;
  45405. }
  45406. #ifdef __LITTLE_ENDIAN__
  45407. __ai uint64x2_t vcvtq_u64_f64(float64x2_t __p0) {
  45408.   uint64x2_t __ret;
  45409.   __ret = (uint64x2_t) __builtin_neon_vcvtq_u64_v((int8x16_t)__p0, 51);
  45410.   return __ret;
  45411. }
  45412. #else
  45413. __ai uint64x2_t vcvtq_u64_f64(float64x2_t __p0) {
  45414.   uint64x2_t __ret;
  45415.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45416.   __ret = (uint64x2_t) __builtin_neon_vcvtq_u64_v((int8x16_t)__rev0, 51);
  45417.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45418.   return __ret;
  45419. }
  45420. #endif
  45421.  
  45422. __ai uint64x1_t vcvt_u64_f64(float64x1_t __p0) {
  45423.   uint64x1_t __ret;
  45424.   __ret = (uint64x1_t) __builtin_neon_vcvt_u64_v((int8x8_t)__p0, 19);
  45425.   return __ret;
  45426. }
  45427. __ai int32_t vcvtas_s32_f32(float32_t __p0) {
  45428.   int32_t __ret;
  45429.   __ret = (int32_t) __builtin_neon_vcvtas_s32_f32(__p0);
  45430.   return __ret;
  45431. }
  45432. #ifdef __LITTLE_ENDIAN__
  45433. __ai int64x2_t vcvtaq_s64_f64(float64x2_t __p0) {
  45434.   int64x2_t __ret;
  45435.   __ret = (int64x2_t) __builtin_neon_vcvtaq_s64_v((int8x16_t)__p0, 35);
  45436.   return __ret;
  45437. }
  45438. #else
  45439. __ai int64x2_t vcvtaq_s64_f64(float64x2_t __p0) {
  45440.   int64x2_t __ret;
  45441.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45442.   __ret = (int64x2_t) __builtin_neon_vcvtaq_s64_v((int8x16_t)__rev0, 35);
  45443.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45444.   return __ret;
  45445. }
  45446. #endif
  45447.  
  45448. __ai int64x1_t vcvta_s64_f64(float64x1_t __p0) {
  45449.   int64x1_t __ret;
  45450.   __ret = (int64x1_t) __builtin_neon_vcvta_s64_v((int8x8_t)__p0, 3);
  45451.   return __ret;
  45452. }
  45453. __ai int64_t vcvtad_s64_f64(float64_t __p0) {
  45454.   int64_t __ret;
  45455.   __ret = (int64_t) __builtin_neon_vcvtad_s64_f64(__p0);
  45456.   return __ret;
  45457. }
  45458. __ai uint32_t vcvtas_u32_f32(float32_t __p0) {
  45459.   uint32_t __ret;
  45460.   __ret = (uint32_t) __builtin_neon_vcvtas_u32_f32(__p0);
  45461.   return __ret;
  45462. }
  45463. #ifdef __LITTLE_ENDIAN__
  45464. __ai uint64x2_t vcvtaq_u64_f64(float64x2_t __p0) {
  45465.   uint64x2_t __ret;
  45466.   __ret = (uint64x2_t) __builtin_neon_vcvtaq_u64_v((int8x16_t)__p0, 51);
  45467.   return __ret;
  45468. }
  45469. #else
  45470. __ai uint64x2_t vcvtaq_u64_f64(float64x2_t __p0) {
  45471.   uint64x2_t __ret;
  45472.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45473.   __ret = (uint64x2_t) __builtin_neon_vcvtaq_u64_v((int8x16_t)__rev0, 51);
  45474.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45475.   return __ret;
  45476. }
  45477. #endif
  45478.  
  45479. __ai uint64x1_t vcvta_u64_f64(float64x1_t __p0) {
  45480.   uint64x1_t __ret;
  45481.   __ret = (uint64x1_t) __builtin_neon_vcvta_u64_v((int8x8_t)__p0, 19);
  45482.   return __ret;
  45483. }
  45484. __ai uint64_t vcvtad_u64_f64(float64_t __p0) {
  45485.   uint64_t __ret;
  45486.   __ret = (uint64_t) __builtin_neon_vcvtad_u64_f64(__p0);
  45487.   return __ret;
  45488. }
  45489. __ai int32_t vcvtms_s32_f32(float32_t __p0) {
  45490.   int32_t __ret;
  45491.   __ret = (int32_t) __builtin_neon_vcvtms_s32_f32(__p0);
  45492.   return __ret;
  45493. }
  45494. #ifdef __LITTLE_ENDIAN__
  45495. __ai int64x2_t vcvtmq_s64_f64(float64x2_t __p0) {
  45496.   int64x2_t __ret;
  45497.   __ret = (int64x2_t) __builtin_neon_vcvtmq_s64_v((int8x16_t)__p0, 35);
  45498.   return __ret;
  45499. }
  45500. #else
  45501. __ai int64x2_t vcvtmq_s64_f64(float64x2_t __p0) {
  45502.   int64x2_t __ret;
  45503.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45504.   __ret = (int64x2_t) __builtin_neon_vcvtmq_s64_v((int8x16_t)__rev0, 35);
  45505.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45506.   return __ret;
  45507. }
  45508. #endif
  45509.  
  45510. __ai int64x1_t vcvtm_s64_f64(float64x1_t __p0) {
  45511.   int64x1_t __ret;
  45512.   __ret = (int64x1_t) __builtin_neon_vcvtm_s64_v((int8x8_t)__p0, 3);
  45513.   return __ret;
  45514. }
  45515. __ai int64_t vcvtmd_s64_f64(float64_t __p0) {
  45516.   int64_t __ret;
  45517.   __ret = (int64_t) __builtin_neon_vcvtmd_s64_f64(__p0);
  45518.   return __ret;
  45519. }
  45520. __ai uint32_t vcvtms_u32_f32(float32_t __p0) {
  45521.   uint32_t __ret;
  45522.   __ret = (uint32_t) __builtin_neon_vcvtms_u32_f32(__p0);
  45523.   return __ret;
  45524. }
  45525. #ifdef __LITTLE_ENDIAN__
  45526. __ai uint64x2_t vcvtmq_u64_f64(float64x2_t __p0) {
  45527.   uint64x2_t __ret;
  45528.   __ret = (uint64x2_t) __builtin_neon_vcvtmq_u64_v((int8x16_t)__p0, 51);
  45529.   return __ret;
  45530. }
  45531. #else
  45532. __ai uint64x2_t vcvtmq_u64_f64(float64x2_t __p0) {
  45533.   uint64x2_t __ret;
  45534.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45535.   __ret = (uint64x2_t) __builtin_neon_vcvtmq_u64_v((int8x16_t)__rev0, 51);
  45536.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45537.   return __ret;
  45538. }
  45539. #endif
  45540.  
  45541. __ai uint64x1_t vcvtm_u64_f64(float64x1_t __p0) {
  45542.   uint64x1_t __ret;
  45543.   __ret = (uint64x1_t) __builtin_neon_vcvtm_u64_v((int8x8_t)__p0, 19);
  45544.   return __ret;
  45545. }
  45546. __ai uint64_t vcvtmd_u64_f64(float64_t __p0) {
  45547.   uint64_t __ret;
  45548.   __ret = (uint64_t) __builtin_neon_vcvtmd_u64_f64(__p0);
  45549.   return __ret;
  45550. }
  45551. __ai int32_t vcvtns_s32_f32(float32_t __p0) {
  45552.   int32_t __ret;
  45553.   __ret = (int32_t) __builtin_neon_vcvtns_s32_f32(__p0);
  45554.   return __ret;
  45555. }
  45556. #ifdef __LITTLE_ENDIAN__
  45557. __ai int64x2_t vcvtnq_s64_f64(float64x2_t __p0) {
  45558.   int64x2_t __ret;
  45559.   __ret = (int64x2_t) __builtin_neon_vcvtnq_s64_v((int8x16_t)__p0, 35);
  45560.   return __ret;
  45561. }
  45562. #else
  45563. __ai int64x2_t vcvtnq_s64_f64(float64x2_t __p0) {
  45564.   int64x2_t __ret;
  45565.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45566.   __ret = (int64x2_t) __builtin_neon_vcvtnq_s64_v((int8x16_t)__rev0, 35);
  45567.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45568.   return __ret;
  45569. }
  45570. #endif
  45571.  
  45572. __ai int64x1_t vcvtn_s64_f64(float64x1_t __p0) {
  45573.   int64x1_t __ret;
  45574.   __ret = (int64x1_t) __builtin_neon_vcvtn_s64_v((int8x8_t)__p0, 3);
  45575.   return __ret;
  45576. }
  45577. __ai int64_t vcvtnd_s64_f64(float64_t __p0) {
  45578.   int64_t __ret;
  45579.   __ret = (int64_t) __builtin_neon_vcvtnd_s64_f64(__p0);
  45580.   return __ret;
  45581. }
  45582. __ai uint32_t vcvtns_u32_f32(float32_t __p0) {
  45583.   uint32_t __ret;
  45584.   __ret = (uint32_t) __builtin_neon_vcvtns_u32_f32(__p0);
  45585.   return __ret;
  45586. }
  45587. #ifdef __LITTLE_ENDIAN__
  45588. __ai uint64x2_t vcvtnq_u64_f64(float64x2_t __p0) {
  45589.   uint64x2_t __ret;
  45590.   __ret = (uint64x2_t) __builtin_neon_vcvtnq_u64_v((int8x16_t)__p0, 51);
  45591.   return __ret;
  45592. }
  45593. #else
  45594. __ai uint64x2_t vcvtnq_u64_f64(float64x2_t __p0) {
  45595.   uint64x2_t __ret;
  45596.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45597.   __ret = (uint64x2_t) __builtin_neon_vcvtnq_u64_v((int8x16_t)__rev0, 51);
  45598.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45599.   return __ret;
  45600. }
  45601. #endif
  45602.  
  45603. __ai uint64x1_t vcvtn_u64_f64(float64x1_t __p0) {
  45604.   uint64x1_t __ret;
  45605.   __ret = (uint64x1_t) __builtin_neon_vcvtn_u64_v((int8x8_t)__p0, 19);
  45606.   return __ret;
  45607. }
  45608. __ai uint64_t vcvtnd_u64_f64(float64_t __p0) {
  45609.   uint64_t __ret;
  45610.   __ret = (uint64_t) __builtin_neon_vcvtnd_u64_f64(__p0);
  45611.   return __ret;
  45612. }
  45613. __ai int32_t vcvtps_s32_f32(float32_t __p0) {
  45614.   int32_t __ret;
  45615.   __ret = (int32_t) __builtin_neon_vcvtps_s32_f32(__p0);
  45616.   return __ret;
  45617. }
  45618. #ifdef __LITTLE_ENDIAN__
  45619. __ai int64x2_t vcvtpq_s64_f64(float64x2_t __p0) {
  45620.   int64x2_t __ret;
  45621.   __ret = (int64x2_t) __builtin_neon_vcvtpq_s64_v((int8x16_t)__p0, 35);
  45622.   return __ret;
  45623. }
  45624. #else
  45625. __ai int64x2_t vcvtpq_s64_f64(float64x2_t __p0) {
  45626.   int64x2_t __ret;
  45627.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45628.   __ret = (int64x2_t) __builtin_neon_vcvtpq_s64_v((int8x16_t)__rev0, 35);
  45629.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45630.   return __ret;
  45631. }
  45632. #endif
  45633.  
  45634. __ai int64x1_t vcvtp_s64_f64(float64x1_t __p0) {
  45635.   int64x1_t __ret;
  45636.   __ret = (int64x1_t) __builtin_neon_vcvtp_s64_v((int8x8_t)__p0, 3);
  45637.   return __ret;
  45638. }
  45639. __ai int64_t vcvtpd_s64_f64(float64_t __p0) {
  45640.   int64_t __ret;
  45641.   __ret = (int64_t) __builtin_neon_vcvtpd_s64_f64(__p0);
  45642.   return __ret;
  45643. }
  45644. __ai uint32_t vcvtps_u32_f32(float32_t __p0) {
  45645.   uint32_t __ret;
  45646.   __ret = (uint32_t) __builtin_neon_vcvtps_u32_f32(__p0);
  45647.   return __ret;
  45648. }
  45649. #ifdef __LITTLE_ENDIAN__
  45650. __ai uint64x2_t vcvtpq_u64_f64(float64x2_t __p0) {
  45651.   uint64x2_t __ret;
  45652.   __ret = (uint64x2_t) __builtin_neon_vcvtpq_u64_v((int8x16_t)__p0, 51);
  45653.   return __ret;
  45654. }
  45655. #else
  45656. __ai uint64x2_t vcvtpq_u64_f64(float64x2_t __p0) {
  45657.   uint64x2_t __ret;
  45658.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45659.   __ret = (uint64x2_t) __builtin_neon_vcvtpq_u64_v((int8x16_t)__rev0, 51);
  45660.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45661.   return __ret;
  45662. }
  45663. #endif
  45664.  
  45665. __ai uint64x1_t vcvtp_u64_f64(float64x1_t __p0) {
  45666.   uint64x1_t __ret;
  45667.   __ret = (uint64x1_t) __builtin_neon_vcvtp_u64_v((int8x8_t)__p0, 19);
  45668.   return __ret;
  45669. }
  45670. __ai uint64_t vcvtpd_u64_f64(float64_t __p0) {
  45671.   uint64_t __ret;
  45672.   __ret = (uint64_t) __builtin_neon_vcvtpd_u64_f64(__p0);
  45673.   return __ret;
  45674. }
  45675. __ai float32_t vcvtxd_f32_f64(float64_t __p0) {
  45676.   float32_t __ret;
  45677.   __ret = (float32_t) __builtin_neon_vcvtxd_f32_f64(__p0);
  45678.   return __ret;
  45679. }
  45680. #ifdef __LITTLE_ENDIAN__
  45681. __ai float32x2_t vcvtx_f32_f64(float64x2_t __p0) {
  45682.   float32x2_t __ret;
  45683.   __ret = (float32x2_t) __builtin_neon_vcvtx_f32_v((int8x16_t)__p0, 42);
  45684.   return __ret;
  45685. }
  45686. #else
  45687. __ai float32x2_t vcvtx_f32_f64(float64x2_t __p0) {
  45688.   float32x2_t __ret;
  45689.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45690.   __ret = (float32x2_t) __builtin_neon_vcvtx_f32_v((int8x16_t)__rev0, 42);
  45691.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45692.   return __ret;
  45693. }
  45694. __ai float32x2_t __noswap_vcvtx_f32_f64(float64x2_t __p0) {
  45695.   float32x2_t __ret;
  45696.   __ret = (float32x2_t) __builtin_neon_vcvtx_f32_v((int8x16_t)__p0, 42);
  45697.   return __ret;
  45698. }
  45699. #endif
  45700.  
  45701. #ifdef __LITTLE_ENDIAN__
  45702. __ai float32x4_t vcvtx_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
  45703.   float32x4_t __ret;
  45704.   __ret = vcombine_f32(__p0, vcvtx_f32_f64(__p1));
  45705.   return __ret;
  45706. }
  45707. #else
  45708. __ai float32x4_t vcvtx_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
  45709.   float32x4_t __ret;
  45710.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45711.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  45712.   __ret = __noswap_vcombine_f32(__rev0, __noswap_vcvtx_f32_f64(__rev1));
  45713.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  45714.   return __ret;
  45715. }
  45716. #endif
  45717.  
  45718. #ifdef __LITTLE_ENDIAN__
  45719. __ai float64x2_t vdivq_f64(float64x2_t __p0, float64x2_t __p1) {
  45720.   float64x2_t __ret;
  45721.   __ret = __p0 / __p1;
  45722.   return __ret;
  45723. }
  45724. #else
  45725. __ai float64x2_t vdivq_f64(float64x2_t __p0, float64x2_t __p1) {
  45726.   float64x2_t __ret;
  45727.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45728.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  45729.   __ret = __rev0 / __rev1;
  45730.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45731.   return __ret;
  45732. }
  45733. #endif
  45734.  
  45735. #ifdef __LITTLE_ENDIAN__
  45736. __ai float32x4_t vdivq_f32(float32x4_t __p0, float32x4_t __p1) {
  45737.   float32x4_t __ret;
  45738.   __ret = __p0 / __p1;
  45739.   return __ret;
  45740. }
  45741. #else
  45742. __ai float32x4_t vdivq_f32(float32x4_t __p0, float32x4_t __p1) {
  45743.   float32x4_t __ret;
  45744.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  45745.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  45746.   __ret = __rev0 / __rev1;
  45747.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  45748.   return __ret;
  45749. }
  45750. #endif
  45751.  
  45752. __ai float64x1_t vdiv_f64(float64x1_t __p0, float64x1_t __p1) {
  45753.   float64x1_t __ret;
  45754.   __ret = __p0 / __p1;
  45755.   return __ret;
  45756. }
  45757. #ifdef __LITTLE_ENDIAN__
  45758. __ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) {
  45759.   float32x2_t __ret;
  45760.   __ret = __p0 / __p1;
  45761.   return __ret;
  45762. }
  45763. #else
  45764. __ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) {
  45765.   float32x2_t __ret;
  45766.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  45767.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  45768.   __ret = __rev0 / __rev1;
  45769.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  45770.   return __ret;
  45771. }
  45772. #endif
  45773.  
  45774. #ifdef __LITTLE_ENDIAN__
  45775. #define vdupb_lane_p8(__p0, __p1) __extension__ ({ \
  45776.   poly8_t __ret; \
  45777.   poly8x8_t __s0 = __p0; \
  45778.   __ret = (poly8_t) __builtin_neon_vdupb_lane_i8((poly8x8_t)__s0, __p1); \
  45779.   __ret; \
  45780. })
  45781. #else
  45782. #define vdupb_lane_p8(__p0, __p1) __extension__ ({ \
  45783.   poly8_t __ret; \
  45784.   poly8x8_t __s0 = __p0; \
  45785.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  45786.   __ret = (poly8_t) __builtin_neon_vdupb_lane_i8((poly8x8_t)__rev0, __p1); \
  45787.   __ret; \
  45788. })
  45789. #endif
  45790.  
  45791. #ifdef __LITTLE_ENDIAN__
  45792. #define vduph_lane_p16(__p0, __p1) __extension__ ({ \
  45793.   poly16_t __ret; \
  45794.   poly16x4_t __s0 = __p0; \
  45795.   __ret = (poly16_t) __builtin_neon_vduph_lane_i16((poly16x4_t)__s0, __p1); \
  45796.   __ret; \
  45797. })
  45798. #else
  45799. #define vduph_lane_p16(__p0, __p1) __extension__ ({ \
  45800.   poly16_t __ret; \
  45801.   poly16x4_t __s0 = __p0; \
  45802.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  45803.   __ret = (poly16_t) __builtin_neon_vduph_lane_i16((poly16x4_t)__rev0, __p1); \
  45804.   __ret; \
  45805. })
  45806. #endif
  45807.  
  45808. #ifdef __LITTLE_ENDIAN__
  45809. #define vdupb_lane_u8(__p0, __p1) __extension__ ({ \
  45810.   uint8_t __ret; \
  45811.   uint8x8_t __s0 = __p0; \
  45812.   __ret = (uint8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__s0, __p1); \
  45813.   __ret; \
  45814. })
  45815. #else
  45816. #define vdupb_lane_u8(__p0, __p1) __extension__ ({ \
  45817.   uint8_t __ret; \
  45818.   uint8x8_t __s0 = __p0; \
  45819.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  45820.   __ret = (uint8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__rev0, __p1); \
  45821.   __ret; \
  45822. })
  45823. #endif
  45824.  
  45825. #ifdef __LITTLE_ENDIAN__
  45826. #define vdups_lane_u32(__p0, __p1) __extension__ ({ \
  45827.   uint32_t __ret; \
  45828.   uint32x2_t __s0 = __p0; \
  45829.   __ret = (uint32_t) __builtin_neon_vdups_lane_i32((int32x2_t)__s0, __p1); \
  45830.   __ret; \
  45831. })
  45832. #else
  45833. #define vdups_lane_u32(__p0, __p1) __extension__ ({ \
  45834.   uint32_t __ret; \
  45835.   uint32x2_t __s0 = __p0; \
  45836.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  45837.   __ret = (uint32_t) __builtin_neon_vdups_lane_i32((int32x2_t)__rev0, __p1); \
  45838.   __ret; \
  45839. })
  45840. #endif
  45841.  
  45842. #define vdupd_lane_u64(__p0, __p1) __extension__ ({ \
  45843.   uint64_t __ret; \
  45844.   uint64x1_t __s0 = __p0; \
  45845.   __ret = (uint64_t) __builtin_neon_vdupd_lane_i64((int64x1_t)__s0, __p1); \
  45846.   __ret; \
  45847. })
  45848. #ifdef __LITTLE_ENDIAN__
  45849. #define vduph_lane_u16(__p0, __p1) __extension__ ({ \
  45850.   uint16_t __ret; \
  45851.   uint16x4_t __s0 = __p0; \
  45852.   __ret = (uint16_t) __builtin_neon_vduph_lane_i16((int16x4_t)__s0, __p1); \
  45853.   __ret; \
  45854. })
  45855. #else
  45856. #define vduph_lane_u16(__p0, __p1) __extension__ ({ \
  45857.   uint16_t __ret; \
  45858.   uint16x4_t __s0 = __p0; \
  45859.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  45860.   __ret = (uint16_t) __builtin_neon_vduph_lane_i16((int16x4_t)__rev0, __p1); \
  45861.   __ret; \
  45862. })
  45863. #endif
  45864.  
  45865. #ifdef __LITTLE_ENDIAN__
  45866. #define vdupb_lane_s8(__p0, __p1) __extension__ ({ \
  45867.   int8_t __ret; \
  45868.   int8x8_t __s0 = __p0; \
  45869.   __ret = (int8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__s0, __p1); \
  45870.   __ret; \
  45871. })
  45872. #else
  45873. #define vdupb_lane_s8(__p0, __p1) __extension__ ({ \
  45874.   int8_t __ret; \
  45875.   int8x8_t __s0 = __p0; \
  45876.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  45877.   __ret = (int8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__rev0, __p1); \
  45878.   __ret; \
  45879. })
  45880. #endif
  45881.  
  45882. #define vdupd_lane_f64(__p0, __p1) __extension__ ({ \
  45883.   float64_t __ret; \
  45884.   float64x1_t __s0 = __p0; \
  45885.   __ret = (float64_t) __builtin_neon_vdupd_lane_f64((float64x1_t)__s0, __p1); \
  45886.   __ret; \
  45887. })
  45888. #ifdef __LITTLE_ENDIAN__
  45889. #define vdups_lane_f32(__p0, __p1) __extension__ ({ \
  45890.   float32_t __ret; \
  45891.   float32x2_t __s0 = __p0; \
  45892.   __ret = (float32_t) __builtin_neon_vdups_lane_f32((float32x2_t)__s0, __p1); \
  45893.   __ret; \
  45894. })
  45895. #else
  45896. #define vdups_lane_f32(__p0, __p1) __extension__ ({ \
  45897.   float32_t __ret; \
  45898.   float32x2_t __s0 = __p0; \
  45899.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  45900.   __ret = (float32_t) __builtin_neon_vdups_lane_f32((float32x2_t)__rev0, __p1); \
  45901.   __ret; \
  45902. })
  45903. #endif
  45904.  
  45905. #ifdef __LITTLE_ENDIAN__
  45906. #define vdups_lane_s32(__p0, __p1) __extension__ ({ \
  45907.   int32_t __ret; \
  45908.   int32x2_t __s0 = __p0; \
  45909.   __ret = (int32_t) __builtin_neon_vdups_lane_i32((int32x2_t)__s0, __p1); \
  45910.   __ret; \
  45911. })
  45912. #else
  45913. #define vdups_lane_s32(__p0, __p1) __extension__ ({ \
  45914.   int32_t __ret; \
  45915.   int32x2_t __s0 = __p0; \
  45916.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  45917.   __ret = (int32_t) __builtin_neon_vdups_lane_i32((int32x2_t)__rev0, __p1); \
  45918.   __ret; \
  45919. })
  45920. #endif
  45921.  
  45922. #define vdupd_lane_s64(__p0, __p1) __extension__ ({ \
  45923.   int64_t __ret; \
  45924.   int64x1_t __s0 = __p0; \
  45925.   __ret = (int64_t) __builtin_neon_vdupd_lane_i64((int64x1_t)__s0, __p1); \
  45926.   __ret; \
  45927. })
  45928. #ifdef __LITTLE_ENDIAN__
  45929. #define vduph_lane_s16(__p0, __p1) __extension__ ({ \
  45930.   int16_t __ret; \
  45931.   int16x4_t __s0 = __p0; \
  45932.   __ret = (int16_t) __builtin_neon_vduph_lane_i16((int16x4_t)__s0, __p1); \
  45933.   __ret; \
  45934. })
  45935. #else
  45936. #define vduph_lane_s16(__p0, __p1) __extension__ ({ \
  45937.   int16_t __ret; \
  45938.   int16x4_t __s0 = __p0; \
  45939.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  45940.   __ret = (int16_t) __builtin_neon_vduph_lane_i16((int16x4_t)__rev0, __p1); \
  45941.   __ret; \
  45942. })
  45943. #endif
  45944.  
  45945. #define vdup_lane_p64(__p0_343, __p1_343) __extension__ ({ \
  45946.   poly64x1_t __ret_343; \
  45947.   poly64x1_t __s0_343 = __p0_343; \
  45948.   __ret_343 = splat_lane_p64(__s0_343, __p1_343); \
  45949.   __ret_343; \
  45950. })
  45951. #ifdef __LITTLE_ENDIAN__
  45952. #define vdupq_lane_p64(__p0_344, __p1_344) __extension__ ({ \
  45953.   poly64x2_t __ret_344; \
  45954.   poly64x1_t __s0_344 = __p0_344; \
  45955.   __ret_344 = splatq_lane_p64(__s0_344, __p1_344); \
  45956.   __ret_344; \
  45957. })
  45958. #else
  45959. #define vdupq_lane_p64(__p0_345, __p1_345) __extension__ ({ \
  45960.   poly64x2_t __ret_345; \
  45961.   poly64x1_t __s0_345 = __p0_345; \
  45962.   __ret_345 = __noswap_splatq_lane_p64(__s0_345, __p1_345); \
  45963.   __ret_345 = __builtin_shufflevector(__ret_345, __ret_345, 1, 0); \
  45964.   __ret_345; \
  45965. })
  45966. #endif
  45967.  
  45968. #ifdef __LITTLE_ENDIAN__
  45969. #define vdupq_lane_f64(__p0_346, __p1_346) __extension__ ({ \
  45970.   float64x2_t __ret_346; \
  45971.   float64x1_t __s0_346 = __p0_346; \
  45972.   __ret_346 = splatq_lane_f64(__s0_346, __p1_346); \
  45973.   __ret_346; \
  45974. })
  45975. #else
  45976. #define vdupq_lane_f64(__p0_347, __p1_347) __extension__ ({ \
  45977.   float64x2_t __ret_347; \
  45978.   float64x1_t __s0_347 = __p0_347; \
  45979.   __ret_347 = __noswap_splatq_lane_f64(__s0_347, __p1_347); \
  45980.   __ret_347 = __builtin_shufflevector(__ret_347, __ret_347, 1, 0); \
  45981.   __ret_347; \
  45982. })
  45983. #endif
  45984.  
  45985. #define vdup_lane_f64(__p0_348, __p1_348) __extension__ ({ \
  45986.   float64x1_t __ret_348; \
  45987.   float64x1_t __s0_348 = __p0_348; \
  45988.   __ret_348 = splat_lane_f64(__s0_348, __p1_348); \
  45989.   __ret_348; \
  45990. })
  45991. #ifdef __LITTLE_ENDIAN__
  45992. #define vdupb_laneq_p8(__p0, __p1) __extension__ ({ \
  45993.   poly8_t __ret; \
  45994.   poly8x16_t __s0 = __p0; \
  45995.   __ret = (poly8_t) __builtin_neon_vdupb_laneq_i8((poly8x16_t)__s0, __p1); \
  45996.   __ret; \
  45997. })
  45998. #else
  45999. #define vdupb_laneq_p8(__p0, __p1) __extension__ ({ \
  46000.   poly8_t __ret; \
  46001.   poly8x16_t __s0 = __p0; \
  46002.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46003.   __ret = (poly8_t) __builtin_neon_vdupb_laneq_i8((poly8x16_t)__rev0, __p1); \
  46004.   __ret; \
  46005. })
  46006. #endif
  46007.  
  46008. #ifdef __LITTLE_ENDIAN__
  46009. #define vduph_laneq_p16(__p0, __p1) __extension__ ({ \
  46010.   poly16_t __ret; \
  46011.   poly16x8_t __s0 = __p0; \
  46012.   __ret = (poly16_t) __builtin_neon_vduph_laneq_i16((poly16x8_t)__s0, __p1); \
  46013.   __ret; \
  46014. })
  46015. #else
  46016. #define vduph_laneq_p16(__p0, __p1) __extension__ ({ \
  46017.   poly16_t __ret; \
  46018.   poly16x8_t __s0 = __p0; \
  46019.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  46020.   __ret = (poly16_t) __builtin_neon_vduph_laneq_i16((poly16x8_t)__rev0, __p1); \
  46021.   __ret; \
  46022. })
  46023. #endif
  46024.  
  46025. #ifdef __LITTLE_ENDIAN__
  46026. #define vdupb_laneq_u8(__p0, __p1) __extension__ ({ \
  46027.   uint8_t __ret; \
  46028.   uint8x16_t __s0 = __p0; \
  46029.   __ret = (uint8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__s0, __p1); \
  46030.   __ret; \
  46031. })
  46032. #else
  46033. #define vdupb_laneq_u8(__p0, __p1) __extension__ ({ \
  46034.   uint8_t __ret; \
  46035.   uint8x16_t __s0 = __p0; \
  46036.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46037.   __ret = (uint8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__rev0, __p1); \
  46038.   __ret; \
  46039. })
  46040. #endif
  46041.  
  46042. #ifdef __LITTLE_ENDIAN__
  46043. #define vdups_laneq_u32(__p0, __p1) __extension__ ({ \
  46044.   uint32_t __ret; \
  46045.   uint32x4_t __s0 = __p0; \
  46046.   __ret = (uint32_t) __builtin_neon_vdups_laneq_i32((int32x4_t)__s0, __p1); \
  46047.   __ret; \
  46048. })
  46049. #else
  46050. #define vdups_laneq_u32(__p0, __p1) __extension__ ({ \
  46051.   uint32_t __ret; \
  46052.   uint32x4_t __s0 = __p0; \
  46053.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  46054.   __ret = (uint32_t) __builtin_neon_vdups_laneq_i32((int32x4_t)__rev0, __p1); \
  46055.   __ret; \
  46056. })
  46057. #endif
  46058.  
  46059. #ifdef __LITTLE_ENDIAN__
  46060. #define vdupd_laneq_u64(__p0, __p1) __extension__ ({ \
  46061.   uint64_t __ret; \
  46062.   uint64x2_t __s0 = __p0; \
  46063.   __ret = (uint64_t) __builtin_neon_vdupd_laneq_i64((int64x2_t)__s0, __p1); \
  46064.   __ret; \
  46065. })
  46066. #else
  46067. #define vdupd_laneq_u64(__p0, __p1) __extension__ ({ \
  46068.   uint64_t __ret; \
  46069.   uint64x2_t __s0 = __p0; \
  46070.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  46071.   __ret = (uint64_t) __builtin_neon_vdupd_laneq_i64((int64x2_t)__rev0, __p1); \
  46072.   __ret; \
  46073. })
  46074. #endif
  46075.  
  46076. #ifdef __LITTLE_ENDIAN__
  46077. #define vduph_laneq_u16(__p0, __p1) __extension__ ({ \
  46078.   uint16_t __ret; \
  46079.   uint16x8_t __s0 = __p0; \
  46080.   __ret = (uint16_t) __builtin_neon_vduph_laneq_i16((int16x8_t)__s0, __p1); \
  46081.   __ret; \
  46082. })
  46083. #else
  46084. #define vduph_laneq_u16(__p0, __p1) __extension__ ({ \
  46085.   uint16_t __ret; \
  46086.   uint16x8_t __s0 = __p0; \
  46087.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  46088.   __ret = (uint16_t) __builtin_neon_vduph_laneq_i16((int16x8_t)__rev0, __p1); \
  46089.   __ret; \
  46090. })
  46091. #endif
  46092.  
  46093. #ifdef __LITTLE_ENDIAN__
  46094. #define vdupb_laneq_s8(__p0, __p1) __extension__ ({ \
  46095.   int8_t __ret; \
  46096.   int8x16_t __s0 = __p0; \
  46097.   __ret = (int8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__s0, __p1); \
  46098.   __ret; \
  46099. })
  46100. #else
  46101. #define vdupb_laneq_s8(__p0, __p1) __extension__ ({ \
  46102.   int8_t __ret; \
  46103.   int8x16_t __s0 = __p0; \
  46104.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46105.   __ret = (int8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__rev0, __p1); \
  46106.   __ret; \
  46107. })
  46108. #endif
  46109.  
  46110. #ifdef __LITTLE_ENDIAN__
  46111. #define vdupd_laneq_f64(__p0, __p1) __extension__ ({ \
  46112.   float64_t __ret; \
  46113.   float64x2_t __s0 = __p0; \
  46114.   __ret = (float64_t) __builtin_neon_vdupd_laneq_f64((float64x2_t)__s0, __p1); \
  46115.   __ret; \
  46116. })
  46117. #else
  46118. #define vdupd_laneq_f64(__p0, __p1) __extension__ ({ \
  46119.   float64_t __ret; \
  46120.   float64x2_t __s0 = __p0; \
  46121.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  46122.   __ret = (float64_t) __builtin_neon_vdupd_laneq_f64((float64x2_t)__rev0, __p1); \
  46123.   __ret; \
  46124. })
  46125. #endif
  46126.  
  46127. #ifdef __LITTLE_ENDIAN__
  46128. #define vdups_laneq_f32(__p0, __p1) __extension__ ({ \
  46129.   float32_t __ret; \
  46130.   float32x4_t __s0 = __p0; \
  46131.   __ret = (float32_t) __builtin_neon_vdups_laneq_f32((float32x4_t)__s0, __p1); \
  46132.   __ret; \
  46133. })
  46134. #else
  46135. #define vdups_laneq_f32(__p0, __p1) __extension__ ({ \
  46136.   float32_t __ret; \
  46137.   float32x4_t __s0 = __p0; \
  46138.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  46139.   __ret = (float32_t) __builtin_neon_vdups_laneq_f32((float32x4_t)__rev0, __p1); \
  46140.   __ret; \
  46141. })
  46142. #endif
  46143.  
  46144. #ifdef __LITTLE_ENDIAN__
  46145. #define vdups_laneq_s32(__p0, __p1) __extension__ ({ \
  46146.   int32_t __ret; \
  46147.   int32x4_t __s0 = __p0; \
  46148.   __ret = (int32_t) __builtin_neon_vdups_laneq_i32((int32x4_t)__s0, __p1); \
  46149.   __ret; \
  46150. })
  46151. #else
  46152. #define vdups_laneq_s32(__p0, __p1) __extension__ ({ \
  46153.   int32_t __ret; \
  46154.   int32x4_t __s0 = __p0; \
  46155.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  46156.   __ret = (int32_t) __builtin_neon_vdups_laneq_i32((int32x4_t)__rev0, __p1); \
  46157.   __ret; \
  46158. })
  46159. #endif
  46160.  
  46161. #ifdef __LITTLE_ENDIAN__
  46162. #define vdupd_laneq_s64(__p0, __p1) __extension__ ({ \
  46163.   int64_t __ret; \
  46164.   int64x2_t __s0 = __p0; \
  46165.   __ret = (int64_t) __builtin_neon_vdupd_laneq_i64((int64x2_t)__s0, __p1); \
  46166.   __ret; \
  46167. })
  46168. #else
  46169. #define vdupd_laneq_s64(__p0, __p1) __extension__ ({ \
  46170.   int64_t __ret; \
  46171.   int64x2_t __s0 = __p0; \
  46172.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  46173.   __ret = (int64_t) __builtin_neon_vdupd_laneq_i64((int64x2_t)__rev0, __p1); \
  46174.   __ret; \
  46175. })
  46176. #endif
  46177.  
  46178. #ifdef __LITTLE_ENDIAN__
  46179. #define vduph_laneq_s16(__p0, __p1) __extension__ ({ \
  46180.   int16_t __ret; \
  46181.   int16x8_t __s0 = __p0; \
  46182.   __ret = (int16_t) __builtin_neon_vduph_laneq_i16((int16x8_t)__s0, __p1); \
  46183.   __ret; \
  46184. })
  46185. #else
  46186. #define vduph_laneq_s16(__p0, __p1) __extension__ ({ \
  46187.   int16_t __ret; \
  46188.   int16x8_t __s0 = __p0; \
  46189.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  46190.   __ret = (int16_t) __builtin_neon_vduph_laneq_i16((int16x8_t)__rev0, __p1); \
  46191.   __ret; \
  46192. })
  46193. #endif
  46194.  
  46195. #ifdef __LITTLE_ENDIAN__
  46196. #define vdup_laneq_p8(__p0_349, __p1_349) __extension__ ({ \
  46197.   poly8x8_t __ret_349; \
  46198.   poly8x16_t __s0_349 = __p0_349; \
  46199.   __ret_349 = splat_laneq_p8(__s0_349, __p1_349); \
  46200.   __ret_349; \
  46201. })
  46202. #else
  46203. #define vdup_laneq_p8(__p0_350, __p1_350) __extension__ ({ \
  46204.   poly8x8_t __ret_350; \
  46205.   poly8x16_t __s0_350 = __p0_350; \
  46206.   poly8x16_t __rev0_350;  __rev0_350 = __builtin_shufflevector(__s0_350, __s0_350, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46207.   __ret_350 = __noswap_splat_laneq_p8(__rev0_350, __p1_350); \
  46208.   __ret_350 = __builtin_shufflevector(__ret_350, __ret_350, 7, 6, 5, 4, 3, 2, 1, 0); \
  46209.   __ret_350; \
  46210. })
  46211. #endif
  46212.  
  46213. #ifdef __LITTLE_ENDIAN__
  46214. #define vdup_laneq_p64(__p0_351, __p1_351) __extension__ ({ \
  46215.   poly64x1_t __ret_351; \
  46216.   poly64x2_t __s0_351 = __p0_351; \
  46217.   __ret_351 = splat_laneq_p64(__s0_351, __p1_351); \
  46218.   __ret_351; \
  46219. })
  46220. #else
  46221. #define vdup_laneq_p64(__p0_352, __p1_352) __extension__ ({ \
  46222.   poly64x1_t __ret_352; \
  46223.   poly64x2_t __s0_352 = __p0_352; \
  46224.   poly64x2_t __rev0_352;  __rev0_352 = __builtin_shufflevector(__s0_352, __s0_352, 1, 0); \
  46225.   __ret_352 = __noswap_splat_laneq_p64(__rev0_352, __p1_352); \
  46226.   __ret_352; \
  46227. })
  46228. #endif
  46229.  
  46230. #ifdef __LITTLE_ENDIAN__
  46231. #define vdup_laneq_p16(__p0_353, __p1_353) __extension__ ({ \
  46232.   poly16x4_t __ret_353; \
  46233.   poly16x8_t __s0_353 = __p0_353; \
  46234.   __ret_353 = splat_laneq_p16(__s0_353, __p1_353); \
  46235.   __ret_353; \
  46236. })
  46237. #else
  46238. #define vdup_laneq_p16(__p0_354, __p1_354) __extension__ ({ \
  46239.   poly16x4_t __ret_354; \
  46240.   poly16x8_t __s0_354 = __p0_354; \
  46241.   poly16x8_t __rev0_354;  __rev0_354 = __builtin_shufflevector(__s0_354, __s0_354, 7, 6, 5, 4, 3, 2, 1, 0); \
  46242.   __ret_354 = __noswap_splat_laneq_p16(__rev0_354, __p1_354); \
  46243.   __ret_354 = __builtin_shufflevector(__ret_354, __ret_354, 3, 2, 1, 0); \
  46244.   __ret_354; \
  46245. })
  46246. #endif
  46247.  
  46248. #ifdef __LITTLE_ENDIAN__
  46249. #define vdupq_laneq_p8(__p0_355, __p1_355) __extension__ ({ \
  46250.   poly8x16_t __ret_355; \
  46251.   poly8x16_t __s0_355 = __p0_355; \
  46252.   __ret_355 = splatq_laneq_p8(__s0_355, __p1_355); \
  46253.   __ret_355; \
  46254. })
  46255. #else
  46256. #define vdupq_laneq_p8(__p0_356, __p1_356) __extension__ ({ \
  46257.   poly8x16_t __ret_356; \
  46258.   poly8x16_t __s0_356 = __p0_356; \
  46259.   poly8x16_t __rev0_356;  __rev0_356 = __builtin_shufflevector(__s0_356, __s0_356, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46260.   __ret_356 = __noswap_splatq_laneq_p8(__rev0_356, __p1_356); \
  46261.   __ret_356 = __builtin_shufflevector(__ret_356, __ret_356, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46262.   __ret_356; \
  46263. })
  46264. #endif
  46265.  
  46266. #ifdef __LITTLE_ENDIAN__
  46267. #define vdupq_laneq_p64(__p0_357, __p1_357) __extension__ ({ \
  46268.   poly64x2_t __ret_357; \
  46269.   poly64x2_t __s0_357 = __p0_357; \
  46270.   __ret_357 = splatq_laneq_p64(__s0_357, __p1_357); \
  46271.   __ret_357; \
  46272. })
  46273. #else
  46274. #define vdupq_laneq_p64(__p0_358, __p1_358) __extension__ ({ \
  46275.   poly64x2_t __ret_358; \
  46276.   poly64x2_t __s0_358 = __p0_358; \
  46277.   poly64x2_t __rev0_358;  __rev0_358 = __builtin_shufflevector(__s0_358, __s0_358, 1, 0); \
  46278.   __ret_358 = __noswap_splatq_laneq_p64(__rev0_358, __p1_358); \
  46279.   __ret_358 = __builtin_shufflevector(__ret_358, __ret_358, 1, 0); \
  46280.   __ret_358; \
  46281. })
  46282. #endif
  46283.  
  46284. #ifdef __LITTLE_ENDIAN__
  46285. #define vdupq_laneq_p16(__p0_359, __p1_359) __extension__ ({ \
  46286.   poly16x8_t __ret_359; \
  46287.   poly16x8_t __s0_359 = __p0_359; \
  46288.   __ret_359 = splatq_laneq_p16(__s0_359, __p1_359); \
  46289.   __ret_359; \
  46290. })
  46291. #else
  46292. #define vdupq_laneq_p16(__p0_360, __p1_360) __extension__ ({ \
  46293.   poly16x8_t __ret_360; \
  46294.   poly16x8_t __s0_360 = __p0_360; \
  46295.   poly16x8_t __rev0_360;  __rev0_360 = __builtin_shufflevector(__s0_360, __s0_360, 7, 6, 5, 4, 3, 2, 1, 0); \
  46296.   __ret_360 = __noswap_splatq_laneq_p16(__rev0_360, __p1_360); \
  46297.   __ret_360 = __builtin_shufflevector(__ret_360, __ret_360, 7, 6, 5, 4, 3, 2, 1, 0); \
  46298.   __ret_360; \
  46299. })
  46300. #endif
  46301.  
  46302. #ifdef __LITTLE_ENDIAN__
  46303. #define vdupq_laneq_u8(__p0_361, __p1_361) __extension__ ({ \
  46304.   uint8x16_t __ret_361; \
  46305.   uint8x16_t __s0_361 = __p0_361; \
  46306.   __ret_361 = splatq_laneq_u8(__s0_361, __p1_361); \
  46307.   __ret_361; \
  46308. })
  46309. #else
  46310. #define vdupq_laneq_u8(__p0_362, __p1_362) __extension__ ({ \
  46311.   uint8x16_t __ret_362; \
  46312.   uint8x16_t __s0_362 = __p0_362; \
  46313.   uint8x16_t __rev0_362;  __rev0_362 = __builtin_shufflevector(__s0_362, __s0_362, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46314.   __ret_362 = __noswap_splatq_laneq_u8(__rev0_362, __p1_362); \
  46315.   __ret_362 = __builtin_shufflevector(__ret_362, __ret_362, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46316.   __ret_362; \
  46317. })
  46318. #endif
  46319.  
  46320. #ifdef __LITTLE_ENDIAN__
  46321. #define vdupq_laneq_u32(__p0_363, __p1_363) __extension__ ({ \
  46322.   uint32x4_t __ret_363; \
  46323.   uint32x4_t __s0_363 = __p0_363; \
  46324.   __ret_363 = splatq_laneq_u32(__s0_363, __p1_363); \
  46325.   __ret_363; \
  46326. })
  46327. #else
  46328. #define vdupq_laneq_u32(__p0_364, __p1_364) __extension__ ({ \
  46329.   uint32x4_t __ret_364; \
  46330.   uint32x4_t __s0_364 = __p0_364; \
  46331.   uint32x4_t __rev0_364;  __rev0_364 = __builtin_shufflevector(__s0_364, __s0_364, 3, 2, 1, 0); \
  46332.   __ret_364 = __noswap_splatq_laneq_u32(__rev0_364, __p1_364); \
  46333.   __ret_364 = __builtin_shufflevector(__ret_364, __ret_364, 3, 2, 1, 0); \
  46334.   __ret_364; \
  46335. })
  46336. #endif
  46337.  
  46338. #ifdef __LITTLE_ENDIAN__
  46339. #define vdupq_laneq_u64(__p0_365, __p1_365) __extension__ ({ \
  46340.   uint64x2_t __ret_365; \
  46341.   uint64x2_t __s0_365 = __p0_365; \
  46342.   __ret_365 = splatq_laneq_u64(__s0_365, __p1_365); \
  46343.   __ret_365; \
  46344. })
  46345. #else
  46346. #define vdupq_laneq_u64(__p0_366, __p1_366) __extension__ ({ \
  46347.   uint64x2_t __ret_366; \
  46348.   uint64x2_t __s0_366 = __p0_366; \
  46349.   uint64x2_t __rev0_366;  __rev0_366 = __builtin_shufflevector(__s0_366, __s0_366, 1, 0); \
  46350.   __ret_366 = __noswap_splatq_laneq_u64(__rev0_366, __p1_366); \
  46351.   __ret_366 = __builtin_shufflevector(__ret_366, __ret_366, 1, 0); \
  46352.   __ret_366; \
  46353. })
  46354. #endif
  46355.  
  46356. #ifdef __LITTLE_ENDIAN__
  46357. #define vdupq_laneq_u16(__p0_367, __p1_367) __extension__ ({ \
  46358.   uint16x8_t __ret_367; \
  46359.   uint16x8_t __s0_367 = __p0_367; \
  46360.   __ret_367 = splatq_laneq_u16(__s0_367, __p1_367); \
  46361.   __ret_367; \
  46362. })
  46363. #else
  46364. #define vdupq_laneq_u16(__p0_368, __p1_368) __extension__ ({ \
  46365.   uint16x8_t __ret_368; \
  46366.   uint16x8_t __s0_368 = __p0_368; \
  46367.   uint16x8_t __rev0_368;  __rev0_368 = __builtin_shufflevector(__s0_368, __s0_368, 7, 6, 5, 4, 3, 2, 1, 0); \
  46368.   __ret_368 = __noswap_splatq_laneq_u16(__rev0_368, __p1_368); \
  46369.   __ret_368 = __builtin_shufflevector(__ret_368, __ret_368, 7, 6, 5, 4, 3, 2, 1, 0); \
  46370.   __ret_368; \
  46371. })
  46372. #endif
  46373.  
  46374. #ifdef __LITTLE_ENDIAN__
  46375. #define vdupq_laneq_s8(__p0_369, __p1_369) __extension__ ({ \
  46376.   int8x16_t __ret_369; \
  46377.   int8x16_t __s0_369 = __p0_369; \
  46378.   __ret_369 = splatq_laneq_s8(__s0_369, __p1_369); \
  46379.   __ret_369; \
  46380. })
  46381. #else
  46382. #define vdupq_laneq_s8(__p0_370, __p1_370) __extension__ ({ \
  46383.   int8x16_t __ret_370; \
  46384.   int8x16_t __s0_370 = __p0_370; \
  46385.   int8x16_t __rev0_370;  __rev0_370 = __builtin_shufflevector(__s0_370, __s0_370, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46386.   __ret_370 = __noswap_splatq_laneq_s8(__rev0_370, __p1_370); \
  46387.   __ret_370 = __builtin_shufflevector(__ret_370, __ret_370, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46388.   __ret_370; \
  46389. })
  46390. #endif
  46391.  
  46392. #ifdef __LITTLE_ENDIAN__
  46393. #define vdupq_laneq_f64(__p0_371, __p1_371) __extension__ ({ \
  46394.   float64x2_t __ret_371; \
  46395.   float64x2_t __s0_371 = __p0_371; \
  46396.   __ret_371 = splatq_laneq_f64(__s0_371, __p1_371); \
  46397.   __ret_371; \
  46398. })
  46399. #else
  46400. #define vdupq_laneq_f64(__p0_372, __p1_372) __extension__ ({ \
  46401.   float64x2_t __ret_372; \
  46402.   float64x2_t __s0_372 = __p0_372; \
  46403.   float64x2_t __rev0_372;  __rev0_372 = __builtin_shufflevector(__s0_372, __s0_372, 1, 0); \
  46404.   __ret_372 = __noswap_splatq_laneq_f64(__rev0_372, __p1_372); \
  46405.   __ret_372 = __builtin_shufflevector(__ret_372, __ret_372, 1, 0); \
  46406.   __ret_372; \
  46407. })
  46408. #endif
  46409.  
  46410. #ifdef __LITTLE_ENDIAN__
  46411. #define vdupq_laneq_f32(__p0_373, __p1_373) __extension__ ({ \
  46412.   float32x4_t __ret_373; \
  46413.   float32x4_t __s0_373 = __p0_373; \
  46414.   __ret_373 = splatq_laneq_f32(__s0_373, __p1_373); \
  46415.   __ret_373; \
  46416. })
  46417. #else
  46418. #define vdupq_laneq_f32(__p0_374, __p1_374) __extension__ ({ \
  46419.   float32x4_t __ret_374; \
  46420.   float32x4_t __s0_374 = __p0_374; \
  46421.   float32x4_t __rev0_374;  __rev0_374 = __builtin_shufflevector(__s0_374, __s0_374, 3, 2, 1, 0); \
  46422.   __ret_374 = __noswap_splatq_laneq_f32(__rev0_374, __p1_374); \
  46423.   __ret_374 = __builtin_shufflevector(__ret_374, __ret_374, 3, 2, 1, 0); \
  46424.   __ret_374; \
  46425. })
  46426. #endif
  46427.  
  46428. #ifdef __LITTLE_ENDIAN__
  46429. #define vdupq_laneq_f16(__p0_375, __p1_375) __extension__ ({ \
  46430.   float16x8_t __ret_375; \
  46431.   float16x8_t __s0_375 = __p0_375; \
  46432.   __ret_375 = splatq_laneq_f16(__s0_375, __p1_375); \
  46433.   __ret_375; \
  46434. })
  46435. #else
  46436. #define vdupq_laneq_f16(__p0_376, __p1_376) __extension__ ({ \
  46437.   float16x8_t __ret_376; \
  46438.   float16x8_t __s0_376 = __p0_376; \
  46439.   float16x8_t __rev0_376;  __rev0_376 = __builtin_shufflevector(__s0_376, __s0_376, 7, 6, 5, 4, 3, 2, 1, 0); \
  46440.   __ret_376 = __noswap_splatq_laneq_f16(__rev0_376, __p1_376); \
  46441.   __ret_376 = __builtin_shufflevector(__ret_376, __ret_376, 7, 6, 5, 4, 3, 2, 1, 0); \
  46442.   __ret_376; \
  46443. })
  46444. #endif
  46445.  
  46446. #ifdef __LITTLE_ENDIAN__
  46447. #define vdupq_laneq_s32(__p0_377, __p1_377) __extension__ ({ \
  46448.   int32x4_t __ret_377; \
  46449.   int32x4_t __s0_377 = __p0_377; \
  46450.   __ret_377 = splatq_laneq_s32(__s0_377, __p1_377); \
  46451.   __ret_377; \
  46452. })
  46453. #else
  46454. #define vdupq_laneq_s32(__p0_378, __p1_378) __extension__ ({ \
  46455.   int32x4_t __ret_378; \
  46456.   int32x4_t __s0_378 = __p0_378; \
  46457.   int32x4_t __rev0_378;  __rev0_378 = __builtin_shufflevector(__s0_378, __s0_378, 3, 2, 1, 0); \
  46458.   __ret_378 = __noswap_splatq_laneq_s32(__rev0_378, __p1_378); \
  46459.   __ret_378 = __builtin_shufflevector(__ret_378, __ret_378, 3, 2, 1, 0); \
  46460.   __ret_378; \
  46461. })
  46462. #endif
  46463.  
  46464. #ifdef __LITTLE_ENDIAN__
  46465. #define vdupq_laneq_s64(__p0_379, __p1_379) __extension__ ({ \
  46466.   int64x2_t __ret_379; \
  46467.   int64x2_t __s0_379 = __p0_379; \
  46468.   __ret_379 = splatq_laneq_s64(__s0_379, __p1_379); \
  46469.   __ret_379; \
  46470. })
  46471. #else
  46472. #define vdupq_laneq_s64(__p0_380, __p1_380) __extension__ ({ \
  46473.   int64x2_t __ret_380; \
  46474.   int64x2_t __s0_380 = __p0_380; \
  46475.   int64x2_t __rev0_380;  __rev0_380 = __builtin_shufflevector(__s0_380, __s0_380, 1, 0); \
  46476.   __ret_380 = __noswap_splatq_laneq_s64(__rev0_380, __p1_380); \
  46477.   __ret_380 = __builtin_shufflevector(__ret_380, __ret_380, 1, 0); \
  46478.   __ret_380; \
  46479. })
  46480. #endif
  46481.  
  46482. #ifdef __LITTLE_ENDIAN__
  46483. #define vdupq_laneq_s16(__p0_381, __p1_381) __extension__ ({ \
  46484.   int16x8_t __ret_381; \
  46485.   int16x8_t __s0_381 = __p0_381; \
  46486.   __ret_381 = splatq_laneq_s16(__s0_381, __p1_381); \
  46487.   __ret_381; \
  46488. })
  46489. #else
  46490. #define vdupq_laneq_s16(__p0_382, __p1_382) __extension__ ({ \
  46491.   int16x8_t __ret_382; \
  46492.   int16x8_t __s0_382 = __p0_382; \
  46493.   int16x8_t __rev0_382;  __rev0_382 = __builtin_shufflevector(__s0_382, __s0_382, 7, 6, 5, 4, 3, 2, 1, 0); \
  46494.   __ret_382 = __noswap_splatq_laneq_s16(__rev0_382, __p1_382); \
  46495.   __ret_382 = __builtin_shufflevector(__ret_382, __ret_382, 7, 6, 5, 4, 3, 2, 1, 0); \
  46496.   __ret_382; \
  46497. })
  46498. #endif
  46499.  
  46500. #ifdef __LITTLE_ENDIAN__
  46501. #define vdup_laneq_u8(__p0_383, __p1_383) __extension__ ({ \
  46502.   uint8x8_t __ret_383; \
  46503.   uint8x16_t __s0_383 = __p0_383; \
  46504.   __ret_383 = splat_laneq_u8(__s0_383, __p1_383); \
  46505.   __ret_383; \
  46506. })
  46507. #else
  46508. #define vdup_laneq_u8(__p0_384, __p1_384) __extension__ ({ \
  46509.   uint8x8_t __ret_384; \
  46510.   uint8x16_t __s0_384 = __p0_384; \
  46511.   uint8x16_t __rev0_384;  __rev0_384 = __builtin_shufflevector(__s0_384, __s0_384, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46512.   __ret_384 = __noswap_splat_laneq_u8(__rev0_384, __p1_384); \
  46513.   __ret_384 = __builtin_shufflevector(__ret_384, __ret_384, 7, 6, 5, 4, 3, 2, 1, 0); \
  46514.   __ret_384; \
  46515. })
  46516. #endif
  46517.  
  46518. #ifdef __LITTLE_ENDIAN__
  46519. #define vdup_laneq_u32(__p0_385, __p1_385) __extension__ ({ \
  46520.   uint32x2_t __ret_385; \
  46521.   uint32x4_t __s0_385 = __p0_385; \
  46522.   __ret_385 = splat_laneq_u32(__s0_385, __p1_385); \
  46523.   __ret_385; \
  46524. })
  46525. #else
  46526. #define vdup_laneq_u32(__p0_386, __p1_386) __extension__ ({ \
  46527.   uint32x2_t __ret_386; \
  46528.   uint32x4_t __s0_386 = __p0_386; \
  46529.   uint32x4_t __rev0_386;  __rev0_386 = __builtin_shufflevector(__s0_386, __s0_386, 3, 2, 1, 0); \
  46530.   __ret_386 = __noswap_splat_laneq_u32(__rev0_386, __p1_386); \
  46531.   __ret_386 = __builtin_shufflevector(__ret_386, __ret_386, 1, 0); \
  46532.   __ret_386; \
  46533. })
  46534. #endif
  46535.  
  46536. #ifdef __LITTLE_ENDIAN__
  46537. #define vdup_laneq_u64(__p0_387, __p1_387) __extension__ ({ \
  46538.   uint64x1_t __ret_387; \
  46539.   uint64x2_t __s0_387 = __p0_387; \
  46540.   __ret_387 = splat_laneq_u64(__s0_387, __p1_387); \
  46541.   __ret_387; \
  46542. })
  46543. #else
  46544. #define vdup_laneq_u64(__p0_388, __p1_388) __extension__ ({ \
  46545.   uint64x1_t __ret_388; \
  46546.   uint64x2_t __s0_388 = __p0_388; \
  46547.   uint64x2_t __rev0_388;  __rev0_388 = __builtin_shufflevector(__s0_388, __s0_388, 1, 0); \
  46548.   __ret_388 = __noswap_splat_laneq_u64(__rev0_388, __p1_388); \
  46549.   __ret_388; \
  46550. })
  46551. #endif
  46552.  
  46553. #ifdef __LITTLE_ENDIAN__
  46554. #define vdup_laneq_u16(__p0_389, __p1_389) __extension__ ({ \
  46555.   uint16x4_t __ret_389; \
  46556.   uint16x8_t __s0_389 = __p0_389; \
  46557.   __ret_389 = splat_laneq_u16(__s0_389, __p1_389); \
  46558.   __ret_389; \
  46559. })
  46560. #else
  46561. #define vdup_laneq_u16(__p0_390, __p1_390) __extension__ ({ \
  46562.   uint16x4_t __ret_390; \
  46563.   uint16x8_t __s0_390 = __p0_390; \
  46564.   uint16x8_t __rev0_390;  __rev0_390 = __builtin_shufflevector(__s0_390, __s0_390, 7, 6, 5, 4, 3, 2, 1, 0); \
  46565.   __ret_390 = __noswap_splat_laneq_u16(__rev0_390, __p1_390); \
  46566.   __ret_390 = __builtin_shufflevector(__ret_390, __ret_390, 3, 2, 1, 0); \
  46567.   __ret_390; \
  46568. })
  46569. #endif
  46570.  
  46571. #ifdef __LITTLE_ENDIAN__
  46572. #define vdup_laneq_s8(__p0_391, __p1_391) __extension__ ({ \
  46573.   int8x8_t __ret_391; \
  46574.   int8x16_t __s0_391 = __p0_391; \
  46575.   __ret_391 = splat_laneq_s8(__s0_391, __p1_391); \
  46576.   __ret_391; \
  46577. })
  46578. #else
  46579. #define vdup_laneq_s8(__p0_392, __p1_392) __extension__ ({ \
  46580.   int8x8_t __ret_392; \
  46581.   int8x16_t __s0_392 = __p0_392; \
  46582.   int8x16_t __rev0_392;  __rev0_392 = __builtin_shufflevector(__s0_392, __s0_392, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  46583.   __ret_392 = __noswap_splat_laneq_s8(__rev0_392, __p1_392); \
  46584.   __ret_392 = __builtin_shufflevector(__ret_392, __ret_392, 7, 6, 5, 4, 3, 2, 1, 0); \
  46585.   __ret_392; \
  46586. })
  46587. #endif
  46588.  
  46589. #ifdef __LITTLE_ENDIAN__
  46590. #define vdup_laneq_f64(__p0_393, __p1_393) __extension__ ({ \
  46591.   float64x1_t __ret_393; \
  46592.   float64x2_t __s0_393 = __p0_393; \
  46593.   __ret_393 = splat_laneq_f64(__s0_393, __p1_393); \
  46594.   __ret_393; \
  46595. })
  46596. #else
  46597. #define vdup_laneq_f64(__p0_394, __p1_394) __extension__ ({ \
  46598.   float64x1_t __ret_394; \
  46599.   float64x2_t __s0_394 = __p0_394; \
  46600.   float64x2_t __rev0_394;  __rev0_394 = __builtin_shufflevector(__s0_394, __s0_394, 1, 0); \
  46601.   __ret_394 = __noswap_splat_laneq_f64(__rev0_394, __p1_394); \
  46602.   __ret_394; \
  46603. })
  46604. #endif
  46605.  
  46606. #ifdef __LITTLE_ENDIAN__
  46607. #define vdup_laneq_f32(__p0_395, __p1_395) __extension__ ({ \
  46608.   float32x2_t __ret_395; \
  46609.   float32x4_t __s0_395 = __p0_395; \
  46610.   __ret_395 = splat_laneq_f32(__s0_395, __p1_395); \
  46611.   __ret_395; \
  46612. })
  46613. #else
  46614. #define vdup_laneq_f32(__p0_396, __p1_396) __extension__ ({ \
  46615.   float32x2_t __ret_396; \
  46616.   float32x4_t __s0_396 = __p0_396; \
  46617.   float32x4_t __rev0_396;  __rev0_396 = __builtin_shufflevector(__s0_396, __s0_396, 3, 2, 1, 0); \
  46618.   __ret_396 = __noswap_splat_laneq_f32(__rev0_396, __p1_396); \
  46619.   __ret_396 = __builtin_shufflevector(__ret_396, __ret_396, 1, 0); \
  46620.   __ret_396; \
  46621. })
  46622. #endif
  46623.  
  46624. #ifdef __LITTLE_ENDIAN__
  46625. #define vdup_laneq_f16(__p0_397, __p1_397) __extension__ ({ \
  46626.   float16x4_t __ret_397; \
  46627.   float16x8_t __s0_397 = __p0_397; \
  46628.   __ret_397 = splat_laneq_f16(__s0_397, __p1_397); \
  46629.   __ret_397; \
  46630. })
  46631. #else
  46632. #define vdup_laneq_f16(__p0_398, __p1_398) __extension__ ({ \
  46633.   float16x4_t __ret_398; \
  46634.   float16x8_t __s0_398 = __p0_398; \
  46635.   float16x8_t __rev0_398;  __rev0_398 = __builtin_shufflevector(__s0_398, __s0_398, 7, 6, 5, 4, 3, 2, 1, 0); \
  46636.   __ret_398 = __noswap_splat_laneq_f16(__rev0_398, __p1_398); \
  46637.   __ret_398 = __builtin_shufflevector(__ret_398, __ret_398, 3, 2, 1, 0); \
  46638.   __ret_398; \
  46639. })
  46640. #endif
  46641.  
  46642. #ifdef __LITTLE_ENDIAN__
  46643. #define vdup_laneq_s32(__p0_399, __p1_399) __extension__ ({ \
  46644.   int32x2_t __ret_399; \
  46645.   int32x4_t __s0_399 = __p0_399; \
  46646.   __ret_399 = splat_laneq_s32(__s0_399, __p1_399); \
  46647.   __ret_399; \
  46648. })
  46649. #else
  46650. #define vdup_laneq_s32(__p0_400, __p1_400) __extension__ ({ \
  46651.   int32x2_t __ret_400; \
  46652.   int32x4_t __s0_400 = __p0_400; \
  46653.   int32x4_t __rev0_400;  __rev0_400 = __builtin_shufflevector(__s0_400, __s0_400, 3, 2, 1, 0); \
  46654.   __ret_400 = __noswap_splat_laneq_s32(__rev0_400, __p1_400); \
  46655.   __ret_400 = __builtin_shufflevector(__ret_400, __ret_400, 1, 0); \
  46656.   __ret_400; \
  46657. })
  46658. #endif
  46659.  
  46660. #ifdef __LITTLE_ENDIAN__
  46661. #define vdup_laneq_s64(__p0_401, __p1_401) __extension__ ({ \
  46662.   int64x1_t __ret_401; \
  46663.   int64x2_t __s0_401 = __p0_401; \
  46664.   __ret_401 = splat_laneq_s64(__s0_401, __p1_401); \
  46665.   __ret_401; \
  46666. })
  46667. #else
  46668. #define vdup_laneq_s64(__p0_402, __p1_402) __extension__ ({ \
  46669.   int64x1_t __ret_402; \
  46670.   int64x2_t __s0_402 = __p0_402; \
  46671.   int64x2_t __rev0_402;  __rev0_402 = __builtin_shufflevector(__s0_402, __s0_402, 1, 0); \
  46672.   __ret_402 = __noswap_splat_laneq_s64(__rev0_402, __p1_402); \
  46673.   __ret_402; \
  46674. })
  46675. #endif
  46676.  
  46677. #ifdef __LITTLE_ENDIAN__
  46678. #define vdup_laneq_s16(__p0_403, __p1_403) __extension__ ({ \
  46679.   int16x4_t __ret_403; \
  46680.   int16x8_t __s0_403 = __p0_403; \
  46681.   __ret_403 = splat_laneq_s16(__s0_403, __p1_403); \
  46682.   __ret_403; \
  46683. })
  46684. #else
  46685. #define vdup_laneq_s16(__p0_404, __p1_404) __extension__ ({ \
  46686.   int16x4_t __ret_404; \
  46687.   int16x8_t __s0_404 = __p0_404; \
  46688.   int16x8_t __rev0_404;  __rev0_404 = __builtin_shufflevector(__s0_404, __s0_404, 7, 6, 5, 4, 3, 2, 1, 0); \
  46689.   __ret_404 = __noswap_splat_laneq_s16(__rev0_404, __p1_404); \
  46690.   __ret_404 = __builtin_shufflevector(__ret_404, __ret_404, 3, 2, 1, 0); \
  46691.   __ret_404; \
  46692. })
  46693. #endif
  46694.  
  46695. __ai poly64x1_t vdup_n_p64(poly64_t __p0) {
  46696.   poly64x1_t __ret;
  46697.   __ret = (poly64x1_t) {__p0};
  46698.   return __ret;
  46699. }
  46700. #ifdef __LITTLE_ENDIAN__
  46701. __ai poly64x2_t vdupq_n_p64(poly64_t __p0) {
  46702.   poly64x2_t __ret;
  46703.   __ret = (poly64x2_t) {__p0, __p0};
  46704.   return __ret;
  46705. }
  46706. #else
  46707. __ai poly64x2_t vdupq_n_p64(poly64_t __p0) {
  46708.   poly64x2_t __ret;
  46709.   __ret = (poly64x2_t) {__p0, __p0};
  46710.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  46711.   return __ret;
  46712. }
  46713. #endif
  46714.  
  46715. #ifdef __LITTLE_ENDIAN__
  46716. __ai float64x2_t vdupq_n_f64(float64_t __p0) {
  46717.   float64x2_t __ret;
  46718.   __ret = (float64x2_t) {__p0, __p0};
  46719.   return __ret;
  46720. }
  46721. #else
  46722. __ai float64x2_t vdupq_n_f64(float64_t __p0) {
  46723.   float64x2_t __ret;
  46724.   __ret = (float64x2_t) {__p0, __p0};
  46725.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  46726.   return __ret;
  46727. }
  46728. #endif
  46729.  
  46730. __ai float64x1_t vdup_n_f64(float64_t __p0) {
  46731.   float64x1_t __ret;
  46732.   __ret = (float64x1_t) {__p0};
  46733.   return __ret;
  46734. }
  46735. #define vext_p64(__p0, __p1, __p2) __extension__ ({ \
  46736.   poly64x1_t __ret; \
  46737.   poly64x1_t __s0 = __p0; \
  46738.   poly64x1_t __s1 = __p1; \
  46739.   __ret = (poly64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
  46740.   __ret; \
  46741. })
  46742. #ifdef __LITTLE_ENDIAN__
  46743. #define vextq_p64(__p0, __p1, __p2) __extension__ ({ \
  46744.   poly64x2_t __ret; \
  46745.   poly64x2_t __s0 = __p0; \
  46746.   poly64x2_t __s1 = __p1; \
  46747.   __ret = (poly64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 38); \
  46748.   __ret; \
  46749. })
  46750. #else
  46751. #define vextq_p64(__p0, __p1, __p2) __extension__ ({ \
  46752.   poly64x2_t __ret; \
  46753.   poly64x2_t __s0 = __p0; \
  46754.   poly64x2_t __s1 = __p1; \
  46755.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  46756.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  46757.   __ret = (poly64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 38); \
  46758.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  46759.   __ret; \
  46760. })
  46761. #endif
  46762.  
  46763. #ifdef __LITTLE_ENDIAN__
  46764. #define vextq_f64(__p0, __p1, __p2) __extension__ ({ \
  46765.   float64x2_t __ret; \
  46766.   float64x2_t __s0 = __p0; \
  46767.   float64x2_t __s1 = __p1; \
  46768.   __ret = (float64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 42); \
  46769.   __ret; \
  46770. })
  46771. #else
  46772. #define vextq_f64(__p0, __p1, __p2) __extension__ ({ \
  46773.   float64x2_t __ret; \
  46774.   float64x2_t __s0 = __p0; \
  46775.   float64x2_t __s1 = __p1; \
  46776.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  46777.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  46778.   __ret = (float64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 42); \
  46779.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  46780.   __ret; \
  46781. })
  46782. #endif
  46783.  
  46784. #define vext_f64(__p0, __p1, __p2) __extension__ ({ \
  46785.   float64x1_t __ret; \
  46786.   float64x1_t __s0 = __p0; \
  46787.   float64x1_t __s1 = __p1; \
  46788.   __ret = (float64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
  46789.   __ret; \
  46790. })
  46791. #ifdef __LITTLE_ENDIAN__
  46792. __ai float64x2_t vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  46793.   float64x2_t __ret;
  46794.   __ret = (float64x2_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  46795.   return __ret;
  46796. }
  46797. #else
  46798. __ai float64x2_t vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  46799.   float64x2_t __ret;
  46800.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  46801.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  46802.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  46803.   __ret = (float64x2_t) __builtin_neon_vfmaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
  46804.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  46805.   return __ret;
  46806. }
  46807. __ai float64x2_t __noswap_vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  46808.   float64x2_t __ret;
  46809.   __ret = (float64x2_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  46810.   return __ret;
  46811. }
  46812. #endif
  46813.  
  46814. __ai float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
  46815.   float64x1_t __ret;
  46816.   __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
  46817.   return __ret;
  46818. }
  46819. #define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  46820.   float64_t __ret; \
  46821.   float64_t __s0 = __p0; \
  46822.   float64_t __s1 = __p1; \
  46823.   float64x1_t __s2 = __p2; \
  46824.   __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (float64x1_t)__s2, __p3); \
  46825.   __ret; \
  46826. })
  46827. #ifdef __LITTLE_ENDIAN__
  46828. #define vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46829.   float32_t __ret; \
  46830.   float32_t __s0 = __p0; \
  46831.   float32_t __s1 = __p1; \
  46832.   float32x2_t __s2 = __p2; \
  46833.   __ret = (float32_t) __builtin_neon_vfmas_lane_f32(__s0, __s1, (float32x2_t)__s2, __p3); \
  46834.   __ret; \
  46835. })
  46836. #else
  46837. #define vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46838.   float32_t __ret; \
  46839.   float32_t __s0 = __p0; \
  46840.   float32_t __s1 = __p1; \
  46841.   float32x2_t __s2 = __p2; \
  46842.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
  46843.   __ret = (float32_t) __builtin_neon_vfmas_lane_f32(__s0, __s1, (float32x2_t)__rev2, __p3); \
  46844.   __ret; \
  46845. })
  46846. #define __noswap_vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46847.   float32_t __ret; \
  46848.   float32_t __s0 = __p0; \
  46849.   float32_t __s1 = __p1; \
  46850.   float32x2_t __s2 = __p2; \
  46851.   __ret = (float32_t) __builtin_neon_vfmas_lane_f32(__s0, __s1, (float32x2_t)__s2, __p3); \
  46852.   __ret; \
  46853. })
  46854. #endif
  46855.  
  46856. #ifdef __LITTLE_ENDIAN__
  46857. #define vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  46858.   float64x2_t __ret; \
  46859.   float64x2_t __s0 = __p0; \
  46860.   float64x2_t __s1 = __p1; \
  46861.   float64x1_t __s2 = __p2; \
  46862.   __ret = (float64x2_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 42); \
  46863.   __ret; \
  46864. })
  46865. #else
  46866. #define vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  46867.   float64x2_t __ret; \
  46868.   float64x2_t __s0 = __p0; \
  46869.   float64x2_t __s1 = __p1; \
  46870.   float64x1_t __s2 = __p2; \
  46871.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  46872.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  46873.   __ret = (float64x2_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__s2, __p3, 42); \
  46874.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  46875.   __ret; \
  46876. })
  46877. #define __noswap_vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  46878.   float64x2_t __ret; \
  46879.   float64x2_t __s0 = __p0; \
  46880.   float64x2_t __s1 = __p1; \
  46881.   float64x1_t __s2 = __p2; \
  46882.   __ret = (float64x2_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 42); \
  46883.   __ret; \
  46884. })
  46885. #endif
  46886.  
  46887. #ifdef __LITTLE_ENDIAN__
  46888. #define vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46889.   float32x4_t __ret; \
  46890.   float32x4_t __s0 = __p0; \
  46891.   float32x4_t __s1 = __p1; \
  46892.   float32x2_t __s2 = __p2; \
  46893.   __ret = (float32x4_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 41); \
  46894.   __ret; \
  46895. })
  46896. #else
  46897. #define vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46898.   float32x4_t __ret; \
  46899.   float32x4_t __s0 = __p0; \
  46900.   float32x4_t __s1 = __p1; \
  46901.   float32x2_t __s2 = __p2; \
  46902.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  46903.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  46904.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
  46905.   __ret = (float32x4_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, __p3, 41); \
  46906.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  46907.   __ret; \
  46908. })
  46909. #define __noswap_vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46910.   float32x4_t __ret; \
  46911.   float32x4_t __s0 = __p0; \
  46912.   float32x4_t __s1 = __p1; \
  46913.   float32x2_t __s2 = __p2; \
  46914.   __ret = (float32x4_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 41); \
  46915.   __ret; \
  46916. })
  46917. #endif
  46918.  
  46919. #define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  46920.   float64x1_t __ret; \
  46921.   float64x1_t __s0 = __p0; \
  46922.   float64x1_t __s1 = __p1; \
  46923.   float64x1_t __s2 = __p2; \
  46924.   __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
  46925.   __ret; \
  46926. })
  46927. #ifdef __LITTLE_ENDIAN__
  46928. #define vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46929.   float32x2_t __ret; \
  46930.   float32x2_t __s0 = __p0; \
  46931.   float32x2_t __s1 = __p1; \
  46932.   float32x2_t __s2 = __p2; \
  46933.   __ret = (float32x2_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 9); \
  46934.   __ret; \
  46935. })
  46936. #else
  46937. #define vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46938.   float32x2_t __ret; \
  46939.   float32x2_t __s0 = __p0; \
  46940.   float32x2_t __s1 = __p1; \
  46941.   float32x2_t __s2 = __p2; \
  46942.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  46943.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  46944.   float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
  46945.   __ret = (float32x2_t) __builtin_neon_vfma_lane_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, __p3, 9); \
  46946.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  46947.   __ret; \
  46948. })
  46949. #define __noswap_vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46950.   float32x2_t __ret; \
  46951.   float32x2_t __s0 = __p0; \
  46952.   float32x2_t __s1 = __p1; \
  46953.   float32x2_t __s2 = __p2; \
  46954.   __ret = (float32x2_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 9); \
  46955.   __ret; \
  46956. })
  46957. #endif
  46958.  
  46959. #ifdef __LITTLE_ENDIAN__
  46960. #define vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  46961.   float64_t __ret; \
  46962.   float64_t __s0 = __p0; \
  46963.   float64_t __s1 = __p1; \
  46964.   float64x2_t __s2 = __p2; \
  46965.   __ret = (float64_t) __builtin_neon_vfmad_laneq_f64(__s0, __s1, (float64x2_t)__s2, __p3); \
  46966.   __ret; \
  46967. })
  46968. #else
  46969. #define vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  46970.   float64_t __ret; \
  46971.   float64_t __s0 = __p0; \
  46972.   float64_t __s1 = __p1; \
  46973.   float64x2_t __s2 = __p2; \
  46974.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
  46975.   __ret = (float64_t) __builtin_neon_vfmad_laneq_f64(__s0, __s1, (float64x2_t)__rev2, __p3); \
  46976.   __ret; \
  46977. })
  46978. #define __noswap_vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  46979.   float64_t __ret; \
  46980.   float64_t __s0 = __p0; \
  46981.   float64_t __s1 = __p1; \
  46982.   float64x2_t __s2 = __p2; \
  46983.   __ret = (float64_t) __builtin_neon_vfmad_laneq_f64(__s0, __s1, (float64x2_t)__s2, __p3); \
  46984.   __ret; \
  46985. })
  46986. #endif
  46987.  
  46988. #ifdef __LITTLE_ENDIAN__
  46989. #define vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46990.   float32_t __ret; \
  46991.   float32_t __s0 = __p0; \
  46992.   float32_t __s1 = __p1; \
  46993.   float32x4_t __s2 = __p2; \
  46994.   __ret = (float32_t) __builtin_neon_vfmas_laneq_f32(__s0, __s1, (float32x4_t)__s2, __p3); \
  46995.   __ret; \
  46996. })
  46997. #else
  46998. #define vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  46999.   float32_t __ret; \
  47000.   float32_t __s0 = __p0; \
  47001.   float32_t __s1 = __p1; \
  47002.   float32x4_t __s2 = __p2; \
  47003.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  47004.   __ret = (float32_t) __builtin_neon_vfmas_laneq_f32(__s0, __s1, (float32x4_t)__rev2, __p3); \
  47005.   __ret; \
  47006. })
  47007. #define __noswap_vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  47008.   float32_t __ret; \
  47009.   float32_t __s0 = __p0; \
  47010.   float32_t __s1 = __p1; \
  47011.   float32x4_t __s2 = __p2; \
  47012.   __ret = (float32_t) __builtin_neon_vfmas_laneq_f32(__s0, __s1, (float32x4_t)__s2, __p3); \
  47013.   __ret; \
  47014. })
  47015. #endif
  47016.  
  47017. #ifdef __LITTLE_ENDIAN__
  47018. #define vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  47019.   float64x2_t __ret; \
  47020.   float64x2_t __s0 = __p0; \
  47021.   float64x2_t __s1 = __p1; \
  47022.   float64x2_t __s2 = __p2; \
  47023.   __ret = (float64x2_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 42); \
  47024.   __ret; \
  47025. })
  47026. #else
  47027. #define vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  47028.   float64x2_t __ret; \
  47029.   float64x2_t __s0 = __p0; \
  47030.   float64x2_t __s1 = __p1; \
  47031.   float64x2_t __s2 = __p2; \
  47032.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  47033.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  47034.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
  47035.   __ret = (float64x2_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 42); \
  47036.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  47037.   __ret; \
  47038. })
  47039. #define __noswap_vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  47040.   float64x2_t __ret; \
  47041.   float64x2_t __s0 = __p0; \
  47042.   float64x2_t __s1 = __p1; \
  47043.   float64x2_t __s2 = __p2; \
  47044.   __ret = (float64x2_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 42); \
  47045.   __ret; \
  47046. })
  47047. #endif
  47048.  
  47049. #ifdef __LITTLE_ENDIAN__
  47050. #define vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  47051.   float32x4_t __ret; \
  47052.   float32x4_t __s0 = __p0; \
  47053.   float32x4_t __s1 = __p1; \
  47054.   float32x4_t __s2 = __p2; \
  47055.   __ret = (float32x4_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 41); \
  47056.   __ret; \
  47057. })
  47058. #else
  47059. #define vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  47060.   float32x4_t __ret; \
  47061.   float32x4_t __s0 = __p0; \
  47062.   float32x4_t __s1 = __p1; \
  47063.   float32x4_t __s2 = __p2; \
  47064.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  47065.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  47066.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  47067.   __ret = (float32x4_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 41); \
  47068.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  47069.   __ret; \
  47070. })
  47071. #define __noswap_vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  47072.   float32x4_t __ret; \
  47073.   float32x4_t __s0 = __p0; \
  47074.   float32x4_t __s1 = __p1; \
  47075.   float32x4_t __s2 = __p2; \
  47076.   __ret = (float32x4_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 41); \
  47077.   __ret; \
  47078. })
  47079. #endif
  47080.  
  47081. #ifdef __LITTLE_ENDIAN__
  47082. #define vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  47083.   float64x1_t __ret; \
  47084.   float64x1_t __s0 = __p0; \
  47085.   float64x1_t __s1 = __p1; \
  47086.   float64x2_t __s2 = __p2; \
  47087.   __ret = (float64x1_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 10); \
  47088.   __ret; \
  47089. })
  47090. #else
  47091. #define vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  47092.   float64x1_t __ret; \
  47093.   float64x1_t __s0 = __p0; \
  47094.   float64x1_t __s1 = __p1; \
  47095.   float64x2_t __s2 = __p2; \
  47096.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
  47097.   __ret = (float64x1_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__rev2, __p3, 10); \
  47098.   __ret; \
  47099. })
  47100. #define __noswap_vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
  47101.   float64x1_t __ret; \
  47102.   float64x1_t __s0 = __p0; \
  47103.   float64x1_t __s1 = __p1; \
  47104.   float64x2_t __s2 = __p2; \
  47105.   __ret = (float64x1_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 10); \
  47106.   __ret; \
  47107. })
  47108. #endif
  47109.  
  47110. #ifdef __LITTLE_ENDIAN__
  47111. #define vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  47112.   float32x2_t __ret; \
  47113.   float32x2_t __s0 = __p0; \
  47114.   float32x2_t __s1 = __p1; \
  47115.   float32x4_t __s2 = __p2; \
  47116.   __ret = (float32x2_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 9); \
  47117.   __ret; \
  47118. })
  47119. #else
  47120. #define vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  47121.   float32x2_t __ret; \
  47122.   float32x2_t __s0 = __p0; \
  47123.   float32x2_t __s1 = __p1; \
  47124.   float32x4_t __s2 = __p2; \
  47125.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  47126.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  47127.   float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  47128.   __ret = (float32x2_t) __builtin_neon_vfma_laneq_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x16_t)__rev2, __p3, 9); \
  47129.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  47130.   __ret; \
  47131. })
  47132. #define __noswap_vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
  47133.   float32x2_t __ret; \
  47134.   float32x2_t __s0 = __p0; \
  47135.   float32x2_t __s1 = __p1; \
  47136.   float32x4_t __s2 = __p2; \
  47137.   __ret = (float32x2_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 9); \
  47138.   __ret; \
  47139. })
  47140. #endif
  47141.  
  47142. #ifdef __LITTLE_ENDIAN__
  47143. __ai float64x2_t vfmaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
  47144.   float64x2_t __ret;
  47145.   __ret = vfmaq_f64(__p0, __p1, (float64x2_t) {__p2, __p2});
  47146.   return __ret;
  47147. }
  47148. #else
  47149. __ai float64x2_t vfmaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
  47150.   float64x2_t __ret;
  47151.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  47152.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  47153.   __ret = __noswap_vfmaq_f64(__rev0, __rev1, (float64x2_t) {__p2, __p2});
  47154.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  47155.   return __ret;
  47156. }
  47157. #endif
  47158.  
  47159. __ai float64x1_t vfma_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
  47160.   float64x1_t __ret;
  47161.   __ret = vfma_f64(__p0, __p1, (float64x1_t) {__p2});
  47162.   return __ret;
  47163. }
  47164. #ifdef __LITTLE_ENDIAN__
  47165. __ai float64x2_t vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  47166.   float64x2_t __ret;
  47167.   __ret = vfmaq_f64(__p0, -__p1, __p2);
  47168.   return __ret;
  47169. }
  47170. #else
  47171. __ai float64x2_t vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  47172.   float64x2_t __ret;
  47173.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  47174.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  47175.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  47176.   __ret = __noswap_vfmaq_f64(__rev0, -__rev1, __rev2);
  47177.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  47178.   return __ret;
  47179. }
  47180. #endif
  47181.  
  47182. __ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
  47183.   float64x1_t __ret;
  47184.   __ret = vfma_f64(__p0, -__p1, __p2);
  47185.   return __ret;
  47186. }
  47187. #define vfmsd_lane_f64(__p0_405, __p1_405, __p2_405, __p3_405) __extension__ ({ \
  47188.   float64_t __ret_405; \
  47189.   float64_t __s0_405 = __p0_405; \
  47190.   float64_t __s1_405 = __p1_405; \
  47191.   float64x1_t __s2_405 = __p2_405; \
  47192.   __ret_405 = vfmad_lane_f64(__s0_405, -__s1_405, __s2_405, __p3_405); \
  47193.   __ret_405; \
  47194. })
  47195. #ifdef __LITTLE_ENDIAN__
  47196. #define vfmss_lane_f32(__p0_406, __p1_406, __p2_406, __p3_406) __extension__ ({ \
  47197.   float32_t __ret_406; \
  47198.   float32_t __s0_406 = __p0_406; \
  47199.   float32_t __s1_406 = __p1_406; \
  47200.   float32x2_t __s2_406 = __p2_406; \
  47201.   __ret_406 = vfmas_lane_f32(__s0_406, -__s1_406, __s2_406, __p3_406); \
  47202.   __ret_406; \
  47203. })
  47204. #else
  47205. #define vfmss_lane_f32(__p0_407, __p1_407, __p2_407, __p3_407) __extension__ ({ \
  47206.   float32_t __ret_407; \
  47207.   float32_t __s0_407 = __p0_407; \
  47208.   float32_t __s1_407 = __p1_407; \
  47209.   float32x2_t __s2_407 = __p2_407; \
  47210.   float32x2_t __rev2_407;  __rev2_407 = __builtin_shufflevector(__s2_407, __s2_407, 1, 0); \
  47211.   __ret_407 = __noswap_vfmas_lane_f32(__s0_407, -__s1_407, __rev2_407, __p3_407); \
  47212.   __ret_407; \
  47213. })
  47214. #endif
  47215.  
  47216. #ifdef __LITTLE_ENDIAN__
  47217. #define vfmsq_lane_f64(__p0_408, __p1_408, __p2_408, __p3_408) __extension__ ({ \
  47218.   float64x2_t __ret_408; \
  47219.   float64x2_t __s0_408 = __p0_408; \
  47220.   float64x2_t __s1_408 = __p1_408; \
  47221.   float64x1_t __s2_408 = __p2_408; \
  47222.   __ret_408 = vfmaq_lane_f64(__s0_408, -__s1_408, __s2_408, __p3_408); \
  47223.   __ret_408; \
  47224. })
  47225. #else
  47226. #define vfmsq_lane_f64(__p0_409, __p1_409, __p2_409, __p3_409) __extension__ ({ \
  47227.   float64x2_t __ret_409; \
  47228.   float64x2_t __s0_409 = __p0_409; \
  47229.   float64x2_t __s1_409 = __p1_409; \
  47230.   float64x1_t __s2_409 = __p2_409; \
  47231.   float64x2_t __rev0_409;  __rev0_409 = __builtin_shufflevector(__s0_409, __s0_409, 1, 0); \
  47232.   float64x2_t __rev1_409;  __rev1_409 = __builtin_shufflevector(__s1_409, __s1_409, 1, 0); \
  47233.   __ret_409 = __noswap_vfmaq_lane_f64(__rev0_409, -__rev1_409, __s2_409, __p3_409); \
  47234.   __ret_409 = __builtin_shufflevector(__ret_409, __ret_409, 1, 0); \
  47235.   __ret_409; \
  47236. })
  47237. #endif
  47238.  
  47239. #ifdef __LITTLE_ENDIAN__
  47240. #define vfmsq_lane_f32(__p0_410, __p1_410, __p2_410, __p3_410) __extension__ ({ \
  47241.   float32x4_t __ret_410; \
  47242.   float32x4_t __s0_410 = __p0_410; \
  47243.   float32x4_t __s1_410 = __p1_410; \
  47244.   float32x2_t __s2_410 = __p2_410; \
  47245.   __ret_410 = vfmaq_lane_f32(__s0_410, -__s1_410, __s2_410, __p3_410); \
  47246.   __ret_410; \
  47247. })
  47248. #else
  47249. #define vfmsq_lane_f32(__p0_411, __p1_411, __p2_411, __p3_411) __extension__ ({ \
  47250.   float32x4_t __ret_411; \
  47251.   float32x4_t __s0_411 = __p0_411; \
  47252.   float32x4_t __s1_411 = __p1_411; \
  47253.   float32x2_t __s2_411 = __p2_411; \
  47254.   float32x4_t __rev0_411;  __rev0_411 = __builtin_shufflevector(__s0_411, __s0_411, 3, 2, 1, 0); \
  47255.   float32x4_t __rev1_411;  __rev1_411 = __builtin_shufflevector(__s1_411, __s1_411, 3, 2, 1, 0); \
  47256.   float32x2_t __rev2_411;  __rev2_411 = __builtin_shufflevector(__s2_411, __s2_411, 1, 0); \
  47257.   __ret_411 = __noswap_vfmaq_lane_f32(__rev0_411, -__rev1_411, __rev2_411, __p3_411); \
  47258.   __ret_411 = __builtin_shufflevector(__ret_411, __ret_411, 3, 2, 1, 0); \
  47259.   __ret_411; \
  47260. })
  47261. #endif
  47262.  
  47263. #define vfms_lane_f64(__p0_412, __p1_412, __p2_412, __p3_412) __extension__ ({ \
  47264.   float64x1_t __ret_412; \
  47265.   float64x1_t __s0_412 = __p0_412; \
  47266.   float64x1_t __s1_412 = __p1_412; \
  47267.   float64x1_t __s2_412 = __p2_412; \
  47268.   __ret_412 = vfma_lane_f64(__s0_412, -__s1_412, __s2_412, __p3_412); \
  47269.   __ret_412; \
  47270. })
  47271. #ifdef __LITTLE_ENDIAN__
  47272. #define vfms_lane_f32(__p0_413, __p1_413, __p2_413, __p3_413) __extension__ ({ \
  47273.   float32x2_t __ret_413; \
  47274.   float32x2_t __s0_413 = __p0_413; \
  47275.   float32x2_t __s1_413 = __p1_413; \
  47276.   float32x2_t __s2_413 = __p2_413; \
  47277.   __ret_413 = vfma_lane_f32(__s0_413, -__s1_413, __s2_413, __p3_413); \
  47278.   __ret_413; \
  47279. })
  47280. #else
  47281. #define vfms_lane_f32(__p0_414, __p1_414, __p2_414, __p3_414) __extension__ ({ \
  47282.   float32x2_t __ret_414; \
  47283.   float32x2_t __s0_414 = __p0_414; \
  47284.   float32x2_t __s1_414 = __p1_414; \
  47285.   float32x2_t __s2_414 = __p2_414; \
  47286.   float32x2_t __rev0_414;  __rev0_414 = __builtin_shufflevector(__s0_414, __s0_414, 1, 0); \
  47287.   float32x2_t __rev1_414;  __rev1_414 = __builtin_shufflevector(__s1_414, __s1_414, 1, 0); \
  47288.   float32x2_t __rev2_414;  __rev2_414 = __builtin_shufflevector(__s2_414, __s2_414, 1, 0); \
  47289.   __ret_414 = __noswap_vfma_lane_f32(__rev0_414, -__rev1_414, __rev2_414, __p3_414); \
  47290.   __ret_414 = __builtin_shufflevector(__ret_414, __ret_414, 1, 0); \
  47291.   __ret_414; \
  47292. })
  47293. #endif
  47294.  
  47295. #ifdef __LITTLE_ENDIAN__
  47296. #define vfmsd_laneq_f64(__p0_415, __p1_415, __p2_415, __p3_415) __extension__ ({ \
  47297.   float64_t __ret_415; \
  47298.   float64_t __s0_415 = __p0_415; \
  47299.   float64_t __s1_415 = __p1_415; \
  47300.   float64x2_t __s2_415 = __p2_415; \
  47301.   __ret_415 = vfmad_laneq_f64(__s0_415, -__s1_415, __s2_415, __p3_415); \
  47302.   __ret_415; \
  47303. })
  47304. #else
  47305. #define vfmsd_laneq_f64(__p0_416, __p1_416, __p2_416, __p3_416) __extension__ ({ \
  47306.   float64_t __ret_416; \
  47307.   float64_t __s0_416 = __p0_416; \
  47308.   float64_t __s1_416 = __p1_416; \
  47309.   float64x2_t __s2_416 = __p2_416; \
  47310.   float64x2_t __rev2_416;  __rev2_416 = __builtin_shufflevector(__s2_416, __s2_416, 1, 0); \
  47311.   __ret_416 = __noswap_vfmad_laneq_f64(__s0_416, -__s1_416, __rev2_416, __p3_416); \
  47312.   __ret_416; \
  47313. })
  47314. #endif
  47315.  
  47316. #ifdef __LITTLE_ENDIAN__
  47317. #define vfmss_laneq_f32(__p0_417, __p1_417, __p2_417, __p3_417) __extension__ ({ \
  47318.   float32_t __ret_417; \
  47319.   float32_t __s0_417 = __p0_417; \
  47320.   float32_t __s1_417 = __p1_417; \
  47321.   float32x4_t __s2_417 = __p2_417; \
  47322.   __ret_417 = vfmas_laneq_f32(__s0_417, -__s1_417, __s2_417, __p3_417); \
  47323.   __ret_417; \
  47324. })
  47325. #else
  47326. #define vfmss_laneq_f32(__p0_418, __p1_418, __p2_418, __p3_418) __extension__ ({ \
  47327.   float32_t __ret_418; \
  47328.   float32_t __s0_418 = __p0_418; \
  47329.   float32_t __s1_418 = __p1_418; \
  47330.   float32x4_t __s2_418 = __p2_418; \
  47331.   float32x4_t __rev2_418;  __rev2_418 = __builtin_shufflevector(__s2_418, __s2_418, 3, 2, 1, 0); \
  47332.   __ret_418 = __noswap_vfmas_laneq_f32(__s0_418, -__s1_418, __rev2_418, __p3_418); \
  47333.   __ret_418; \
  47334. })
  47335. #endif
  47336.  
  47337. #ifdef __LITTLE_ENDIAN__
  47338. #define vfmsq_laneq_f64(__p0_419, __p1_419, __p2_419, __p3_419) __extension__ ({ \
  47339.   float64x2_t __ret_419; \
  47340.   float64x2_t __s0_419 = __p0_419; \
  47341.   float64x2_t __s1_419 = __p1_419; \
  47342.   float64x2_t __s2_419 = __p2_419; \
  47343.   __ret_419 = vfmaq_laneq_f64(__s0_419, -__s1_419, __s2_419, __p3_419); \
  47344.   __ret_419; \
  47345. })
  47346. #else
  47347. #define vfmsq_laneq_f64(__p0_420, __p1_420, __p2_420, __p3_420) __extension__ ({ \
  47348.   float64x2_t __ret_420; \
  47349.   float64x2_t __s0_420 = __p0_420; \
  47350.   float64x2_t __s1_420 = __p1_420; \
  47351.   float64x2_t __s2_420 = __p2_420; \
  47352.   float64x2_t __rev0_420;  __rev0_420 = __builtin_shufflevector(__s0_420, __s0_420, 1, 0); \
  47353.   float64x2_t __rev1_420;  __rev1_420 = __builtin_shufflevector(__s1_420, __s1_420, 1, 0); \
  47354.   float64x2_t __rev2_420;  __rev2_420 = __builtin_shufflevector(__s2_420, __s2_420, 1, 0); \
  47355.   __ret_420 = __noswap_vfmaq_laneq_f64(__rev0_420, -__rev1_420, __rev2_420, __p3_420); \
  47356.   __ret_420 = __builtin_shufflevector(__ret_420, __ret_420, 1, 0); \
  47357.   __ret_420; \
  47358. })
  47359. #endif
  47360.  
  47361. #ifdef __LITTLE_ENDIAN__
  47362. #define vfmsq_laneq_f32(__p0_421, __p1_421, __p2_421, __p3_421) __extension__ ({ \
  47363.   float32x4_t __ret_421; \
  47364.   float32x4_t __s0_421 = __p0_421; \
  47365.   float32x4_t __s1_421 = __p1_421; \
  47366.   float32x4_t __s2_421 = __p2_421; \
  47367.   __ret_421 = vfmaq_laneq_f32(__s0_421, -__s1_421, __s2_421, __p3_421); \
  47368.   __ret_421; \
  47369. })
  47370. #else
  47371. #define vfmsq_laneq_f32(__p0_422, __p1_422, __p2_422, __p3_422) __extension__ ({ \
  47372.   float32x4_t __ret_422; \
  47373.   float32x4_t __s0_422 = __p0_422; \
  47374.   float32x4_t __s1_422 = __p1_422; \
  47375.   float32x4_t __s2_422 = __p2_422; \
  47376.   float32x4_t __rev0_422;  __rev0_422 = __builtin_shufflevector(__s0_422, __s0_422, 3, 2, 1, 0); \
  47377.   float32x4_t __rev1_422;  __rev1_422 = __builtin_shufflevector(__s1_422, __s1_422, 3, 2, 1, 0); \
  47378.   float32x4_t __rev2_422;  __rev2_422 = __builtin_shufflevector(__s2_422, __s2_422, 3, 2, 1, 0); \
  47379.   __ret_422 = __noswap_vfmaq_laneq_f32(__rev0_422, -__rev1_422, __rev2_422, __p3_422); \
  47380.   __ret_422 = __builtin_shufflevector(__ret_422, __ret_422, 3, 2, 1, 0); \
  47381.   __ret_422; \
  47382. })
  47383. #endif
  47384.  
  47385. #ifdef __LITTLE_ENDIAN__
  47386. #define vfms_laneq_f64(__p0_423, __p1_423, __p2_423, __p3_423) __extension__ ({ \
  47387.   float64x1_t __ret_423; \
  47388.   float64x1_t __s0_423 = __p0_423; \
  47389.   float64x1_t __s1_423 = __p1_423; \
  47390.   float64x2_t __s2_423 = __p2_423; \
  47391.   __ret_423 = vfma_laneq_f64(__s0_423, -__s1_423, __s2_423, __p3_423); \
  47392.   __ret_423; \
  47393. })
  47394. #else
  47395. #define vfms_laneq_f64(__p0_424, __p1_424, __p2_424, __p3_424) __extension__ ({ \
  47396.   float64x1_t __ret_424; \
  47397.   float64x1_t __s0_424 = __p0_424; \
  47398.   float64x1_t __s1_424 = __p1_424; \
  47399.   float64x2_t __s2_424 = __p2_424; \
  47400.   float64x2_t __rev2_424;  __rev2_424 = __builtin_shufflevector(__s2_424, __s2_424, 1, 0); \
  47401.   __ret_424 = __noswap_vfma_laneq_f64(__s0_424, -__s1_424, __rev2_424, __p3_424); \
  47402.   __ret_424; \
  47403. })
  47404. #endif
  47405.  
  47406. #ifdef __LITTLE_ENDIAN__
  47407. #define vfms_laneq_f32(__p0_425, __p1_425, __p2_425, __p3_425) __extension__ ({ \
  47408.   float32x2_t __ret_425; \
  47409.   float32x2_t __s0_425 = __p0_425; \
  47410.   float32x2_t __s1_425 = __p1_425; \
  47411.   float32x4_t __s2_425 = __p2_425; \
  47412.   __ret_425 = vfma_laneq_f32(__s0_425, -__s1_425, __s2_425, __p3_425); \
  47413.   __ret_425; \
  47414. })
  47415. #else
  47416. #define vfms_laneq_f32(__p0_426, __p1_426, __p2_426, __p3_426) __extension__ ({ \
  47417.   float32x2_t __ret_426; \
  47418.   float32x2_t __s0_426 = __p0_426; \
  47419.   float32x2_t __s1_426 = __p1_426; \
  47420.   float32x4_t __s2_426 = __p2_426; \
  47421.   float32x2_t __rev0_426;  __rev0_426 = __builtin_shufflevector(__s0_426, __s0_426, 1, 0); \
  47422.   float32x2_t __rev1_426;  __rev1_426 = __builtin_shufflevector(__s1_426, __s1_426, 1, 0); \
  47423.   float32x4_t __rev2_426;  __rev2_426 = __builtin_shufflevector(__s2_426, __s2_426, 3, 2, 1, 0); \
  47424.   __ret_426 = __noswap_vfma_laneq_f32(__rev0_426, -__rev1_426, __rev2_426, __p3_426); \
  47425.   __ret_426 = __builtin_shufflevector(__ret_426, __ret_426, 1, 0); \
  47426.   __ret_426; \
  47427. })
  47428. #endif
  47429.  
  47430. #ifdef __LITTLE_ENDIAN__
  47431. __ai float64x2_t vfmsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
  47432.   float64x2_t __ret;
  47433.   __ret = vfmaq_f64(__p0, -__p1, (float64x2_t) {__p2, __p2});
  47434.   return __ret;
  47435. }
  47436. #else
  47437. __ai float64x2_t vfmsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
  47438.   float64x2_t __ret;
  47439.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  47440.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  47441.   __ret = __noswap_vfmaq_f64(__rev0, -__rev1, (float64x2_t) {__p2, __p2});
  47442.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  47443.   return __ret;
  47444. }
  47445. #endif
  47446.  
  47447. #ifdef __LITTLE_ENDIAN__
  47448. __ai float32x4_t vfmsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
  47449.   float32x4_t __ret;
  47450.   __ret = vfmaq_f32(__p0, -__p1, (float32x4_t) {__p2, __p2, __p2, __p2});
  47451.   return __ret;
  47452. }
  47453. #else
  47454. __ai float32x4_t vfmsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
  47455.   float32x4_t __ret;
  47456.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  47457.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  47458.   __ret = __noswap_vfmaq_f32(__rev0, -__rev1, (float32x4_t) {__p2, __p2, __p2, __p2});
  47459.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  47460.   return __ret;
  47461. }
  47462. #endif
  47463.  
  47464. __ai float64x1_t vfms_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
  47465.   float64x1_t __ret;
  47466.   __ret = vfma_f64(__p0, -__p1, (float64x1_t) {__p2});
  47467.   return __ret;
  47468. }
  47469. #ifdef __LITTLE_ENDIAN__
  47470. __ai float32x2_t vfms_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
  47471.   float32x2_t __ret;
  47472.   __ret = vfma_f32(__p0, -__p1, (float32x2_t) {__p2, __p2});
  47473.   return __ret;
  47474. }
  47475. #else
  47476. __ai float32x2_t vfms_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
  47477.   float32x2_t __ret;
  47478.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  47479.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  47480.   __ret = __noswap_vfma_f32(__rev0, -__rev1, (float32x2_t) {__p2, __p2});
  47481.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  47482.   return __ret;
  47483. }
  47484. #endif
  47485.  
  47486. #ifdef __LITTLE_ENDIAN__
  47487. __ai poly64x1_t vget_high_p64(poly64x2_t __p0) {
  47488.   poly64x1_t __ret;
  47489.   __ret = __builtin_shufflevector(__p0, __p0, 1);
  47490.   return __ret;
  47491. }
  47492. #else
  47493. __ai poly64x1_t vget_high_p64(poly64x2_t __p0) {
  47494.   poly64x1_t __ret;
  47495.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  47496.   __ret = __builtin_shufflevector(__rev0, __rev0, 1);
  47497.   return __ret;
  47498. }
  47499. __ai poly64x1_t __noswap_vget_high_p64(poly64x2_t __p0) {
  47500.   poly64x1_t __ret;
  47501.   __ret = __builtin_shufflevector(__p0, __p0, 1);
  47502.   return __ret;
  47503. }
  47504. #endif
  47505.  
  47506. #ifdef __LITTLE_ENDIAN__
  47507. __ai float64x1_t vget_high_f64(float64x2_t __p0) {
  47508.   float64x1_t __ret;
  47509.   __ret = __builtin_shufflevector(__p0, __p0, 1);
  47510.   return __ret;
  47511. }
  47512. #else
  47513. __ai float64x1_t vget_high_f64(float64x2_t __p0) {
  47514.   float64x1_t __ret;
  47515.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  47516.   __ret = __builtin_shufflevector(__rev0, __rev0, 1);
  47517.   return __ret;
  47518. }
  47519. #endif
  47520.  
  47521. #define vget_lane_p64(__p0, __p1) __extension__ ({ \
  47522.   poly64_t __ret; \
  47523.   poly64x1_t __s0 = __p0; \
  47524.   __ret = (poly64_t) __builtin_neon_vget_lane_i64((poly64x1_t)__s0, __p1); \
  47525.   __ret; \
  47526. })
  47527. #ifdef __LITTLE_ENDIAN__
  47528. #define vgetq_lane_p64(__p0, __p1) __extension__ ({ \
  47529.   poly64_t __ret; \
  47530.   poly64x2_t __s0 = __p0; \
  47531.   __ret = (poly64_t) __builtin_neon_vgetq_lane_i64((poly64x2_t)__s0, __p1); \
  47532.   __ret; \
  47533. })
  47534. #else
  47535. #define vgetq_lane_p64(__p0, __p1) __extension__ ({ \
  47536.   poly64_t __ret; \
  47537.   poly64x2_t __s0 = __p0; \
  47538.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  47539.   __ret = (poly64_t) __builtin_neon_vgetq_lane_i64((poly64x2_t)__rev0, __p1); \
  47540.   __ret; \
  47541. })
  47542. #define __noswap_vgetq_lane_p64(__p0, __p1) __extension__ ({ \
  47543.   poly64_t __ret; \
  47544.   poly64x2_t __s0 = __p0; \
  47545.   __ret = (poly64_t) __builtin_neon_vgetq_lane_i64((poly64x2_t)__s0, __p1); \
  47546.   __ret; \
  47547. })
  47548. #endif
  47549.  
  47550. #ifdef __LITTLE_ENDIAN__
  47551. #define vgetq_lane_f64(__p0, __p1) __extension__ ({ \
  47552.   float64_t __ret; \
  47553.   float64x2_t __s0 = __p0; \
  47554.   __ret = (float64_t) __builtin_neon_vgetq_lane_f64((float64x2_t)__s0, __p1); \
  47555.   __ret; \
  47556. })
  47557. #else
  47558. #define vgetq_lane_f64(__p0, __p1) __extension__ ({ \
  47559.   float64_t __ret; \
  47560.   float64x2_t __s0 = __p0; \
  47561.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  47562.   __ret = (float64_t) __builtin_neon_vgetq_lane_f64((float64x2_t)__rev0, __p1); \
  47563.   __ret; \
  47564. })
  47565. #define __noswap_vgetq_lane_f64(__p0, __p1) __extension__ ({ \
  47566.   float64_t __ret; \
  47567.   float64x2_t __s0 = __p0; \
  47568.   __ret = (float64_t) __builtin_neon_vgetq_lane_f64((float64x2_t)__s0, __p1); \
  47569.   __ret; \
  47570. })
  47571. #endif
  47572.  
  47573. #define vget_lane_f64(__p0, __p1) __extension__ ({ \
  47574.   float64_t __ret; \
  47575.   float64x1_t __s0 = __p0; \
  47576.   __ret = (float64_t) __builtin_neon_vget_lane_f64((float64x1_t)__s0, __p1); \
  47577.   __ret; \
  47578. })
  47579. #ifdef __LITTLE_ENDIAN__
  47580. __ai poly64x1_t vget_low_p64(poly64x2_t __p0) {
  47581.   poly64x1_t __ret;
  47582.   __ret = __builtin_shufflevector(__p0, __p0, 0);
  47583.   return __ret;
  47584. }
  47585. #else
  47586. __ai poly64x1_t vget_low_p64(poly64x2_t __p0) {
  47587.   poly64x1_t __ret;
  47588.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  47589.   __ret = __builtin_shufflevector(__rev0, __rev0, 0);
  47590.   return __ret;
  47591. }
  47592. #endif
  47593.  
  47594. #ifdef __LITTLE_ENDIAN__
  47595. __ai float64x1_t vget_low_f64(float64x2_t __p0) {
  47596.   float64x1_t __ret;
  47597.   __ret = __builtin_shufflevector(__p0, __p0, 0);
  47598.   return __ret;
  47599. }
  47600. #else
  47601. __ai float64x1_t vget_low_f64(float64x2_t __p0) {
  47602.   float64x1_t __ret;
  47603.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  47604.   __ret = __builtin_shufflevector(__rev0, __rev0, 0);
  47605.   return __ret;
  47606. }
  47607. #endif
  47608.  
  47609. #define vld1_p64(__p0) __extension__ ({ \
  47610.   poly64x1_t __ret; \
  47611.   __ret = (poly64x1_t) __builtin_neon_vld1_v(__p0, 6); \
  47612.   __ret; \
  47613. })
  47614. #ifdef __LITTLE_ENDIAN__
  47615. #define vld1q_p64(__p0) __extension__ ({ \
  47616.   poly64x2_t __ret; \
  47617.   __ret = (poly64x2_t) __builtin_neon_vld1q_v(__p0, 38); \
  47618.   __ret; \
  47619. })
  47620. #else
  47621. #define vld1q_p64(__p0) __extension__ ({ \
  47622.   poly64x2_t __ret; \
  47623.   __ret = (poly64x2_t) __builtin_neon_vld1q_v(__p0, 38); \
  47624.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  47625.   __ret; \
  47626. })
  47627. #endif
  47628.  
  47629. #ifdef __LITTLE_ENDIAN__
  47630. #define vld1q_f64(__p0) __extension__ ({ \
  47631.   float64x2_t __ret; \
  47632.   __ret = (float64x2_t) __builtin_neon_vld1q_v(__p0, 42); \
  47633.   __ret; \
  47634. })
  47635. #else
  47636. #define vld1q_f64(__p0) __extension__ ({ \
  47637.   float64x2_t __ret; \
  47638.   __ret = (float64x2_t) __builtin_neon_vld1q_v(__p0, 42); \
  47639.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  47640.   __ret; \
  47641. })
  47642. #endif
  47643.  
  47644. #define vld1_f64(__p0) __extension__ ({ \
  47645.   float64x1_t __ret; \
  47646.   __ret = (float64x1_t) __builtin_neon_vld1_v(__p0, 10); \
  47647.   __ret; \
  47648. })
  47649. #define vld1_dup_p64(__p0) __extension__ ({ \
  47650.   poly64x1_t __ret; \
  47651.   __ret = (poly64x1_t) __builtin_neon_vld1_dup_v(__p0, 6); \
  47652.   __ret; \
  47653. })
  47654. #ifdef __LITTLE_ENDIAN__
  47655. #define vld1q_dup_p64(__p0) __extension__ ({ \
  47656.   poly64x2_t __ret; \
  47657.   __ret = (poly64x2_t) __builtin_neon_vld1q_dup_v(__p0, 38); \
  47658.   __ret; \
  47659. })
  47660. #else
  47661. #define vld1q_dup_p64(__p0) __extension__ ({ \
  47662.   poly64x2_t __ret; \
  47663.   __ret = (poly64x2_t) __builtin_neon_vld1q_dup_v(__p0, 38); \
  47664.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  47665.   __ret; \
  47666. })
  47667. #endif
  47668.  
  47669. #ifdef __LITTLE_ENDIAN__
  47670. #define vld1q_dup_f64(__p0) __extension__ ({ \
  47671.   float64x2_t __ret; \
  47672.   __ret = (float64x2_t) __builtin_neon_vld1q_dup_v(__p0, 42); \
  47673.   __ret; \
  47674. })
  47675. #else
  47676. #define vld1q_dup_f64(__p0) __extension__ ({ \
  47677.   float64x2_t __ret; \
  47678.   __ret = (float64x2_t) __builtin_neon_vld1q_dup_v(__p0, 42); \
  47679.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  47680.   __ret; \
  47681. })
  47682. #endif
  47683.  
  47684. #define vld1_dup_f64(__p0) __extension__ ({ \
  47685.   float64x1_t __ret; \
  47686.   __ret = (float64x1_t) __builtin_neon_vld1_dup_v(__p0, 10); \
  47687.   __ret; \
  47688. })
  47689. #define vld1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  47690.   poly64x1_t __ret; \
  47691.   poly64x1_t __s1 = __p1; \
  47692.   __ret = (poly64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
  47693.   __ret; \
  47694. })
  47695. #ifdef __LITTLE_ENDIAN__
  47696. #define vld1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  47697.   poly64x2_t __ret; \
  47698.   poly64x2_t __s1 = __p1; \
  47699.   __ret = (poly64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 38); \
  47700.   __ret; \
  47701. })
  47702. #else
  47703. #define vld1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  47704.   poly64x2_t __ret; \
  47705.   poly64x2_t __s1 = __p1; \
  47706.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  47707.   __ret = (poly64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 38); \
  47708.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  47709.   __ret; \
  47710. })
  47711. #endif
  47712.  
  47713. #ifdef __LITTLE_ENDIAN__
  47714. #define vld1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  47715.   float64x2_t __ret; \
  47716.   float64x2_t __s1 = __p1; \
  47717.   __ret = (float64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 42); \
  47718.   __ret; \
  47719. })
  47720. #else
  47721. #define vld1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  47722.   float64x2_t __ret; \
  47723.   float64x2_t __s1 = __p1; \
  47724.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  47725.   __ret = (float64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 42); \
  47726.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  47727.   __ret; \
  47728. })
  47729. #endif
  47730.  
  47731. #define vld1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  47732.   float64x1_t __ret; \
  47733.   float64x1_t __s1 = __p1; \
  47734.   __ret = (float64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
  47735.   __ret; \
  47736. })
  47737. #define vld1_p64_x2(__p0) __extension__ ({ \
  47738.   poly64x1x2_t __ret; \
  47739.   __builtin_neon_vld1_x2_v(&__ret, __p0, 6); \
  47740.   __ret; \
  47741. })
  47742. #ifdef __LITTLE_ENDIAN__
  47743. #define vld1q_p64_x2(__p0) __extension__ ({ \
  47744.   poly64x2x2_t __ret; \
  47745.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 38); \
  47746.   __ret; \
  47747. })
  47748. #else
  47749. #define vld1q_p64_x2(__p0) __extension__ ({ \
  47750.   poly64x2x2_t __ret; \
  47751.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 38); \
  47752.  \
  47753.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47754.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47755.   __ret; \
  47756. })
  47757. #endif
  47758.  
  47759. #ifdef __LITTLE_ENDIAN__
  47760. #define vld1q_f64_x2(__p0) __extension__ ({ \
  47761.   float64x2x2_t __ret; \
  47762.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 42); \
  47763.   __ret; \
  47764. })
  47765. #else
  47766. #define vld1q_f64_x2(__p0) __extension__ ({ \
  47767.   float64x2x2_t __ret; \
  47768.   __builtin_neon_vld1q_x2_v(&__ret, __p0, 42); \
  47769.  \
  47770.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47771.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47772.   __ret; \
  47773. })
  47774. #endif
  47775.  
  47776. #define vld1_f64_x2(__p0) __extension__ ({ \
  47777.   float64x1x2_t __ret; \
  47778.   __builtin_neon_vld1_x2_v(&__ret, __p0, 10); \
  47779.   __ret; \
  47780. })
  47781. #define vld1_p64_x3(__p0) __extension__ ({ \
  47782.   poly64x1x3_t __ret; \
  47783.   __builtin_neon_vld1_x3_v(&__ret, __p0, 6); \
  47784.   __ret; \
  47785. })
  47786. #ifdef __LITTLE_ENDIAN__
  47787. #define vld1q_p64_x3(__p0) __extension__ ({ \
  47788.   poly64x2x3_t __ret; \
  47789.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 38); \
  47790.   __ret; \
  47791. })
  47792. #else
  47793. #define vld1q_p64_x3(__p0) __extension__ ({ \
  47794.   poly64x2x3_t __ret; \
  47795.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 38); \
  47796.  \
  47797.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47798.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47799.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  47800.   __ret; \
  47801. })
  47802. #endif
  47803.  
  47804. #ifdef __LITTLE_ENDIAN__
  47805. #define vld1q_f64_x3(__p0) __extension__ ({ \
  47806.   float64x2x3_t __ret; \
  47807.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 42); \
  47808.   __ret; \
  47809. })
  47810. #else
  47811. #define vld1q_f64_x3(__p0) __extension__ ({ \
  47812.   float64x2x3_t __ret; \
  47813.   __builtin_neon_vld1q_x3_v(&__ret, __p0, 42); \
  47814.  \
  47815.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47816.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47817.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  47818.   __ret; \
  47819. })
  47820. #endif
  47821.  
  47822. #define vld1_f64_x3(__p0) __extension__ ({ \
  47823.   float64x1x3_t __ret; \
  47824.   __builtin_neon_vld1_x3_v(&__ret, __p0, 10); \
  47825.   __ret; \
  47826. })
  47827. #define vld1_p64_x4(__p0) __extension__ ({ \
  47828.   poly64x1x4_t __ret; \
  47829.   __builtin_neon_vld1_x4_v(&__ret, __p0, 6); \
  47830.   __ret; \
  47831. })
  47832. #ifdef __LITTLE_ENDIAN__
  47833. #define vld1q_p64_x4(__p0) __extension__ ({ \
  47834.   poly64x2x4_t __ret; \
  47835.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 38); \
  47836.   __ret; \
  47837. })
  47838. #else
  47839. #define vld1q_p64_x4(__p0) __extension__ ({ \
  47840.   poly64x2x4_t __ret; \
  47841.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 38); \
  47842.  \
  47843.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47844.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47845.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  47846.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  47847.   __ret; \
  47848. })
  47849. #endif
  47850.  
  47851. #ifdef __LITTLE_ENDIAN__
  47852. #define vld1q_f64_x4(__p0) __extension__ ({ \
  47853.   float64x2x4_t __ret; \
  47854.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 42); \
  47855.   __ret; \
  47856. })
  47857. #else
  47858. #define vld1q_f64_x4(__p0) __extension__ ({ \
  47859.   float64x2x4_t __ret; \
  47860.   __builtin_neon_vld1q_x4_v(&__ret, __p0, 42); \
  47861.  \
  47862.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47863.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47864.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  47865.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  47866.   __ret; \
  47867. })
  47868. #endif
  47869.  
  47870. #define vld1_f64_x4(__p0) __extension__ ({ \
  47871.   float64x1x4_t __ret; \
  47872.   __builtin_neon_vld1_x4_v(&__ret, __p0, 10); \
  47873.   __ret; \
  47874. })
  47875. #define vld2_p64(__p0) __extension__ ({ \
  47876.   poly64x1x2_t __ret; \
  47877.   __builtin_neon_vld2_v(&__ret, __p0, 6); \
  47878.   __ret; \
  47879. })
  47880. #ifdef __LITTLE_ENDIAN__
  47881. #define vld2q_p64(__p0) __extension__ ({ \
  47882.   poly64x2x2_t __ret; \
  47883.   __builtin_neon_vld2q_v(&__ret, __p0, 38); \
  47884.   __ret; \
  47885. })
  47886. #else
  47887. #define vld2q_p64(__p0) __extension__ ({ \
  47888.   poly64x2x2_t __ret; \
  47889.   __builtin_neon_vld2q_v(&__ret, __p0, 38); \
  47890.  \
  47891.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47892.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47893.   __ret; \
  47894. })
  47895. #endif
  47896.  
  47897. #ifdef __LITTLE_ENDIAN__
  47898. #define vld2q_u64(__p0) __extension__ ({ \
  47899.   uint64x2x2_t __ret; \
  47900.   __builtin_neon_vld2q_v(&__ret, __p0, 51); \
  47901.   __ret; \
  47902. })
  47903. #else
  47904. #define vld2q_u64(__p0) __extension__ ({ \
  47905.   uint64x2x2_t __ret; \
  47906.   __builtin_neon_vld2q_v(&__ret, __p0, 51); \
  47907.  \
  47908.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47909.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47910.   __ret; \
  47911. })
  47912. #endif
  47913.  
  47914. #ifdef __LITTLE_ENDIAN__
  47915. #define vld2q_f64(__p0) __extension__ ({ \
  47916.   float64x2x2_t __ret; \
  47917.   __builtin_neon_vld2q_v(&__ret, __p0, 42); \
  47918.   __ret; \
  47919. })
  47920. #else
  47921. #define vld2q_f64(__p0) __extension__ ({ \
  47922.   float64x2x2_t __ret; \
  47923.   __builtin_neon_vld2q_v(&__ret, __p0, 42); \
  47924.  \
  47925.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47926.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47927.   __ret; \
  47928. })
  47929. #endif
  47930.  
  47931. #ifdef __LITTLE_ENDIAN__
  47932. #define vld2q_s64(__p0) __extension__ ({ \
  47933.   int64x2x2_t __ret; \
  47934.   __builtin_neon_vld2q_v(&__ret, __p0, 35); \
  47935.   __ret; \
  47936. })
  47937. #else
  47938. #define vld2q_s64(__p0) __extension__ ({ \
  47939.   int64x2x2_t __ret; \
  47940.   __builtin_neon_vld2q_v(&__ret, __p0, 35); \
  47941.  \
  47942.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47943.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47944.   __ret; \
  47945. })
  47946. #endif
  47947.  
  47948. #define vld2_f64(__p0) __extension__ ({ \
  47949.   float64x1x2_t __ret; \
  47950.   __builtin_neon_vld2_v(&__ret, __p0, 10); \
  47951.   __ret; \
  47952. })
  47953. #define vld2_dup_p64(__p0) __extension__ ({ \
  47954.   poly64x1x2_t __ret; \
  47955.   __builtin_neon_vld2_dup_v(&__ret, __p0, 6); \
  47956.   __ret; \
  47957. })
  47958. #ifdef __LITTLE_ENDIAN__
  47959. #define vld2q_dup_p64(__p0) __extension__ ({ \
  47960.   poly64x2x2_t __ret; \
  47961.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 38); \
  47962.   __ret; \
  47963. })
  47964. #else
  47965. #define vld2q_dup_p64(__p0) __extension__ ({ \
  47966.   poly64x2x2_t __ret; \
  47967.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 38); \
  47968.  \
  47969.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47970.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47971.   __ret; \
  47972. })
  47973. #endif
  47974.  
  47975. #ifdef __LITTLE_ENDIAN__
  47976. #define vld2q_dup_f64(__p0) __extension__ ({ \
  47977.   float64x2x2_t __ret; \
  47978.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 42); \
  47979.   __ret; \
  47980. })
  47981. #else
  47982. #define vld2q_dup_f64(__p0) __extension__ ({ \
  47983.   float64x2x2_t __ret; \
  47984.   __builtin_neon_vld2q_dup_v(&__ret, __p0, 42); \
  47985.  \
  47986.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  47987.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  47988.   __ret; \
  47989. })
  47990. #endif
  47991.  
  47992. #define vld2_dup_f64(__p0) __extension__ ({ \
  47993.   float64x1x2_t __ret; \
  47994.   __builtin_neon_vld2_dup_v(&__ret, __p0, 10); \
  47995.   __ret; \
  47996. })
  47997. #define vld2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  47998.   poly64x1x2_t __ret; \
  47999.   poly64x1x2_t __s1 = __p1; \
  48000.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
  48001.   __ret; \
  48002. })
  48003. #ifdef __LITTLE_ENDIAN__
  48004. #define vld2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  48005.   poly8x16x2_t __ret; \
  48006.   poly8x16x2_t __s1 = __p1; \
  48007.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 36); \
  48008.   __ret; \
  48009. })
  48010. #else
  48011. #define vld2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  48012.   poly8x16x2_t __ret; \
  48013.   poly8x16x2_t __s1 = __p1; \
  48014.   poly8x16x2_t __rev1; \
  48015.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48016.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48017.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 36); \
  48018.  \
  48019.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48020.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48021.   __ret; \
  48022. })
  48023. #endif
  48024.  
  48025. #ifdef __LITTLE_ENDIAN__
  48026. #define vld2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  48027.   poly64x2x2_t __ret; \
  48028.   poly64x2x2_t __s1 = __p1; \
  48029.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 38); \
  48030.   __ret; \
  48031. })
  48032. #else
  48033. #define vld2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  48034.   poly64x2x2_t __ret; \
  48035.   poly64x2x2_t __s1 = __p1; \
  48036.   poly64x2x2_t __rev1; \
  48037.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48038.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48039.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 38); \
  48040.  \
  48041.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48042.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48043.   __ret; \
  48044. })
  48045. #endif
  48046.  
  48047. #ifdef __LITTLE_ENDIAN__
  48048. #define vld2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  48049.   uint8x16x2_t __ret; \
  48050.   uint8x16x2_t __s1 = __p1; \
  48051.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 48); \
  48052.   __ret; \
  48053. })
  48054. #else
  48055. #define vld2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  48056.   uint8x16x2_t __ret; \
  48057.   uint8x16x2_t __s1 = __p1; \
  48058.   uint8x16x2_t __rev1; \
  48059.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48060.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48061.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 48); \
  48062.  \
  48063.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48064.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48065.   __ret; \
  48066. })
  48067. #endif
  48068.  
  48069. #ifdef __LITTLE_ENDIAN__
  48070. #define vld2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  48071.   uint64x2x2_t __ret; \
  48072.   uint64x2x2_t __s1 = __p1; \
  48073.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 51); \
  48074.   __ret; \
  48075. })
  48076. #else
  48077. #define vld2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  48078.   uint64x2x2_t __ret; \
  48079.   uint64x2x2_t __s1 = __p1; \
  48080.   uint64x2x2_t __rev1; \
  48081.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48082.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48083.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 51); \
  48084.  \
  48085.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48086.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48087.   __ret; \
  48088. })
  48089. #endif
  48090.  
  48091. #ifdef __LITTLE_ENDIAN__
  48092. #define vld2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  48093.   int8x16x2_t __ret; \
  48094.   int8x16x2_t __s1 = __p1; \
  48095.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 32); \
  48096.   __ret; \
  48097. })
  48098. #else
  48099. #define vld2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  48100.   int8x16x2_t __ret; \
  48101.   int8x16x2_t __s1 = __p1; \
  48102.   int8x16x2_t __rev1; \
  48103.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48104.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48105.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 32); \
  48106.  \
  48107.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48108.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48109.   __ret; \
  48110. })
  48111. #endif
  48112.  
  48113. #ifdef __LITTLE_ENDIAN__
  48114. #define vld2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  48115.   float64x2x2_t __ret; \
  48116.   float64x2x2_t __s1 = __p1; \
  48117.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 42); \
  48118.   __ret; \
  48119. })
  48120. #else
  48121. #define vld2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  48122.   float64x2x2_t __ret; \
  48123.   float64x2x2_t __s1 = __p1; \
  48124.   float64x2x2_t __rev1; \
  48125.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48126.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48127.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 42); \
  48128.  \
  48129.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48130.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48131.   __ret; \
  48132. })
  48133. #endif
  48134.  
  48135. #ifdef __LITTLE_ENDIAN__
  48136. #define vld2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  48137.   int64x2x2_t __ret; \
  48138.   int64x2x2_t __s1 = __p1; \
  48139.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 35); \
  48140.   __ret; \
  48141. })
  48142. #else
  48143. #define vld2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  48144.   int64x2x2_t __ret; \
  48145.   int64x2x2_t __s1 = __p1; \
  48146.   int64x2x2_t __rev1; \
  48147.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48148.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48149.   __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 35); \
  48150.  \
  48151.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48152.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48153.   __ret; \
  48154. })
  48155. #endif
  48156.  
  48157. #define vld2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  48158.   uint64x1x2_t __ret; \
  48159.   uint64x1x2_t __s1 = __p1; \
  48160.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
  48161.   __ret; \
  48162. })
  48163. #define vld2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  48164.   float64x1x2_t __ret; \
  48165.   float64x1x2_t __s1 = __p1; \
  48166.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 10); \
  48167.   __ret; \
  48168. })
  48169. #define vld2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  48170.   int64x1x2_t __ret; \
  48171.   int64x1x2_t __s1 = __p1; \
  48172.   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 3); \
  48173.   __ret; \
  48174. })
  48175. #define vld3_p64(__p0) __extension__ ({ \
  48176.   poly64x1x3_t __ret; \
  48177.   __builtin_neon_vld3_v(&__ret, __p0, 6); \
  48178.   __ret; \
  48179. })
  48180. #ifdef __LITTLE_ENDIAN__
  48181. #define vld3q_p64(__p0) __extension__ ({ \
  48182.   poly64x2x3_t __ret; \
  48183.   __builtin_neon_vld3q_v(&__ret, __p0, 38); \
  48184.   __ret; \
  48185. })
  48186. #else
  48187. #define vld3q_p64(__p0) __extension__ ({ \
  48188.   poly64x2x3_t __ret; \
  48189.   __builtin_neon_vld3q_v(&__ret, __p0, 38); \
  48190.  \
  48191.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48192.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48193.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48194.   __ret; \
  48195. })
  48196. #endif
  48197.  
  48198. #ifdef __LITTLE_ENDIAN__
  48199. #define vld3q_u64(__p0) __extension__ ({ \
  48200.   uint64x2x3_t __ret; \
  48201.   __builtin_neon_vld3q_v(&__ret, __p0, 51); \
  48202.   __ret; \
  48203. })
  48204. #else
  48205. #define vld3q_u64(__p0) __extension__ ({ \
  48206.   uint64x2x3_t __ret; \
  48207.   __builtin_neon_vld3q_v(&__ret, __p0, 51); \
  48208.  \
  48209.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48210.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48211.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48212.   __ret; \
  48213. })
  48214. #endif
  48215.  
  48216. #ifdef __LITTLE_ENDIAN__
  48217. #define vld3q_f64(__p0) __extension__ ({ \
  48218.   float64x2x3_t __ret; \
  48219.   __builtin_neon_vld3q_v(&__ret, __p0, 42); \
  48220.   __ret; \
  48221. })
  48222. #else
  48223. #define vld3q_f64(__p0) __extension__ ({ \
  48224.   float64x2x3_t __ret; \
  48225.   __builtin_neon_vld3q_v(&__ret, __p0, 42); \
  48226.  \
  48227.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48228.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48229.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48230.   __ret; \
  48231. })
  48232. #endif
  48233.  
  48234. #ifdef __LITTLE_ENDIAN__
  48235. #define vld3q_s64(__p0) __extension__ ({ \
  48236.   int64x2x3_t __ret; \
  48237.   __builtin_neon_vld3q_v(&__ret, __p0, 35); \
  48238.   __ret; \
  48239. })
  48240. #else
  48241. #define vld3q_s64(__p0) __extension__ ({ \
  48242.   int64x2x3_t __ret; \
  48243.   __builtin_neon_vld3q_v(&__ret, __p0, 35); \
  48244.  \
  48245.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48246.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48247.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48248.   __ret; \
  48249. })
  48250. #endif
  48251.  
  48252. #define vld3_f64(__p0) __extension__ ({ \
  48253.   float64x1x3_t __ret; \
  48254.   __builtin_neon_vld3_v(&__ret, __p0, 10); \
  48255.   __ret; \
  48256. })
  48257. #define vld3_dup_p64(__p0) __extension__ ({ \
  48258.   poly64x1x3_t __ret; \
  48259.   __builtin_neon_vld3_dup_v(&__ret, __p0, 6); \
  48260.   __ret; \
  48261. })
  48262. #ifdef __LITTLE_ENDIAN__
  48263. #define vld3q_dup_p64(__p0) __extension__ ({ \
  48264.   poly64x2x3_t __ret; \
  48265.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 38); \
  48266.   __ret; \
  48267. })
  48268. #else
  48269. #define vld3q_dup_p64(__p0) __extension__ ({ \
  48270.   poly64x2x3_t __ret; \
  48271.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 38); \
  48272.  \
  48273.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48274.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48275.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48276.   __ret; \
  48277. })
  48278. #endif
  48279.  
  48280. #ifdef __LITTLE_ENDIAN__
  48281. #define vld3q_dup_f64(__p0) __extension__ ({ \
  48282.   float64x2x3_t __ret; \
  48283.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 42); \
  48284.   __ret; \
  48285. })
  48286. #else
  48287. #define vld3q_dup_f64(__p0) __extension__ ({ \
  48288.   float64x2x3_t __ret; \
  48289.   __builtin_neon_vld3q_dup_v(&__ret, __p0, 42); \
  48290.  \
  48291.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48292.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48293.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48294.   __ret; \
  48295. })
  48296. #endif
  48297.  
  48298. #define vld3_dup_f64(__p0) __extension__ ({ \
  48299.   float64x1x3_t __ret; \
  48300.   __builtin_neon_vld3_dup_v(&__ret, __p0, 10); \
  48301.   __ret; \
  48302. })
  48303. #define vld3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  48304.   poly64x1x3_t __ret; \
  48305.   poly64x1x3_t __s1 = __p1; \
  48306.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
  48307.   __ret; \
  48308. })
  48309. #ifdef __LITTLE_ENDIAN__
  48310. #define vld3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  48311.   poly8x16x3_t __ret; \
  48312.   poly8x16x3_t __s1 = __p1; \
  48313.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 36); \
  48314.   __ret; \
  48315. })
  48316. #else
  48317. #define vld3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  48318.   poly8x16x3_t __ret; \
  48319.   poly8x16x3_t __s1 = __p1; \
  48320.   poly8x16x3_t __rev1; \
  48321.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48322.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48323.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48324.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 36); \
  48325.  \
  48326.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48327.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48328.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48329.   __ret; \
  48330. })
  48331. #endif
  48332.  
  48333. #ifdef __LITTLE_ENDIAN__
  48334. #define vld3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  48335.   poly64x2x3_t __ret; \
  48336.   poly64x2x3_t __s1 = __p1; \
  48337.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 38); \
  48338.   __ret; \
  48339. })
  48340. #else
  48341. #define vld3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  48342.   poly64x2x3_t __ret; \
  48343.   poly64x2x3_t __s1 = __p1; \
  48344.   poly64x2x3_t __rev1; \
  48345.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48346.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48347.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  48348.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 38); \
  48349.  \
  48350.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48351.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48352.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48353.   __ret; \
  48354. })
  48355. #endif
  48356.  
  48357. #ifdef __LITTLE_ENDIAN__
  48358. #define vld3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  48359.   uint8x16x3_t __ret; \
  48360.   uint8x16x3_t __s1 = __p1; \
  48361.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 48); \
  48362.   __ret; \
  48363. })
  48364. #else
  48365. #define vld3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  48366.   uint8x16x3_t __ret; \
  48367.   uint8x16x3_t __s1 = __p1; \
  48368.   uint8x16x3_t __rev1; \
  48369.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48370.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48371.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48372.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 48); \
  48373.  \
  48374.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48375.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48376.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48377.   __ret; \
  48378. })
  48379. #endif
  48380.  
  48381. #ifdef __LITTLE_ENDIAN__
  48382. #define vld3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  48383.   uint64x2x3_t __ret; \
  48384.   uint64x2x3_t __s1 = __p1; \
  48385.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 51); \
  48386.   __ret; \
  48387. })
  48388. #else
  48389. #define vld3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  48390.   uint64x2x3_t __ret; \
  48391.   uint64x2x3_t __s1 = __p1; \
  48392.   uint64x2x3_t __rev1; \
  48393.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48394.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48395.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  48396.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 51); \
  48397.  \
  48398.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48399.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48400.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48401.   __ret; \
  48402. })
  48403. #endif
  48404.  
  48405. #ifdef __LITTLE_ENDIAN__
  48406. #define vld3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  48407.   int8x16x3_t __ret; \
  48408.   int8x16x3_t __s1 = __p1; \
  48409.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 32); \
  48410.   __ret; \
  48411. })
  48412. #else
  48413. #define vld3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  48414.   int8x16x3_t __ret; \
  48415.   int8x16x3_t __s1 = __p1; \
  48416.   int8x16x3_t __rev1; \
  48417.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48418.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48419.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48420.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 32); \
  48421.  \
  48422.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48423.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48424.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48425.   __ret; \
  48426. })
  48427. #endif
  48428.  
  48429. #ifdef __LITTLE_ENDIAN__
  48430. #define vld3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  48431.   float64x2x3_t __ret; \
  48432.   float64x2x3_t __s1 = __p1; \
  48433.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 42); \
  48434.   __ret; \
  48435. })
  48436. #else
  48437. #define vld3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  48438.   float64x2x3_t __ret; \
  48439.   float64x2x3_t __s1 = __p1; \
  48440.   float64x2x3_t __rev1; \
  48441.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48442.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48443.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  48444.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 42); \
  48445.  \
  48446.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48447.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48448.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48449.   __ret; \
  48450. })
  48451. #endif
  48452.  
  48453. #ifdef __LITTLE_ENDIAN__
  48454. #define vld3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  48455.   int64x2x3_t __ret; \
  48456.   int64x2x3_t __s1 = __p1; \
  48457.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 35); \
  48458.   __ret; \
  48459. })
  48460. #else
  48461. #define vld3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  48462.   int64x2x3_t __ret; \
  48463.   int64x2x3_t __s1 = __p1; \
  48464.   int64x2x3_t __rev1; \
  48465.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48466.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48467.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  48468.   __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 35); \
  48469.  \
  48470.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48471.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48472.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48473.   __ret; \
  48474. })
  48475. #endif
  48476.  
  48477. #define vld3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  48478.   uint64x1x3_t __ret; \
  48479.   uint64x1x3_t __s1 = __p1; \
  48480.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
  48481.   __ret; \
  48482. })
  48483. #define vld3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  48484.   float64x1x3_t __ret; \
  48485.   float64x1x3_t __s1 = __p1; \
  48486.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 10); \
  48487.   __ret; \
  48488. })
  48489. #define vld3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  48490.   int64x1x3_t __ret; \
  48491.   int64x1x3_t __s1 = __p1; \
  48492.   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 3); \
  48493.   __ret; \
  48494. })
  48495. #define vld4_p64(__p0) __extension__ ({ \
  48496.   poly64x1x4_t __ret; \
  48497.   __builtin_neon_vld4_v(&__ret, __p0, 6); \
  48498.   __ret; \
  48499. })
  48500. #ifdef __LITTLE_ENDIAN__
  48501. #define vld4q_p64(__p0) __extension__ ({ \
  48502.   poly64x2x4_t __ret; \
  48503.   __builtin_neon_vld4q_v(&__ret, __p0, 38); \
  48504.   __ret; \
  48505. })
  48506. #else
  48507. #define vld4q_p64(__p0) __extension__ ({ \
  48508.   poly64x2x4_t __ret; \
  48509.   __builtin_neon_vld4q_v(&__ret, __p0, 38); \
  48510.  \
  48511.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48512.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48513.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48514.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48515.   __ret; \
  48516. })
  48517. #endif
  48518.  
  48519. #ifdef __LITTLE_ENDIAN__
  48520. #define vld4q_u64(__p0) __extension__ ({ \
  48521.   uint64x2x4_t __ret; \
  48522.   __builtin_neon_vld4q_v(&__ret, __p0, 51); \
  48523.   __ret; \
  48524. })
  48525. #else
  48526. #define vld4q_u64(__p0) __extension__ ({ \
  48527.   uint64x2x4_t __ret; \
  48528.   __builtin_neon_vld4q_v(&__ret, __p0, 51); \
  48529.  \
  48530.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48531.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48532.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48533.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48534.   __ret; \
  48535. })
  48536. #endif
  48537.  
  48538. #ifdef __LITTLE_ENDIAN__
  48539. #define vld4q_f64(__p0) __extension__ ({ \
  48540.   float64x2x4_t __ret; \
  48541.   __builtin_neon_vld4q_v(&__ret, __p0, 42); \
  48542.   __ret; \
  48543. })
  48544. #else
  48545. #define vld4q_f64(__p0) __extension__ ({ \
  48546.   float64x2x4_t __ret; \
  48547.   __builtin_neon_vld4q_v(&__ret, __p0, 42); \
  48548.  \
  48549.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48550.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48551.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48552.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48553.   __ret; \
  48554. })
  48555. #endif
  48556.  
  48557. #ifdef __LITTLE_ENDIAN__
  48558. #define vld4q_s64(__p0) __extension__ ({ \
  48559.   int64x2x4_t __ret; \
  48560.   __builtin_neon_vld4q_v(&__ret, __p0, 35); \
  48561.   __ret; \
  48562. })
  48563. #else
  48564. #define vld4q_s64(__p0) __extension__ ({ \
  48565.   int64x2x4_t __ret; \
  48566.   __builtin_neon_vld4q_v(&__ret, __p0, 35); \
  48567.  \
  48568.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48569.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48570.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48571.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48572.   __ret; \
  48573. })
  48574. #endif
  48575.  
  48576. #define vld4_f64(__p0) __extension__ ({ \
  48577.   float64x1x4_t __ret; \
  48578.   __builtin_neon_vld4_v(&__ret, __p0, 10); \
  48579.   __ret; \
  48580. })
  48581. #define vld4_dup_p64(__p0) __extension__ ({ \
  48582.   poly64x1x4_t __ret; \
  48583.   __builtin_neon_vld4_dup_v(&__ret, __p0, 6); \
  48584.   __ret; \
  48585. })
  48586. #ifdef __LITTLE_ENDIAN__
  48587. #define vld4q_dup_p64(__p0) __extension__ ({ \
  48588.   poly64x2x4_t __ret; \
  48589.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 38); \
  48590.   __ret; \
  48591. })
  48592. #else
  48593. #define vld4q_dup_p64(__p0) __extension__ ({ \
  48594.   poly64x2x4_t __ret; \
  48595.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 38); \
  48596.  \
  48597.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48598.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48599.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48600.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48601.   __ret; \
  48602. })
  48603. #endif
  48604.  
  48605. #ifdef __LITTLE_ENDIAN__
  48606. #define vld4q_dup_f64(__p0) __extension__ ({ \
  48607.   float64x2x4_t __ret; \
  48608.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 42); \
  48609.   __ret; \
  48610. })
  48611. #else
  48612. #define vld4q_dup_f64(__p0) __extension__ ({ \
  48613.   float64x2x4_t __ret; \
  48614.   __builtin_neon_vld4q_dup_v(&__ret, __p0, 42); \
  48615.  \
  48616.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48617.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48618.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48619.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48620.   __ret; \
  48621. })
  48622. #endif
  48623.  
  48624. #define vld4_dup_f64(__p0) __extension__ ({ \
  48625.   float64x1x4_t __ret; \
  48626.   __builtin_neon_vld4_dup_v(&__ret, __p0, 10); \
  48627.   __ret; \
  48628. })
  48629. #define vld4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  48630.   poly64x1x4_t __ret; \
  48631.   poly64x1x4_t __s1 = __p1; \
  48632.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
  48633.   __ret; \
  48634. })
  48635. #ifdef __LITTLE_ENDIAN__
  48636. #define vld4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  48637.   poly8x16x4_t __ret; \
  48638.   poly8x16x4_t __s1 = __p1; \
  48639.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 36); \
  48640.   __ret; \
  48641. })
  48642. #else
  48643. #define vld4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  48644.   poly8x16x4_t __ret; \
  48645.   poly8x16x4_t __s1 = __p1; \
  48646.   poly8x16x4_t __rev1; \
  48647.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48648.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48649.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48650.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48651.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 36); \
  48652.  \
  48653.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48654.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48655.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48656.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48657.   __ret; \
  48658. })
  48659. #endif
  48660.  
  48661. #ifdef __LITTLE_ENDIAN__
  48662. #define vld4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  48663.   poly64x2x4_t __ret; \
  48664.   poly64x2x4_t __s1 = __p1; \
  48665.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 38); \
  48666.   __ret; \
  48667. })
  48668. #else
  48669. #define vld4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  48670.   poly64x2x4_t __ret; \
  48671.   poly64x2x4_t __s1 = __p1; \
  48672.   poly64x2x4_t __rev1; \
  48673.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48674.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48675.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  48676.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  48677.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 38); \
  48678.  \
  48679.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48680.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48681.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48682.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48683.   __ret; \
  48684. })
  48685. #endif
  48686.  
  48687. #ifdef __LITTLE_ENDIAN__
  48688. #define vld4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  48689.   uint8x16x4_t __ret; \
  48690.   uint8x16x4_t __s1 = __p1; \
  48691.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 48); \
  48692.   __ret; \
  48693. })
  48694. #else
  48695. #define vld4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  48696.   uint8x16x4_t __ret; \
  48697.   uint8x16x4_t __s1 = __p1; \
  48698.   uint8x16x4_t __rev1; \
  48699.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48700.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48701.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48702.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48703.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 48); \
  48704.  \
  48705.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48706.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48707.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48708.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48709.   __ret; \
  48710. })
  48711. #endif
  48712.  
  48713. #ifdef __LITTLE_ENDIAN__
  48714. #define vld4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  48715.   uint64x2x4_t __ret; \
  48716.   uint64x2x4_t __s1 = __p1; \
  48717.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 51); \
  48718.   __ret; \
  48719. })
  48720. #else
  48721. #define vld4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  48722.   uint64x2x4_t __ret; \
  48723.   uint64x2x4_t __s1 = __p1; \
  48724.   uint64x2x4_t __rev1; \
  48725.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48726.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48727.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  48728.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  48729.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 51); \
  48730.  \
  48731.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48732.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48733.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48734.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48735.   __ret; \
  48736. })
  48737. #endif
  48738.  
  48739. #ifdef __LITTLE_ENDIAN__
  48740. #define vld4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  48741.   int8x16x4_t __ret; \
  48742.   int8x16x4_t __s1 = __p1; \
  48743.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 32); \
  48744.   __ret; \
  48745. })
  48746. #else
  48747. #define vld4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  48748.   int8x16x4_t __ret; \
  48749.   int8x16x4_t __s1 = __p1; \
  48750.   int8x16x4_t __rev1; \
  48751.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48752.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48753.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48754.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48755.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 32); \
  48756.  \
  48757.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48758.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48759.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48760.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  48761.   __ret; \
  48762. })
  48763. #endif
  48764.  
  48765. #ifdef __LITTLE_ENDIAN__
  48766. #define vld4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  48767.   float64x2x4_t __ret; \
  48768.   float64x2x4_t __s1 = __p1; \
  48769.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 42); \
  48770.   __ret; \
  48771. })
  48772. #else
  48773. #define vld4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  48774.   float64x2x4_t __ret; \
  48775.   float64x2x4_t __s1 = __p1; \
  48776.   float64x2x4_t __rev1; \
  48777.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48778.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48779.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  48780.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  48781.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 42); \
  48782.  \
  48783.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48784.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48785.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48786.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48787.   __ret; \
  48788. })
  48789. #endif
  48790.  
  48791. #ifdef __LITTLE_ENDIAN__
  48792. #define vld4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  48793.   int64x2x4_t __ret; \
  48794.   int64x2x4_t __s1 = __p1; \
  48795.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 35); \
  48796.   __ret; \
  48797. })
  48798. #else
  48799. #define vld4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  48800.   int64x2x4_t __ret; \
  48801.   int64x2x4_t __s1 = __p1; \
  48802.   int64x2x4_t __rev1; \
  48803.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  48804.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  48805.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  48806.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  48807.   __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 35); \
  48808.  \
  48809.   __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
  48810.   __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
  48811.   __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
  48812.   __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
  48813.   __ret; \
  48814. })
  48815. #endif
  48816.  
  48817. #define vld4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  48818.   uint64x1x4_t __ret; \
  48819.   uint64x1x4_t __s1 = __p1; \
  48820.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
  48821.   __ret; \
  48822. })
  48823. #define vld4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  48824.   float64x1x4_t __ret; \
  48825.   float64x1x4_t __s1 = __p1; \
  48826.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 10); \
  48827.   __ret; \
  48828. })
  48829. #define vld4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  48830.   int64x1x4_t __ret; \
  48831.   int64x1x4_t __s1 = __p1; \
  48832.   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 3); \
  48833.   __ret; \
  48834. })
  48835. #define vldrq_p128(__p0) __extension__ ({ \
  48836.   poly128_t __ret; \
  48837.   __ret = (poly128_t) __builtin_neon_vldrq_p128(__p0); \
  48838.   __ret; \
  48839. })
  48840. #ifdef __LITTLE_ENDIAN__
  48841. __ai float64x2_t vmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
  48842.   float64x2_t __ret;
  48843.   __ret = (float64x2_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  48844.   return __ret;
  48845. }
  48846. #else
  48847. __ai float64x2_t vmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
  48848.   float64x2_t __ret;
  48849.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  48850.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  48851.   __ret = (float64x2_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  48852.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  48853.   return __ret;
  48854. }
  48855. #endif
  48856.  
  48857. __ai float64x1_t vmax_f64(float64x1_t __p0, float64x1_t __p1) {
  48858.   float64x1_t __ret;
  48859.   __ret = (float64x1_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
  48860.   return __ret;
  48861. }
  48862. #ifdef __LITTLE_ENDIAN__
  48863. __ai float64_t vmaxnmvq_f64(float64x2_t __p0) {
  48864.   float64_t __ret;
  48865.   __ret = (float64_t) __builtin_neon_vmaxnmvq_f64(__p0);
  48866.   return __ret;
  48867. }
  48868. #else
  48869. __ai float64_t vmaxnmvq_f64(float64x2_t __p0) {
  48870.   float64_t __ret;
  48871.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  48872.   __ret = (float64_t) __builtin_neon_vmaxnmvq_f64(__rev0);
  48873.   return __ret;
  48874. }
  48875. #endif
  48876.  
  48877. #ifdef __LITTLE_ENDIAN__
  48878. __ai float32_t vmaxnmvq_f32(float32x4_t __p0) {
  48879.   float32_t __ret;
  48880.   __ret = (float32_t) __builtin_neon_vmaxnmvq_f32(__p0);
  48881.   return __ret;
  48882. }
  48883. #else
  48884. __ai float32_t vmaxnmvq_f32(float32x4_t __p0) {
  48885.   float32_t __ret;
  48886.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  48887.   __ret = (float32_t) __builtin_neon_vmaxnmvq_f32(__rev0);
  48888.   return __ret;
  48889. }
  48890. #endif
  48891.  
  48892. #ifdef __LITTLE_ENDIAN__
  48893. __ai float32_t vmaxnmv_f32(float32x2_t __p0) {
  48894.   float32_t __ret;
  48895.   __ret = (float32_t) __builtin_neon_vmaxnmv_f32(__p0);
  48896.   return __ret;
  48897. }
  48898. #else
  48899. __ai float32_t vmaxnmv_f32(float32x2_t __p0) {
  48900.   float32_t __ret;
  48901.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  48902.   __ret = (float32_t) __builtin_neon_vmaxnmv_f32(__rev0);
  48903.   return __ret;
  48904. }
  48905. #endif
  48906.  
  48907. #ifdef __LITTLE_ENDIAN__
  48908. __ai uint8_t vmaxvq_u8(uint8x16_t __p0) {
  48909.   uint8_t __ret;
  48910.   __ret = (uint8_t) __builtin_neon_vmaxvq_u8(__p0);
  48911.   return __ret;
  48912. }
  48913. #else
  48914. __ai uint8_t vmaxvq_u8(uint8x16_t __p0) {
  48915.   uint8_t __ret;
  48916.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  48917.   __ret = (uint8_t) __builtin_neon_vmaxvq_u8(__rev0);
  48918.   return __ret;
  48919. }
  48920. #endif
  48921.  
  48922. #ifdef __LITTLE_ENDIAN__
  48923. __ai uint32_t vmaxvq_u32(uint32x4_t __p0) {
  48924.   uint32_t __ret;
  48925.   __ret = (uint32_t) __builtin_neon_vmaxvq_u32(__p0);
  48926.   return __ret;
  48927. }
  48928. #else
  48929. __ai uint32_t vmaxvq_u32(uint32x4_t __p0) {
  48930.   uint32_t __ret;
  48931.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  48932.   __ret = (uint32_t) __builtin_neon_vmaxvq_u32(__rev0);
  48933.   return __ret;
  48934. }
  48935. #endif
  48936.  
  48937. #ifdef __LITTLE_ENDIAN__
  48938. __ai uint16_t vmaxvq_u16(uint16x8_t __p0) {
  48939.   uint16_t __ret;
  48940.   __ret = (uint16_t) __builtin_neon_vmaxvq_u16(__p0);
  48941.   return __ret;
  48942. }
  48943. #else
  48944. __ai uint16_t vmaxvq_u16(uint16x8_t __p0) {
  48945.   uint16_t __ret;
  48946.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  48947.   __ret = (uint16_t) __builtin_neon_vmaxvq_u16(__rev0);
  48948.   return __ret;
  48949. }
  48950. #endif
  48951.  
  48952. #ifdef __LITTLE_ENDIAN__
  48953. __ai int8_t vmaxvq_s8(int8x16_t __p0) {
  48954.   int8_t __ret;
  48955.   __ret = (int8_t) __builtin_neon_vmaxvq_s8(__p0);
  48956.   return __ret;
  48957. }
  48958. #else
  48959. __ai int8_t vmaxvq_s8(int8x16_t __p0) {
  48960.   int8_t __ret;
  48961.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  48962.   __ret = (int8_t) __builtin_neon_vmaxvq_s8(__rev0);
  48963.   return __ret;
  48964. }
  48965. #endif
  48966.  
  48967. #ifdef __LITTLE_ENDIAN__
  48968. __ai float64_t vmaxvq_f64(float64x2_t __p0) {
  48969.   float64_t __ret;
  48970.   __ret = (float64_t) __builtin_neon_vmaxvq_f64(__p0);
  48971.   return __ret;
  48972. }
  48973. #else
  48974. __ai float64_t vmaxvq_f64(float64x2_t __p0) {
  48975.   float64_t __ret;
  48976.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  48977.   __ret = (float64_t) __builtin_neon_vmaxvq_f64(__rev0);
  48978.   return __ret;
  48979. }
  48980. #endif
  48981.  
  48982. #ifdef __LITTLE_ENDIAN__
  48983. __ai float32_t vmaxvq_f32(float32x4_t __p0) {
  48984.   float32_t __ret;
  48985.   __ret = (float32_t) __builtin_neon_vmaxvq_f32(__p0);
  48986.   return __ret;
  48987. }
  48988. #else
  48989. __ai float32_t vmaxvq_f32(float32x4_t __p0) {
  48990.   float32_t __ret;
  48991.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  48992.   __ret = (float32_t) __builtin_neon_vmaxvq_f32(__rev0);
  48993.   return __ret;
  48994. }
  48995. #endif
  48996.  
  48997. #ifdef __LITTLE_ENDIAN__
  48998. __ai int32_t vmaxvq_s32(int32x4_t __p0) {
  48999.   int32_t __ret;
  49000.   __ret = (int32_t) __builtin_neon_vmaxvq_s32(__p0);
  49001.   return __ret;
  49002. }
  49003. #else
  49004. __ai int32_t vmaxvq_s32(int32x4_t __p0) {
  49005.   int32_t __ret;
  49006.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  49007.   __ret = (int32_t) __builtin_neon_vmaxvq_s32(__rev0);
  49008.   return __ret;
  49009. }
  49010. #endif
  49011.  
  49012. #ifdef __LITTLE_ENDIAN__
  49013. __ai int16_t vmaxvq_s16(int16x8_t __p0) {
  49014.   int16_t __ret;
  49015.   __ret = (int16_t) __builtin_neon_vmaxvq_s16(__p0);
  49016.   return __ret;
  49017. }
  49018. #else
  49019. __ai int16_t vmaxvq_s16(int16x8_t __p0) {
  49020.   int16_t __ret;
  49021.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  49022.   __ret = (int16_t) __builtin_neon_vmaxvq_s16(__rev0);
  49023.   return __ret;
  49024. }
  49025. #endif
  49026.  
  49027. #ifdef __LITTLE_ENDIAN__
  49028. __ai uint8_t vmaxv_u8(uint8x8_t __p0) {
  49029.   uint8_t __ret;
  49030.   __ret = (uint8_t) __builtin_neon_vmaxv_u8(__p0);
  49031.   return __ret;
  49032. }
  49033. #else
  49034. __ai uint8_t vmaxv_u8(uint8x8_t __p0) {
  49035.   uint8_t __ret;
  49036.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  49037.   __ret = (uint8_t) __builtin_neon_vmaxv_u8(__rev0);
  49038.   return __ret;
  49039. }
  49040. #endif
  49041.  
  49042. #ifdef __LITTLE_ENDIAN__
  49043. __ai uint32_t vmaxv_u32(uint32x2_t __p0) {
  49044.   uint32_t __ret;
  49045.   __ret = (uint32_t) __builtin_neon_vmaxv_u32(__p0);
  49046.   return __ret;
  49047. }
  49048. #else
  49049. __ai uint32_t vmaxv_u32(uint32x2_t __p0) {
  49050.   uint32_t __ret;
  49051.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49052.   __ret = (uint32_t) __builtin_neon_vmaxv_u32(__rev0);
  49053.   return __ret;
  49054. }
  49055. #endif
  49056.  
  49057. #ifdef __LITTLE_ENDIAN__
  49058. __ai uint16_t vmaxv_u16(uint16x4_t __p0) {
  49059.   uint16_t __ret;
  49060.   __ret = (uint16_t) __builtin_neon_vmaxv_u16(__p0);
  49061.   return __ret;
  49062. }
  49063. #else
  49064. __ai uint16_t vmaxv_u16(uint16x4_t __p0) {
  49065.   uint16_t __ret;
  49066.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  49067.   __ret = (uint16_t) __builtin_neon_vmaxv_u16(__rev0);
  49068.   return __ret;
  49069. }
  49070. #endif
  49071.  
  49072. #ifdef __LITTLE_ENDIAN__
  49073. __ai int8_t vmaxv_s8(int8x8_t __p0) {
  49074.   int8_t __ret;
  49075.   __ret = (int8_t) __builtin_neon_vmaxv_s8(__p0);
  49076.   return __ret;
  49077. }
  49078. #else
  49079. __ai int8_t vmaxv_s8(int8x8_t __p0) {
  49080.   int8_t __ret;
  49081.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  49082.   __ret = (int8_t) __builtin_neon_vmaxv_s8(__rev0);
  49083.   return __ret;
  49084. }
  49085. #endif
  49086.  
  49087. #ifdef __LITTLE_ENDIAN__
  49088. __ai float32_t vmaxv_f32(float32x2_t __p0) {
  49089.   float32_t __ret;
  49090.   __ret = (float32_t) __builtin_neon_vmaxv_f32(__p0);
  49091.   return __ret;
  49092. }
  49093. #else
  49094. __ai float32_t vmaxv_f32(float32x2_t __p0) {
  49095.   float32_t __ret;
  49096.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49097.   __ret = (float32_t) __builtin_neon_vmaxv_f32(__rev0);
  49098.   return __ret;
  49099. }
  49100. #endif
  49101.  
  49102. #ifdef __LITTLE_ENDIAN__
  49103. __ai int32_t vmaxv_s32(int32x2_t __p0) {
  49104.   int32_t __ret;
  49105.   __ret = (int32_t) __builtin_neon_vmaxv_s32(__p0);
  49106.   return __ret;
  49107. }
  49108. #else
  49109. __ai int32_t vmaxv_s32(int32x2_t __p0) {
  49110.   int32_t __ret;
  49111.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49112.   __ret = (int32_t) __builtin_neon_vmaxv_s32(__rev0);
  49113.   return __ret;
  49114. }
  49115. #endif
  49116.  
  49117. #ifdef __LITTLE_ENDIAN__
  49118. __ai int16_t vmaxv_s16(int16x4_t __p0) {
  49119.   int16_t __ret;
  49120.   __ret = (int16_t) __builtin_neon_vmaxv_s16(__p0);
  49121.   return __ret;
  49122. }
  49123. #else
  49124. __ai int16_t vmaxv_s16(int16x4_t __p0) {
  49125.   int16_t __ret;
  49126.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  49127.   __ret = (int16_t) __builtin_neon_vmaxv_s16(__rev0);
  49128.   return __ret;
  49129. }
  49130. #endif
  49131.  
  49132. #ifdef __LITTLE_ENDIAN__
  49133. __ai float64x2_t vminq_f64(float64x2_t __p0, float64x2_t __p1) {
  49134.   float64x2_t __ret;
  49135.   __ret = (float64x2_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  49136.   return __ret;
  49137. }
  49138. #else
  49139. __ai float64x2_t vminq_f64(float64x2_t __p0, float64x2_t __p1) {
  49140.   float64x2_t __ret;
  49141.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49142.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  49143.   __ret = (float64x2_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  49144.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  49145.   return __ret;
  49146. }
  49147. #endif
  49148.  
  49149. __ai float64x1_t vmin_f64(float64x1_t __p0, float64x1_t __p1) {
  49150.   float64x1_t __ret;
  49151.   __ret = (float64x1_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
  49152.   return __ret;
  49153. }
  49154. #ifdef __LITTLE_ENDIAN__
  49155. __ai float64_t vminnmvq_f64(float64x2_t __p0) {
  49156.   float64_t __ret;
  49157.   __ret = (float64_t) __builtin_neon_vminnmvq_f64(__p0);
  49158.   return __ret;
  49159. }
  49160. #else
  49161. __ai float64_t vminnmvq_f64(float64x2_t __p0) {
  49162.   float64_t __ret;
  49163.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49164.   __ret = (float64_t) __builtin_neon_vminnmvq_f64(__rev0);
  49165.   return __ret;
  49166. }
  49167. #endif
  49168.  
  49169. #ifdef __LITTLE_ENDIAN__
  49170. __ai float32_t vminnmvq_f32(float32x4_t __p0) {
  49171.   float32_t __ret;
  49172.   __ret = (float32_t) __builtin_neon_vminnmvq_f32(__p0);
  49173.   return __ret;
  49174. }
  49175. #else
  49176. __ai float32_t vminnmvq_f32(float32x4_t __p0) {
  49177.   float32_t __ret;
  49178.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  49179.   __ret = (float32_t) __builtin_neon_vminnmvq_f32(__rev0);
  49180.   return __ret;
  49181. }
  49182. #endif
  49183.  
  49184. #ifdef __LITTLE_ENDIAN__
  49185. __ai float32_t vminnmv_f32(float32x2_t __p0) {
  49186.   float32_t __ret;
  49187.   __ret = (float32_t) __builtin_neon_vminnmv_f32(__p0);
  49188.   return __ret;
  49189. }
  49190. #else
  49191. __ai float32_t vminnmv_f32(float32x2_t __p0) {
  49192.   float32_t __ret;
  49193.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49194.   __ret = (float32_t) __builtin_neon_vminnmv_f32(__rev0);
  49195.   return __ret;
  49196. }
  49197. #endif
  49198.  
  49199. #ifdef __LITTLE_ENDIAN__
  49200. __ai uint8_t vminvq_u8(uint8x16_t __p0) {
  49201.   uint8_t __ret;
  49202.   __ret = (uint8_t) __builtin_neon_vminvq_u8(__p0);
  49203.   return __ret;
  49204. }
  49205. #else
  49206. __ai uint8_t vminvq_u8(uint8x16_t __p0) {
  49207.   uint8_t __ret;
  49208.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  49209.   __ret = (uint8_t) __builtin_neon_vminvq_u8(__rev0);
  49210.   return __ret;
  49211. }
  49212. #endif
  49213.  
  49214. #ifdef __LITTLE_ENDIAN__
  49215. __ai uint32_t vminvq_u32(uint32x4_t __p0) {
  49216.   uint32_t __ret;
  49217.   __ret = (uint32_t) __builtin_neon_vminvq_u32(__p0);
  49218.   return __ret;
  49219. }
  49220. #else
  49221. __ai uint32_t vminvq_u32(uint32x4_t __p0) {
  49222.   uint32_t __ret;
  49223.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  49224.   __ret = (uint32_t) __builtin_neon_vminvq_u32(__rev0);
  49225.   return __ret;
  49226. }
  49227. #endif
  49228.  
  49229. #ifdef __LITTLE_ENDIAN__
  49230. __ai uint16_t vminvq_u16(uint16x8_t __p0) {
  49231.   uint16_t __ret;
  49232.   __ret = (uint16_t) __builtin_neon_vminvq_u16(__p0);
  49233.   return __ret;
  49234. }
  49235. #else
  49236. __ai uint16_t vminvq_u16(uint16x8_t __p0) {
  49237.   uint16_t __ret;
  49238.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  49239.   __ret = (uint16_t) __builtin_neon_vminvq_u16(__rev0);
  49240.   return __ret;
  49241. }
  49242. #endif
  49243.  
  49244. #ifdef __LITTLE_ENDIAN__
  49245. __ai int8_t vminvq_s8(int8x16_t __p0) {
  49246.   int8_t __ret;
  49247.   __ret = (int8_t) __builtin_neon_vminvq_s8(__p0);
  49248.   return __ret;
  49249. }
  49250. #else
  49251. __ai int8_t vminvq_s8(int8x16_t __p0) {
  49252.   int8_t __ret;
  49253.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  49254.   __ret = (int8_t) __builtin_neon_vminvq_s8(__rev0);
  49255.   return __ret;
  49256. }
  49257. #endif
  49258.  
  49259. #ifdef __LITTLE_ENDIAN__
  49260. __ai float64_t vminvq_f64(float64x2_t __p0) {
  49261.   float64_t __ret;
  49262.   __ret = (float64_t) __builtin_neon_vminvq_f64(__p0);
  49263.   return __ret;
  49264. }
  49265. #else
  49266. __ai float64_t vminvq_f64(float64x2_t __p0) {
  49267.   float64_t __ret;
  49268.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49269.   __ret = (float64_t) __builtin_neon_vminvq_f64(__rev0);
  49270.   return __ret;
  49271. }
  49272. #endif
  49273.  
  49274. #ifdef __LITTLE_ENDIAN__
  49275. __ai float32_t vminvq_f32(float32x4_t __p0) {
  49276.   float32_t __ret;
  49277.   __ret = (float32_t) __builtin_neon_vminvq_f32(__p0);
  49278.   return __ret;
  49279. }
  49280. #else
  49281. __ai float32_t vminvq_f32(float32x4_t __p0) {
  49282.   float32_t __ret;
  49283.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  49284.   __ret = (float32_t) __builtin_neon_vminvq_f32(__rev0);
  49285.   return __ret;
  49286. }
  49287. #endif
  49288.  
  49289. #ifdef __LITTLE_ENDIAN__
  49290. __ai int32_t vminvq_s32(int32x4_t __p0) {
  49291.   int32_t __ret;
  49292.   __ret = (int32_t) __builtin_neon_vminvq_s32(__p0);
  49293.   return __ret;
  49294. }
  49295. #else
  49296. __ai int32_t vminvq_s32(int32x4_t __p0) {
  49297.   int32_t __ret;
  49298.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  49299.   __ret = (int32_t) __builtin_neon_vminvq_s32(__rev0);
  49300.   return __ret;
  49301. }
  49302. #endif
  49303.  
  49304. #ifdef __LITTLE_ENDIAN__
  49305. __ai int16_t vminvq_s16(int16x8_t __p0) {
  49306.   int16_t __ret;
  49307.   __ret = (int16_t) __builtin_neon_vminvq_s16(__p0);
  49308.   return __ret;
  49309. }
  49310. #else
  49311. __ai int16_t vminvq_s16(int16x8_t __p0) {
  49312.   int16_t __ret;
  49313.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  49314.   __ret = (int16_t) __builtin_neon_vminvq_s16(__rev0);
  49315.   return __ret;
  49316. }
  49317. #endif
  49318.  
  49319. #ifdef __LITTLE_ENDIAN__
  49320. __ai uint8_t vminv_u8(uint8x8_t __p0) {
  49321.   uint8_t __ret;
  49322.   __ret = (uint8_t) __builtin_neon_vminv_u8(__p0);
  49323.   return __ret;
  49324. }
  49325. #else
  49326. __ai uint8_t vminv_u8(uint8x8_t __p0) {
  49327.   uint8_t __ret;
  49328.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  49329.   __ret = (uint8_t) __builtin_neon_vminv_u8(__rev0);
  49330.   return __ret;
  49331. }
  49332. #endif
  49333.  
  49334. #ifdef __LITTLE_ENDIAN__
  49335. __ai uint32_t vminv_u32(uint32x2_t __p0) {
  49336.   uint32_t __ret;
  49337.   __ret = (uint32_t) __builtin_neon_vminv_u32(__p0);
  49338.   return __ret;
  49339. }
  49340. #else
  49341. __ai uint32_t vminv_u32(uint32x2_t __p0) {
  49342.   uint32_t __ret;
  49343.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49344.   __ret = (uint32_t) __builtin_neon_vminv_u32(__rev0);
  49345.   return __ret;
  49346. }
  49347. #endif
  49348.  
  49349. #ifdef __LITTLE_ENDIAN__
  49350. __ai uint16_t vminv_u16(uint16x4_t __p0) {
  49351.   uint16_t __ret;
  49352.   __ret = (uint16_t) __builtin_neon_vminv_u16(__p0);
  49353.   return __ret;
  49354. }
  49355. #else
  49356. __ai uint16_t vminv_u16(uint16x4_t __p0) {
  49357.   uint16_t __ret;
  49358.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  49359.   __ret = (uint16_t) __builtin_neon_vminv_u16(__rev0);
  49360.   return __ret;
  49361. }
  49362. #endif
  49363.  
  49364. #ifdef __LITTLE_ENDIAN__
  49365. __ai int8_t vminv_s8(int8x8_t __p0) {
  49366.   int8_t __ret;
  49367.   __ret = (int8_t) __builtin_neon_vminv_s8(__p0);
  49368.   return __ret;
  49369. }
  49370. #else
  49371. __ai int8_t vminv_s8(int8x8_t __p0) {
  49372.   int8_t __ret;
  49373.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  49374.   __ret = (int8_t) __builtin_neon_vminv_s8(__rev0);
  49375.   return __ret;
  49376. }
  49377. #endif
  49378.  
  49379. #ifdef __LITTLE_ENDIAN__
  49380. __ai float32_t vminv_f32(float32x2_t __p0) {
  49381.   float32_t __ret;
  49382.   __ret = (float32_t) __builtin_neon_vminv_f32(__p0);
  49383.   return __ret;
  49384. }
  49385. #else
  49386. __ai float32_t vminv_f32(float32x2_t __p0) {
  49387.   float32_t __ret;
  49388.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49389.   __ret = (float32_t) __builtin_neon_vminv_f32(__rev0);
  49390.   return __ret;
  49391. }
  49392. #endif
  49393.  
  49394. #ifdef __LITTLE_ENDIAN__
  49395. __ai int32_t vminv_s32(int32x2_t __p0) {
  49396.   int32_t __ret;
  49397.   __ret = (int32_t) __builtin_neon_vminv_s32(__p0);
  49398.   return __ret;
  49399. }
  49400. #else
  49401. __ai int32_t vminv_s32(int32x2_t __p0) {
  49402.   int32_t __ret;
  49403.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49404.   __ret = (int32_t) __builtin_neon_vminv_s32(__rev0);
  49405.   return __ret;
  49406. }
  49407. #endif
  49408.  
  49409. #ifdef __LITTLE_ENDIAN__
  49410. __ai int16_t vminv_s16(int16x4_t __p0) {
  49411.   int16_t __ret;
  49412.   __ret = (int16_t) __builtin_neon_vminv_s16(__p0);
  49413.   return __ret;
  49414. }
  49415. #else
  49416. __ai int16_t vminv_s16(int16x4_t __p0) {
  49417.   int16_t __ret;
  49418.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  49419.   __ret = (int16_t) __builtin_neon_vminv_s16(__rev0);
  49420.   return __ret;
  49421. }
  49422. #endif
  49423.  
  49424. #ifdef __LITTLE_ENDIAN__
  49425. __ai float64x2_t vmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  49426.   float64x2_t __ret;
  49427.   __ret = __p0 + __p1 * __p2;
  49428.   return __ret;
  49429. }
  49430. #else
  49431. __ai float64x2_t vmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  49432.   float64x2_t __ret;
  49433.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49434.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  49435.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  49436.   __ret = __rev0 + __rev1 * __rev2;
  49437.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  49438.   return __ret;
  49439. }
  49440. #endif
  49441.  
  49442. __ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
  49443.   float64x1_t __ret;
  49444.   __ret = __p0 + __p1 * __p2;
  49445.   return __ret;
  49446. }
  49447. #ifdef __LITTLE_ENDIAN__
  49448. #define vmlaq_laneq_u32(__p0_427, __p1_427, __p2_427, __p3_427) __extension__ ({ \
  49449.   uint32x4_t __ret_427; \
  49450.   uint32x4_t __s0_427 = __p0_427; \
  49451.   uint32x4_t __s1_427 = __p1_427; \
  49452.   uint32x4_t __s2_427 = __p2_427; \
  49453.   __ret_427 = __s0_427 + __s1_427 * splatq_laneq_u32(__s2_427, __p3_427); \
  49454.   __ret_427; \
  49455. })
  49456. #else
  49457. #define vmlaq_laneq_u32(__p0_428, __p1_428, __p2_428, __p3_428) __extension__ ({ \
  49458.   uint32x4_t __ret_428; \
  49459.   uint32x4_t __s0_428 = __p0_428; \
  49460.   uint32x4_t __s1_428 = __p1_428; \
  49461.   uint32x4_t __s2_428 = __p2_428; \
  49462.   uint32x4_t __rev0_428;  __rev0_428 = __builtin_shufflevector(__s0_428, __s0_428, 3, 2, 1, 0); \
  49463.   uint32x4_t __rev1_428;  __rev1_428 = __builtin_shufflevector(__s1_428, __s1_428, 3, 2, 1, 0); \
  49464.   uint32x4_t __rev2_428;  __rev2_428 = __builtin_shufflevector(__s2_428, __s2_428, 3, 2, 1, 0); \
  49465.   __ret_428 = __rev0_428 + __rev1_428 * __noswap_splatq_laneq_u32(__rev2_428, __p3_428); \
  49466.   __ret_428 = __builtin_shufflevector(__ret_428, __ret_428, 3, 2, 1, 0); \
  49467.   __ret_428; \
  49468. })
  49469. #endif
  49470.  
  49471. #ifdef __LITTLE_ENDIAN__
  49472. #define vmlaq_laneq_u16(__p0_429, __p1_429, __p2_429, __p3_429) __extension__ ({ \
  49473.   uint16x8_t __ret_429; \
  49474.   uint16x8_t __s0_429 = __p0_429; \
  49475.   uint16x8_t __s1_429 = __p1_429; \
  49476.   uint16x8_t __s2_429 = __p2_429; \
  49477.   __ret_429 = __s0_429 + __s1_429 * splatq_laneq_u16(__s2_429, __p3_429); \
  49478.   __ret_429; \
  49479. })
  49480. #else
  49481. #define vmlaq_laneq_u16(__p0_430, __p1_430, __p2_430, __p3_430) __extension__ ({ \
  49482.   uint16x8_t __ret_430; \
  49483.   uint16x8_t __s0_430 = __p0_430; \
  49484.   uint16x8_t __s1_430 = __p1_430; \
  49485.   uint16x8_t __s2_430 = __p2_430; \
  49486.   uint16x8_t __rev0_430;  __rev0_430 = __builtin_shufflevector(__s0_430, __s0_430, 7, 6, 5, 4, 3, 2, 1, 0); \
  49487.   uint16x8_t __rev1_430;  __rev1_430 = __builtin_shufflevector(__s1_430, __s1_430, 7, 6, 5, 4, 3, 2, 1, 0); \
  49488.   uint16x8_t __rev2_430;  __rev2_430 = __builtin_shufflevector(__s2_430, __s2_430, 7, 6, 5, 4, 3, 2, 1, 0); \
  49489.   __ret_430 = __rev0_430 + __rev1_430 * __noswap_splatq_laneq_u16(__rev2_430, __p3_430); \
  49490.   __ret_430 = __builtin_shufflevector(__ret_430, __ret_430, 7, 6, 5, 4, 3, 2, 1, 0); \
  49491.   __ret_430; \
  49492. })
  49493. #endif
  49494.  
  49495. #ifdef __LITTLE_ENDIAN__
  49496. #define vmlaq_laneq_f32(__p0_431, __p1_431, __p2_431, __p3_431) __extension__ ({ \
  49497.   float32x4_t __ret_431; \
  49498.   float32x4_t __s0_431 = __p0_431; \
  49499.   float32x4_t __s1_431 = __p1_431; \
  49500.   float32x4_t __s2_431 = __p2_431; \
  49501.   __ret_431 = __s0_431 + __s1_431 * splatq_laneq_f32(__s2_431, __p3_431); \
  49502.   __ret_431; \
  49503. })
  49504. #else
  49505. #define vmlaq_laneq_f32(__p0_432, __p1_432, __p2_432, __p3_432) __extension__ ({ \
  49506.   float32x4_t __ret_432; \
  49507.   float32x4_t __s0_432 = __p0_432; \
  49508.   float32x4_t __s1_432 = __p1_432; \
  49509.   float32x4_t __s2_432 = __p2_432; \
  49510.   float32x4_t __rev0_432;  __rev0_432 = __builtin_shufflevector(__s0_432, __s0_432, 3, 2, 1, 0); \
  49511.   float32x4_t __rev1_432;  __rev1_432 = __builtin_shufflevector(__s1_432, __s1_432, 3, 2, 1, 0); \
  49512.   float32x4_t __rev2_432;  __rev2_432 = __builtin_shufflevector(__s2_432, __s2_432, 3, 2, 1, 0); \
  49513.   __ret_432 = __rev0_432 + __rev1_432 * __noswap_splatq_laneq_f32(__rev2_432, __p3_432); \
  49514.   __ret_432 = __builtin_shufflevector(__ret_432, __ret_432, 3, 2, 1, 0); \
  49515.   __ret_432; \
  49516. })
  49517. #endif
  49518.  
  49519. #ifdef __LITTLE_ENDIAN__
  49520. #define vmlaq_laneq_s32(__p0_433, __p1_433, __p2_433, __p3_433) __extension__ ({ \
  49521.   int32x4_t __ret_433; \
  49522.   int32x4_t __s0_433 = __p0_433; \
  49523.   int32x4_t __s1_433 = __p1_433; \
  49524.   int32x4_t __s2_433 = __p2_433; \
  49525.   __ret_433 = __s0_433 + __s1_433 * splatq_laneq_s32(__s2_433, __p3_433); \
  49526.   __ret_433; \
  49527. })
  49528. #else
  49529. #define vmlaq_laneq_s32(__p0_434, __p1_434, __p2_434, __p3_434) __extension__ ({ \
  49530.   int32x4_t __ret_434; \
  49531.   int32x4_t __s0_434 = __p0_434; \
  49532.   int32x4_t __s1_434 = __p1_434; \
  49533.   int32x4_t __s2_434 = __p2_434; \
  49534.   int32x4_t __rev0_434;  __rev0_434 = __builtin_shufflevector(__s0_434, __s0_434, 3, 2, 1, 0); \
  49535.   int32x4_t __rev1_434;  __rev1_434 = __builtin_shufflevector(__s1_434, __s1_434, 3, 2, 1, 0); \
  49536.   int32x4_t __rev2_434;  __rev2_434 = __builtin_shufflevector(__s2_434, __s2_434, 3, 2, 1, 0); \
  49537.   __ret_434 = __rev0_434 + __rev1_434 * __noswap_splatq_laneq_s32(__rev2_434, __p3_434); \
  49538.   __ret_434 = __builtin_shufflevector(__ret_434, __ret_434, 3, 2, 1, 0); \
  49539.   __ret_434; \
  49540. })
  49541. #endif
  49542.  
  49543. #ifdef __LITTLE_ENDIAN__
  49544. #define vmlaq_laneq_s16(__p0_435, __p1_435, __p2_435, __p3_435) __extension__ ({ \
  49545.   int16x8_t __ret_435; \
  49546.   int16x8_t __s0_435 = __p0_435; \
  49547.   int16x8_t __s1_435 = __p1_435; \
  49548.   int16x8_t __s2_435 = __p2_435; \
  49549.   __ret_435 = __s0_435 + __s1_435 * splatq_laneq_s16(__s2_435, __p3_435); \
  49550.   __ret_435; \
  49551. })
  49552. #else
  49553. #define vmlaq_laneq_s16(__p0_436, __p1_436, __p2_436, __p3_436) __extension__ ({ \
  49554.   int16x8_t __ret_436; \
  49555.   int16x8_t __s0_436 = __p0_436; \
  49556.   int16x8_t __s1_436 = __p1_436; \
  49557.   int16x8_t __s2_436 = __p2_436; \
  49558.   int16x8_t __rev0_436;  __rev0_436 = __builtin_shufflevector(__s0_436, __s0_436, 7, 6, 5, 4, 3, 2, 1, 0); \
  49559.   int16x8_t __rev1_436;  __rev1_436 = __builtin_shufflevector(__s1_436, __s1_436, 7, 6, 5, 4, 3, 2, 1, 0); \
  49560.   int16x8_t __rev2_436;  __rev2_436 = __builtin_shufflevector(__s2_436, __s2_436, 7, 6, 5, 4, 3, 2, 1, 0); \
  49561.   __ret_436 = __rev0_436 + __rev1_436 * __noswap_splatq_laneq_s16(__rev2_436, __p3_436); \
  49562.   __ret_436 = __builtin_shufflevector(__ret_436, __ret_436, 7, 6, 5, 4, 3, 2, 1, 0); \
  49563.   __ret_436; \
  49564. })
  49565. #endif
  49566.  
  49567. #ifdef __LITTLE_ENDIAN__
  49568. #define vmla_laneq_u32(__p0_437, __p1_437, __p2_437, __p3_437) __extension__ ({ \
  49569.   uint32x2_t __ret_437; \
  49570.   uint32x2_t __s0_437 = __p0_437; \
  49571.   uint32x2_t __s1_437 = __p1_437; \
  49572.   uint32x4_t __s2_437 = __p2_437; \
  49573.   __ret_437 = __s0_437 + __s1_437 * splat_laneq_u32(__s2_437, __p3_437); \
  49574.   __ret_437; \
  49575. })
  49576. #else
  49577. #define vmla_laneq_u32(__p0_438, __p1_438, __p2_438, __p3_438) __extension__ ({ \
  49578.   uint32x2_t __ret_438; \
  49579.   uint32x2_t __s0_438 = __p0_438; \
  49580.   uint32x2_t __s1_438 = __p1_438; \
  49581.   uint32x4_t __s2_438 = __p2_438; \
  49582.   uint32x2_t __rev0_438;  __rev0_438 = __builtin_shufflevector(__s0_438, __s0_438, 1, 0); \
  49583.   uint32x2_t __rev1_438;  __rev1_438 = __builtin_shufflevector(__s1_438, __s1_438, 1, 0); \
  49584.   uint32x4_t __rev2_438;  __rev2_438 = __builtin_shufflevector(__s2_438, __s2_438, 3, 2, 1, 0); \
  49585.   __ret_438 = __rev0_438 + __rev1_438 * __noswap_splat_laneq_u32(__rev2_438, __p3_438); \
  49586.   __ret_438 = __builtin_shufflevector(__ret_438, __ret_438, 1, 0); \
  49587.   __ret_438; \
  49588. })
  49589. #endif
  49590.  
  49591. #ifdef __LITTLE_ENDIAN__
  49592. #define vmla_laneq_u16(__p0_439, __p1_439, __p2_439, __p3_439) __extension__ ({ \
  49593.   uint16x4_t __ret_439; \
  49594.   uint16x4_t __s0_439 = __p0_439; \
  49595.   uint16x4_t __s1_439 = __p1_439; \
  49596.   uint16x8_t __s2_439 = __p2_439; \
  49597.   __ret_439 = __s0_439 + __s1_439 * splat_laneq_u16(__s2_439, __p3_439); \
  49598.   __ret_439; \
  49599. })
  49600. #else
  49601. #define vmla_laneq_u16(__p0_440, __p1_440, __p2_440, __p3_440) __extension__ ({ \
  49602.   uint16x4_t __ret_440; \
  49603.   uint16x4_t __s0_440 = __p0_440; \
  49604.   uint16x4_t __s1_440 = __p1_440; \
  49605.   uint16x8_t __s2_440 = __p2_440; \
  49606.   uint16x4_t __rev0_440;  __rev0_440 = __builtin_shufflevector(__s0_440, __s0_440, 3, 2, 1, 0); \
  49607.   uint16x4_t __rev1_440;  __rev1_440 = __builtin_shufflevector(__s1_440, __s1_440, 3, 2, 1, 0); \
  49608.   uint16x8_t __rev2_440;  __rev2_440 = __builtin_shufflevector(__s2_440, __s2_440, 7, 6, 5, 4, 3, 2, 1, 0); \
  49609.   __ret_440 = __rev0_440 + __rev1_440 * __noswap_splat_laneq_u16(__rev2_440, __p3_440); \
  49610.   __ret_440 = __builtin_shufflevector(__ret_440, __ret_440, 3, 2, 1, 0); \
  49611.   __ret_440; \
  49612. })
  49613. #endif
  49614.  
  49615. #ifdef __LITTLE_ENDIAN__
  49616. #define vmla_laneq_f32(__p0_441, __p1_441, __p2_441, __p3_441) __extension__ ({ \
  49617.   float32x2_t __ret_441; \
  49618.   float32x2_t __s0_441 = __p0_441; \
  49619.   float32x2_t __s1_441 = __p1_441; \
  49620.   float32x4_t __s2_441 = __p2_441; \
  49621.   __ret_441 = __s0_441 + __s1_441 * splat_laneq_f32(__s2_441, __p3_441); \
  49622.   __ret_441; \
  49623. })
  49624. #else
  49625. #define vmla_laneq_f32(__p0_442, __p1_442, __p2_442, __p3_442) __extension__ ({ \
  49626.   float32x2_t __ret_442; \
  49627.   float32x2_t __s0_442 = __p0_442; \
  49628.   float32x2_t __s1_442 = __p1_442; \
  49629.   float32x4_t __s2_442 = __p2_442; \
  49630.   float32x2_t __rev0_442;  __rev0_442 = __builtin_shufflevector(__s0_442, __s0_442, 1, 0); \
  49631.   float32x2_t __rev1_442;  __rev1_442 = __builtin_shufflevector(__s1_442, __s1_442, 1, 0); \
  49632.   float32x4_t __rev2_442;  __rev2_442 = __builtin_shufflevector(__s2_442, __s2_442, 3, 2, 1, 0); \
  49633.   __ret_442 = __rev0_442 + __rev1_442 * __noswap_splat_laneq_f32(__rev2_442, __p3_442); \
  49634.   __ret_442 = __builtin_shufflevector(__ret_442, __ret_442, 1, 0); \
  49635.   __ret_442; \
  49636. })
  49637. #endif
  49638.  
  49639. #ifdef __LITTLE_ENDIAN__
  49640. #define vmla_laneq_s32(__p0_443, __p1_443, __p2_443, __p3_443) __extension__ ({ \
  49641.   int32x2_t __ret_443; \
  49642.   int32x2_t __s0_443 = __p0_443; \
  49643.   int32x2_t __s1_443 = __p1_443; \
  49644.   int32x4_t __s2_443 = __p2_443; \
  49645.   __ret_443 = __s0_443 + __s1_443 * splat_laneq_s32(__s2_443, __p3_443); \
  49646.   __ret_443; \
  49647. })
  49648. #else
  49649. #define vmla_laneq_s32(__p0_444, __p1_444, __p2_444, __p3_444) __extension__ ({ \
  49650.   int32x2_t __ret_444; \
  49651.   int32x2_t __s0_444 = __p0_444; \
  49652.   int32x2_t __s1_444 = __p1_444; \
  49653.   int32x4_t __s2_444 = __p2_444; \
  49654.   int32x2_t __rev0_444;  __rev0_444 = __builtin_shufflevector(__s0_444, __s0_444, 1, 0); \
  49655.   int32x2_t __rev1_444;  __rev1_444 = __builtin_shufflevector(__s1_444, __s1_444, 1, 0); \
  49656.   int32x4_t __rev2_444;  __rev2_444 = __builtin_shufflevector(__s2_444, __s2_444, 3, 2, 1, 0); \
  49657.   __ret_444 = __rev0_444 + __rev1_444 * __noswap_splat_laneq_s32(__rev2_444, __p3_444); \
  49658.   __ret_444 = __builtin_shufflevector(__ret_444, __ret_444, 1, 0); \
  49659.   __ret_444; \
  49660. })
  49661. #endif
  49662.  
  49663. #ifdef __LITTLE_ENDIAN__
  49664. #define vmla_laneq_s16(__p0_445, __p1_445, __p2_445, __p3_445) __extension__ ({ \
  49665.   int16x4_t __ret_445; \
  49666.   int16x4_t __s0_445 = __p0_445; \
  49667.   int16x4_t __s1_445 = __p1_445; \
  49668.   int16x8_t __s2_445 = __p2_445; \
  49669.   __ret_445 = __s0_445 + __s1_445 * splat_laneq_s16(__s2_445, __p3_445); \
  49670.   __ret_445; \
  49671. })
  49672. #else
  49673. #define vmla_laneq_s16(__p0_446, __p1_446, __p2_446, __p3_446) __extension__ ({ \
  49674.   int16x4_t __ret_446; \
  49675.   int16x4_t __s0_446 = __p0_446; \
  49676.   int16x4_t __s1_446 = __p1_446; \
  49677.   int16x8_t __s2_446 = __p2_446; \
  49678.   int16x4_t __rev0_446;  __rev0_446 = __builtin_shufflevector(__s0_446, __s0_446, 3, 2, 1, 0); \
  49679.   int16x4_t __rev1_446;  __rev1_446 = __builtin_shufflevector(__s1_446, __s1_446, 3, 2, 1, 0); \
  49680.   int16x8_t __rev2_446;  __rev2_446 = __builtin_shufflevector(__s2_446, __s2_446, 7, 6, 5, 4, 3, 2, 1, 0); \
  49681.   __ret_446 = __rev0_446 + __rev1_446 * __noswap_splat_laneq_s16(__rev2_446, __p3_446); \
  49682.   __ret_446 = __builtin_shufflevector(__ret_446, __ret_446, 3, 2, 1, 0); \
  49683.   __ret_446; \
  49684. })
  49685. #endif
  49686.  
  49687. #ifdef __LITTLE_ENDIAN__
  49688. #define vmlal_high_lane_u32(__p0_447, __p1_447, __p2_447, __p3_447) __extension__ ({ \
  49689.   uint64x2_t __ret_447; \
  49690.   uint64x2_t __s0_447 = __p0_447; \
  49691.   uint32x4_t __s1_447 = __p1_447; \
  49692.   uint32x2_t __s2_447 = __p2_447; \
  49693.   __ret_447 = __s0_447 + vmull_u32(vget_high_u32(__s1_447), splat_lane_u32(__s2_447, __p3_447)); \
  49694.   __ret_447; \
  49695. })
  49696. #else
  49697. #define vmlal_high_lane_u32(__p0_448, __p1_448, __p2_448, __p3_448) __extension__ ({ \
  49698.   uint64x2_t __ret_448; \
  49699.   uint64x2_t __s0_448 = __p0_448; \
  49700.   uint32x4_t __s1_448 = __p1_448; \
  49701.   uint32x2_t __s2_448 = __p2_448; \
  49702.   uint64x2_t __rev0_448;  __rev0_448 = __builtin_shufflevector(__s0_448, __s0_448, 1, 0); \
  49703.   uint32x4_t __rev1_448;  __rev1_448 = __builtin_shufflevector(__s1_448, __s1_448, 3, 2, 1, 0); \
  49704.   uint32x2_t __rev2_448;  __rev2_448 = __builtin_shufflevector(__s2_448, __s2_448, 1, 0); \
  49705.   __ret_448 = __rev0_448 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_448), __noswap_splat_lane_u32(__rev2_448, __p3_448)); \
  49706.   __ret_448 = __builtin_shufflevector(__ret_448, __ret_448, 1, 0); \
  49707.   __ret_448; \
  49708. })
  49709. #endif
  49710.  
  49711. #ifdef __LITTLE_ENDIAN__
  49712. #define vmlal_high_lane_u16(__p0_449, __p1_449, __p2_449, __p3_449) __extension__ ({ \
  49713.   uint32x4_t __ret_449; \
  49714.   uint32x4_t __s0_449 = __p0_449; \
  49715.   uint16x8_t __s1_449 = __p1_449; \
  49716.   uint16x4_t __s2_449 = __p2_449; \
  49717.   __ret_449 = __s0_449 + vmull_u16(vget_high_u16(__s1_449), splat_lane_u16(__s2_449, __p3_449)); \
  49718.   __ret_449; \
  49719. })
  49720. #else
  49721. #define vmlal_high_lane_u16(__p0_450, __p1_450, __p2_450, __p3_450) __extension__ ({ \
  49722.   uint32x4_t __ret_450; \
  49723.   uint32x4_t __s0_450 = __p0_450; \
  49724.   uint16x8_t __s1_450 = __p1_450; \
  49725.   uint16x4_t __s2_450 = __p2_450; \
  49726.   uint32x4_t __rev0_450;  __rev0_450 = __builtin_shufflevector(__s0_450, __s0_450, 3, 2, 1, 0); \
  49727.   uint16x8_t __rev1_450;  __rev1_450 = __builtin_shufflevector(__s1_450, __s1_450, 7, 6, 5, 4, 3, 2, 1, 0); \
  49728.   uint16x4_t __rev2_450;  __rev2_450 = __builtin_shufflevector(__s2_450, __s2_450, 3, 2, 1, 0); \
  49729.   __ret_450 = __rev0_450 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_450), __noswap_splat_lane_u16(__rev2_450, __p3_450)); \
  49730.   __ret_450 = __builtin_shufflevector(__ret_450, __ret_450, 3, 2, 1, 0); \
  49731.   __ret_450; \
  49732. })
  49733. #endif
  49734.  
  49735. #ifdef __LITTLE_ENDIAN__
  49736. #define vmlal_high_lane_s32(__p0_451, __p1_451, __p2_451, __p3_451) __extension__ ({ \
  49737.   int64x2_t __ret_451; \
  49738.   int64x2_t __s0_451 = __p0_451; \
  49739.   int32x4_t __s1_451 = __p1_451; \
  49740.   int32x2_t __s2_451 = __p2_451; \
  49741.   __ret_451 = __s0_451 + vmull_s32(vget_high_s32(__s1_451), splat_lane_s32(__s2_451, __p3_451)); \
  49742.   __ret_451; \
  49743. })
  49744. #else
  49745. #define vmlal_high_lane_s32(__p0_452, __p1_452, __p2_452, __p3_452) __extension__ ({ \
  49746.   int64x2_t __ret_452; \
  49747.   int64x2_t __s0_452 = __p0_452; \
  49748.   int32x4_t __s1_452 = __p1_452; \
  49749.   int32x2_t __s2_452 = __p2_452; \
  49750.   int64x2_t __rev0_452;  __rev0_452 = __builtin_shufflevector(__s0_452, __s0_452, 1, 0); \
  49751.   int32x4_t __rev1_452;  __rev1_452 = __builtin_shufflevector(__s1_452, __s1_452, 3, 2, 1, 0); \
  49752.   int32x2_t __rev2_452;  __rev2_452 = __builtin_shufflevector(__s2_452, __s2_452, 1, 0); \
  49753.   __ret_452 = __rev0_452 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_452), __noswap_splat_lane_s32(__rev2_452, __p3_452)); \
  49754.   __ret_452 = __builtin_shufflevector(__ret_452, __ret_452, 1, 0); \
  49755.   __ret_452; \
  49756. })
  49757. #endif
  49758.  
  49759. #ifdef __LITTLE_ENDIAN__
  49760. #define vmlal_high_lane_s16(__p0_453, __p1_453, __p2_453, __p3_453) __extension__ ({ \
  49761.   int32x4_t __ret_453; \
  49762.   int32x4_t __s0_453 = __p0_453; \
  49763.   int16x8_t __s1_453 = __p1_453; \
  49764.   int16x4_t __s2_453 = __p2_453; \
  49765.   __ret_453 = __s0_453 + vmull_s16(vget_high_s16(__s1_453), splat_lane_s16(__s2_453, __p3_453)); \
  49766.   __ret_453; \
  49767. })
  49768. #else
  49769. #define vmlal_high_lane_s16(__p0_454, __p1_454, __p2_454, __p3_454) __extension__ ({ \
  49770.   int32x4_t __ret_454; \
  49771.   int32x4_t __s0_454 = __p0_454; \
  49772.   int16x8_t __s1_454 = __p1_454; \
  49773.   int16x4_t __s2_454 = __p2_454; \
  49774.   int32x4_t __rev0_454;  __rev0_454 = __builtin_shufflevector(__s0_454, __s0_454, 3, 2, 1, 0); \
  49775.   int16x8_t __rev1_454;  __rev1_454 = __builtin_shufflevector(__s1_454, __s1_454, 7, 6, 5, 4, 3, 2, 1, 0); \
  49776.   int16x4_t __rev2_454;  __rev2_454 = __builtin_shufflevector(__s2_454, __s2_454, 3, 2, 1, 0); \
  49777.   __ret_454 = __rev0_454 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_454), __noswap_splat_lane_s16(__rev2_454, __p3_454)); \
  49778.   __ret_454 = __builtin_shufflevector(__ret_454, __ret_454, 3, 2, 1, 0); \
  49779.   __ret_454; \
  49780. })
  49781. #endif
  49782.  
  49783. #ifdef __LITTLE_ENDIAN__
  49784. #define vmlal_high_laneq_u32(__p0_455, __p1_455, __p2_455, __p3_455) __extension__ ({ \
  49785.   uint64x2_t __ret_455; \
  49786.   uint64x2_t __s0_455 = __p0_455; \
  49787.   uint32x4_t __s1_455 = __p1_455; \
  49788.   uint32x4_t __s2_455 = __p2_455; \
  49789.   __ret_455 = __s0_455 + vmull_u32(vget_high_u32(__s1_455), splat_laneq_u32(__s2_455, __p3_455)); \
  49790.   __ret_455; \
  49791. })
  49792. #else
  49793. #define vmlal_high_laneq_u32(__p0_456, __p1_456, __p2_456, __p3_456) __extension__ ({ \
  49794.   uint64x2_t __ret_456; \
  49795.   uint64x2_t __s0_456 = __p0_456; \
  49796.   uint32x4_t __s1_456 = __p1_456; \
  49797.   uint32x4_t __s2_456 = __p2_456; \
  49798.   uint64x2_t __rev0_456;  __rev0_456 = __builtin_shufflevector(__s0_456, __s0_456, 1, 0); \
  49799.   uint32x4_t __rev1_456;  __rev1_456 = __builtin_shufflevector(__s1_456, __s1_456, 3, 2, 1, 0); \
  49800.   uint32x4_t __rev2_456;  __rev2_456 = __builtin_shufflevector(__s2_456, __s2_456, 3, 2, 1, 0); \
  49801.   __ret_456 = __rev0_456 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_456), __noswap_splat_laneq_u32(__rev2_456, __p3_456)); \
  49802.   __ret_456 = __builtin_shufflevector(__ret_456, __ret_456, 1, 0); \
  49803.   __ret_456; \
  49804. })
  49805. #endif
  49806.  
  49807. #ifdef __LITTLE_ENDIAN__
  49808. #define vmlal_high_laneq_u16(__p0_457, __p1_457, __p2_457, __p3_457) __extension__ ({ \
  49809.   uint32x4_t __ret_457; \
  49810.   uint32x4_t __s0_457 = __p0_457; \
  49811.   uint16x8_t __s1_457 = __p1_457; \
  49812.   uint16x8_t __s2_457 = __p2_457; \
  49813.   __ret_457 = __s0_457 + vmull_u16(vget_high_u16(__s1_457), splat_laneq_u16(__s2_457, __p3_457)); \
  49814.   __ret_457; \
  49815. })
  49816. #else
  49817. #define vmlal_high_laneq_u16(__p0_458, __p1_458, __p2_458, __p3_458) __extension__ ({ \
  49818.   uint32x4_t __ret_458; \
  49819.   uint32x4_t __s0_458 = __p0_458; \
  49820.   uint16x8_t __s1_458 = __p1_458; \
  49821.   uint16x8_t __s2_458 = __p2_458; \
  49822.   uint32x4_t __rev0_458;  __rev0_458 = __builtin_shufflevector(__s0_458, __s0_458, 3, 2, 1, 0); \
  49823.   uint16x8_t __rev1_458;  __rev1_458 = __builtin_shufflevector(__s1_458, __s1_458, 7, 6, 5, 4, 3, 2, 1, 0); \
  49824.   uint16x8_t __rev2_458;  __rev2_458 = __builtin_shufflevector(__s2_458, __s2_458, 7, 6, 5, 4, 3, 2, 1, 0); \
  49825.   __ret_458 = __rev0_458 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_458), __noswap_splat_laneq_u16(__rev2_458, __p3_458)); \
  49826.   __ret_458 = __builtin_shufflevector(__ret_458, __ret_458, 3, 2, 1, 0); \
  49827.   __ret_458; \
  49828. })
  49829. #endif
  49830.  
  49831. #ifdef __LITTLE_ENDIAN__
  49832. #define vmlal_high_laneq_s32(__p0_459, __p1_459, __p2_459, __p3_459) __extension__ ({ \
  49833.   int64x2_t __ret_459; \
  49834.   int64x2_t __s0_459 = __p0_459; \
  49835.   int32x4_t __s1_459 = __p1_459; \
  49836.   int32x4_t __s2_459 = __p2_459; \
  49837.   __ret_459 = __s0_459 + vmull_s32(vget_high_s32(__s1_459), splat_laneq_s32(__s2_459, __p3_459)); \
  49838.   __ret_459; \
  49839. })
  49840. #else
  49841. #define vmlal_high_laneq_s32(__p0_460, __p1_460, __p2_460, __p3_460) __extension__ ({ \
  49842.   int64x2_t __ret_460; \
  49843.   int64x2_t __s0_460 = __p0_460; \
  49844.   int32x4_t __s1_460 = __p1_460; \
  49845.   int32x4_t __s2_460 = __p2_460; \
  49846.   int64x2_t __rev0_460;  __rev0_460 = __builtin_shufflevector(__s0_460, __s0_460, 1, 0); \
  49847.   int32x4_t __rev1_460;  __rev1_460 = __builtin_shufflevector(__s1_460, __s1_460, 3, 2, 1, 0); \
  49848.   int32x4_t __rev2_460;  __rev2_460 = __builtin_shufflevector(__s2_460, __s2_460, 3, 2, 1, 0); \
  49849.   __ret_460 = __rev0_460 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_460), __noswap_splat_laneq_s32(__rev2_460, __p3_460)); \
  49850.   __ret_460 = __builtin_shufflevector(__ret_460, __ret_460, 1, 0); \
  49851.   __ret_460; \
  49852. })
  49853. #endif
  49854.  
  49855. #ifdef __LITTLE_ENDIAN__
  49856. #define vmlal_high_laneq_s16(__p0_461, __p1_461, __p2_461, __p3_461) __extension__ ({ \
  49857.   int32x4_t __ret_461; \
  49858.   int32x4_t __s0_461 = __p0_461; \
  49859.   int16x8_t __s1_461 = __p1_461; \
  49860.   int16x8_t __s2_461 = __p2_461; \
  49861.   __ret_461 = __s0_461 + vmull_s16(vget_high_s16(__s1_461), splat_laneq_s16(__s2_461, __p3_461)); \
  49862.   __ret_461; \
  49863. })
  49864. #else
  49865. #define vmlal_high_laneq_s16(__p0_462, __p1_462, __p2_462, __p3_462) __extension__ ({ \
  49866.   int32x4_t __ret_462; \
  49867.   int32x4_t __s0_462 = __p0_462; \
  49868.   int16x8_t __s1_462 = __p1_462; \
  49869.   int16x8_t __s2_462 = __p2_462; \
  49870.   int32x4_t __rev0_462;  __rev0_462 = __builtin_shufflevector(__s0_462, __s0_462, 3, 2, 1, 0); \
  49871.   int16x8_t __rev1_462;  __rev1_462 = __builtin_shufflevector(__s1_462, __s1_462, 7, 6, 5, 4, 3, 2, 1, 0); \
  49872.   int16x8_t __rev2_462;  __rev2_462 = __builtin_shufflevector(__s2_462, __s2_462, 7, 6, 5, 4, 3, 2, 1, 0); \
  49873.   __ret_462 = __rev0_462 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_462), __noswap_splat_laneq_s16(__rev2_462, __p3_462)); \
  49874.   __ret_462 = __builtin_shufflevector(__ret_462, __ret_462, 3, 2, 1, 0); \
  49875.   __ret_462; \
  49876. })
  49877. #endif
  49878.  
  49879. #ifdef __LITTLE_ENDIAN__
  49880. #define vmlal_laneq_u32(__p0_463, __p1_463, __p2_463, __p3_463) __extension__ ({ \
  49881.   uint64x2_t __ret_463; \
  49882.   uint64x2_t __s0_463 = __p0_463; \
  49883.   uint32x2_t __s1_463 = __p1_463; \
  49884.   uint32x4_t __s2_463 = __p2_463; \
  49885.   __ret_463 = __s0_463 + vmull_u32(__s1_463, splat_laneq_u32(__s2_463, __p3_463)); \
  49886.   __ret_463; \
  49887. })
  49888. #else
  49889. #define vmlal_laneq_u32(__p0_464, __p1_464, __p2_464, __p3_464) __extension__ ({ \
  49890.   uint64x2_t __ret_464; \
  49891.   uint64x2_t __s0_464 = __p0_464; \
  49892.   uint32x2_t __s1_464 = __p1_464; \
  49893.   uint32x4_t __s2_464 = __p2_464; \
  49894.   uint64x2_t __rev0_464;  __rev0_464 = __builtin_shufflevector(__s0_464, __s0_464, 1, 0); \
  49895.   uint32x2_t __rev1_464;  __rev1_464 = __builtin_shufflevector(__s1_464, __s1_464, 1, 0); \
  49896.   uint32x4_t __rev2_464;  __rev2_464 = __builtin_shufflevector(__s2_464, __s2_464, 3, 2, 1, 0); \
  49897.   __ret_464 = __rev0_464 + __noswap_vmull_u32(__rev1_464, __noswap_splat_laneq_u32(__rev2_464, __p3_464)); \
  49898.   __ret_464 = __builtin_shufflevector(__ret_464, __ret_464, 1, 0); \
  49899.   __ret_464; \
  49900. })
  49901. #endif
  49902.  
  49903. #ifdef __LITTLE_ENDIAN__
  49904. #define vmlal_laneq_u16(__p0_465, __p1_465, __p2_465, __p3_465) __extension__ ({ \
  49905.   uint32x4_t __ret_465; \
  49906.   uint32x4_t __s0_465 = __p0_465; \
  49907.   uint16x4_t __s1_465 = __p1_465; \
  49908.   uint16x8_t __s2_465 = __p2_465; \
  49909.   __ret_465 = __s0_465 + vmull_u16(__s1_465, splat_laneq_u16(__s2_465, __p3_465)); \
  49910.   __ret_465; \
  49911. })
  49912. #else
  49913. #define vmlal_laneq_u16(__p0_466, __p1_466, __p2_466, __p3_466) __extension__ ({ \
  49914.   uint32x4_t __ret_466; \
  49915.   uint32x4_t __s0_466 = __p0_466; \
  49916.   uint16x4_t __s1_466 = __p1_466; \
  49917.   uint16x8_t __s2_466 = __p2_466; \
  49918.   uint32x4_t __rev0_466;  __rev0_466 = __builtin_shufflevector(__s0_466, __s0_466, 3, 2, 1, 0); \
  49919.   uint16x4_t __rev1_466;  __rev1_466 = __builtin_shufflevector(__s1_466, __s1_466, 3, 2, 1, 0); \
  49920.   uint16x8_t __rev2_466;  __rev2_466 = __builtin_shufflevector(__s2_466, __s2_466, 7, 6, 5, 4, 3, 2, 1, 0); \
  49921.   __ret_466 = __rev0_466 + __noswap_vmull_u16(__rev1_466, __noswap_splat_laneq_u16(__rev2_466, __p3_466)); \
  49922.   __ret_466 = __builtin_shufflevector(__ret_466, __ret_466, 3, 2, 1, 0); \
  49923.   __ret_466; \
  49924. })
  49925. #endif
  49926.  
  49927. #ifdef __LITTLE_ENDIAN__
  49928. #define vmlal_laneq_s32(__p0_467, __p1_467, __p2_467, __p3_467) __extension__ ({ \
  49929.   int64x2_t __ret_467; \
  49930.   int64x2_t __s0_467 = __p0_467; \
  49931.   int32x2_t __s1_467 = __p1_467; \
  49932.   int32x4_t __s2_467 = __p2_467; \
  49933.   __ret_467 = __s0_467 + vmull_s32(__s1_467, splat_laneq_s32(__s2_467, __p3_467)); \
  49934.   __ret_467; \
  49935. })
  49936. #else
  49937. #define vmlal_laneq_s32(__p0_468, __p1_468, __p2_468, __p3_468) __extension__ ({ \
  49938.   int64x2_t __ret_468; \
  49939.   int64x2_t __s0_468 = __p0_468; \
  49940.   int32x2_t __s1_468 = __p1_468; \
  49941.   int32x4_t __s2_468 = __p2_468; \
  49942.   int64x2_t __rev0_468;  __rev0_468 = __builtin_shufflevector(__s0_468, __s0_468, 1, 0); \
  49943.   int32x2_t __rev1_468;  __rev1_468 = __builtin_shufflevector(__s1_468, __s1_468, 1, 0); \
  49944.   int32x4_t __rev2_468;  __rev2_468 = __builtin_shufflevector(__s2_468, __s2_468, 3, 2, 1, 0); \
  49945.   __ret_468 = __rev0_468 + __noswap_vmull_s32(__rev1_468, __noswap_splat_laneq_s32(__rev2_468, __p3_468)); \
  49946.   __ret_468 = __builtin_shufflevector(__ret_468, __ret_468, 1, 0); \
  49947.   __ret_468; \
  49948. })
  49949. #endif
  49950.  
  49951. #ifdef __LITTLE_ENDIAN__
  49952. #define vmlal_laneq_s16(__p0_469, __p1_469, __p2_469, __p3_469) __extension__ ({ \
  49953.   int32x4_t __ret_469; \
  49954.   int32x4_t __s0_469 = __p0_469; \
  49955.   int16x4_t __s1_469 = __p1_469; \
  49956.   int16x8_t __s2_469 = __p2_469; \
  49957.   __ret_469 = __s0_469 + vmull_s16(__s1_469, splat_laneq_s16(__s2_469, __p3_469)); \
  49958.   __ret_469; \
  49959. })
  49960. #else
  49961. #define vmlal_laneq_s16(__p0_470, __p1_470, __p2_470, __p3_470) __extension__ ({ \
  49962.   int32x4_t __ret_470; \
  49963.   int32x4_t __s0_470 = __p0_470; \
  49964.   int16x4_t __s1_470 = __p1_470; \
  49965.   int16x8_t __s2_470 = __p2_470; \
  49966.   int32x4_t __rev0_470;  __rev0_470 = __builtin_shufflevector(__s0_470, __s0_470, 3, 2, 1, 0); \
  49967.   int16x4_t __rev1_470;  __rev1_470 = __builtin_shufflevector(__s1_470, __s1_470, 3, 2, 1, 0); \
  49968.   int16x8_t __rev2_470;  __rev2_470 = __builtin_shufflevector(__s2_470, __s2_470, 7, 6, 5, 4, 3, 2, 1, 0); \
  49969.   __ret_470 = __rev0_470 + __noswap_vmull_s16(__rev1_470, __noswap_splat_laneq_s16(__rev2_470, __p3_470)); \
  49970.   __ret_470 = __builtin_shufflevector(__ret_470, __ret_470, 3, 2, 1, 0); \
  49971.   __ret_470; \
  49972. })
  49973. #endif
  49974.  
  49975. #ifdef __LITTLE_ENDIAN__
  49976. __ai float64x2_t vmlsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  49977.   float64x2_t __ret;
  49978.   __ret = __p0 - __p1 * __p2;
  49979.   return __ret;
  49980. }
  49981. #else
  49982. __ai float64x2_t vmlsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  49983.   float64x2_t __ret;
  49984.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  49985.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  49986.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  49987.   __ret = __rev0 - __rev1 * __rev2;
  49988.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  49989.   return __ret;
  49990. }
  49991. #endif
  49992.  
  49993. __ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
  49994.   float64x1_t __ret;
  49995.   __ret = __p0 - __p1 * __p2;
  49996.   return __ret;
  49997. }
  49998. #ifdef __LITTLE_ENDIAN__
  49999. #define vmlsq_laneq_u32(__p0_471, __p1_471, __p2_471, __p3_471) __extension__ ({ \
  50000.   uint32x4_t __ret_471; \
  50001.   uint32x4_t __s0_471 = __p0_471; \
  50002.   uint32x4_t __s1_471 = __p1_471; \
  50003.   uint32x4_t __s2_471 = __p2_471; \
  50004.   __ret_471 = __s0_471 - __s1_471 * splatq_laneq_u32(__s2_471, __p3_471); \
  50005.   __ret_471; \
  50006. })
  50007. #else
  50008. #define vmlsq_laneq_u32(__p0_472, __p1_472, __p2_472, __p3_472) __extension__ ({ \
  50009.   uint32x4_t __ret_472; \
  50010.   uint32x4_t __s0_472 = __p0_472; \
  50011.   uint32x4_t __s1_472 = __p1_472; \
  50012.   uint32x4_t __s2_472 = __p2_472; \
  50013.   uint32x4_t __rev0_472;  __rev0_472 = __builtin_shufflevector(__s0_472, __s0_472, 3, 2, 1, 0); \
  50014.   uint32x4_t __rev1_472;  __rev1_472 = __builtin_shufflevector(__s1_472, __s1_472, 3, 2, 1, 0); \
  50015.   uint32x4_t __rev2_472;  __rev2_472 = __builtin_shufflevector(__s2_472, __s2_472, 3, 2, 1, 0); \
  50016.   __ret_472 = __rev0_472 - __rev1_472 * __noswap_splatq_laneq_u32(__rev2_472, __p3_472); \
  50017.   __ret_472 = __builtin_shufflevector(__ret_472, __ret_472, 3, 2, 1, 0); \
  50018.   __ret_472; \
  50019. })
  50020. #endif
  50021.  
  50022. #ifdef __LITTLE_ENDIAN__
  50023. #define vmlsq_laneq_u16(__p0_473, __p1_473, __p2_473, __p3_473) __extension__ ({ \
  50024.   uint16x8_t __ret_473; \
  50025.   uint16x8_t __s0_473 = __p0_473; \
  50026.   uint16x8_t __s1_473 = __p1_473; \
  50027.   uint16x8_t __s2_473 = __p2_473; \
  50028.   __ret_473 = __s0_473 - __s1_473 * splatq_laneq_u16(__s2_473, __p3_473); \
  50029.   __ret_473; \
  50030. })
  50031. #else
  50032. #define vmlsq_laneq_u16(__p0_474, __p1_474, __p2_474, __p3_474) __extension__ ({ \
  50033.   uint16x8_t __ret_474; \
  50034.   uint16x8_t __s0_474 = __p0_474; \
  50035.   uint16x8_t __s1_474 = __p1_474; \
  50036.   uint16x8_t __s2_474 = __p2_474; \
  50037.   uint16x8_t __rev0_474;  __rev0_474 = __builtin_shufflevector(__s0_474, __s0_474, 7, 6, 5, 4, 3, 2, 1, 0); \
  50038.   uint16x8_t __rev1_474;  __rev1_474 = __builtin_shufflevector(__s1_474, __s1_474, 7, 6, 5, 4, 3, 2, 1, 0); \
  50039.   uint16x8_t __rev2_474;  __rev2_474 = __builtin_shufflevector(__s2_474, __s2_474, 7, 6, 5, 4, 3, 2, 1, 0); \
  50040.   __ret_474 = __rev0_474 - __rev1_474 * __noswap_splatq_laneq_u16(__rev2_474, __p3_474); \
  50041.   __ret_474 = __builtin_shufflevector(__ret_474, __ret_474, 7, 6, 5, 4, 3, 2, 1, 0); \
  50042.   __ret_474; \
  50043. })
  50044. #endif
  50045.  
  50046. #ifdef __LITTLE_ENDIAN__
  50047. #define vmlsq_laneq_f32(__p0_475, __p1_475, __p2_475, __p3_475) __extension__ ({ \
  50048.   float32x4_t __ret_475; \
  50049.   float32x4_t __s0_475 = __p0_475; \
  50050.   float32x4_t __s1_475 = __p1_475; \
  50051.   float32x4_t __s2_475 = __p2_475; \
  50052.   __ret_475 = __s0_475 - __s1_475 * splatq_laneq_f32(__s2_475, __p3_475); \
  50053.   __ret_475; \
  50054. })
  50055. #else
  50056. #define vmlsq_laneq_f32(__p0_476, __p1_476, __p2_476, __p3_476) __extension__ ({ \
  50057.   float32x4_t __ret_476; \
  50058.   float32x4_t __s0_476 = __p0_476; \
  50059.   float32x4_t __s1_476 = __p1_476; \
  50060.   float32x4_t __s2_476 = __p2_476; \
  50061.   float32x4_t __rev0_476;  __rev0_476 = __builtin_shufflevector(__s0_476, __s0_476, 3, 2, 1, 0); \
  50062.   float32x4_t __rev1_476;  __rev1_476 = __builtin_shufflevector(__s1_476, __s1_476, 3, 2, 1, 0); \
  50063.   float32x4_t __rev2_476;  __rev2_476 = __builtin_shufflevector(__s2_476, __s2_476, 3, 2, 1, 0); \
  50064.   __ret_476 = __rev0_476 - __rev1_476 * __noswap_splatq_laneq_f32(__rev2_476, __p3_476); \
  50065.   __ret_476 = __builtin_shufflevector(__ret_476, __ret_476, 3, 2, 1, 0); \
  50066.   __ret_476; \
  50067. })
  50068. #endif
  50069.  
  50070. #ifdef __LITTLE_ENDIAN__
  50071. #define vmlsq_laneq_s32(__p0_477, __p1_477, __p2_477, __p3_477) __extension__ ({ \
  50072.   int32x4_t __ret_477; \
  50073.   int32x4_t __s0_477 = __p0_477; \
  50074.   int32x4_t __s1_477 = __p1_477; \
  50075.   int32x4_t __s2_477 = __p2_477; \
  50076.   __ret_477 = __s0_477 - __s1_477 * splatq_laneq_s32(__s2_477, __p3_477); \
  50077.   __ret_477; \
  50078. })
  50079. #else
  50080. #define vmlsq_laneq_s32(__p0_478, __p1_478, __p2_478, __p3_478) __extension__ ({ \
  50081.   int32x4_t __ret_478; \
  50082.   int32x4_t __s0_478 = __p0_478; \
  50083.   int32x4_t __s1_478 = __p1_478; \
  50084.   int32x4_t __s2_478 = __p2_478; \
  50085.   int32x4_t __rev0_478;  __rev0_478 = __builtin_shufflevector(__s0_478, __s0_478, 3, 2, 1, 0); \
  50086.   int32x4_t __rev1_478;  __rev1_478 = __builtin_shufflevector(__s1_478, __s1_478, 3, 2, 1, 0); \
  50087.   int32x4_t __rev2_478;  __rev2_478 = __builtin_shufflevector(__s2_478, __s2_478, 3, 2, 1, 0); \
  50088.   __ret_478 = __rev0_478 - __rev1_478 * __noswap_splatq_laneq_s32(__rev2_478, __p3_478); \
  50089.   __ret_478 = __builtin_shufflevector(__ret_478, __ret_478, 3, 2, 1, 0); \
  50090.   __ret_478; \
  50091. })
  50092. #endif
  50093.  
  50094. #ifdef __LITTLE_ENDIAN__
  50095. #define vmlsq_laneq_s16(__p0_479, __p1_479, __p2_479, __p3_479) __extension__ ({ \
  50096.   int16x8_t __ret_479; \
  50097.   int16x8_t __s0_479 = __p0_479; \
  50098.   int16x8_t __s1_479 = __p1_479; \
  50099.   int16x8_t __s2_479 = __p2_479; \
  50100.   __ret_479 = __s0_479 - __s1_479 * splatq_laneq_s16(__s2_479, __p3_479); \
  50101.   __ret_479; \
  50102. })
  50103. #else
  50104. #define vmlsq_laneq_s16(__p0_480, __p1_480, __p2_480, __p3_480) __extension__ ({ \
  50105.   int16x8_t __ret_480; \
  50106.   int16x8_t __s0_480 = __p0_480; \
  50107.   int16x8_t __s1_480 = __p1_480; \
  50108.   int16x8_t __s2_480 = __p2_480; \
  50109.   int16x8_t __rev0_480;  __rev0_480 = __builtin_shufflevector(__s0_480, __s0_480, 7, 6, 5, 4, 3, 2, 1, 0); \
  50110.   int16x8_t __rev1_480;  __rev1_480 = __builtin_shufflevector(__s1_480, __s1_480, 7, 6, 5, 4, 3, 2, 1, 0); \
  50111.   int16x8_t __rev2_480;  __rev2_480 = __builtin_shufflevector(__s2_480, __s2_480, 7, 6, 5, 4, 3, 2, 1, 0); \
  50112.   __ret_480 = __rev0_480 - __rev1_480 * __noswap_splatq_laneq_s16(__rev2_480, __p3_480); \
  50113.   __ret_480 = __builtin_shufflevector(__ret_480, __ret_480, 7, 6, 5, 4, 3, 2, 1, 0); \
  50114.   __ret_480; \
  50115. })
  50116. #endif
  50117.  
  50118. #ifdef __LITTLE_ENDIAN__
  50119. #define vmls_laneq_u32(__p0_481, __p1_481, __p2_481, __p3_481) __extension__ ({ \
  50120.   uint32x2_t __ret_481; \
  50121.   uint32x2_t __s0_481 = __p0_481; \
  50122.   uint32x2_t __s1_481 = __p1_481; \
  50123.   uint32x4_t __s2_481 = __p2_481; \
  50124.   __ret_481 = __s0_481 - __s1_481 * splat_laneq_u32(__s2_481, __p3_481); \
  50125.   __ret_481; \
  50126. })
  50127. #else
  50128. #define vmls_laneq_u32(__p0_482, __p1_482, __p2_482, __p3_482) __extension__ ({ \
  50129.   uint32x2_t __ret_482; \
  50130.   uint32x2_t __s0_482 = __p0_482; \
  50131.   uint32x2_t __s1_482 = __p1_482; \
  50132.   uint32x4_t __s2_482 = __p2_482; \
  50133.   uint32x2_t __rev0_482;  __rev0_482 = __builtin_shufflevector(__s0_482, __s0_482, 1, 0); \
  50134.   uint32x2_t __rev1_482;  __rev1_482 = __builtin_shufflevector(__s1_482, __s1_482, 1, 0); \
  50135.   uint32x4_t __rev2_482;  __rev2_482 = __builtin_shufflevector(__s2_482, __s2_482, 3, 2, 1, 0); \
  50136.   __ret_482 = __rev0_482 - __rev1_482 * __noswap_splat_laneq_u32(__rev2_482, __p3_482); \
  50137.   __ret_482 = __builtin_shufflevector(__ret_482, __ret_482, 1, 0); \
  50138.   __ret_482; \
  50139. })
  50140. #endif
  50141.  
  50142. #ifdef __LITTLE_ENDIAN__
  50143. #define vmls_laneq_u16(__p0_483, __p1_483, __p2_483, __p3_483) __extension__ ({ \
  50144.   uint16x4_t __ret_483; \
  50145.   uint16x4_t __s0_483 = __p0_483; \
  50146.   uint16x4_t __s1_483 = __p1_483; \
  50147.   uint16x8_t __s2_483 = __p2_483; \
  50148.   __ret_483 = __s0_483 - __s1_483 * splat_laneq_u16(__s2_483, __p3_483); \
  50149.   __ret_483; \
  50150. })
  50151. #else
  50152. #define vmls_laneq_u16(__p0_484, __p1_484, __p2_484, __p3_484) __extension__ ({ \
  50153.   uint16x4_t __ret_484; \
  50154.   uint16x4_t __s0_484 = __p0_484; \
  50155.   uint16x4_t __s1_484 = __p1_484; \
  50156.   uint16x8_t __s2_484 = __p2_484; \
  50157.   uint16x4_t __rev0_484;  __rev0_484 = __builtin_shufflevector(__s0_484, __s0_484, 3, 2, 1, 0); \
  50158.   uint16x4_t __rev1_484;  __rev1_484 = __builtin_shufflevector(__s1_484, __s1_484, 3, 2, 1, 0); \
  50159.   uint16x8_t __rev2_484;  __rev2_484 = __builtin_shufflevector(__s2_484, __s2_484, 7, 6, 5, 4, 3, 2, 1, 0); \
  50160.   __ret_484 = __rev0_484 - __rev1_484 * __noswap_splat_laneq_u16(__rev2_484, __p3_484); \
  50161.   __ret_484 = __builtin_shufflevector(__ret_484, __ret_484, 3, 2, 1, 0); \
  50162.   __ret_484; \
  50163. })
  50164. #endif
  50165.  
  50166. #ifdef __LITTLE_ENDIAN__
  50167. #define vmls_laneq_f32(__p0_485, __p1_485, __p2_485, __p3_485) __extension__ ({ \
  50168.   float32x2_t __ret_485; \
  50169.   float32x2_t __s0_485 = __p0_485; \
  50170.   float32x2_t __s1_485 = __p1_485; \
  50171.   float32x4_t __s2_485 = __p2_485; \
  50172.   __ret_485 = __s0_485 - __s1_485 * splat_laneq_f32(__s2_485, __p3_485); \
  50173.   __ret_485; \
  50174. })
  50175. #else
  50176. #define vmls_laneq_f32(__p0_486, __p1_486, __p2_486, __p3_486) __extension__ ({ \
  50177.   float32x2_t __ret_486; \
  50178.   float32x2_t __s0_486 = __p0_486; \
  50179.   float32x2_t __s1_486 = __p1_486; \
  50180.   float32x4_t __s2_486 = __p2_486; \
  50181.   float32x2_t __rev0_486;  __rev0_486 = __builtin_shufflevector(__s0_486, __s0_486, 1, 0); \
  50182.   float32x2_t __rev1_486;  __rev1_486 = __builtin_shufflevector(__s1_486, __s1_486, 1, 0); \
  50183.   float32x4_t __rev2_486;  __rev2_486 = __builtin_shufflevector(__s2_486, __s2_486, 3, 2, 1, 0); \
  50184.   __ret_486 = __rev0_486 - __rev1_486 * __noswap_splat_laneq_f32(__rev2_486, __p3_486); \
  50185.   __ret_486 = __builtin_shufflevector(__ret_486, __ret_486, 1, 0); \
  50186.   __ret_486; \
  50187. })
  50188. #endif
  50189.  
  50190. #ifdef __LITTLE_ENDIAN__
  50191. #define vmls_laneq_s32(__p0_487, __p1_487, __p2_487, __p3_487) __extension__ ({ \
  50192.   int32x2_t __ret_487; \
  50193.   int32x2_t __s0_487 = __p0_487; \
  50194.   int32x2_t __s1_487 = __p1_487; \
  50195.   int32x4_t __s2_487 = __p2_487; \
  50196.   __ret_487 = __s0_487 - __s1_487 * splat_laneq_s32(__s2_487, __p3_487); \
  50197.   __ret_487; \
  50198. })
  50199. #else
  50200. #define vmls_laneq_s32(__p0_488, __p1_488, __p2_488, __p3_488) __extension__ ({ \
  50201.   int32x2_t __ret_488; \
  50202.   int32x2_t __s0_488 = __p0_488; \
  50203.   int32x2_t __s1_488 = __p1_488; \
  50204.   int32x4_t __s2_488 = __p2_488; \
  50205.   int32x2_t __rev0_488;  __rev0_488 = __builtin_shufflevector(__s0_488, __s0_488, 1, 0); \
  50206.   int32x2_t __rev1_488;  __rev1_488 = __builtin_shufflevector(__s1_488, __s1_488, 1, 0); \
  50207.   int32x4_t __rev2_488;  __rev2_488 = __builtin_shufflevector(__s2_488, __s2_488, 3, 2, 1, 0); \
  50208.   __ret_488 = __rev0_488 - __rev1_488 * __noswap_splat_laneq_s32(__rev2_488, __p3_488); \
  50209.   __ret_488 = __builtin_shufflevector(__ret_488, __ret_488, 1, 0); \
  50210.   __ret_488; \
  50211. })
  50212. #endif
  50213.  
  50214. #ifdef __LITTLE_ENDIAN__
  50215. #define vmls_laneq_s16(__p0_489, __p1_489, __p2_489, __p3_489) __extension__ ({ \
  50216.   int16x4_t __ret_489; \
  50217.   int16x4_t __s0_489 = __p0_489; \
  50218.   int16x4_t __s1_489 = __p1_489; \
  50219.   int16x8_t __s2_489 = __p2_489; \
  50220.   __ret_489 = __s0_489 - __s1_489 * splat_laneq_s16(__s2_489, __p3_489); \
  50221.   __ret_489; \
  50222. })
  50223. #else
  50224. #define vmls_laneq_s16(__p0_490, __p1_490, __p2_490, __p3_490) __extension__ ({ \
  50225.   int16x4_t __ret_490; \
  50226.   int16x4_t __s0_490 = __p0_490; \
  50227.   int16x4_t __s1_490 = __p1_490; \
  50228.   int16x8_t __s2_490 = __p2_490; \
  50229.   int16x4_t __rev0_490;  __rev0_490 = __builtin_shufflevector(__s0_490, __s0_490, 3, 2, 1, 0); \
  50230.   int16x4_t __rev1_490;  __rev1_490 = __builtin_shufflevector(__s1_490, __s1_490, 3, 2, 1, 0); \
  50231.   int16x8_t __rev2_490;  __rev2_490 = __builtin_shufflevector(__s2_490, __s2_490, 7, 6, 5, 4, 3, 2, 1, 0); \
  50232.   __ret_490 = __rev0_490 - __rev1_490 * __noswap_splat_laneq_s16(__rev2_490, __p3_490); \
  50233.   __ret_490 = __builtin_shufflevector(__ret_490, __ret_490, 3, 2, 1, 0); \
  50234.   __ret_490; \
  50235. })
  50236. #endif
  50237.  
  50238. #ifdef __LITTLE_ENDIAN__
  50239. #define vmlsl_high_lane_u32(__p0_491, __p1_491, __p2_491, __p3_491) __extension__ ({ \
  50240.   uint64x2_t __ret_491; \
  50241.   uint64x2_t __s0_491 = __p0_491; \
  50242.   uint32x4_t __s1_491 = __p1_491; \
  50243.   uint32x2_t __s2_491 = __p2_491; \
  50244.   __ret_491 = __s0_491 - vmull_u32(vget_high_u32(__s1_491), splat_lane_u32(__s2_491, __p3_491)); \
  50245.   __ret_491; \
  50246. })
  50247. #else
  50248. #define vmlsl_high_lane_u32(__p0_492, __p1_492, __p2_492, __p3_492) __extension__ ({ \
  50249.   uint64x2_t __ret_492; \
  50250.   uint64x2_t __s0_492 = __p0_492; \
  50251.   uint32x4_t __s1_492 = __p1_492; \
  50252.   uint32x2_t __s2_492 = __p2_492; \
  50253.   uint64x2_t __rev0_492;  __rev0_492 = __builtin_shufflevector(__s0_492, __s0_492, 1, 0); \
  50254.   uint32x4_t __rev1_492;  __rev1_492 = __builtin_shufflevector(__s1_492, __s1_492, 3, 2, 1, 0); \
  50255.   uint32x2_t __rev2_492;  __rev2_492 = __builtin_shufflevector(__s2_492, __s2_492, 1, 0); \
  50256.   __ret_492 = __rev0_492 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_492), __noswap_splat_lane_u32(__rev2_492, __p3_492)); \
  50257.   __ret_492 = __builtin_shufflevector(__ret_492, __ret_492, 1, 0); \
  50258.   __ret_492; \
  50259. })
  50260. #endif
  50261.  
  50262. #ifdef __LITTLE_ENDIAN__
  50263. #define vmlsl_high_lane_u16(__p0_493, __p1_493, __p2_493, __p3_493) __extension__ ({ \
  50264.   uint32x4_t __ret_493; \
  50265.   uint32x4_t __s0_493 = __p0_493; \
  50266.   uint16x8_t __s1_493 = __p1_493; \
  50267.   uint16x4_t __s2_493 = __p2_493; \
  50268.   __ret_493 = __s0_493 - vmull_u16(vget_high_u16(__s1_493), splat_lane_u16(__s2_493, __p3_493)); \
  50269.   __ret_493; \
  50270. })
  50271. #else
  50272. #define vmlsl_high_lane_u16(__p0_494, __p1_494, __p2_494, __p3_494) __extension__ ({ \
  50273.   uint32x4_t __ret_494; \
  50274.   uint32x4_t __s0_494 = __p0_494; \
  50275.   uint16x8_t __s1_494 = __p1_494; \
  50276.   uint16x4_t __s2_494 = __p2_494; \
  50277.   uint32x4_t __rev0_494;  __rev0_494 = __builtin_shufflevector(__s0_494, __s0_494, 3, 2, 1, 0); \
  50278.   uint16x8_t __rev1_494;  __rev1_494 = __builtin_shufflevector(__s1_494, __s1_494, 7, 6, 5, 4, 3, 2, 1, 0); \
  50279.   uint16x4_t __rev2_494;  __rev2_494 = __builtin_shufflevector(__s2_494, __s2_494, 3, 2, 1, 0); \
  50280.   __ret_494 = __rev0_494 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_494), __noswap_splat_lane_u16(__rev2_494, __p3_494)); \
  50281.   __ret_494 = __builtin_shufflevector(__ret_494, __ret_494, 3, 2, 1, 0); \
  50282.   __ret_494; \
  50283. })
  50284. #endif
  50285.  
  50286. #ifdef __LITTLE_ENDIAN__
  50287. #define vmlsl_high_lane_s32(__p0_495, __p1_495, __p2_495, __p3_495) __extension__ ({ \
  50288.   int64x2_t __ret_495; \
  50289.   int64x2_t __s0_495 = __p0_495; \
  50290.   int32x4_t __s1_495 = __p1_495; \
  50291.   int32x2_t __s2_495 = __p2_495; \
  50292.   __ret_495 = __s0_495 - vmull_s32(vget_high_s32(__s1_495), splat_lane_s32(__s2_495, __p3_495)); \
  50293.   __ret_495; \
  50294. })
  50295. #else
  50296. #define vmlsl_high_lane_s32(__p0_496, __p1_496, __p2_496, __p3_496) __extension__ ({ \
  50297.   int64x2_t __ret_496; \
  50298.   int64x2_t __s0_496 = __p0_496; \
  50299.   int32x4_t __s1_496 = __p1_496; \
  50300.   int32x2_t __s2_496 = __p2_496; \
  50301.   int64x2_t __rev0_496;  __rev0_496 = __builtin_shufflevector(__s0_496, __s0_496, 1, 0); \
  50302.   int32x4_t __rev1_496;  __rev1_496 = __builtin_shufflevector(__s1_496, __s1_496, 3, 2, 1, 0); \
  50303.   int32x2_t __rev2_496;  __rev2_496 = __builtin_shufflevector(__s2_496, __s2_496, 1, 0); \
  50304.   __ret_496 = __rev0_496 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_496), __noswap_splat_lane_s32(__rev2_496, __p3_496)); \
  50305.   __ret_496 = __builtin_shufflevector(__ret_496, __ret_496, 1, 0); \
  50306.   __ret_496; \
  50307. })
  50308. #endif
  50309.  
  50310. #ifdef __LITTLE_ENDIAN__
  50311. #define vmlsl_high_lane_s16(__p0_497, __p1_497, __p2_497, __p3_497) __extension__ ({ \
  50312.   int32x4_t __ret_497; \
  50313.   int32x4_t __s0_497 = __p0_497; \
  50314.   int16x8_t __s1_497 = __p1_497; \
  50315.   int16x4_t __s2_497 = __p2_497; \
  50316.   __ret_497 = __s0_497 - vmull_s16(vget_high_s16(__s1_497), splat_lane_s16(__s2_497, __p3_497)); \
  50317.   __ret_497; \
  50318. })
  50319. #else
  50320. #define vmlsl_high_lane_s16(__p0_498, __p1_498, __p2_498, __p3_498) __extension__ ({ \
  50321.   int32x4_t __ret_498; \
  50322.   int32x4_t __s0_498 = __p0_498; \
  50323.   int16x8_t __s1_498 = __p1_498; \
  50324.   int16x4_t __s2_498 = __p2_498; \
  50325.   int32x4_t __rev0_498;  __rev0_498 = __builtin_shufflevector(__s0_498, __s0_498, 3, 2, 1, 0); \
  50326.   int16x8_t __rev1_498;  __rev1_498 = __builtin_shufflevector(__s1_498, __s1_498, 7, 6, 5, 4, 3, 2, 1, 0); \
  50327.   int16x4_t __rev2_498;  __rev2_498 = __builtin_shufflevector(__s2_498, __s2_498, 3, 2, 1, 0); \
  50328.   __ret_498 = __rev0_498 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_498), __noswap_splat_lane_s16(__rev2_498, __p3_498)); \
  50329.   __ret_498 = __builtin_shufflevector(__ret_498, __ret_498, 3, 2, 1, 0); \
  50330.   __ret_498; \
  50331. })
  50332. #endif
  50333.  
  50334. #ifdef __LITTLE_ENDIAN__
  50335. #define vmlsl_high_laneq_u32(__p0_499, __p1_499, __p2_499, __p3_499) __extension__ ({ \
  50336.   uint64x2_t __ret_499; \
  50337.   uint64x2_t __s0_499 = __p0_499; \
  50338.   uint32x4_t __s1_499 = __p1_499; \
  50339.   uint32x4_t __s2_499 = __p2_499; \
  50340.   __ret_499 = __s0_499 - vmull_u32(vget_high_u32(__s1_499), splat_laneq_u32(__s2_499, __p3_499)); \
  50341.   __ret_499; \
  50342. })
  50343. #else
  50344. #define vmlsl_high_laneq_u32(__p0_500, __p1_500, __p2_500, __p3_500) __extension__ ({ \
  50345.   uint64x2_t __ret_500; \
  50346.   uint64x2_t __s0_500 = __p0_500; \
  50347.   uint32x4_t __s1_500 = __p1_500; \
  50348.   uint32x4_t __s2_500 = __p2_500; \
  50349.   uint64x2_t __rev0_500;  __rev0_500 = __builtin_shufflevector(__s0_500, __s0_500, 1, 0); \
  50350.   uint32x4_t __rev1_500;  __rev1_500 = __builtin_shufflevector(__s1_500, __s1_500, 3, 2, 1, 0); \
  50351.   uint32x4_t __rev2_500;  __rev2_500 = __builtin_shufflevector(__s2_500, __s2_500, 3, 2, 1, 0); \
  50352.   __ret_500 = __rev0_500 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_500), __noswap_splat_laneq_u32(__rev2_500, __p3_500)); \
  50353.   __ret_500 = __builtin_shufflevector(__ret_500, __ret_500, 1, 0); \
  50354.   __ret_500; \
  50355. })
  50356. #endif
  50357.  
  50358. #ifdef __LITTLE_ENDIAN__
  50359. #define vmlsl_high_laneq_u16(__p0_501, __p1_501, __p2_501, __p3_501) __extension__ ({ \
  50360.   uint32x4_t __ret_501; \
  50361.   uint32x4_t __s0_501 = __p0_501; \
  50362.   uint16x8_t __s1_501 = __p1_501; \
  50363.   uint16x8_t __s2_501 = __p2_501; \
  50364.   __ret_501 = __s0_501 - vmull_u16(vget_high_u16(__s1_501), splat_laneq_u16(__s2_501, __p3_501)); \
  50365.   __ret_501; \
  50366. })
  50367. #else
  50368. #define vmlsl_high_laneq_u16(__p0_502, __p1_502, __p2_502, __p3_502) __extension__ ({ \
  50369.   uint32x4_t __ret_502; \
  50370.   uint32x4_t __s0_502 = __p0_502; \
  50371.   uint16x8_t __s1_502 = __p1_502; \
  50372.   uint16x8_t __s2_502 = __p2_502; \
  50373.   uint32x4_t __rev0_502;  __rev0_502 = __builtin_shufflevector(__s0_502, __s0_502, 3, 2, 1, 0); \
  50374.   uint16x8_t __rev1_502;  __rev1_502 = __builtin_shufflevector(__s1_502, __s1_502, 7, 6, 5, 4, 3, 2, 1, 0); \
  50375.   uint16x8_t __rev2_502;  __rev2_502 = __builtin_shufflevector(__s2_502, __s2_502, 7, 6, 5, 4, 3, 2, 1, 0); \
  50376.   __ret_502 = __rev0_502 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_502), __noswap_splat_laneq_u16(__rev2_502, __p3_502)); \
  50377.   __ret_502 = __builtin_shufflevector(__ret_502, __ret_502, 3, 2, 1, 0); \
  50378.   __ret_502; \
  50379. })
  50380. #endif
  50381.  
  50382. #ifdef __LITTLE_ENDIAN__
  50383. #define vmlsl_high_laneq_s32(__p0_503, __p1_503, __p2_503, __p3_503) __extension__ ({ \
  50384.   int64x2_t __ret_503; \
  50385.   int64x2_t __s0_503 = __p0_503; \
  50386.   int32x4_t __s1_503 = __p1_503; \
  50387.   int32x4_t __s2_503 = __p2_503; \
  50388.   __ret_503 = __s0_503 - vmull_s32(vget_high_s32(__s1_503), splat_laneq_s32(__s2_503, __p3_503)); \
  50389.   __ret_503; \
  50390. })
  50391. #else
  50392. #define vmlsl_high_laneq_s32(__p0_504, __p1_504, __p2_504, __p3_504) __extension__ ({ \
  50393.   int64x2_t __ret_504; \
  50394.   int64x2_t __s0_504 = __p0_504; \
  50395.   int32x4_t __s1_504 = __p1_504; \
  50396.   int32x4_t __s2_504 = __p2_504; \
  50397.   int64x2_t __rev0_504;  __rev0_504 = __builtin_shufflevector(__s0_504, __s0_504, 1, 0); \
  50398.   int32x4_t __rev1_504;  __rev1_504 = __builtin_shufflevector(__s1_504, __s1_504, 3, 2, 1, 0); \
  50399.   int32x4_t __rev2_504;  __rev2_504 = __builtin_shufflevector(__s2_504, __s2_504, 3, 2, 1, 0); \
  50400.   __ret_504 = __rev0_504 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_504), __noswap_splat_laneq_s32(__rev2_504, __p3_504)); \
  50401.   __ret_504 = __builtin_shufflevector(__ret_504, __ret_504, 1, 0); \
  50402.   __ret_504; \
  50403. })
  50404. #endif
  50405.  
  50406. #ifdef __LITTLE_ENDIAN__
  50407. #define vmlsl_high_laneq_s16(__p0_505, __p1_505, __p2_505, __p3_505) __extension__ ({ \
  50408.   int32x4_t __ret_505; \
  50409.   int32x4_t __s0_505 = __p0_505; \
  50410.   int16x8_t __s1_505 = __p1_505; \
  50411.   int16x8_t __s2_505 = __p2_505; \
  50412.   __ret_505 = __s0_505 - vmull_s16(vget_high_s16(__s1_505), splat_laneq_s16(__s2_505, __p3_505)); \
  50413.   __ret_505; \
  50414. })
  50415. #else
  50416. #define vmlsl_high_laneq_s16(__p0_506, __p1_506, __p2_506, __p3_506) __extension__ ({ \
  50417.   int32x4_t __ret_506; \
  50418.   int32x4_t __s0_506 = __p0_506; \
  50419.   int16x8_t __s1_506 = __p1_506; \
  50420.   int16x8_t __s2_506 = __p2_506; \
  50421.   int32x4_t __rev0_506;  __rev0_506 = __builtin_shufflevector(__s0_506, __s0_506, 3, 2, 1, 0); \
  50422.   int16x8_t __rev1_506;  __rev1_506 = __builtin_shufflevector(__s1_506, __s1_506, 7, 6, 5, 4, 3, 2, 1, 0); \
  50423.   int16x8_t __rev2_506;  __rev2_506 = __builtin_shufflevector(__s2_506, __s2_506, 7, 6, 5, 4, 3, 2, 1, 0); \
  50424.   __ret_506 = __rev0_506 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_506), __noswap_splat_laneq_s16(__rev2_506, __p3_506)); \
  50425.   __ret_506 = __builtin_shufflevector(__ret_506, __ret_506, 3, 2, 1, 0); \
  50426.   __ret_506; \
  50427. })
  50428. #endif
  50429.  
  50430. #ifdef __LITTLE_ENDIAN__
  50431. #define vmlsl_laneq_u32(__p0_507, __p1_507, __p2_507, __p3_507) __extension__ ({ \
  50432.   uint64x2_t __ret_507; \
  50433.   uint64x2_t __s0_507 = __p0_507; \
  50434.   uint32x2_t __s1_507 = __p1_507; \
  50435.   uint32x4_t __s2_507 = __p2_507; \
  50436.   __ret_507 = __s0_507 - vmull_u32(__s1_507, splat_laneq_u32(__s2_507, __p3_507)); \
  50437.   __ret_507; \
  50438. })
  50439. #else
  50440. #define vmlsl_laneq_u32(__p0_508, __p1_508, __p2_508, __p3_508) __extension__ ({ \
  50441.   uint64x2_t __ret_508; \
  50442.   uint64x2_t __s0_508 = __p0_508; \
  50443.   uint32x2_t __s1_508 = __p1_508; \
  50444.   uint32x4_t __s2_508 = __p2_508; \
  50445.   uint64x2_t __rev0_508;  __rev0_508 = __builtin_shufflevector(__s0_508, __s0_508, 1, 0); \
  50446.   uint32x2_t __rev1_508;  __rev1_508 = __builtin_shufflevector(__s1_508, __s1_508, 1, 0); \
  50447.   uint32x4_t __rev2_508;  __rev2_508 = __builtin_shufflevector(__s2_508, __s2_508, 3, 2, 1, 0); \
  50448.   __ret_508 = __rev0_508 - __noswap_vmull_u32(__rev1_508, __noswap_splat_laneq_u32(__rev2_508, __p3_508)); \
  50449.   __ret_508 = __builtin_shufflevector(__ret_508, __ret_508, 1, 0); \
  50450.   __ret_508; \
  50451. })
  50452. #endif
  50453.  
  50454. #ifdef __LITTLE_ENDIAN__
  50455. #define vmlsl_laneq_u16(__p0_509, __p1_509, __p2_509, __p3_509) __extension__ ({ \
  50456.   uint32x4_t __ret_509; \
  50457.   uint32x4_t __s0_509 = __p0_509; \
  50458.   uint16x4_t __s1_509 = __p1_509; \
  50459.   uint16x8_t __s2_509 = __p2_509; \
  50460.   __ret_509 = __s0_509 - vmull_u16(__s1_509, splat_laneq_u16(__s2_509, __p3_509)); \
  50461.   __ret_509; \
  50462. })
  50463. #else
  50464. #define vmlsl_laneq_u16(__p0_510, __p1_510, __p2_510, __p3_510) __extension__ ({ \
  50465.   uint32x4_t __ret_510; \
  50466.   uint32x4_t __s0_510 = __p0_510; \
  50467.   uint16x4_t __s1_510 = __p1_510; \
  50468.   uint16x8_t __s2_510 = __p2_510; \
  50469.   uint32x4_t __rev0_510;  __rev0_510 = __builtin_shufflevector(__s0_510, __s0_510, 3, 2, 1, 0); \
  50470.   uint16x4_t __rev1_510;  __rev1_510 = __builtin_shufflevector(__s1_510, __s1_510, 3, 2, 1, 0); \
  50471.   uint16x8_t __rev2_510;  __rev2_510 = __builtin_shufflevector(__s2_510, __s2_510, 7, 6, 5, 4, 3, 2, 1, 0); \
  50472.   __ret_510 = __rev0_510 - __noswap_vmull_u16(__rev1_510, __noswap_splat_laneq_u16(__rev2_510, __p3_510)); \
  50473.   __ret_510 = __builtin_shufflevector(__ret_510, __ret_510, 3, 2, 1, 0); \
  50474.   __ret_510; \
  50475. })
  50476. #endif
  50477.  
  50478. #ifdef __LITTLE_ENDIAN__
  50479. #define vmlsl_laneq_s32(__p0_511, __p1_511, __p2_511, __p3_511) __extension__ ({ \
  50480.   int64x2_t __ret_511; \
  50481.   int64x2_t __s0_511 = __p0_511; \
  50482.   int32x2_t __s1_511 = __p1_511; \
  50483.   int32x4_t __s2_511 = __p2_511; \
  50484.   __ret_511 = __s0_511 - vmull_s32(__s1_511, splat_laneq_s32(__s2_511, __p3_511)); \
  50485.   __ret_511; \
  50486. })
  50487. #else
  50488. #define vmlsl_laneq_s32(__p0_512, __p1_512, __p2_512, __p3_512) __extension__ ({ \
  50489.   int64x2_t __ret_512; \
  50490.   int64x2_t __s0_512 = __p0_512; \
  50491.   int32x2_t __s1_512 = __p1_512; \
  50492.   int32x4_t __s2_512 = __p2_512; \
  50493.   int64x2_t __rev0_512;  __rev0_512 = __builtin_shufflevector(__s0_512, __s0_512, 1, 0); \
  50494.   int32x2_t __rev1_512;  __rev1_512 = __builtin_shufflevector(__s1_512, __s1_512, 1, 0); \
  50495.   int32x4_t __rev2_512;  __rev2_512 = __builtin_shufflevector(__s2_512, __s2_512, 3, 2, 1, 0); \
  50496.   __ret_512 = __rev0_512 - __noswap_vmull_s32(__rev1_512, __noswap_splat_laneq_s32(__rev2_512, __p3_512)); \
  50497.   __ret_512 = __builtin_shufflevector(__ret_512, __ret_512, 1, 0); \
  50498.   __ret_512; \
  50499. })
  50500. #endif
  50501.  
  50502. #ifdef __LITTLE_ENDIAN__
  50503. #define vmlsl_laneq_s16(__p0_513, __p1_513, __p2_513, __p3_513) __extension__ ({ \
  50504.   int32x4_t __ret_513; \
  50505.   int32x4_t __s0_513 = __p0_513; \
  50506.   int16x4_t __s1_513 = __p1_513; \
  50507.   int16x8_t __s2_513 = __p2_513; \
  50508.   __ret_513 = __s0_513 - vmull_s16(__s1_513, splat_laneq_s16(__s2_513, __p3_513)); \
  50509.   __ret_513; \
  50510. })
  50511. #else
  50512. #define vmlsl_laneq_s16(__p0_514, __p1_514, __p2_514, __p3_514) __extension__ ({ \
  50513.   int32x4_t __ret_514; \
  50514.   int32x4_t __s0_514 = __p0_514; \
  50515.   int16x4_t __s1_514 = __p1_514; \
  50516.   int16x8_t __s2_514 = __p2_514; \
  50517.   int32x4_t __rev0_514;  __rev0_514 = __builtin_shufflevector(__s0_514, __s0_514, 3, 2, 1, 0); \
  50518.   int16x4_t __rev1_514;  __rev1_514 = __builtin_shufflevector(__s1_514, __s1_514, 3, 2, 1, 0); \
  50519.   int16x8_t __rev2_514;  __rev2_514 = __builtin_shufflevector(__s2_514, __s2_514, 7, 6, 5, 4, 3, 2, 1, 0); \
  50520.   __ret_514 = __rev0_514 - __noswap_vmull_s16(__rev1_514, __noswap_splat_laneq_s16(__rev2_514, __p3_514)); \
  50521.   __ret_514 = __builtin_shufflevector(__ret_514, __ret_514, 3, 2, 1, 0); \
  50522.   __ret_514; \
  50523. })
  50524. #endif
  50525.  
  50526. __ai poly64x1_t vmov_n_p64(poly64_t __p0) {
  50527.   poly64x1_t __ret;
  50528.   __ret = (poly64x1_t) {__p0};
  50529.   return __ret;
  50530. }
  50531. #ifdef __LITTLE_ENDIAN__
  50532. __ai poly64x2_t vmovq_n_p64(poly64_t __p0) {
  50533.   poly64x2_t __ret;
  50534.   __ret = (poly64x2_t) {__p0, __p0};
  50535.   return __ret;
  50536. }
  50537. #else
  50538. __ai poly64x2_t vmovq_n_p64(poly64_t __p0) {
  50539.   poly64x2_t __ret;
  50540.   __ret = (poly64x2_t) {__p0, __p0};
  50541.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  50542.   return __ret;
  50543. }
  50544. #endif
  50545.  
  50546. #ifdef __LITTLE_ENDIAN__
  50547. __ai float64x2_t vmovq_n_f64(float64_t __p0) {
  50548.   float64x2_t __ret;
  50549.   __ret = (float64x2_t) {__p0, __p0};
  50550.   return __ret;
  50551. }
  50552. #else
  50553. __ai float64x2_t vmovq_n_f64(float64_t __p0) {
  50554.   float64x2_t __ret;
  50555.   __ret = (float64x2_t) {__p0, __p0};
  50556.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  50557.   return __ret;
  50558. }
  50559. #endif
  50560.  
  50561. __ai float64x1_t vmov_n_f64(float64_t __p0) {
  50562.   float64x1_t __ret;
  50563.   __ret = (float64x1_t) {__p0};
  50564.   return __ret;
  50565. }
  50566. #ifdef __LITTLE_ENDIAN__
  50567. __ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_515) {
  50568.   uint16x8_t __ret_515;
  50569.   uint8x8_t __a1_515 = vget_high_u8(__p0_515);
  50570.   __ret_515 = (uint16x8_t)(vshll_n_u8(__a1_515, 0));
  50571.   return __ret_515;
  50572. }
  50573. #else
  50574. __ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_516) {
  50575.   uint16x8_t __ret_516;
  50576.   uint8x16_t __rev0_516;  __rev0_516 = __builtin_shufflevector(__p0_516, __p0_516, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  50577.   uint8x8_t __a1_516 = __noswap_vget_high_u8(__rev0_516);
  50578.   __ret_516 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_516, 0));
  50579.   __ret_516 = __builtin_shufflevector(__ret_516, __ret_516, 7, 6, 5, 4, 3, 2, 1, 0);
  50580.   return __ret_516;
  50581. }
  50582. __ai uint16x8_t __noswap_vmovl_high_u8(uint8x16_t __p0_517) {
  50583.   uint16x8_t __ret_517;
  50584.   uint8x8_t __a1_517 = __noswap_vget_high_u8(__p0_517);
  50585.   __ret_517 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_517, 0));
  50586.   return __ret_517;
  50587. }
  50588. #endif
  50589.  
  50590. #ifdef __LITTLE_ENDIAN__
  50591. __ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_518) {
  50592.   uint64x2_t __ret_518;
  50593.   uint32x2_t __a1_518 = vget_high_u32(__p0_518);
  50594.   __ret_518 = (uint64x2_t)(vshll_n_u32(__a1_518, 0));
  50595.   return __ret_518;
  50596. }
  50597. #else
  50598. __ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_519) {
  50599.   uint64x2_t __ret_519;
  50600.   uint32x4_t __rev0_519;  __rev0_519 = __builtin_shufflevector(__p0_519, __p0_519, 3, 2, 1, 0);
  50601.   uint32x2_t __a1_519 = __noswap_vget_high_u32(__rev0_519);
  50602.   __ret_519 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_519, 0));
  50603.   __ret_519 = __builtin_shufflevector(__ret_519, __ret_519, 1, 0);
  50604.   return __ret_519;
  50605. }
  50606. __ai uint64x2_t __noswap_vmovl_high_u32(uint32x4_t __p0_520) {
  50607.   uint64x2_t __ret_520;
  50608.   uint32x2_t __a1_520 = __noswap_vget_high_u32(__p0_520);
  50609.   __ret_520 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_520, 0));
  50610.   return __ret_520;
  50611. }
  50612. #endif
  50613.  
  50614. #ifdef __LITTLE_ENDIAN__
  50615. __ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_521) {
  50616.   uint32x4_t __ret_521;
  50617.   uint16x4_t __a1_521 = vget_high_u16(__p0_521);
  50618.   __ret_521 = (uint32x4_t)(vshll_n_u16(__a1_521, 0));
  50619.   return __ret_521;
  50620. }
  50621. #else
  50622. __ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_522) {
  50623.   uint32x4_t __ret_522;
  50624.   uint16x8_t __rev0_522;  __rev0_522 = __builtin_shufflevector(__p0_522, __p0_522, 7, 6, 5, 4, 3, 2, 1, 0);
  50625.   uint16x4_t __a1_522 = __noswap_vget_high_u16(__rev0_522);
  50626.   __ret_522 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_522, 0));
  50627.   __ret_522 = __builtin_shufflevector(__ret_522, __ret_522, 3, 2, 1, 0);
  50628.   return __ret_522;
  50629. }
  50630. __ai uint32x4_t __noswap_vmovl_high_u16(uint16x8_t __p0_523) {
  50631.   uint32x4_t __ret_523;
  50632.   uint16x4_t __a1_523 = __noswap_vget_high_u16(__p0_523);
  50633.   __ret_523 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_523, 0));
  50634.   return __ret_523;
  50635. }
  50636. #endif
  50637.  
  50638. #ifdef __LITTLE_ENDIAN__
  50639. __ai int16x8_t vmovl_high_s8(int8x16_t __p0_524) {
  50640.   int16x8_t __ret_524;
  50641.   int8x8_t __a1_524 = vget_high_s8(__p0_524);
  50642.   __ret_524 = (int16x8_t)(vshll_n_s8(__a1_524, 0));
  50643.   return __ret_524;
  50644. }
  50645. #else
  50646. __ai int16x8_t vmovl_high_s8(int8x16_t __p0_525) {
  50647.   int16x8_t __ret_525;
  50648.   int8x16_t __rev0_525;  __rev0_525 = __builtin_shufflevector(__p0_525, __p0_525, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  50649.   int8x8_t __a1_525 = __noswap_vget_high_s8(__rev0_525);
  50650.   __ret_525 = (int16x8_t)(__noswap_vshll_n_s8(__a1_525, 0));
  50651.   __ret_525 = __builtin_shufflevector(__ret_525, __ret_525, 7, 6, 5, 4, 3, 2, 1, 0);
  50652.   return __ret_525;
  50653. }
  50654. __ai int16x8_t __noswap_vmovl_high_s8(int8x16_t __p0_526) {
  50655.   int16x8_t __ret_526;
  50656.   int8x8_t __a1_526 = __noswap_vget_high_s8(__p0_526);
  50657.   __ret_526 = (int16x8_t)(__noswap_vshll_n_s8(__a1_526, 0));
  50658.   return __ret_526;
  50659. }
  50660. #endif
  50661.  
  50662. #ifdef __LITTLE_ENDIAN__
  50663. __ai int64x2_t vmovl_high_s32(int32x4_t __p0_527) {
  50664.   int64x2_t __ret_527;
  50665.   int32x2_t __a1_527 = vget_high_s32(__p0_527);
  50666.   __ret_527 = (int64x2_t)(vshll_n_s32(__a1_527, 0));
  50667.   return __ret_527;
  50668. }
  50669. #else
  50670. __ai int64x2_t vmovl_high_s32(int32x4_t __p0_528) {
  50671.   int64x2_t __ret_528;
  50672.   int32x4_t __rev0_528;  __rev0_528 = __builtin_shufflevector(__p0_528, __p0_528, 3, 2, 1, 0);
  50673.   int32x2_t __a1_528 = __noswap_vget_high_s32(__rev0_528);
  50674.   __ret_528 = (int64x2_t)(__noswap_vshll_n_s32(__a1_528, 0));
  50675.   __ret_528 = __builtin_shufflevector(__ret_528, __ret_528, 1, 0);
  50676.   return __ret_528;
  50677. }
  50678. __ai int64x2_t __noswap_vmovl_high_s32(int32x4_t __p0_529) {
  50679.   int64x2_t __ret_529;
  50680.   int32x2_t __a1_529 = __noswap_vget_high_s32(__p0_529);
  50681.   __ret_529 = (int64x2_t)(__noswap_vshll_n_s32(__a1_529, 0));
  50682.   return __ret_529;
  50683. }
  50684. #endif
  50685.  
  50686. #ifdef __LITTLE_ENDIAN__
  50687. __ai int32x4_t vmovl_high_s16(int16x8_t __p0_530) {
  50688.   int32x4_t __ret_530;
  50689.   int16x4_t __a1_530 = vget_high_s16(__p0_530);
  50690.   __ret_530 = (int32x4_t)(vshll_n_s16(__a1_530, 0));
  50691.   return __ret_530;
  50692. }
  50693. #else
  50694. __ai int32x4_t vmovl_high_s16(int16x8_t __p0_531) {
  50695.   int32x4_t __ret_531;
  50696.   int16x8_t __rev0_531;  __rev0_531 = __builtin_shufflevector(__p0_531, __p0_531, 7, 6, 5, 4, 3, 2, 1, 0);
  50697.   int16x4_t __a1_531 = __noswap_vget_high_s16(__rev0_531);
  50698.   __ret_531 = (int32x4_t)(__noswap_vshll_n_s16(__a1_531, 0));
  50699.   __ret_531 = __builtin_shufflevector(__ret_531, __ret_531, 3, 2, 1, 0);
  50700.   return __ret_531;
  50701. }
  50702. __ai int32x4_t __noswap_vmovl_high_s16(int16x8_t __p0_532) {
  50703.   int32x4_t __ret_532;
  50704.   int16x4_t __a1_532 = __noswap_vget_high_s16(__p0_532);
  50705.   __ret_532 = (int32x4_t)(__noswap_vshll_n_s16(__a1_532, 0));
  50706.   return __ret_532;
  50707. }
  50708. #endif
  50709.  
  50710. #ifdef __LITTLE_ENDIAN__
  50711. __ai uint16x8_t vmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
  50712.   uint16x8_t __ret;
  50713.   __ret = vcombine_u16(__p0, vmovn_u32(__p1));
  50714.   return __ret;
  50715. }
  50716. #else
  50717. __ai uint16x8_t vmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
  50718.   uint16x8_t __ret;
  50719.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  50720.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  50721.   __ret = __noswap_vcombine_u16(__rev0, __noswap_vmovn_u32(__rev1));
  50722.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  50723.   return __ret;
  50724. }
  50725. #endif
  50726.  
  50727. #ifdef __LITTLE_ENDIAN__
  50728. __ai uint32x4_t vmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
  50729.   uint32x4_t __ret;
  50730.   __ret = vcombine_u32(__p0, vmovn_u64(__p1));
  50731.   return __ret;
  50732. }
  50733. #else
  50734. __ai uint32x4_t vmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
  50735.   uint32x4_t __ret;
  50736.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  50737.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  50738.   __ret = __noswap_vcombine_u32(__rev0, __noswap_vmovn_u64(__rev1));
  50739.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  50740.   return __ret;
  50741. }
  50742. #endif
  50743.  
  50744. #ifdef __LITTLE_ENDIAN__
  50745. __ai uint8x16_t vmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
  50746.   uint8x16_t __ret;
  50747.   __ret = vcombine_u8(__p0, vmovn_u16(__p1));
  50748.   return __ret;
  50749. }
  50750. #else
  50751. __ai uint8x16_t vmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
  50752.   uint8x16_t __ret;
  50753.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  50754.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  50755.   __ret = __noswap_vcombine_u8(__rev0, __noswap_vmovn_u16(__rev1));
  50756.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  50757.   return __ret;
  50758. }
  50759. #endif
  50760.  
  50761. #ifdef __LITTLE_ENDIAN__
  50762. __ai int16x8_t vmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
  50763.   int16x8_t __ret;
  50764.   __ret = vcombine_s16(__p0, vmovn_s32(__p1));
  50765.   return __ret;
  50766. }
  50767. #else
  50768. __ai int16x8_t vmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
  50769.   int16x8_t __ret;
  50770.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  50771.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  50772.   __ret = __noswap_vcombine_s16(__rev0, __noswap_vmovn_s32(__rev1));
  50773.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  50774.   return __ret;
  50775. }
  50776. #endif
  50777.  
  50778. #ifdef __LITTLE_ENDIAN__
  50779. __ai int32x4_t vmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
  50780.   int32x4_t __ret;
  50781.   __ret = vcombine_s32(__p0, vmovn_s64(__p1));
  50782.   return __ret;
  50783. }
  50784. #else
  50785. __ai int32x4_t vmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
  50786.   int32x4_t __ret;
  50787.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  50788.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  50789.   __ret = __noswap_vcombine_s32(__rev0, __noswap_vmovn_s64(__rev1));
  50790.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  50791.   return __ret;
  50792. }
  50793. #endif
  50794.  
  50795. #ifdef __LITTLE_ENDIAN__
  50796. __ai int8x16_t vmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
  50797.   int8x16_t __ret;
  50798.   __ret = vcombine_s8(__p0, vmovn_s16(__p1));
  50799.   return __ret;
  50800. }
  50801. #else
  50802. __ai int8x16_t vmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
  50803.   int8x16_t __ret;
  50804.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  50805.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  50806.   __ret = __noswap_vcombine_s8(__rev0, __noswap_vmovn_s16(__rev1));
  50807.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  50808.   return __ret;
  50809. }
  50810. #endif
  50811.  
  50812. #ifdef __LITTLE_ENDIAN__
  50813. __ai float64x2_t vmulq_f64(float64x2_t __p0, float64x2_t __p1) {
  50814.   float64x2_t __ret;
  50815.   __ret = __p0 * __p1;
  50816.   return __ret;
  50817. }
  50818. #else
  50819. __ai float64x2_t vmulq_f64(float64x2_t __p0, float64x2_t __p1) {
  50820.   float64x2_t __ret;
  50821.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  50822.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  50823.   __ret = __rev0 * __rev1;
  50824.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  50825.   return __ret;
  50826. }
  50827. #endif
  50828.  
  50829. __ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) {
  50830.   float64x1_t __ret;
  50831.   __ret = __p0 * __p1;
  50832.   return __ret;
  50833. }
  50834. #define vmuld_lane_f64(__p0_533, __p1_533, __p2_533) __extension__ ({ \
  50835.   float64_t __ret_533; \
  50836.   float64_t __s0_533 = __p0_533; \
  50837.   float64x1_t __s1_533 = __p1_533; \
  50838.   __ret_533 = __s0_533 * vget_lane_f64(__s1_533, __p2_533); \
  50839.   __ret_533; \
  50840. })
  50841. #ifdef __LITTLE_ENDIAN__
  50842. #define vmuls_lane_f32(__p0_534, __p1_534, __p2_534) __extension__ ({ \
  50843.   float32_t __ret_534; \
  50844.   float32_t __s0_534 = __p0_534; \
  50845.   float32x2_t __s1_534 = __p1_534; \
  50846.   __ret_534 = __s0_534 * vget_lane_f32(__s1_534, __p2_534); \
  50847.   __ret_534; \
  50848. })
  50849. #else
  50850. #define vmuls_lane_f32(__p0_535, __p1_535, __p2_535) __extension__ ({ \
  50851.   float32_t __ret_535; \
  50852.   float32_t __s0_535 = __p0_535; \
  50853.   float32x2_t __s1_535 = __p1_535; \
  50854.   float32x2_t __rev1_535;  __rev1_535 = __builtin_shufflevector(__s1_535, __s1_535, 1, 0); \
  50855.   __ret_535 = __s0_535 * __noswap_vget_lane_f32(__rev1_535, __p2_535); \
  50856.   __ret_535; \
  50857. })
  50858. #endif
  50859.  
  50860. #define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  50861.   float64x1_t __ret; \
  50862.   float64x1_t __s0 = __p0; \
  50863.   float64x1_t __s1 = __p1; \
  50864.   __ret = (float64x1_t) __builtin_neon_vmul_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
  50865.   __ret; \
  50866. })
  50867. #ifdef __LITTLE_ENDIAN__
  50868. #define vmulq_lane_f64(__p0_536, __p1_536, __p2_536) __extension__ ({ \
  50869.   float64x2_t __ret_536; \
  50870.   float64x2_t __s0_536 = __p0_536; \
  50871.   float64x1_t __s1_536 = __p1_536; \
  50872.   __ret_536 = __s0_536 * splatq_lane_f64(__s1_536, __p2_536); \
  50873.   __ret_536; \
  50874. })
  50875. #else
  50876. #define vmulq_lane_f64(__p0_537, __p1_537, __p2_537) __extension__ ({ \
  50877.   float64x2_t __ret_537; \
  50878.   float64x2_t __s0_537 = __p0_537; \
  50879.   float64x1_t __s1_537 = __p1_537; \
  50880.   float64x2_t __rev0_537;  __rev0_537 = __builtin_shufflevector(__s0_537, __s0_537, 1, 0); \
  50881.   __ret_537 = __rev0_537 * __noswap_splatq_lane_f64(__s1_537, __p2_537); \
  50882.   __ret_537 = __builtin_shufflevector(__ret_537, __ret_537, 1, 0); \
  50883.   __ret_537; \
  50884. })
  50885. #endif
  50886.  
  50887. #ifdef __LITTLE_ENDIAN__
  50888. #define vmuld_laneq_f64(__p0_538, __p1_538, __p2_538) __extension__ ({ \
  50889.   float64_t __ret_538; \
  50890.   float64_t __s0_538 = __p0_538; \
  50891.   float64x2_t __s1_538 = __p1_538; \
  50892.   __ret_538 = __s0_538 * vgetq_lane_f64(__s1_538, __p2_538); \
  50893.   __ret_538; \
  50894. })
  50895. #else
  50896. #define vmuld_laneq_f64(__p0_539, __p1_539, __p2_539) __extension__ ({ \
  50897.   float64_t __ret_539; \
  50898.   float64_t __s0_539 = __p0_539; \
  50899.   float64x2_t __s1_539 = __p1_539; \
  50900.   float64x2_t __rev1_539;  __rev1_539 = __builtin_shufflevector(__s1_539, __s1_539, 1, 0); \
  50901.   __ret_539 = __s0_539 * __noswap_vgetq_lane_f64(__rev1_539, __p2_539); \
  50902.   __ret_539; \
  50903. })
  50904. #endif
  50905.  
  50906. #ifdef __LITTLE_ENDIAN__
  50907. #define vmuls_laneq_f32(__p0_540, __p1_540, __p2_540) __extension__ ({ \
  50908.   float32_t __ret_540; \
  50909.   float32_t __s0_540 = __p0_540; \
  50910.   float32x4_t __s1_540 = __p1_540; \
  50911.   __ret_540 = __s0_540 * vgetq_lane_f32(__s1_540, __p2_540); \
  50912.   __ret_540; \
  50913. })
  50914. #else
  50915. #define vmuls_laneq_f32(__p0_541, __p1_541, __p2_541) __extension__ ({ \
  50916.   float32_t __ret_541; \
  50917.   float32_t __s0_541 = __p0_541; \
  50918.   float32x4_t __s1_541 = __p1_541; \
  50919.   float32x4_t __rev1_541;  __rev1_541 = __builtin_shufflevector(__s1_541, __s1_541, 3, 2, 1, 0); \
  50920.   __ret_541 = __s0_541 * __noswap_vgetq_lane_f32(__rev1_541, __p2_541); \
  50921.   __ret_541; \
  50922. })
  50923. #endif
  50924.  
  50925. #ifdef __LITTLE_ENDIAN__
  50926. #define vmul_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
  50927.   float64x1_t __ret; \
  50928.   float64x1_t __s0 = __p0; \
  50929.   float64x2_t __s1 = __p1; \
  50930.   __ret = (float64x1_t) __builtin_neon_vmul_laneq_v((int8x8_t)__s0, (int8x16_t)__s1, __p2, 10); \
  50931.   __ret; \
  50932. })
  50933. #else
  50934. #define vmul_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
  50935.   float64x1_t __ret; \
  50936.   float64x1_t __s0 = __p0; \
  50937.   float64x2_t __s1 = __p1; \
  50938.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  50939.   __ret = (float64x1_t) __builtin_neon_vmul_laneq_v((int8x8_t)__s0, (int8x16_t)__rev1, __p2, 10); \
  50940.   __ret; \
  50941. })
  50942. #endif
  50943.  
  50944. #ifdef __LITTLE_ENDIAN__
  50945. #define vmulq_laneq_u32(__p0_542, __p1_542, __p2_542) __extension__ ({ \
  50946.   uint32x4_t __ret_542; \
  50947.   uint32x4_t __s0_542 = __p0_542; \
  50948.   uint32x4_t __s1_542 = __p1_542; \
  50949.   __ret_542 = __s0_542 * splatq_laneq_u32(__s1_542, __p2_542); \
  50950.   __ret_542; \
  50951. })
  50952. #else
  50953. #define vmulq_laneq_u32(__p0_543, __p1_543, __p2_543) __extension__ ({ \
  50954.   uint32x4_t __ret_543; \
  50955.   uint32x4_t __s0_543 = __p0_543; \
  50956.   uint32x4_t __s1_543 = __p1_543; \
  50957.   uint32x4_t __rev0_543;  __rev0_543 = __builtin_shufflevector(__s0_543, __s0_543, 3, 2, 1, 0); \
  50958.   uint32x4_t __rev1_543;  __rev1_543 = __builtin_shufflevector(__s1_543, __s1_543, 3, 2, 1, 0); \
  50959.   __ret_543 = __rev0_543 * __noswap_splatq_laneq_u32(__rev1_543, __p2_543); \
  50960.   __ret_543 = __builtin_shufflevector(__ret_543, __ret_543, 3, 2, 1, 0); \
  50961.   __ret_543; \
  50962. })
  50963. #endif
  50964.  
  50965. #ifdef __LITTLE_ENDIAN__
  50966. #define vmulq_laneq_u16(__p0_544, __p1_544, __p2_544) __extension__ ({ \
  50967.   uint16x8_t __ret_544; \
  50968.   uint16x8_t __s0_544 = __p0_544; \
  50969.   uint16x8_t __s1_544 = __p1_544; \
  50970.   __ret_544 = __s0_544 * splatq_laneq_u16(__s1_544, __p2_544); \
  50971.   __ret_544; \
  50972. })
  50973. #else
  50974. #define vmulq_laneq_u16(__p0_545, __p1_545, __p2_545) __extension__ ({ \
  50975.   uint16x8_t __ret_545; \
  50976.   uint16x8_t __s0_545 = __p0_545; \
  50977.   uint16x8_t __s1_545 = __p1_545; \
  50978.   uint16x8_t __rev0_545;  __rev0_545 = __builtin_shufflevector(__s0_545, __s0_545, 7, 6, 5, 4, 3, 2, 1, 0); \
  50979.   uint16x8_t __rev1_545;  __rev1_545 = __builtin_shufflevector(__s1_545, __s1_545, 7, 6, 5, 4, 3, 2, 1, 0); \
  50980.   __ret_545 = __rev0_545 * __noswap_splatq_laneq_u16(__rev1_545, __p2_545); \
  50981.   __ret_545 = __builtin_shufflevector(__ret_545, __ret_545, 7, 6, 5, 4, 3, 2, 1, 0); \
  50982.   __ret_545; \
  50983. })
  50984. #endif
  50985.  
  50986. #ifdef __LITTLE_ENDIAN__
  50987. #define vmulq_laneq_f64(__p0_546, __p1_546, __p2_546) __extension__ ({ \
  50988.   float64x2_t __ret_546; \
  50989.   float64x2_t __s0_546 = __p0_546; \
  50990.   float64x2_t __s1_546 = __p1_546; \
  50991.   __ret_546 = __s0_546 * splatq_laneq_f64(__s1_546, __p2_546); \
  50992.   __ret_546; \
  50993. })
  50994. #else
  50995. #define vmulq_laneq_f64(__p0_547, __p1_547, __p2_547) __extension__ ({ \
  50996.   float64x2_t __ret_547; \
  50997.   float64x2_t __s0_547 = __p0_547; \
  50998.   float64x2_t __s1_547 = __p1_547; \
  50999.   float64x2_t __rev0_547;  __rev0_547 = __builtin_shufflevector(__s0_547, __s0_547, 1, 0); \
  51000.   float64x2_t __rev1_547;  __rev1_547 = __builtin_shufflevector(__s1_547, __s1_547, 1, 0); \
  51001.   __ret_547 = __rev0_547 * __noswap_splatq_laneq_f64(__rev1_547, __p2_547); \
  51002.   __ret_547 = __builtin_shufflevector(__ret_547, __ret_547, 1, 0); \
  51003.   __ret_547; \
  51004. })
  51005. #endif
  51006.  
  51007. #ifdef __LITTLE_ENDIAN__
  51008. #define vmulq_laneq_f32(__p0_548, __p1_548, __p2_548) __extension__ ({ \
  51009.   float32x4_t __ret_548; \
  51010.   float32x4_t __s0_548 = __p0_548; \
  51011.   float32x4_t __s1_548 = __p1_548; \
  51012.   __ret_548 = __s0_548 * splatq_laneq_f32(__s1_548, __p2_548); \
  51013.   __ret_548; \
  51014. })
  51015. #else
  51016. #define vmulq_laneq_f32(__p0_549, __p1_549, __p2_549) __extension__ ({ \
  51017.   float32x4_t __ret_549; \
  51018.   float32x4_t __s0_549 = __p0_549; \
  51019.   float32x4_t __s1_549 = __p1_549; \
  51020.   float32x4_t __rev0_549;  __rev0_549 = __builtin_shufflevector(__s0_549, __s0_549, 3, 2, 1, 0); \
  51021.   float32x4_t __rev1_549;  __rev1_549 = __builtin_shufflevector(__s1_549, __s1_549, 3, 2, 1, 0); \
  51022.   __ret_549 = __rev0_549 * __noswap_splatq_laneq_f32(__rev1_549, __p2_549); \
  51023.   __ret_549 = __builtin_shufflevector(__ret_549, __ret_549, 3, 2, 1, 0); \
  51024.   __ret_549; \
  51025. })
  51026. #endif
  51027.  
  51028. #ifdef __LITTLE_ENDIAN__
  51029. #define vmulq_laneq_s32(__p0_550, __p1_550, __p2_550) __extension__ ({ \
  51030.   int32x4_t __ret_550; \
  51031.   int32x4_t __s0_550 = __p0_550; \
  51032.   int32x4_t __s1_550 = __p1_550; \
  51033.   __ret_550 = __s0_550 * splatq_laneq_s32(__s1_550, __p2_550); \
  51034.   __ret_550; \
  51035. })
  51036. #else
  51037. #define vmulq_laneq_s32(__p0_551, __p1_551, __p2_551) __extension__ ({ \
  51038.   int32x4_t __ret_551; \
  51039.   int32x4_t __s0_551 = __p0_551; \
  51040.   int32x4_t __s1_551 = __p1_551; \
  51041.   int32x4_t __rev0_551;  __rev0_551 = __builtin_shufflevector(__s0_551, __s0_551, 3, 2, 1, 0); \
  51042.   int32x4_t __rev1_551;  __rev1_551 = __builtin_shufflevector(__s1_551, __s1_551, 3, 2, 1, 0); \
  51043.   __ret_551 = __rev0_551 * __noswap_splatq_laneq_s32(__rev1_551, __p2_551); \
  51044.   __ret_551 = __builtin_shufflevector(__ret_551, __ret_551, 3, 2, 1, 0); \
  51045.   __ret_551; \
  51046. })
  51047. #endif
  51048.  
  51049. #ifdef __LITTLE_ENDIAN__
  51050. #define vmulq_laneq_s16(__p0_552, __p1_552, __p2_552) __extension__ ({ \
  51051.   int16x8_t __ret_552; \
  51052.   int16x8_t __s0_552 = __p0_552; \
  51053.   int16x8_t __s1_552 = __p1_552; \
  51054.   __ret_552 = __s0_552 * splatq_laneq_s16(__s1_552, __p2_552); \
  51055.   __ret_552; \
  51056. })
  51057. #else
  51058. #define vmulq_laneq_s16(__p0_553, __p1_553, __p2_553) __extension__ ({ \
  51059.   int16x8_t __ret_553; \
  51060.   int16x8_t __s0_553 = __p0_553; \
  51061.   int16x8_t __s1_553 = __p1_553; \
  51062.   int16x8_t __rev0_553;  __rev0_553 = __builtin_shufflevector(__s0_553, __s0_553, 7, 6, 5, 4, 3, 2, 1, 0); \
  51063.   int16x8_t __rev1_553;  __rev1_553 = __builtin_shufflevector(__s1_553, __s1_553, 7, 6, 5, 4, 3, 2, 1, 0); \
  51064.   __ret_553 = __rev0_553 * __noswap_splatq_laneq_s16(__rev1_553, __p2_553); \
  51065.   __ret_553 = __builtin_shufflevector(__ret_553, __ret_553, 7, 6, 5, 4, 3, 2, 1, 0); \
  51066.   __ret_553; \
  51067. })
  51068. #endif
  51069.  
  51070. #ifdef __LITTLE_ENDIAN__
  51071. #define vmul_laneq_u32(__p0_554, __p1_554, __p2_554) __extension__ ({ \
  51072.   uint32x2_t __ret_554; \
  51073.   uint32x2_t __s0_554 = __p0_554; \
  51074.   uint32x4_t __s1_554 = __p1_554; \
  51075.   __ret_554 = __s0_554 * splat_laneq_u32(__s1_554, __p2_554); \
  51076.   __ret_554; \
  51077. })
  51078. #else
  51079. #define vmul_laneq_u32(__p0_555, __p1_555, __p2_555) __extension__ ({ \
  51080.   uint32x2_t __ret_555; \
  51081.   uint32x2_t __s0_555 = __p0_555; \
  51082.   uint32x4_t __s1_555 = __p1_555; \
  51083.   uint32x2_t __rev0_555;  __rev0_555 = __builtin_shufflevector(__s0_555, __s0_555, 1, 0); \
  51084.   uint32x4_t __rev1_555;  __rev1_555 = __builtin_shufflevector(__s1_555, __s1_555, 3, 2, 1, 0); \
  51085.   __ret_555 = __rev0_555 * __noswap_splat_laneq_u32(__rev1_555, __p2_555); \
  51086.   __ret_555 = __builtin_shufflevector(__ret_555, __ret_555, 1, 0); \
  51087.   __ret_555; \
  51088. })
  51089. #endif
  51090.  
  51091. #ifdef __LITTLE_ENDIAN__
  51092. #define vmul_laneq_u16(__p0_556, __p1_556, __p2_556) __extension__ ({ \
  51093.   uint16x4_t __ret_556; \
  51094.   uint16x4_t __s0_556 = __p0_556; \
  51095.   uint16x8_t __s1_556 = __p1_556; \
  51096.   __ret_556 = __s0_556 * splat_laneq_u16(__s1_556, __p2_556); \
  51097.   __ret_556; \
  51098. })
  51099. #else
  51100. #define vmul_laneq_u16(__p0_557, __p1_557, __p2_557) __extension__ ({ \
  51101.   uint16x4_t __ret_557; \
  51102.   uint16x4_t __s0_557 = __p0_557; \
  51103.   uint16x8_t __s1_557 = __p1_557; \
  51104.   uint16x4_t __rev0_557;  __rev0_557 = __builtin_shufflevector(__s0_557, __s0_557, 3, 2, 1, 0); \
  51105.   uint16x8_t __rev1_557;  __rev1_557 = __builtin_shufflevector(__s1_557, __s1_557, 7, 6, 5, 4, 3, 2, 1, 0); \
  51106.   __ret_557 = __rev0_557 * __noswap_splat_laneq_u16(__rev1_557, __p2_557); \
  51107.   __ret_557 = __builtin_shufflevector(__ret_557, __ret_557, 3, 2, 1, 0); \
  51108.   __ret_557; \
  51109. })
  51110. #endif
  51111.  
  51112. #ifdef __LITTLE_ENDIAN__
  51113. #define vmul_laneq_f32(__p0_558, __p1_558, __p2_558) __extension__ ({ \
  51114.   float32x2_t __ret_558; \
  51115.   float32x2_t __s0_558 = __p0_558; \
  51116.   float32x4_t __s1_558 = __p1_558; \
  51117.   __ret_558 = __s0_558 * splat_laneq_f32(__s1_558, __p2_558); \
  51118.   __ret_558; \
  51119. })
  51120. #else
  51121. #define vmul_laneq_f32(__p0_559, __p1_559, __p2_559) __extension__ ({ \
  51122.   float32x2_t __ret_559; \
  51123.   float32x2_t __s0_559 = __p0_559; \
  51124.   float32x4_t __s1_559 = __p1_559; \
  51125.   float32x2_t __rev0_559;  __rev0_559 = __builtin_shufflevector(__s0_559, __s0_559, 1, 0); \
  51126.   float32x4_t __rev1_559;  __rev1_559 = __builtin_shufflevector(__s1_559, __s1_559, 3, 2, 1, 0); \
  51127.   __ret_559 = __rev0_559 * __noswap_splat_laneq_f32(__rev1_559, __p2_559); \
  51128.   __ret_559 = __builtin_shufflevector(__ret_559, __ret_559, 1, 0); \
  51129.   __ret_559; \
  51130. })
  51131. #endif
  51132.  
  51133. #ifdef __LITTLE_ENDIAN__
  51134. #define vmul_laneq_s32(__p0_560, __p1_560, __p2_560) __extension__ ({ \
  51135.   int32x2_t __ret_560; \
  51136.   int32x2_t __s0_560 = __p0_560; \
  51137.   int32x4_t __s1_560 = __p1_560; \
  51138.   __ret_560 = __s0_560 * splat_laneq_s32(__s1_560, __p2_560); \
  51139.   __ret_560; \
  51140. })
  51141. #else
  51142. #define vmul_laneq_s32(__p0_561, __p1_561, __p2_561) __extension__ ({ \
  51143.   int32x2_t __ret_561; \
  51144.   int32x2_t __s0_561 = __p0_561; \
  51145.   int32x4_t __s1_561 = __p1_561; \
  51146.   int32x2_t __rev0_561;  __rev0_561 = __builtin_shufflevector(__s0_561, __s0_561, 1, 0); \
  51147.   int32x4_t __rev1_561;  __rev1_561 = __builtin_shufflevector(__s1_561, __s1_561, 3, 2, 1, 0); \
  51148.   __ret_561 = __rev0_561 * __noswap_splat_laneq_s32(__rev1_561, __p2_561); \
  51149.   __ret_561 = __builtin_shufflevector(__ret_561, __ret_561, 1, 0); \
  51150.   __ret_561; \
  51151. })
  51152. #endif
  51153.  
  51154. #ifdef __LITTLE_ENDIAN__
  51155. #define vmul_laneq_s16(__p0_562, __p1_562, __p2_562) __extension__ ({ \
  51156.   int16x4_t __ret_562; \
  51157.   int16x4_t __s0_562 = __p0_562; \
  51158.   int16x8_t __s1_562 = __p1_562; \
  51159.   __ret_562 = __s0_562 * splat_laneq_s16(__s1_562, __p2_562); \
  51160.   __ret_562; \
  51161. })
  51162. #else
  51163. #define vmul_laneq_s16(__p0_563, __p1_563, __p2_563) __extension__ ({ \
  51164.   int16x4_t __ret_563; \
  51165.   int16x4_t __s0_563 = __p0_563; \
  51166.   int16x8_t __s1_563 = __p1_563; \
  51167.   int16x4_t __rev0_563;  __rev0_563 = __builtin_shufflevector(__s0_563, __s0_563, 3, 2, 1, 0); \
  51168.   int16x8_t __rev1_563;  __rev1_563 = __builtin_shufflevector(__s1_563, __s1_563, 7, 6, 5, 4, 3, 2, 1, 0); \
  51169.   __ret_563 = __rev0_563 * __noswap_splat_laneq_s16(__rev1_563, __p2_563); \
  51170.   __ret_563 = __builtin_shufflevector(__ret_563, __ret_563, 3, 2, 1, 0); \
  51171.   __ret_563; \
  51172. })
  51173. #endif
  51174.  
  51175. __ai float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1) {
  51176.   float64x1_t __ret;
  51177.   __ret = (float64x1_t) __builtin_neon_vmul_n_f64((float64x1_t)__p0, __p1);
  51178.   return __ret;
  51179. }
  51180. #ifdef __LITTLE_ENDIAN__
  51181. __ai float64x2_t vmulq_n_f64(float64x2_t __p0, float64_t __p1) {
  51182.   float64x2_t __ret;
  51183.   __ret = __p0 * (float64x2_t) {__p1, __p1};
  51184.   return __ret;
  51185. }
  51186. #else
  51187. __ai float64x2_t vmulq_n_f64(float64x2_t __p0, float64_t __p1) {
  51188.   float64x2_t __ret;
  51189.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  51190.   __ret = __rev0 * (float64x2_t) {__p1, __p1};
  51191.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51192.   return __ret;
  51193. }
  51194. #endif
  51195.  
  51196. #ifdef __LITTLE_ENDIAN__
  51197. __ai poly16x8_t vmull_high_p8(poly8x16_t __p0, poly8x16_t __p1) {
  51198.   poly16x8_t __ret;
  51199.   __ret = vmull_p8(vget_high_p8(__p0), vget_high_p8(__p1));
  51200.   return __ret;
  51201. }
  51202. #else
  51203. __ai poly16x8_t vmull_high_p8(poly8x16_t __p0, poly8x16_t __p1) {
  51204.   poly16x8_t __ret;
  51205.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  51206.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  51207.   __ret = __noswap_vmull_p8(__noswap_vget_high_p8(__rev0), __noswap_vget_high_p8(__rev1));
  51208.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  51209.   return __ret;
  51210. }
  51211. #endif
  51212.  
  51213. #ifdef __LITTLE_ENDIAN__
  51214. __ai uint16x8_t vmull_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
  51215.   uint16x8_t __ret;
  51216.   __ret = vmull_u8(vget_high_u8(__p0), vget_high_u8(__p1));
  51217.   return __ret;
  51218. }
  51219. #else
  51220. __ai uint16x8_t vmull_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
  51221.   uint16x8_t __ret;
  51222.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  51223.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  51224.   __ret = __noswap_vmull_u8(__noswap_vget_high_u8(__rev0), __noswap_vget_high_u8(__rev1));
  51225.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  51226.   return __ret;
  51227. }
  51228. #endif
  51229.  
  51230. #ifdef __LITTLE_ENDIAN__
  51231. __ai uint64x2_t vmull_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
  51232.   uint64x2_t __ret;
  51233.   __ret = vmull_u32(vget_high_u32(__p0), vget_high_u32(__p1));
  51234.   return __ret;
  51235. }
  51236. #else
  51237. __ai uint64x2_t vmull_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
  51238.   uint64x2_t __ret;
  51239.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  51240.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  51241.   __ret = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0), __noswap_vget_high_u32(__rev1));
  51242.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51243.   return __ret;
  51244. }
  51245. #endif
  51246.  
  51247. #ifdef __LITTLE_ENDIAN__
  51248. __ai uint32x4_t vmull_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
  51249.   uint32x4_t __ret;
  51250.   __ret = vmull_u16(vget_high_u16(__p0), vget_high_u16(__p1));
  51251.   return __ret;
  51252. }
  51253. #else
  51254. __ai uint32x4_t vmull_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
  51255.   uint32x4_t __ret;
  51256.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  51257.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  51258.   __ret = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0), __noswap_vget_high_u16(__rev1));
  51259.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  51260.   return __ret;
  51261. }
  51262. #endif
  51263.  
  51264. #ifdef __LITTLE_ENDIAN__
  51265. __ai int16x8_t vmull_high_s8(int8x16_t __p0, int8x16_t __p1) {
  51266.   int16x8_t __ret;
  51267.   __ret = vmull_s8(vget_high_s8(__p0), vget_high_s8(__p1));
  51268.   return __ret;
  51269. }
  51270. #else
  51271. __ai int16x8_t vmull_high_s8(int8x16_t __p0, int8x16_t __p1) {
  51272.   int16x8_t __ret;
  51273.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  51274.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  51275.   __ret = __noswap_vmull_s8(__noswap_vget_high_s8(__rev0), __noswap_vget_high_s8(__rev1));
  51276.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  51277.   return __ret;
  51278. }
  51279. #endif
  51280.  
  51281. #ifdef __LITTLE_ENDIAN__
  51282. __ai int64x2_t vmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
  51283.   int64x2_t __ret;
  51284.   __ret = vmull_s32(vget_high_s32(__p0), vget_high_s32(__p1));
  51285.   return __ret;
  51286. }
  51287. #else
  51288. __ai int64x2_t vmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
  51289.   int64x2_t __ret;
  51290.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  51291.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  51292.   __ret = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0), __noswap_vget_high_s32(__rev1));
  51293.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51294.   return __ret;
  51295. }
  51296. #endif
  51297.  
  51298. #ifdef __LITTLE_ENDIAN__
  51299. __ai int32x4_t vmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
  51300.   int32x4_t __ret;
  51301.   __ret = vmull_s16(vget_high_s16(__p0), vget_high_s16(__p1));
  51302.   return __ret;
  51303. }
  51304. #else
  51305. __ai int32x4_t vmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
  51306.   int32x4_t __ret;
  51307.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  51308.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  51309.   __ret = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0), __noswap_vget_high_s16(__rev1));
  51310.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  51311.   return __ret;
  51312. }
  51313. #endif
  51314.  
  51315. #ifdef __LITTLE_ENDIAN__
  51316. #define vmull_high_lane_u32(__p0_564, __p1_564, __p2_564) __extension__ ({ \
  51317.   uint64x2_t __ret_564; \
  51318.   uint32x4_t __s0_564 = __p0_564; \
  51319.   uint32x2_t __s1_564 = __p1_564; \
  51320.   __ret_564 = vmull_u32(vget_high_u32(__s0_564), splat_lane_u32(__s1_564, __p2_564)); \
  51321.   __ret_564; \
  51322. })
  51323. #else
  51324. #define vmull_high_lane_u32(__p0_565, __p1_565, __p2_565) __extension__ ({ \
  51325.   uint64x2_t __ret_565; \
  51326.   uint32x4_t __s0_565 = __p0_565; \
  51327.   uint32x2_t __s1_565 = __p1_565; \
  51328.   uint32x4_t __rev0_565;  __rev0_565 = __builtin_shufflevector(__s0_565, __s0_565, 3, 2, 1, 0); \
  51329.   uint32x2_t __rev1_565;  __rev1_565 = __builtin_shufflevector(__s1_565, __s1_565, 1, 0); \
  51330.   __ret_565 = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0_565), __noswap_splat_lane_u32(__rev1_565, __p2_565)); \
  51331.   __ret_565 = __builtin_shufflevector(__ret_565, __ret_565, 1, 0); \
  51332.   __ret_565; \
  51333. })
  51334. #endif
  51335.  
  51336. #ifdef __LITTLE_ENDIAN__
  51337. #define vmull_high_lane_u16(__p0_566, __p1_566, __p2_566) __extension__ ({ \
  51338.   uint32x4_t __ret_566; \
  51339.   uint16x8_t __s0_566 = __p0_566; \
  51340.   uint16x4_t __s1_566 = __p1_566; \
  51341.   __ret_566 = vmull_u16(vget_high_u16(__s0_566), splat_lane_u16(__s1_566, __p2_566)); \
  51342.   __ret_566; \
  51343. })
  51344. #else
  51345. #define vmull_high_lane_u16(__p0_567, __p1_567, __p2_567) __extension__ ({ \
  51346.   uint32x4_t __ret_567; \
  51347.   uint16x8_t __s0_567 = __p0_567; \
  51348.   uint16x4_t __s1_567 = __p1_567; \
  51349.   uint16x8_t __rev0_567;  __rev0_567 = __builtin_shufflevector(__s0_567, __s0_567, 7, 6, 5, 4, 3, 2, 1, 0); \
  51350.   uint16x4_t __rev1_567;  __rev1_567 = __builtin_shufflevector(__s1_567, __s1_567, 3, 2, 1, 0); \
  51351.   __ret_567 = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0_567), __noswap_splat_lane_u16(__rev1_567, __p2_567)); \
  51352.   __ret_567 = __builtin_shufflevector(__ret_567, __ret_567, 3, 2, 1, 0); \
  51353.   __ret_567; \
  51354. })
  51355. #endif
  51356.  
  51357. #ifdef __LITTLE_ENDIAN__
  51358. #define vmull_high_lane_s32(__p0_568, __p1_568, __p2_568) __extension__ ({ \
  51359.   int64x2_t __ret_568; \
  51360.   int32x4_t __s0_568 = __p0_568; \
  51361.   int32x2_t __s1_568 = __p1_568; \
  51362.   __ret_568 = vmull_s32(vget_high_s32(__s0_568), splat_lane_s32(__s1_568, __p2_568)); \
  51363.   __ret_568; \
  51364. })
  51365. #else
  51366. #define vmull_high_lane_s32(__p0_569, __p1_569, __p2_569) __extension__ ({ \
  51367.   int64x2_t __ret_569; \
  51368.   int32x4_t __s0_569 = __p0_569; \
  51369.   int32x2_t __s1_569 = __p1_569; \
  51370.   int32x4_t __rev0_569;  __rev0_569 = __builtin_shufflevector(__s0_569, __s0_569, 3, 2, 1, 0); \
  51371.   int32x2_t __rev1_569;  __rev1_569 = __builtin_shufflevector(__s1_569, __s1_569, 1, 0); \
  51372.   __ret_569 = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0_569), __noswap_splat_lane_s32(__rev1_569, __p2_569)); \
  51373.   __ret_569 = __builtin_shufflevector(__ret_569, __ret_569, 1, 0); \
  51374.   __ret_569; \
  51375. })
  51376. #endif
  51377.  
  51378. #ifdef __LITTLE_ENDIAN__
  51379. #define vmull_high_lane_s16(__p0_570, __p1_570, __p2_570) __extension__ ({ \
  51380.   int32x4_t __ret_570; \
  51381.   int16x8_t __s0_570 = __p0_570; \
  51382.   int16x4_t __s1_570 = __p1_570; \
  51383.   __ret_570 = vmull_s16(vget_high_s16(__s0_570), splat_lane_s16(__s1_570, __p2_570)); \
  51384.   __ret_570; \
  51385. })
  51386. #else
  51387. #define vmull_high_lane_s16(__p0_571, __p1_571, __p2_571) __extension__ ({ \
  51388.   int32x4_t __ret_571; \
  51389.   int16x8_t __s0_571 = __p0_571; \
  51390.   int16x4_t __s1_571 = __p1_571; \
  51391.   int16x8_t __rev0_571;  __rev0_571 = __builtin_shufflevector(__s0_571, __s0_571, 7, 6, 5, 4, 3, 2, 1, 0); \
  51392.   int16x4_t __rev1_571;  __rev1_571 = __builtin_shufflevector(__s1_571, __s1_571, 3, 2, 1, 0); \
  51393.   __ret_571 = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0_571), __noswap_splat_lane_s16(__rev1_571, __p2_571)); \
  51394.   __ret_571 = __builtin_shufflevector(__ret_571, __ret_571, 3, 2, 1, 0); \
  51395.   __ret_571; \
  51396. })
  51397. #endif
  51398.  
  51399. #ifdef __LITTLE_ENDIAN__
  51400. #define vmull_high_laneq_u32(__p0_572, __p1_572, __p2_572) __extension__ ({ \
  51401.   uint64x2_t __ret_572; \
  51402.   uint32x4_t __s0_572 = __p0_572; \
  51403.   uint32x4_t __s1_572 = __p1_572; \
  51404.   __ret_572 = vmull_u32(vget_high_u32(__s0_572), splat_laneq_u32(__s1_572, __p2_572)); \
  51405.   __ret_572; \
  51406. })
  51407. #else
  51408. #define vmull_high_laneq_u32(__p0_573, __p1_573, __p2_573) __extension__ ({ \
  51409.   uint64x2_t __ret_573; \
  51410.   uint32x4_t __s0_573 = __p0_573; \
  51411.   uint32x4_t __s1_573 = __p1_573; \
  51412.   uint32x4_t __rev0_573;  __rev0_573 = __builtin_shufflevector(__s0_573, __s0_573, 3, 2, 1, 0); \
  51413.   uint32x4_t __rev1_573;  __rev1_573 = __builtin_shufflevector(__s1_573, __s1_573, 3, 2, 1, 0); \
  51414.   __ret_573 = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0_573), __noswap_splat_laneq_u32(__rev1_573, __p2_573)); \
  51415.   __ret_573 = __builtin_shufflevector(__ret_573, __ret_573, 1, 0); \
  51416.   __ret_573; \
  51417. })
  51418. #endif
  51419.  
  51420. #ifdef __LITTLE_ENDIAN__
  51421. #define vmull_high_laneq_u16(__p0_574, __p1_574, __p2_574) __extension__ ({ \
  51422.   uint32x4_t __ret_574; \
  51423.   uint16x8_t __s0_574 = __p0_574; \
  51424.   uint16x8_t __s1_574 = __p1_574; \
  51425.   __ret_574 = vmull_u16(vget_high_u16(__s0_574), splat_laneq_u16(__s1_574, __p2_574)); \
  51426.   __ret_574; \
  51427. })
  51428. #else
  51429. #define vmull_high_laneq_u16(__p0_575, __p1_575, __p2_575) __extension__ ({ \
  51430.   uint32x4_t __ret_575; \
  51431.   uint16x8_t __s0_575 = __p0_575; \
  51432.   uint16x8_t __s1_575 = __p1_575; \
  51433.   uint16x8_t __rev0_575;  __rev0_575 = __builtin_shufflevector(__s0_575, __s0_575, 7, 6, 5, 4, 3, 2, 1, 0); \
  51434.   uint16x8_t __rev1_575;  __rev1_575 = __builtin_shufflevector(__s1_575, __s1_575, 7, 6, 5, 4, 3, 2, 1, 0); \
  51435.   __ret_575 = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0_575), __noswap_splat_laneq_u16(__rev1_575, __p2_575)); \
  51436.   __ret_575 = __builtin_shufflevector(__ret_575, __ret_575, 3, 2, 1, 0); \
  51437.   __ret_575; \
  51438. })
  51439. #endif
  51440.  
  51441. #ifdef __LITTLE_ENDIAN__
  51442. #define vmull_high_laneq_s32(__p0_576, __p1_576, __p2_576) __extension__ ({ \
  51443.   int64x2_t __ret_576; \
  51444.   int32x4_t __s0_576 = __p0_576; \
  51445.   int32x4_t __s1_576 = __p1_576; \
  51446.   __ret_576 = vmull_s32(vget_high_s32(__s0_576), splat_laneq_s32(__s1_576, __p2_576)); \
  51447.   __ret_576; \
  51448. })
  51449. #else
  51450. #define vmull_high_laneq_s32(__p0_577, __p1_577, __p2_577) __extension__ ({ \
  51451.   int64x2_t __ret_577; \
  51452.   int32x4_t __s0_577 = __p0_577; \
  51453.   int32x4_t __s1_577 = __p1_577; \
  51454.   int32x4_t __rev0_577;  __rev0_577 = __builtin_shufflevector(__s0_577, __s0_577, 3, 2, 1, 0); \
  51455.   int32x4_t __rev1_577;  __rev1_577 = __builtin_shufflevector(__s1_577, __s1_577, 3, 2, 1, 0); \
  51456.   __ret_577 = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0_577), __noswap_splat_laneq_s32(__rev1_577, __p2_577)); \
  51457.   __ret_577 = __builtin_shufflevector(__ret_577, __ret_577, 1, 0); \
  51458.   __ret_577; \
  51459. })
  51460. #endif
  51461.  
  51462. #ifdef __LITTLE_ENDIAN__
  51463. #define vmull_high_laneq_s16(__p0_578, __p1_578, __p2_578) __extension__ ({ \
  51464.   int32x4_t __ret_578; \
  51465.   int16x8_t __s0_578 = __p0_578; \
  51466.   int16x8_t __s1_578 = __p1_578; \
  51467.   __ret_578 = vmull_s16(vget_high_s16(__s0_578), splat_laneq_s16(__s1_578, __p2_578)); \
  51468.   __ret_578; \
  51469. })
  51470. #else
  51471. #define vmull_high_laneq_s16(__p0_579, __p1_579, __p2_579) __extension__ ({ \
  51472.   int32x4_t __ret_579; \
  51473.   int16x8_t __s0_579 = __p0_579; \
  51474.   int16x8_t __s1_579 = __p1_579; \
  51475.   int16x8_t __rev0_579;  __rev0_579 = __builtin_shufflevector(__s0_579, __s0_579, 7, 6, 5, 4, 3, 2, 1, 0); \
  51476.   int16x8_t __rev1_579;  __rev1_579 = __builtin_shufflevector(__s1_579, __s1_579, 7, 6, 5, 4, 3, 2, 1, 0); \
  51477.   __ret_579 = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0_579), __noswap_splat_laneq_s16(__rev1_579, __p2_579)); \
  51478.   __ret_579 = __builtin_shufflevector(__ret_579, __ret_579, 3, 2, 1, 0); \
  51479.   __ret_579; \
  51480. })
  51481. #endif
  51482.  
  51483. #ifdef __LITTLE_ENDIAN__
  51484. __ai uint64x2_t vmull_high_n_u32(uint32x4_t __p0, uint32_t __p1) {
  51485.   uint64x2_t __ret;
  51486.   __ret = vmull_n_u32(vget_high_u32(__p0), __p1);
  51487.   return __ret;
  51488. }
  51489. #else
  51490. __ai uint64x2_t vmull_high_n_u32(uint32x4_t __p0, uint32_t __p1) {
  51491.   uint64x2_t __ret;
  51492.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  51493.   __ret = __noswap_vmull_n_u32(__noswap_vget_high_u32(__rev0), __p1);
  51494.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51495.   return __ret;
  51496. }
  51497. #endif
  51498.  
  51499. #ifdef __LITTLE_ENDIAN__
  51500. __ai uint32x4_t vmull_high_n_u16(uint16x8_t __p0, uint16_t __p1) {
  51501.   uint32x4_t __ret;
  51502.   __ret = vmull_n_u16(vget_high_u16(__p0), __p1);
  51503.   return __ret;
  51504. }
  51505. #else
  51506. __ai uint32x4_t vmull_high_n_u16(uint16x8_t __p0, uint16_t __p1) {
  51507.   uint32x4_t __ret;
  51508.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  51509.   __ret = __noswap_vmull_n_u16(__noswap_vget_high_u16(__rev0), __p1);
  51510.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  51511.   return __ret;
  51512. }
  51513. #endif
  51514.  
  51515. #ifdef __LITTLE_ENDIAN__
  51516. __ai int64x2_t vmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
  51517.   int64x2_t __ret;
  51518.   __ret = vmull_n_s32(vget_high_s32(__p0), __p1);
  51519.   return __ret;
  51520. }
  51521. #else
  51522. __ai int64x2_t vmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
  51523.   int64x2_t __ret;
  51524.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  51525.   __ret = __noswap_vmull_n_s32(__noswap_vget_high_s32(__rev0), __p1);
  51526.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51527.   return __ret;
  51528. }
  51529. #endif
  51530.  
  51531. #ifdef __LITTLE_ENDIAN__
  51532. __ai int32x4_t vmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
  51533.   int32x4_t __ret;
  51534.   __ret = vmull_n_s16(vget_high_s16(__p0), __p1);
  51535.   return __ret;
  51536. }
  51537. #else
  51538. __ai int32x4_t vmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
  51539.   int32x4_t __ret;
  51540.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  51541.   __ret = __noswap_vmull_n_s16(__noswap_vget_high_s16(__rev0), __p1);
  51542.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  51543.   return __ret;
  51544. }
  51545. #endif
  51546.  
  51547. #ifdef __LITTLE_ENDIAN__
  51548. #define vmull_laneq_u32(__p0_580, __p1_580, __p2_580) __extension__ ({ \
  51549.   uint64x2_t __ret_580; \
  51550.   uint32x2_t __s0_580 = __p0_580; \
  51551.   uint32x4_t __s1_580 = __p1_580; \
  51552.   __ret_580 = vmull_u32(__s0_580, splat_laneq_u32(__s1_580, __p2_580)); \
  51553.   __ret_580; \
  51554. })
  51555. #else
  51556. #define vmull_laneq_u32(__p0_581, __p1_581, __p2_581) __extension__ ({ \
  51557.   uint64x2_t __ret_581; \
  51558.   uint32x2_t __s0_581 = __p0_581; \
  51559.   uint32x4_t __s1_581 = __p1_581; \
  51560.   uint32x2_t __rev0_581;  __rev0_581 = __builtin_shufflevector(__s0_581, __s0_581, 1, 0); \
  51561.   uint32x4_t __rev1_581;  __rev1_581 = __builtin_shufflevector(__s1_581, __s1_581, 3, 2, 1, 0); \
  51562.   __ret_581 = __noswap_vmull_u32(__rev0_581, __noswap_splat_laneq_u32(__rev1_581, __p2_581)); \
  51563.   __ret_581 = __builtin_shufflevector(__ret_581, __ret_581, 1, 0); \
  51564.   __ret_581; \
  51565. })
  51566. #endif
  51567.  
  51568. #ifdef __LITTLE_ENDIAN__
  51569. #define vmull_laneq_u16(__p0_582, __p1_582, __p2_582) __extension__ ({ \
  51570.   uint32x4_t __ret_582; \
  51571.   uint16x4_t __s0_582 = __p0_582; \
  51572.   uint16x8_t __s1_582 = __p1_582; \
  51573.   __ret_582 = vmull_u16(__s0_582, splat_laneq_u16(__s1_582, __p2_582)); \
  51574.   __ret_582; \
  51575. })
  51576. #else
  51577. #define vmull_laneq_u16(__p0_583, __p1_583, __p2_583) __extension__ ({ \
  51578.   uint32x4_t __ret_583; \
  51579.   uint16x4_t __s0_583 = __p0_583; \
  51580.   uint16x8_t __s1_583 = __p1_583; \
  51581.   uint16x4_t __rev0_583;  __rev0_583 = __builtin_shufflevector(__s0_583, __s0_583, 3, 2, 1, 0); \
  51582.   uint16x8_t __rev1_583;  __rev1_583 = __builtin_shufflevector(__s1_583, __s1_583, 7, 6, 5, 4, 3, 2, 1, 0); \
  51583.   __ret_583 = __noswap_vmull_u16(__rev0_583, __noswap_splat_laneq_u16(__rev1_583, __p2_583)); \
  51584.   __ret_583 = __builtin_shufflevector(__ret_583, __ret_583, 3, 2, 1, 0); \
  51585.   __ret_583; \
  51586. })
  51587. #endif
  51588.  
  51589. #ifdef __LITTLE_ENDIAN__
  51590. #define vmull_laneq_s32(__p0_584, __p1_584, __p2_584) __extension__ ({ \
  51591.   int64x2_t __ret_584; \
  51592.   int32x2_t __s0_584 = __p0_584; \
  51593.   int32x4_t __s1_584 = __p1_584; \
  51594.   __ret_584 = vmull_s32(__s0_584, splat_laneq_s32(__s1_584, __p2_584)); \
  51595.   __ret_584; \
  51596. })
  51597. #else
  51598. #define vmull_laneq_s32(__p0_585, __p1_585, __p2_585) __extension__ ({ \
  51599.   int64x2_t __ret_585; \
  51600.   int32x2_t __s0_585 = __p0_585; \
  51601.   int32x4_t __s1_585 = __p1_585; \
  51602.   int32x2_t __rev0_585;  __rev0_585 = __builtin_shufflevector(__s0_585, __s0_585, 1, 0); \
  51603.   int32x4_t __rev1_585;  __rev1_585 = __builtin_shufflevector(__s1_585, __s1_585, 3, 2, 1, 0); \
  51604.   __ret_585 = __noswap_vmull_s32(__rev0_585, __noswap_splat_laneq_s32(__rev1_585, __p2_585)); \
  51605.   __ret_585 = __builtin_shufflevector(__ret_585, __ret_585, 1, 0); \
  51606.   __ret_585; \
  51607. })
  51608. #endif
  51609.  
  51610. #ifdef __LITTLE_ENDIAN__
  51611. #define vmull_laneq_s16(__p0_586, __p1_586, __p2_586) __extension__ ({ \
  51612.   int32x4_t __ret_586; \
  51613.   int16x4_t __s0_586 = __p0_586; \
  51614.   int16x8_t __s1_586 = __p1_586; \
  51615.   __ret_586 = vmull_s16(__s0_586, splat_laneq_s16(__s1_586, __p2_586)); \
  51616.   __ret_586; \
  51617. })
  51618. #else
  51619. #define vmull_laneq_s16(__p0_587, __p1_587, __p2_587) __extension__ ({ \
  51620.   int32x4_t __ret_587; \
  51621.   int16x4_t __s0_587 = __p0_587; \
  51622.   int16x8_t __s1_587 = __p1_587; \
  51623.   int16x4_t __rev0_587;  __rev0_587 = __builtin_shufflevector(__s0_587, __s0_587, 3, 2, 1, 0); \
  51624.   int16x8_t __rev1_587;  __rev1_587 = __builtin_shufflevector(__s1_587, __s1_587, 7, 6, 5, 4, 3, 2, 1, 0); \
  51625.   __ret_587 = __noswap_vmull_s16(__rev0_587, __noswap_splat_laneq_s16(__rev1_587, __p2_587)); \
  51626.   __ret_587 = __builtin_shufflevector(__ret_587, __ret_587, 3, 2, 1, 0); \
  51627.   __ret_587; \
  51628. })
  51629. #endif
  51630.  
  51631. #ifdef __LITTLE_ENDIAN__
  51632. __ai float64x2_t vmulxq_f64(float64x2_t __p0, float64x2_t __p1) {
  51633.   float64x2_t __ret;
  51634.   __ret = (float64x2_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  51635.   return __ret;
  51636. }
  51637. #else
  51638. __ai float64x2_t vmulxq_f64(float64x2_t __p0, float64x2_t __p1) {
  51639.   float64x2_t __ret;
  51640.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  51641.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  51642.   __ret = (float64x2_t) __builtin_neon_vmulxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  51643.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51644.   return __ret;
  51645. }
  51646. __ai float64x2_t __noswap_vmulxq_f64(float64x2_t __p0, float64x2_t __p1) {
  51647.   float64x2_t __ret;
  51648.   __ret = (float64x2_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  51649.   return __ret;
  51650. }
  51651. #endif
  51652.  
  51653. #ifdef __LITTLE_ENDIAN__
  51654. __ai float32x4_t vmulxq_f32(float32x4_t __p0, float32x4_t __p1) {
  51655.   float32x4_t __ret;
  51656.   __ret = (float32x4_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  51657.   return __ret;
  51658. }
  51659. #else
  51660. __ai float32x4_t vmulxq_f32(float32x4_t __p0, float32x4_t __p1) {
  51661.   float32x4_t __ret;
  51662.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  51663.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  51664.   __ret = (float32x4_t) __builtin_neon_vmulxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  51665.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  51666.   return __ret;
  51667. }
  51668. __ai float32x4_t __noswap_vmulxq_f32(float32x4_t __p0, float32x4_t __p1) {
  51669.   float32x4_t __ret;
  51670.   __ret = (float32x4_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  51671.   return __ret;
  51672. }
  51673. #endif
  51674.  
  51675. __ai float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1) {
  51676.   float64x1_t __ret;
  51677.   __ret = (float64x1_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
  51678.   return __ret;
  51679. }
  51680. #ifdef __LITTLE_ENDIAN__
  51681. __ai float32x2_t vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
  51682.   float32x2_t __ret;
  51683.   __ret = (float32x2_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  51684.   return __ret;
  51685. }
  51686. #else
  51687. __ai float32x2_t vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
  51688.   float32x2_t __ret;
  51689.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  51690.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  51691.   __ret = (float32x2_t) __builtin_neon_vmulx_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  51692.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51693.   return __ret;
  51694. }
  51695. __ai float32x2_t __noswap_vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
  51696.   float32x2_t __ret;
  51697.   __ret = (float32x2_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  51698.   return __ret;
  51699. }
  51700. #endif
  51701.  
  51702. __ai float64_t vmulxd_f64(float64_t __p0, float64_t __p1) {
  51703.   float64_t __ret;
  51704.   __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
  51705.   return __ret;
  51706. }
  51707. __ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) {
  51708.   float32_t __ret;
  51709.   __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
  51710.   return __ret;
  51711. }
  51712. #define vmulxd_lane_f64(__p0_588, __p1_588, __p2_588) __extension__ ({ \
  51713.   float64_t __ret_588; \
  51714.   float64_t __s0_588 = __p0_588; \
  51715.   float64x1_t __s1_588 = __p1_588; \
  51716.   __ret_588 = vmulxd_f64(__s0_588, vget_lane_f64(__s1_588, __p2_588)); \
  51717.   __ret_588; \
  51718. })
  51719. #ifdef __LITTLE_ENDIAN__
  51720. #define vmulxs_lane_f32(__p0_589, __p1_589, __p2_589) __extension__ ({ \
  51721.   float32_t __ret_589; \
  51722.   float32_t __s0_589 = __p0_589; \
  51723.   float32x2_t __s1_589 = __p1_589; \
  51724.   __ret_589 = vmulxs_f32(__s0_589, vget_lane_f32(__s1_589, __p2_589)); \
  51725.   __ret_589; \
  51726. })
  51727. #else
  51728. #define vmulxs_lane_f32(__p0_590, __p1_590, __p2_590) __extension__ ({ \
  51729.   float32_t __ret_590; \
  51730.   float32_t __s0_590 = __p0_590; \
  51731.   float32x2_t __s1_590 = __p1_590; \
  51732.   float32x2_t __rev1_590;  __rev1_590 = __builtin_shufflevector(__s1_590, __s1_590, 1, 0); \
  51733.   __ret_590 = vmulxs_f32(__s0_590, __noswap_vget_lane_f32(__rev1_590, __p2_590)); \
  51734.   __ret_590; \
  51735. })
  51736. #endif
  51737.  
  51738. #ifdef __LITTLE_ENDIAN__
  51739. #define vmulxq_lane_f64(__p0_591, __p1_591, __p2_591) __extension__ ({ \
  51740.   float64x2_t __ret_591; \
  51741.   float64x2_t __s0_591 = __p0_591; \
  51742.   float64x1_t __s1_591 = __p1_591; \
  51743.   __ret_591 = vmulxq_f64(__s0_591, splatq_lane_f64(__s1_591, __p2_591)); \
  51744.   __ret_591; \
  51745. })
  51746. #else
  51747. #define vmulxq_lane_f64(__p0_592, __p1_592, __p2_592) __extension__ ({ \
  51748.   float64x2_t __ret_592; \
  51749.   float64x2_t __s0_592 = __p0_592; \
  51750.   float64x1_t __s1_592 = __p1_592; \
  51751.   float64x2_t __rev0_592;  __rev0_592 = __builtin_shufflevector(__s0_592, __s0_592, 1, 0); \
  51752.   __ret_592 = __noswap_vmulxq_f64(__rev0_592, __noswap_splatq_lane_f64(__s1_592, __p2_592)); \
  51753.   __ret_592 = __builtin_shufflevector(__ret_592, __ret_592, 1, 0); \
  51754.   __ret_592; \
  51755. })
  51756. #endif
  51757.  
  51758. #ifdef __LITTLE_ENDIAN__
  51759. #define vmulxq_lane_f32(__p0_593, __p1_593, __p2_593) __extension__ ({ \
  51760.   float32x4_t __ret_593; \
  51761.   float32x4_t __s0_593 = __p0_593; \
  51762.   float32x2_t __s1_593 = __p1_593; \
  51763.   __ret_593 = vmulxq_f32(__s0_593, splatq_lane_f32(__s1_593, __p2_593)); \
  51764.   __ret_593; \
  51765. })
  51766. #else
  51767. #define vmulxq_lane_f32(__p0_594, __p1_594, __p2_594) __extension__ ({ \
  51768.   float32x4_t __ret_594; \
  51769.   float32x4_t __s0_594 = __p0_594; \
  51770.   float32x2_t __s1_594 = __p1_594; \
  51771.   float32x4_t __rev0_594;  __rev0_594 = __builtin_shufflevector(__s0_594, __s0_594, 3, 2, 1, 0); \
  51772.   float32x2_t __rev1_594;  __rev1_594 = __builtin_shufflevector(__s1_594, __s1_594, 1, 0); \
  51773.   __ret_594 = __noswap_vmulxq_f32(__rev0_594, __noswap_splatq_lane_f32(__rev1_594, __p2_594)); \
  51774.   __ret_594 = __builtin_shufflevector(__ret_594, __ret_594, 3, 2, 1, 0); \
  51775.   __ret_594; \
  51776. })
  51777. #endif
  51778.  
  51779. #ifdef __LITTLE_ENDIAN__
  51780. #define vmulx_lane_f32(__p0_595, __p1_595, __p2_595) __extension__ ({ \
  51781.   float32x2_t __ret_595; \
  51782.   float32x2_t __s0_595 = __p0_595; \
  51783.   float32x2_t __s1_595 = __p1_595; \
  51784.   __ret_595 = vmulx_f32(__s0_595, splat_lane_f32(__s1_595, __p2_595)); \
  51785.   __ret_595; \
  51786. })
  51787. #else
  51788. #define vmulx_lane_f32(__p0_596, __p1_596, __p2_596) __extension__ ({ \
  51789.   float32x2_t __ret_596; \
  51790.   float32x2_t __s0_596 = __p0_596; \
  51791.   float32x2_t __s1_596 = __p1_596; \
  51792.   float32x2_t __rev0_596;  __rev0_596 = __builtin_shufflevector(__s0_596, __s0_596, 1, 0); \
  51793.   float32x2_t __rev1_596;  __rev1_596 = __builtin_shufflevector(__s1_596, __s1_596, 1, 0); \
  51794.   __ret_596 = __noswap_vmulx_f32(__rev0_596, __noswap_splat_lane_f32(__rev1_596, __p2_596)); \
  51795.   __ret_596 = __builtin_shufflevector(__ret_596, __ret_596, 1, 0); \
  51796.   __ret_596; \
  51797. })
  51798. #endif
  51799.  
  51800. #ifdef __LITTLE_ENDIAN__
  51801. #define vmulxd_laneq_f64(__p0_597, __p1_597, __p2_597) __extension__ ({ \
  51802.   float64_t __ret_597; \
  51803.   float64_t __s0_597 = __p0_597; \
  51804.   float64x2_t __s1_597 = __p1_597; \
  51805.   __ret_597 = vmulxd_f64(__s0_597, vgetq_lane_f64(__s1_597, __p2_597)); \
  51806.   __ret_597; \
  51807. })
  51808. #else
  51809. #define vmulxd_laneq_f64(__p0_598, __p1_598, __p2_598) __extension__ ({ \
  51810.   float64_t __ret_598; \
  51811.   float64_t __s0_598 = __p0_598; \
  51812.   float64x2_t __s1_598 = __p1_598; \
  51813.   float64x2_t __rev1_598;  __rev1_598 = __builtin_shufflevector(__s1_598, __s1_598, 1, 0); \
  51814.   __ret_598 = vmulxd_f64(__s0_598, __noswap_vgetq_lane_f64(__rev1_598, __p2_598)); \
  51815.   __ret_598; \
  51816. })
  51817. #endif
  51818.  
  51819. #ifdef __LITTLE_ENDIAN__
  51820. #define vmulxs_laneq_f32(__p0_599, __p1_599, __p2_599) __extension__ ({ \
  51821.   float32_t __ret_599; \
  51822.   float32_t __s0_599 = __p0_599; \
  51823.   float32x4_t __s1_599 = __p1_599; \
  51824.   __ret_599 = vmulxs_f32(__s0_599, vgetq_lane_f32(__s1_599, __p2_599)); \
  51825.   __ret_599; \
  51826. })
  51827. #else
  51828. #define vmulxs_laneq_f32(__p0_600, __p1_600, __p2_600) __extension__ ({ \
  51829.   float32_t __ret_600; \
  51830.   float32_t __s0_600 = __p0_600; \
  51831.   float32x4_t __s1_600 = __p1_600; \
  51832.   float32x4_t __rev1_600;  __rev1_600 = __builtin_shufflevector(__s1_600, __s1_600, 3, 2, 1, 0); \
  51833.   __ret_600 = vmulxs_f32(__s0_600, __noswap_vgetq_lane_f32(__rev1_600, __p2_600)); \
  51834.   __ret_600; \
  51835. })
  51836. #endif
  51837.  
  51838. #ifdef __LITTLE_ENDIAN__
  51839. #define vmulxq_laneq_f64(__p0_601, __p1_601, __p2_601) __extension__ ({ \
  51840.   float64x2_t __ret_601; \
  51841.   float64x2_t __s0_601 = __p0_601; \
  51842.   float64x2_t __s1_601 = __p1_601; \
  51843.   __ret_601 = vmulxq_f64(__s0_601, splatq_laneq_f64(__s1_601, __p2_601)); \
  51844.   __ret_601; \
  51845. })
  51846. #else
  51847. #define vmulxq_laneq_f64(__p0_602, __p1_602, __p2_602) __extension__ ({ \
  51848.   float64x2_t __ret_602; \
  51849.   float64x2_t __s0_602 = __p0_602; \
  51850.   float64x2_t __s1_602 = __p1_602; \
  51851.   float64x2_t __rev0_602;  __rev0_602 = __builtin_shufflevector(__s0_602, __s0_602, 1, 0); \
  51852.   float64x2_t __rev1_602;  __rev1_602 = __builtin_shufflevector(__s1_602, __s1_602, 1, 0); \
  51853.   __ret_602 = __noswap_vmulxq_f64(__rev0_602, __noswap_splatq_laneq_f64(__rev1_602, __p2_602)); \
  51854.   __ret_602 = __builtin_shufflevector(__ret_602, __ret_602, 1, 0); \
  51855.   __ret_602; \
  51856. })
  51857. #endif
  51858.  
  51859. #ifdef __LITTLE_ENDIAN__
  51860. #define vmulxq_laneq_f32(__p0_603, __p1_603, __p2_603) __extension__ ({ \
  51861.   float32x4_t __ret_603; \
  51862.   float32x4_t __s0_603 = __p0_603; \
  51863.   float32x4_t __s1_603 = __p1_603; \
  51864.   __ret_603 = vmulxq_f32(__s0_603, splatq_laneq_f32(__s1_603, __p2_603)); \
  51865.   __ret_603; \
  51866. })
  51867. #else
  51868. #define vmulxq_laneq_f32(__p0_604, __p1_604, __p2_604) __extension__ ({ \
  51869.   float32x4_t __ret_604; \
  51870.   float32x4_t __s0_604 = __p0_604; \
  51871.   float32x4_t __s1_604 = __p1_604; \
  51872.   float32x4_t __rev0_604;  __rev0_604 = __builtin_shufflevector(__s0_604, __s0_604, 3, 2, 1, 0); \
  51873.   float32x4_t __rev1_604;  __rev1_604 = __builtin_shufflevector(__s1_604, __s1_604, 3, 2, 1, 0); \
  51874.   __ret_604 = __noswap_vmulxq_f32(__rev0_604, __noswap_splatq_laneq_f32(__rev1_604, __p2_604)); \
  51875.   __ret_604 = __builtin_shufflevector(__ret_604, __ret_604, 3, 2, 1, 0); \
  51876.   __ret_604; \
  51877. })
  51878. #endif
  51879.  
  51880. #ifdef __LITTLE_ENDIAN__
  51881. #define vmulx_laneq_f32(__p0_605, __p1_605, __p2_605) __extension__ ({ \
  51882.   float32x2_t __ret_605; \
  51883.   float32x2_t __s0_605 = __p0_605; \
  51884.   float32x4_t __s1_605 = __p1_605; \
  51885.   __ret_605 = vmulx_f32(__s0_605, splat_laneq_f32(__s1_605, __p2_605)); \
  51886.   __ret_605; \
  51887. })
  51888. #else
  51889. #define vmulx_laneq_f32(__p0_606, __p1_606, __p2_606) __extension__ ({ \
  51890.   float32x2_t __ret_606; \
  51891.   float32x2_t __s0_606 = __p0_606; \
  51892.   float32x4_t __s1_606 = __p1_606; \
  51893.   float32x2_t __rev0_606;  __rev0_606 = __builtin_shufflevector(__s0_606, __s0_606, 1, 0); \
  51894.   float32x4_t __rev1_606;  __rev1_606 = __builtin_shufflevector(__s1_606, __s1_606, 3, 2, 1, 0); \
  51895.   __ret_606 = __noswap_vmulx_f32(__rev0_606, __noswap_splat_laneq_f32(__rev1_606, __p2_606)); \
  51896.   __ret_606 = __builtin_shufflevector(__ret_606, __ret_606, 1, 0); \
  51897.   __ret_606; \
  51898. })
  51899. #endif
  51900.  
  51901. #ifdef __LITTLE_ENDIAN__
  51902. __ai float64x2_t vnegq_f64(float64x2_t __p0) {
  51903.   float64x2_t __ret;
  51904.   __ret = -__p0;
  51905.   return __ret;
  51906. }
  51907. #else
  51908. __ai float64x2_t vnegq_f64(float64x2_t __p0) {
  51909.   float64x2_t __ret;
  51910.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  51911.   __ret = -__rev0;
  51912.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51913.   return __ret;
  51914. }
  51915. #endif
  51916.  
  51917. #ifdef __LITTLE_ENDIAN__
  51918. __ai int64x2_t vnegq_s64(int64x2_t __p0) {
  51919.   int64x2_t __ret;
  51920.   __ret = -__p0;
  51921.   return __ret;
  51922. }
  51923. #else
  51924. __ai int64x2_t vnegq_s64(int64x2_t __p0) {
  51925.   int64x2_t __ret;
  51926.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  51927.   __ret = -__rev0;
  51928.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51929.   return __ret;
  51930. }
  51931. #endif
  51932.  
  51933. __ai float64x1_t vneg_f64(float64x1_t __p0) {
  51934.   float64x1_t __ret;
  51935.   __ret = -__p0;
  51936.   return __ret;
  51937. }
  51938. __ai int64x1_t vneg_s64(int64x1_t __p0) {
  51939.   int64x1_t __ret;
  51940.   __ret = -__p0;
  51941.   return __ret;
  51942. }
  51943. __ai int64_t vnegd_s64(int64_t __p0) {
  51944.   int64_t __ret;
  51945.   __ret = (int64_t) __builtin_neon_vnegd_s64(__p0);
  51946.   return __ret;
  51947. }
  51948. #ifdef __LITTLE_ENDIAN__
  51949. __ai uint8x16_t vpaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  51950.   uint8x16_t __ret;
  51951.   __ret = (uint8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  51952.   return __ret;
  51953. }
  51954. #else
  51955. __ai uint8x16_t vpaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  51956.   uint8x16_t __ret;
  51957.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  51958.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  51959.   __ret = (uint8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  51960.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  51961.   return __ret;
  51962. }
  51963. #endif
  51964.  
  51965. #ifdef __LITTLE_ENDIAN__
  51966. __ai uint32x4_t vpaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  51967.   uint32x4_t __ret;
  51968.   __ret = (uint32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  51969.   return __ret;
  51970. }
  51971. #else
  51972. __ai uint32x4_t vpaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  51973.   uint32x4_t __ret;
  51974.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  51975.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  51976.   __ret = (uint32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  51977.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  51978.   return __ret;
  51979. }
  51980. #endif
  51981.  
  51982. #ifdef __LITTLE_ENDIAN__
  51983. __ai uint64x2_t vpaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  51984.   uint64x2_t __ret;
  51985.   __ret = (uint64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  51986.   return __ret;
  51987. }
  51988. #else
  51989. __ai uint64x2_t vpaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  51990.   uint64x2_t __ret;
  51991.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  51992.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  51993.   __ret = (uint64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  51994.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  51995.   return __ret;
  51996. }
  51997. #endif
  51998.  
  51999. #ifdef __LITTLE_ENDIAN__
  52000. __ai uint16x8_t vpaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  52001.   uint16x8_t __ret;
  52002.   __ret = (uint16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  52003.   return __ret;
  52004. }
  52005. #else
  52006. __ai uint16x8_t vpaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  52007.   uint16x8_t __ret;
  52008.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  52009.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  52010.   __ret = (uint16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  52011.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  52012.   return __ret;
  52013. }
  52014. #endif
  52015.  
  52016. #ifdef __LITTLE_ENDIAN__
  52017. __ai int8x16_t vpaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  52018.   int8x16_t __ret;
  52019.   __ret = (int8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  52020.   return __ret;
  52021. }
  52022. #else
  52023. __ai int8x16_t vpaddq_s8(int8x16_t __p0, int8x16_t __p1) {
  52024.   int8x16_t __ret;
  52025.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52026.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52027.   __ret = (int8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  52028.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52029.   return __ret;
  52030. }
  52031. #endif
  52032.  
  52033. #ifdef __LITTLE_ENDIAN__
  52034. __ai float64x2_t vpaddq_f64(float64x2_t __p0, float64x2_t __p1) {
  52035.   float64x2_t __ret;
  52036.   __ret = (float64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  52037.   return __ret;
  52038. }
  52039. #else
  52040. __ai float64x2_t vpaddq_f64(float64x2_t __p0, float64x2_t __p1) {
  52041.   float64x2_t __ret;
  52042.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52043.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  52044.   __ret = (float64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  52045.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52046.   return __ret;
  52047. }
  52048. #endif
  52049.  
  52050. #ifdef __LITTLE_ENDIAN__
  52051. __ai float32x4_t vpaddq_f32(float32x4_t __p0, float32x4_t __p1) {
  52052.   float32x4_t __ret;
  52053.   __ret = (float32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  52054.   return __ret;
  52055. }
  52056. #else
  52057. __ai float32x4_t vpaddq_f32(float32x4_t __p0, float32x4_t __p1) {
  52058.   float32x4_t __ret;
  52059.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52060.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52061.   __ret = (float32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  52062.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52063.   return __ret;
  52064. }
  52065. #endif
  52066.  
  52067. #ifdef __LITTLE_ENDIAN__
  52068. __ai int32x4_t vpaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  52069.   int32x4_t __ret;
  52070.   __ret = (int32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  52071.   return __ret;
  52072. }
  52073. #else
  52074. __ai int32x4_t vpaddq_s32(int32x4_t __p0, int32x4_t __p1) {
  52075.   int32x4_t __ret;
  52076.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52077.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52078.   __ret = (int32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  52079.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52080.   return __ret;
  52081. }
  52082. #endif
  52083.  
  52084. #ifdef __LITTLE_ENDIAN__
  52085. __ai int64x2_t vpaddq_s64(int64x2_t __p0, int64x2_t __p1) {
  52086.   int64x2_t __ret;
  52087.   __ret = (int64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
  52088.   return __ret;
  52089. }
  52090. #else
  52091. __ai int64x2_t vpaddq_s64(int64x2_t __p0, int64x2_t __p1) {
  52092.   int64x2_t __ret;
  52093.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52094.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  52095.   __ret = (int64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
  52096.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52097.   return __ret;
  52098. }
  52099. #endif
  52100.  
  52101. #ifdef __LITTLE_ENDIAN__
  52102. __ai int16x8_t vpaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  52103.   int16x8_t __ret;
  52104.   __ret = (int16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  52105.   return __ret;
  52106. }
  52107. #else
  52108. __ai int16x8_t vpaddq_s16(int16x8_t __p0, int16x8_t __p1) {
  52109.   int16x8_t __ret;
  52110.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  52111.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  52112.   __ret = (int16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  52113.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  52114.   return __ret;
  52115. }
  52116. #endif
  52117.  
  52118. #ifdef __LITTLE_ENDIAN__
  52119. __ai uint64_t vpaddd_u64(uint64x2_t __p0) {
  52120.   uint64_t __ret;
  52121.   __ret = (uint64_t) __builtin_neon_vpaddd_u64(__p0);
  52122.   return __ret;
  52123. }
  52124. #else
  52125. __ai uint64_t vpaddd_u64(uint64x2_t __p0) {
  52126.   uint64_t __ret;
  52127.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52128.   __ret = (uint64_t) __builtin_neon_vpaddd_u64(__rev0);
  52129.   return __ret;
  52130. }
  52131. #endif
  52132.  
  52133. #ifdef __LITTLE_ENDIAN__
  52134. __ai float64_t vpaddd_f64(float64x2_t __p0) {
  52135.   float64_t __ret;
  52136.   __ret = (float64_t) __builtin_neon_vpaddd_f64(__p0);
  52137.   return __ret;
  52138. }
  52139. #else
  52140. __ai float64_t vpaddd_f64(float64x2_t __p0) {
  52141.   float64_t __ret;
  52142.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52143.   __ret = (float64_t) __builtin_neon_vpaddd_f64(__rev0);
  52144.   return __ret;
  52145. }
  52146. #endif
  52147.  
  52148. #ifdef __LITTLE_ENDIAN__
  52149. __ai int64_t vpaddd_s64(int64x2_t __p0) {
  52150.   int64_t __ret;
  52151.   __ret = (int64_t) __builtin_neon_vpaddd_s64(__p0);
  52152.   return __ret;
  52153. }
  52154. #else
  52155. __ai int64_t vpaddd_s64(int64x2_t __p0) {
  52156.   int64_t __ret;
  52157.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52158.   __ret = (int64_t) __builtin_neon_vpaddd_s64(__rev0);
  52159.   return __ret;
  52160. }
  52161. #endif
  52162.  
  52163. #ifdef __LITTLE_ENDIAN__
  52164. __ai float32_t vpadds_f32(float32x2_t __p0) {
  52165.   float32_t __ret;
  52166.   __ret = (float32_t) __builtin_neon_vpadds_f32(__p0);
  52167.   return __ret;
  52168. }
  52169. #else
  52170. __ai float32_t vpadds_f32(float32x2_t __p0) {
  52171.   float32_t __ret;
  52172.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52173.   __ret = (float32_t) __builtin_neon_vpadds_f32(__rev0);
  52174.   return __ret;
  52175. }
  52176. #endif
  52177.  
  52178. #ifdef __LITTLE_ENDIAN__
  52179. __ai uint8x16_t vpmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  52180.   uint8x16_t __ret;
  52181.   __ret = (uint8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  52182.   return __ret;
  52183. }
  52184. #else
  52185. __ai uint8x16_t vpmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  52186.   uint8x16_t __ret;
  52187.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52188.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52189.   __ret = (uint8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  52190.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52191.   return __ret;
  52192. }
  52193. #endif
  52194.  
  52195. #ifdef __LITTLE_ENDIAN__
  52196. __ai uint32x4_t vpmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  52197.   uint32x4_t __ret;
  52198.   __ret = (uint32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  52199.   return __ret;
  52200. }
  52201. #else
  52202. __ai uint32x4_t vpmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  52203.   uint32x4_t __ret;
  52204.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52205.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52206.   __ret = (uint32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  52207.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52208.   return __ret;
  52209. }
  52210. #endif
  52211.  
  52212. #ifdef __LITTLE_ENDIAN__
  52213. __ai uint16x8_t vpmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  52214.   uint16x8_t __ret;
  52215.   __ret = (uint16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  52216.   return __ret;
  52217. }
  52218. #else
  52219. __ai uint16x8_t vpmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  52220.   uint16x8_t __ret;
  52221.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  52222.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  52223.   __ret = (uint16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  52224.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  52225.   return __ret;
  52226. }
  52227. #endif
  52228.  
  52229. #ifdef __LITTLE_ENDIAN__
  52230. __ai int8x16_t vpmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
  52231.   int8x16_t __ret;
  52232.   __ret = (int8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  52233.   return __ret;
  52234. }
  52235. #else
  52236. __ai int8x16_t vpmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
  52237.   int8x16_t __ret;
  52238.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52239.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52240.   __ret = (int8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  52241.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52242.   return __ret;
  52243. }
  52244. #endif
  52245.  
  52246. #ifdef __LITTLE_ENDIAN__
  52247. __ai float64x2_t vpmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
  52248.   float64x2_t __ret;
  52249.   __ret = (float64x2_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  52250.   return __ret;
  52251. }
  52252. #else
  52253. __ai float64x2_t vpmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
  52254.   float64x2_t __ret;
  52255.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52256.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  52257.   __ret = (float64x2_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  52258.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52259.   return __ret;
  52260. }
  52261. #endif
  52262.  
  52263. #ifdef __LITTLE_ENDIAN__
  52264. __ai float32x4_t vpmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
  52265.   float32x4_t __ret;
  52266.   __ret = (float32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  52267.   return __ret;
  52268. }
  52269. #else
  52270. __ai float32x4_t vpmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
  52271.   float32x4_t __ret;
  52272.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52273.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52274.   __ret = (float32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  52275.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52276.   return __ret;
  52277. }
  52278. #endif
  52279.  
  52280. #ifdef __LITTLE_ENDIAN__
  52281. __ai int32x4_t vpmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
  52282.   int32x4_t __ret;
  52283.   __ret = (int32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  52284.   return __ret;
  52285. }
  52286. #else
  52287. __ai int32x4_t vpmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
  52288.   int32x4_t __ret;
  52289.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52290.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52291.   __ret = (int32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  52292.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52293.   return __ret;
  52294. }
  52295. #endif
  52296.  
  52297. #ifdef __LITTLE_ENDIAN__
  52298. __ai int16x8_t vpmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
  52299.   int16x8_t __ret;
  52300.   __ret = (int16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  52301.   return __ret;
  52302. }
  52303. #else
  52304. __ai int16x8_t vpmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
  52305.   int16x8_t __ret;
  52306.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  52307.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  52308.   __ret = (int16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  52309.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  52310.   return __ret;
  52311. }
  52312. #endif
  52313.  
  52314. #ifdef __LITTLE_ENDIAN__
  52315. __ai float64_t vpmaxqd_f64(float64x2_t __p0) {
  52316.   float64_t __ret;
  52317.   __ret = (float64_t) __builtin_neon_vpmaxqd_f64(__p0);
  52318.   return __ret;
  52319. }
  52320. #else
  52321. __ai float64_t vpmaxqd_f64(float64x2_t __p0) {
  52322.   float64_t __ret;
  52323.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52324.   __ret = (float64_t) __builtin_neon_vpmaxqd_f64(__rev0);
  52325.   return __ret;
  52326. }
  52327. #endif
  52328.  
  52329. #ifdef __LITTLE_ENDIAN__
  52330. __ai float32_t vpmaxs_f32(float32x2_t __p0) {
  52331.   float32_t __ret;
  52332.   __ret = (float32_t) __builtin_neon_vpmaxs_f32(__p0);
  52333.   return __ret;
  52334. }
  52335. #else
  52336. __ai float32_t vpmaxs_f32(float32x2_t __p0) {
  52337.   float32_t __ret;
  52338.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52339.   __ret = (float32_t) __builtin_neon_vpmaxs_f32(__rev0);
  52340.   return __ret;
  52341. }
  52342. #endif
  52343.  
  52344. #ifdef __LITTLE_ENDIAN__
  52345. __ai float64x2_t vpmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
  52346.   float64x2_t __ret;
  52347.   __ret = (float64x2_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  52348.   return __ret;
  52349. }
  52350. #else
  52351. __ai float64x2_t vpmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
  52352.   float64x2_t __ret;
  52353.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52354.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  52355.   __ret = (float64x2_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  52356.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52357.   return __ret;
  52358. }
  52359. #endif
  52360.  
  52361. #ifdef __LITTLE_ENDIAN__
  52362. __ai float32x4_t vpmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
  52363.   float32x4_t __ret;
  52364.   __ret = (float32x4_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  52365.   return __ret;
  52366. }
  52367. #else
  52368. __ai float32x4_t vpmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
  52369.   float32x4_t __ret;
  52370.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52371.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52372.   __ret = (float32x4_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  52373.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52374.   return __ret;
  52375. }
  52376. #endif
  52377.  
  52378. #ifdef __LITTLE_ENDIAN__
  52379. __ai float32x2_t vpmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
  52380.   float32x2_t __ret;
  52381.   __ret = (float32x2_t) __builtin_neon_vpmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  52382.   return __ret;
  52383. }
  52384. #else
  52385. __ai float32x2_t vpmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
  52386.   float32x2_t __ret;
  52387.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52388.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  52389.   __ret = (float32x2_t) __builtin_neon_vpmaxnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  52390.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52391.   return __ret;
  52392. }
  52393. #endif
  52394.  
  52395. #ifdef __LITTLE_ENDIAN__
  52396. __ai float64_t vpmaxnmqd_f64(float64x2_t __p0) {
  52397.   float64_t __ret;
  52398.   __ret = (float64_t) __builtin_neon_vpmaxnmqd_f64(__p0);
  52399.   return __ret;
  52400. }
  52401. #else
  52402. __ai float64_t vpmaxnmqd_f64(float64x2_t __p0) {
  52403.   float64_t __ret;
  52404.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52405.   __ret = (float64_t) __builtin_neon_vpmaxnmqd_f64(__rev0);
  52406.   return __ret;
  52407. }
  52408. #endif
  52409.  
  52410. #ifdef __LITTLE_ENDIAN__
  52411. __ai float32_t vpmaxnms_f32(float32x2_t __p0) {
  52412.   float32_t __ret;
  52413.   __ret = (float32_t) __builtin_neon_vpmaxnms_f32(__p0);
  52414.   return __ret;
  52415. }
  52416. #else
  52417. __ai float32_t vpmaxnms_f32(float32x2_t __p0) {
  52418.   float32_t __ret;
  52419.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52420.   __ret = (float32_t) __builtin_neon_vpmaxnms_f32(__rev0);
  52421.   return __ret;
  52422. }
  52423. #endif
  52424.  
  52425. #ifdef __LITTLE_ENDIAN__
  52426. __ai uint8x16_t vpminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  52427.   uint8x16_t __ret;
  52428.   __ret = (uint8x16_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  52429.   return __ret;
  52430. }
  52431. #else
  52432. __ai uint8x16_t vpminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
  52433.   uint8x16_t __ret;
  52434.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52435.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52436.   __ret = (uint8x16_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  52437.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52438.   return __ret;
  52439. }
  52440. #endif
  52441.  
  52442. #ifdef __LITTLE_ENDIAN__
  52443. __ai uint32x4_t vpminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  52444.   uint32x4_t __ret;
  52445.   __ret = (uint32x4_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  52446.   return __ret;
  52447. }
  52448. #else
  52449. __ai uint32x4_t vpminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  52450.   uint32x4_t __ret;
  52451.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52452.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52453.   __ret = (uint32x4_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  52454.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52455.   return __ret;
  52456. }
  52457. #endif
  52458.  
  52459. #ifdef __LITTLE_ENDIAN__
  52460. __ai uint16x8_t vpminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  52461.   uint16x8_t __ret;
  52462.   __ret = (uint16x8_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  52463.   return __ret;
  52464. }
  52465. #else
  52466. __ai uint16x8_t vpminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
  52467.   uint16x8_t __ret;
  52468.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  52469.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  52470.   __ret = (uint16x8_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  52471.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  52472.   return __ret;
  52473. }
  52474. #endif
  52475.  
  52476. #ifdef __LITTLE_ENDIAN__
  52477. __ai int8x16_t vpminq_s8(int8x16_t __p0, int8x16_t __p1) {
  52478.   int8x16_t __ret;
  52479.   __ret = (int8x16_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  52480.   return __ret;
  52481. }
  52482. #else
  52483. __ai int8x16_t vpminq_s8(int8x16_t __p0, int8x16_t __p1) {
  52484.   int8x16_t __ret;
  52485.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52486.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52487.   __ret = (int8x16_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  52488.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  52489.   return __ret;
  52490. }
  52491. #endif
  52492.  
  52493. #ifdef __LITTLE_ENDIAN__
  52494. __ai float64x2_t vpminq_f64(float64x2_t __p0, float64x2_t __p1) {
  52495.   float64x2_t __ret;
  52496.   __ret = (float64x2_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  52497.   return __ret;
  52498. }
  52499. #else
  52500. __ai float64x2_t vpminq_f64(float64x2_t __p0, float64x2_t __p1) {
  52501.   float64x2_t __ret;
  52502.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52503.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  52504.   __ret = (float64x2_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  52505.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52506.   return __ret;
  52507. }
  52508. #endif
  52509.  
  52510. #ifdef __LITTLE_ENDIAN__
  52511. __ai float32x4_t vpminq_f32(float32x4_t __p0, float32x4_t __p1) {
  52512.   float32x4_t __ret;
  52513.   __ret = (float32x4_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  52514.   return __ret;
  52515. }
  52516. #else
  52517. __ai float32x4_t vpminq_f32(float32x4_t __p0, float32x4_t __p1) {
  52518.   float32x4_t __ret;
  52519.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52520.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52521.   __ret = (float32x4_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  52522.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52523.   return __ret;
  52524. }
  52525. #endif
  52526.  
  52527. #ifdef __LITTLE_ENDIAN__
  52528. __ai int32x4_t vpminq_s32(int32x4_t __p0, int32x4_t __p1) {
  52529.   int32x4_t __ret;
  52530.   __ret = (int32x4_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  52531.   return __ret;
  52532. }
  52533. #else
  52534. __ai int32x4_t vpminq_s32(int32x4_t __p0, int32x4_t __p1) {
  52535.   int32x4_t __ret;
  52536.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52537.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52538.   __ret = (int32x4_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  52539.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52540.   return __ret;
  52541. }
  52542. #endif
  52543.  
  52544. #ifdef __LITTLE_ENDIAN__
  52545. __ai int16x8_t vpminq_s16(int16x8_t __p0, int16x8_t __p1) {
  52546.   int16x8_t __ret;
  52547.   __ret = (int16x8_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  52548.   return __ret;
  52549. }
  52550. #else
  52551. __ai int16x8_t vpminq_s16(int16x8_t __p0, int16x8_t __p1) {
  52552.   int16x8_t __ret;
  52553.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  52554.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  52555.   __ret = (int16x8_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  52556.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  52557.   return __ret;
  52558. }
  52559. #endif
  52560.  
  52561. #ifdef __LITTLE_ENDIAN__
  52562. __ai float64_t vpminqd_f64(float64x2_t __p0) {
  52563.   float64_t __ret;
  52564.   __ret = (float64_t) __builtin_neon_vpminqd_f64(__p0);
  52565.   return __ret;
  52566. }
  52567. #else
  52568. __ai float64_t vpminqd_f64(float64x2_t __p0) {
  52569.   float64_t __ret;
  52570.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52571.   __ret = (float64_t) __builtin_neon_vpminqd_f64(__rev0);
  52572.   return __ret;
  52573. }
  52574. #endif
  52575.  
  52576. #ifdef __LITTLE_ENDIAN__
  52577. __ai float32_t vpmins_f32(float32x2_t __p0) {
  52578.   float32_t __ret;
  52579.   __ret = (float32_t) __builtin_neon_vpmins_f32(__p0);
  52580.   return __ret;
  52581. }
  52582. #else
  52583. __ai float32_t vpmins_f32(float32x2_t __p0) {
  52584.   float32_t __ret;
  52585.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52586.   __ret = (float32_t) __builtin_neon_vpmins_f32(__rev0);
  52587.   return __ret;
  52588. }
  52589. #endif
  52590.  
  52591. #ifdef __LITTLE_ENDIAN__
  52592. __ai float64x2_t vpminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
  52593.   float64x2_t __ret;
  52594.   __ret = (float64x2_t) __builtin_neon_vpminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  52595.   return __ret;
  52596. }
  52597. #else
  52598. __ai float64x2_t vpminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
  52599.   float64x2_t __ret;
  52600.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52601.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  52602.   __ret = (float64x2_t) __builtin_neon_vpminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  52603.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52604.   return __ret;
  52605. }
  52606. #endif
  52607.  
  52608. #ifdef __LITTLE_ENDIAN__
  52609. __ai float32x4_t vpminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
  52610.   float32x4_t __ret;
  52611.   __ret = (float32x4_t) __builtin_neon_vpminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
  52612.   return __ret;
  52613. }
  52614. #else
  52615. __ai float32x4_t vpminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
  52616.   float32x4_t __ret;
  52617.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52618.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52619.   __ret = (float32x4_t) __builtin_neon_vpminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
  52620.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52621.   return __ret;
  52622. }
  52623. #endif
  52624.  
  52625. #ifdef __LITTLE_ENDIAN__
  52626. __ai float32x2_t vpminnm_f32(float32x2_t __p0, float32x2_t __p1) {
  52627.   float32x2_t __ret;
  52628.   __ret = (float32x2_t) __builtin_neon_vpminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
  52629.   return __ret;
  52630. }
  52631. #else
  52632. __ai float32x2_t vpminnm_f32(float32x2_t __p0, float32x2_t __p1) {
  52633.   float32x2_t __ret;
  52634.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52635.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  52636.   __ret = (float32x2_t) __builtin_neon_vpminnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
  52637.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52638.   return __ret;
  52639. }
  52640. #endif
  52641.  
  52642. #ifdef __LITTLE_ENDIAN__
  52643. __ai float64_t vpminnmqd_f64(float64x2_t __p0) {
  52644.   float64_t __ret;
  52645.   __ret = (float64_t) __builtin_neon_vpminnmqd_f64(__p0);
  52646.   return __ret;
  52647. }
  52648. #else
  52649. __ai float64_t vpminnmqd_f64(float64x2_t __p0) {
  52650.   float64_t __ret;
  52651.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52652.   __ret = (float64_t) __builtin_neon_vpminnmqd_f64(__rev0);
  52653.   return __ret;
  52654. }
  52655. #endif
  52656.  
  52657. #ifdef __LITTLE_ENDIAN__
  52658. __ai float32_t vpminnms_f32(float32x2_t __p0) {
  52659.   float32_t __ret;
  52660.   __ret = (float32_t) __builtin_neon_vpminnms_f32(__p0);
  52661.   return __ret;
  52662. }
  52663. #else
  52664. __ai float32_t vpminnms_f32(float32x2_t __p0) {
  52665.   float32_t __ret;
  52666.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52667.   __ret = (float32_t) __builtin_neon_vpminnms_f32(__rev0);
  52668.   return __ret;
  52669. }
  52670. #endif
  52671.  
  52672. #ifdef __LITTLE_ENDIAN__
  52673. __ai int64x2_t vqabsq_s64(int64x2_t __p0) {
  52674.   int64x2_t __ret;
  52675.   __ret = (int64x2_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 35);
  52676.   return __ret;
  52677. }
  52678. #else
  52679. __ai int64x2_t vqabsq_s64(int64x2_t __p0) {
  52680.   int64x2_t __ret;
  52681.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52682.   __ret = (int64x2_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 35);
  52683.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52684.   return __ret;
  52685. }
  52686. #endif
  52687.  
  52688. __ai int64x1_t vqabs_s64(int64x1_t __p0) {
  52689.   int64x1_t __ret;
  52690.   __ret = (int64x1_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 3);
  52691.   return __ret;
  52692. }
  52693. __ai int8_t vqabsb_s8(int8_t __p0) {
  52694.   int8_t __ret;
  52695.   __ret = (int8_t) __builtin_neon_vqabsb_s8(__p0);
  52696.   return __ret;
  52697. }
  52698. __ai int32_t vqabss_s32(int32_t __p0) {
  52699.   int32_t __ret;
  52700.   __ret = (int32_t) __builtin_neon_vqabss_s32(__p0);
  52701.   return __ret;
  52702. }
  52703. __ai int64_t vqabsd_s64(int64_t __p0) {
  52704.   int64_t __ret;
  52705.   __ret = (int64_t) __builtin_neon_vqabsd_s64(__p0);
  52706.   return __ret;
  52707. }
  52708. __ai int16_t vqabsh_s16(int16_t __p0) {
  52709.   int16_t __ret;
  52710.   __ret = (int16_t) __builtin_neon_vqabsh_s16(__p0);
  52711.   return __ret;
  52712. }
  52713. __ai uint8_t vqaddb_u8(uint8_t __p0, uint8_t __p1) {
  52714.   uint8_t __ret;
  52715.   __ret = (uint8_t) __builtin_neon_vqaddb_u8(__p0, __p1);
  52716.   return __ret;
  52717. }
  52718. __ai uint32_t vqadds_u32(uint32_t __p0, uint32_t __p1) {
  52719.   uint32_t __ret;
  52720.   __ret = (uint32_t) __builtin_neon_vqadds_u32(__p0, __p1);
  52721.   return __ret;
  52722. }
  52723. __ai uint64_t vqaddd_u64(uint64_t __p0, uint64_t __p1) {
  52724.   uint64_t __ret;
  52725.   __ret = (uint64_t) __builtin_neon_vqaddd_u64(__p0, __p1);
  52726.   return __ret;
  52727. }
  52728. __ai uint16_t vqaddh_u16(uint16_t __p0, uint16_t __p1) {
  52729.   uint16_t __ret;
  52730.   __ret = (uint16_t) __builtin_neon_vqaddh_u16(__p0, __p1);
  52731.   return __ret;
  52732. }
  52733. __ai int8_t vqaddb_s8(int8_t __p0, int8_t __p1) {
  52734.   int8_t __ret;
  52735.   __ret = (int8_t) __builtin_neon_vqaddb_s8(__p0, __p1);
  52736.   return __ret;
  52737. }
  52738. __ai int32_t vqadds_s32(int32_t __p0, int32_t __p1) {
  52739.   int32_t __ret;
  52740.   __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
  52741.   return __ret;
  52742. }
  52743. __ai int64_t vqaddd_s64(int64_t __p0, int64_t __p1) {
  52744.   int64_t __ret;
  52745.   __ret = (int64_t) __builtin_neon_vqaddd_s64(__p0, __p1);
  52746.   return __ret;
  52747. }
  52748. __ai int16_t vqaddh_s16(int16_t __p0, int16_t __p1) {
  52749.   int16_t __ret;
  52750.   __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
  52751.   return __ret;
  52752. }
  52753. __ai int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
  52754.   int64_t __ret;
  52755.   __ret = (int64_t) __builtin_neon_vqdmlals_s32(__p0, __p1, __p2);
  52756.   return __ret;
  52757. }
  52758. __ai int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
  52759.   int32_t __ret;
  52760.   __ret = (int32_t) __builtin_neon_vqdmlalh_s16(__p0, __p1, __p2);
  52761.   return __ret;
  52762. }
  52763. #ifdef __LITTLE_ENDIAN__
  52764. __ai int64x2_t vqdmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  52765.   int64x2_t __ret;
  52766.   __ret = vqdmlal_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
  52767.   return __ret;
  52768. }
  52769. #else
  52770. __ai int64x2_t vqdmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  52771.   int64x2_t __ret;
  52772.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52773.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52774.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  52775.   __ret = __noswap_vqdmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
  52776.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52777.   return __ret;
  52778. }
  52779. #endif
  52780.  
  52781. #ifdef __LITTLE_ENDIAN__
  52782. __ai int32x4_t vqdmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  52783.   int32x4_t __ret;
  52784.   __ret = vqdmlal_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
  52785.   return __ret;
  52786. }
  52787. #else
  52788. __ai int32x4_t vqdmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  52789.   int32x4_t __ret;
  52790.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52791.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  52792.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  52793.   __ret = __noswap_vqdmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
  52794.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52795.   return __ret;
  52796. }
  52797. #endif
  52798.  
  52799. #ifdef __LITTLE_ENDIAN__
  52800. #define vqdmlal_high_lane_s32(__p0_607, __p1_607, __p2_607, __p3_607) __extension__ ({ \
  52801.   int64x2_t __ret_607; \
  52802.   int64x2_t __s0_607 = __p0_607; \
  52803.   int32x4_t __s1_607 = __p1_607; \
  52804.   int32x2_t __s2_607 = __p2_607; \
  52805.   __ret_607 = vqdmlal_s32(__s0_607, vget_high_s32(__s1_607), splat_lane_s32(__s2_607, __p3_607)); \
  52806.   __ret_607; \
  52807. })
  52808. #else
  52809. #define vqdmlal_high_lane_s32(__p0_608, __p1_608, __p2_608, __p3_608) __extension__ ({ \
  52810.   int64x2_t __ret_608; \
  52811.   int64x2_t __s0_608 = __p0_608; \
  52812.   int32x4_t __s1_608 = __p1_608; \
  52813.   int32x2_t __s2_608 = __p2_608; \
  52814.   int64x2_t __rev0_608;  __rev0_608 = __builtin_shufflevector(__s0_608, __s0_608, 1, 0); \
  52815.   int32x4_t __rev1_608;  __rev1_608 = __builtin_shufflevector(__s1_608, __s1_608, 3, 2, 1, 0); \
  52816.   int32x2_t __rev2_608;  __rev2_608 = __builtin_shufflevector(__s2_608, __s2_608, 1, 0); \
  52817.   __ret_608 = __noswap_vqdmlal_s32(__rev0_608, __noswap_vget_high_s32(__rev1_608), __noswap_splat_lane_s32(__rev2_608, __p3_608)); \
  52818.   __ret_608 = __builtin_shufflevector(__ret_608, __ret_608, 1, 0); \
  52819.   __ret_608; \
  52820. })
  52821. #endif
  52822.  
  52823. #ifdef __LITTLE_ENDIAN__
  52824. #define vqdmlal_high_lane_s16(__p0_609, __p1_609, __p2_609, __p3_609) __extension__ ({ \
  52825.   int32x4_t __ret_609; \
  52826.   int32x4_t __s0_609 = __p0_609; \
  52827.   int16x8_t __s1_609 = __p1_609; \
  52828.   int16x4_t __s2_609 = __p2_609; \
  52829.   __ret_609 = vqdmlal_s16(__s0_609, vget_high_s16(__s1_609), splat_lane_s16(__s2_609, __p3_609)); \
  52830.   __ret_609; \
  52831. })
  52832. #else
  52833. #define vqdmlal_high_lane_s16(__p0_610, __p1_610, __p2_610, __p3_610) __extension__ ({ \
  52834.   int32x4_t __ret_610; \
  52835.   int32x4_t __s0_610 = __p0_610; \
  52836.   int16x8_t __s1_610 = __p1_610; \
  52837.   int16x4_t __s2_610 = __p2_610; \
  52838.   int32x4_t __rev0_610;  __rev0_610 = __builtin_shufflevector(__s0_610, __s0_610, 3, 2, 1, 0); \
  52839.   int16x8_t __rev1_610;  __rev1_610 = __builtin_shufflevector(__s1_610, __s1_610, 7, 6, 5, 4, 3, 2, 1, 0); \
  52840.   int16x4_t __rev2_610;  __rev2_610 = __builtin_shufflevector(__s2_610, __s2_610, 3, 2, 1, 0); \
  52841.   __ret_610 = __noswap_vqdmlal_s16(__rev0_610, __noswap_vget_high_s16(__rev1_610), __noswap_splat_lane_s16(__rev2_610, __p3_610)); \
  52842.   __ret_610 = __builtin_shufflevector(__ret_610, __ret_610, 3, 2, 1, 0); \
  52843.   __ret_610; \
  52844. })
  52845. #endif
  52846.  
  52847. #ifdef __LITTLE_ENDIAN__
  52848. #define vqdmlal_high_laneq_s32(__p0_611, __p1_611, __p2_611, __p3_611) __extension__ ({ \
  52849.   int64x2_t __ret_611; \
  52850.   int64x2_t __s0_611 = __p0_611; \
  52851.   int32x4_t __s1_611 = __p1_611; \
  52852.   int32x4_t __s2_611 = __p2_611; \
  52853.   __ret_611 = vqdmlal_s32(__s0_611, vget_high_s32(__s1_611), splat_laneq_s32(__s2_611, __p3_611)); \
  52854.   __ret_611; \
  52855. })
  52856. #else
  52857. #define vqdmlal_high_laneq_s32(__p0_612, __p1_612, __p2_612, __p3_612) __extension__ ({ \
  52858.   int64x2_t __ret_612; \
  52859.   int64x2_t __s0_612 = __p0_612; \
  52860.   int32x4_t __s1_612 = __p1_612; \
  52861.   int32x4_t __s2_612 = __p2_612; \
  52862.   int64x2_t __rev0_612;  __rev0_612 = __builtin_shufflevector(__s0_612, __s0_612, 1, 0); \
  52863.   int32x4_t __rev1_612;  __rev1_612 = __builtin_shufflevector(__s1_612, __s1_612, 3, 2, 1, 0); \
  52864.   int32x4_t __rev2_612;  __rev2_612 = __builtin_shufflevector(__s2_612, __s2_612, 3, 2, 1, 0); \
  52865.   __ret_612 = __noswap_vqdmlal_s32(__rev0_612, __noswap_vget_high_s32(__rev1_612), __noswap_splat_laneq_s32(__rev2_612, __p3_612)); \
  52866.   __ret_612 = __builtin_shufflevector(__ret_612, __ret_612, 1, 0); \
  52867.   __ret_612; \
  52868. })
  52869. #endif
  52870.  
  52871. #ifdef __LITTLE_ENDIAN__
  52872. #define vqdmlal_high_laneq_s16(__p0_613, __p1_613, __p2_613, __p3_613) __extension__ ({ \
  52873.   int32x4_t __ret_613; \
  52874.   int32x4_t __s0_613 = __p0_613; \
  52875.   int16x8_t __s1_613 = __p1_613; \
  52876.   int16x8_t __s2_613 = __p2_613; \
  52877.   __ret_613 = vqdmlal_s16(__s0_613, vget_high_s16(__s1_613), splat_laneq_s16(__s2_613, __p3_613)); \
  52878.   __ret_613; \
  52879. })
  52880. #else
  52881. #define vqdmlal_high_laneq_s16(__p0_614, __p1_614, __p2_614, __p3_614) __extension__ ({ \
  52882.   int32x4_t __ret_614; \
  52883.   int32x4_t __s0_614 = __p0_614; \
  52884.   int16x8_t __s1_614 = __p1_614; \
  52885.   int16x8_t __s2_614 = __p2_614; \
  52886.   int32x4_t __rev0_614;  __rev0_614 = __builtin_shufflevector(__s0_614, __s0_614, 3, 2, 1, 0); \
  52887.   int16x8_t __rev1_614;  __rev1_614 = __builtin_shufflevector(__s1_614, __s1_614, 7, 6, 5, 4, 3, 2, 1, 0); \
  52888.   int16x8_t __rev2_614;  __rev2_614 = __builtin_shufflevector(__s2_614, __s2_614, 7, 6, 5, 4, 3, 2, 1, 0); \
  52889.   __ret_614 = __noswap_vqdmlal_s16(__rev0_614, __noswap_vget_high_s16(__rev1_614), __noswap_splat_laneq_s16(__rev2_614, __p3_614)); \
  52890.   __ret_614 = __builtin_shufflevector(__ret_614, __ret_614, 3, 2, 1, 0); \
  52891.   __ret_614; \
  52892. })
  52893. #endif
  52894.  
  52895. #ifdef __LITTLE_ENDIAN__
  52896. __ai int64x2_t vqdmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
  52897.   int64x2_t __ret;
  52898.   __ret = vqdmlal_n_s32(__p0, vget_high_s32(__p1), __p2);
  52899.   return __ret;
  52900. }
  52901. #else
  52902. __ai int64x2_t vqdmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
  52903.   int64x2_t __ret;
  52904.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  52905.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  52906.   __ret = __noswap_vqdmlal_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
  52907.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  52908.   return __ret;
  52909. }
  52910. #endif
  52911.  
  52912. #ifdef __LITTLE_ENDIAN__
  52913. __ai int32x4_t vqdmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
  52914.   int32x4_t __ret;
  52915.   __ret = vqdmlal_n_s16(__p0, vget_high_s16(__p1), __p2);
  52916.   return __ret;
  52917. }
  52918. #else
  52919. __ai int32x4_t vqdmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
  52920.   int32x4_t __ret;
  52921.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  52922.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  52923.   __ret = __noswap_vqdmlal_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
  52924.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  52925.   return __ret;
  52926. }
  52927. #endif
  52928.  
  52929. #ifdef __LITTLE_ENDIAN__
  52930. #define vqdmlals_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
  52931.   int64_t __ret; \
  52932.   int64_t __s0 = __p0; \
  52933.   int32_t __s1 = __p1; \
  52934.   int32x2_t __s2 = __p2; \
  52935.   __ret = (int64_t) __builtin_neon_vqdmlals_lane_s32(__s0, __s1, __s2, __p3); \
  52936.   __ret; \
  52937. })
  52938. #else
  52939. #define vqdmlals_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
  52940.   int64_t __ret; \
  52941.   int64_t __s0 = __p0; \
  52942.   int32_t __s1 = __p1; \
  52943.   int32x2_t __s2 = __p2; \
  52944.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
  52945.   __ret = (int64_t) __builtin_neon_vqdmlals_lane_s32(__s0, __s1, __rev2, __p3); \
  52946.   __ret; \
  52947. })
  52948. #endif
  52949.  
  52950. #ifdef __LITTLE_ENDIAN__
  52951. #define vqdmlalh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
  52952.   int32_t __ret; \
  52953.   int32_t __s0 = __p0; \
  52954.   int16_t __s1 = __p1; \
  52955.   int16x4_t __s2 = __p2; \
  52956.   __ret = (int32_t) __builtin_neon_vqdmlalh_lane_s16(__s0, __s1, __s2, __p3); \
  52957.   __ret; \
  52958. })
  52959. #else
  52960. #define vqdmlalh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
  52961.   int32_t __ret; \
  52962.   int32_t __s0 = __p0; \
  52963.   int16_t __s1 = __p1; \
  52964.   int16x4_t __s2 = __p2; \
  52965.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  52966.   __ret = (int32_t) __builtin_neon_vqdmlalh_lane_s16(__s0, __s1, __rev2, __p3); \
  52967.   __ret; \
  52968. })
  52969. #endif
  52970.  
  52971. #ifdef __LITTLE_ENDIAN__
  52972. #define vqdmlals_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
  52973.   int64_t __ret; \
  52974.   int64_t __s0 = __p0; \
  52975.   int32_t __s1 = __p1; \
  52976.   int32x4_t __s2 = __p2; \
  52977.   __ret = (int64_t) __builtin_neon_vqdmlals_laneq_s32(__s0, __s1, __s2, __p3); \
  52978.   __ret; \
  52979. })
  52980. #else
  52981. #define vqdmlals_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
  52982.   int64_t __ret; \
  52983.   int64_t __s0 = __p0; \
  52984.   int32_t __s1 = __p1; \
  52985.   int32x4_t __s2 = __p2; \
  52986.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  52987.   __ret = (int64_t) __builtin_neon_vqdmlals_laneq_s32(__s0, __s1, __rev2, __p3); \
  52988.   __ret; \
  52989. })
  52990. #endif
  52991.  
  52992. #ifdef __LITTLE_ENDIAN__
  52993. #define vqdmlalh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
  52994.   int32_t __ret; \
  52995.   int32_t __s0 = __p0; \
  52996.   int16_t __s1 = __p1; \
  52997.   int16x8_t __s2 = __p2; \
  52998.   __ret = (int32_t) __builtin_neon_vqdmlalh_laneq_s16(__s0, __s1, __s2, __p3); \
  52999.   __ret; \
  53000. })
  53001. #else
  53002. #define vqdmlalh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
  53003.   int32_t __ret; \
  53004.   int32_t __s0 = __p0; \
  53005.   int16_t __s1 = __p1; \
  53006.   int16x8_t __s2 = __p2; \
  53007.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
  53008.   __ret = (int32_t) __builtin_neon_vqdmlalh_laneq_s16(__s0, __s1, __rev2, __p3); \
  53009.   __ret; \
  53010. })
  53011. #endif
  53012.  
  53013. #ifdef __LITTLE_ENDIAN__
  53014. #define vqdmlal_laneq_s32(__p0_615, __p1_615, __p2_615, __p3_615) __extension__ ({ \
  53015.   int64x2_t __ret_615; \
  53016.   int64x2_t __s0_615 = __p0_615; \
  53017.   int32x2_t __s1_615 = __p1_615; \
  53018.   int32x4_t __s2_615 = __p2_615; \
  53019.   __ret_615 = vqdmlal_s32(__s0_615, __s1_615, splat_laneq_s32(__s2_615, __p3_615)); \
  53020.   __ret_615; \
  53021. })
  53022. #else
  53023. #define vqdmlal_laneq_s32(__p0_616, __p1_616, __p2_616, __p3_616) __extension__ ({ \
  53024.   int64x2_t __ret_616; \
  53025.   int64x2_t __s0_616 = __p0_616; \
  53026.   int32x2_t __s1_616 = __p1_616; \
  53027.   int32x4_t __s2_616 = __p2_616; \
  53028.   int64x2_t __rev0_616;  __rev0_616 = __builtin_shufflevector(__s0_616, __s0_616, 1, 0); \
  53029.   int32x2_t __rev1_616;  __rev1_616 = __builtin_shufflevector(__s1_616, __s1_616, 1, 0); \
  53030.   int32x4_t __rev2_616;  __rev2_616 = __builtin_shufflevector(__s2_616, __s2_616, 3, 2, 1, 0); \
  53031.   __ret_616 = __noswap_vqdmlal_s32(__rev0_616, __rev1_616, __noswap_splat_laneq_s32(__rev2_616, __p3_616)); \
  53032.   __ret_616 = __builtin_shufflevector(__ret_616, __ret_616, 1, 0); \
  53033.   __ret_616; \
  53034. })
  53035. #endif
  53036.  
  53037. #ifdef __LITTLE_ENDIAN__
  53038. #define vqdmlal_laneq_s16(__p0_617, __p1_617, __p2_617, __p3_617) __extension__ ({ \
  53039.   int32x4_t __ret_617; \
  53040.   int32x4_t __s0_617 = __p0_617; \
  53041.   int16x4_t __s1_617 = __p1_617; \
  53042.   int16x8_t __s2_617 = __p2_617; \
  53043.   __ret_617 = vqdmlal_s16(__s0_617, __s1_617, splat_laneq_s16(__s2_617, __p3_617)); \
  53044.   __ret_617; \
  53045. })
  53046. #else
  53047. #define vqdmlal_laneq_s16(__p0_618, __p1_618, __p2_618, __p3_618) __extension__ ({ \
  53048.   int32x4_t __ret_618; \
  53049.   int32x4_t __s0_618 = __p0_618; \
  53050.   int16x4_t __s1_618 = __p1_618; \
  53051.   int16x8_t __s2_618 = __p2_618; \
  53052.   int32x4_t __rev0_618;  __rev0_618 = __builtin_shufflevector(__s0_618, __s0_618, 3, 2, 1, 0); \
  53053.   int16x4_t __rev1_618;  __rev1_618 = __builtin_shufflevector(__s1_618, __s1_618, 3, 2, 1, 0); \
  53054.   int16x8_t __rev2_618;  __rev2_618 = __builtin_shufflevector(__s2_618, __s2_618, 7, 6, 5, 4, 3, 2, 1, 0); \
  53055.   __ret_618 = __noswap_vqdmlal_s16(__rev0_618, __rev1_618, __noswap_splat_laneq_s16(__rev2_618, __p3_618)); \
  53056.   __ret_618 = __builtin_shufflevector(__ret_618, __ret_618, 3, 2, 1, 0); \
  53057.   __ret_618; \
  53058. })
  53059. #endif
  53060.  
  53061. __ai int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
  53062.   int64_t __ret;
  53063.   __ret = (int64_t) __builtin_neon_vqdmlsls_s32(__p0, __p1, __p2);
  53064.   return __ret;
  53065. }
  53066. __ai int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
  53067.   int32_t __ret;
  53068.   __ret = (int32_t) __builtin_neon_vqdmlslh_s16(__p0, __p1, __p2);
  53069.   return __ret;
  53070. }
  53071. #ifdef __LITTLE_ENDIAN__
  53072. __ai int64x2_t vqdmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  53073.   int64x2_t __ret;
  53074.   __ret = vqdmlsl_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
  53075.   return __ret;
  53076. }
  53077. #else
  53078. __ai int64x2_t vqdmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  53079.   int64x2_t __ret;
  53080.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  53081.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  53082.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  53083.   __ret = __noswap_vqdmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
  53084.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  53085.   return __ret;
  53086. }
  53087. #endif
  53088.  
  53089. #ifdef __LITTLE_ENDIAN__
  53090. __ai int32x4_t vqdmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  53091.   int32x4_t __ret;
  53092.   __ret = vqdmlsl_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
  53093.   return __ret;
  53094. }
  53095. #else
  53096. __ai int32x4_t vqdmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  53097.   int32x4_t __ret;
  53098.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  53099.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  53100.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  53101.   __ret = __noswap_vqdmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
  53102.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  53103.   return __ret;
  53104. }
  53105. #endif
  53106.  
  53107. #ifdef __LITTLE_ENDIAN__
  53108. #define vqdmlsl_high_lane_s32(__p0_619, __p1_619, __p2_619, __p3_619) __extension__ ({ \
  53109.   int64x2_t __ret_619; \
  53110.   int64x2_t __s0_619 = __p0_619; \
  53111.   int32x4_t __s1_619 = __p1_619; \
  53112.   int32x2_t __s2_619 = __p2_619; \
  53113.   __ret_619 = vqdmlsl_s32(__s0_619, vget_high_s32(__s1_619), splat_lane_s32(__s2_619, __p3_619)); \
  53114.   __ret_619; \
  53115. })
  53116. #else
  53117. #define vqdmlsl_high_lane_s32(__p0_620, __p1_620, __p2_620, __p3_620) __extension__ ({ \
  53118.   int64x2_t __ret_620; \
  53119.   int64x2_t __s0_620 = __p0_620; \
  53120.   int32x4_t __s1_620 = __p1_620; \
  53121.   int32x2_t __s2_620 = __p2_620; \
  53122.   int64x2_t __rev0_620;  __rev0_620 = __builtin_shufflevector(__s0_620, __s0_620, 1, 0); \
  53123.   int32x4_t __rev1_620;  __rev1_620 = __builtin_shufflevector(__s1_620, __s1_620, 3, 2, 1, 0); \
  53124.   int32x2_t __rev2_620;  __rev2_620 = __builtin_shufflevector(__s2_620, __s2_620, 1, 0); \
  53125.   __ret_620 = __noswap_vqdmlsl_s32(__rev0_620, __noswap_vget_high_s32(__rev1_620), __noswap_splat_lane_s32(__rev2_620, __p3_620)); \
  53126.   __ret_620 = __builtin_shufflevector(__ret_620, __ret_620, 1, 0); \
  53127.   __ret_620; \
  53128. })
  53129. #endif
  53130.  
  53131. #ifdef __LITTLE_ENDIAN__
  53132. #define vqdmlsl_high_lane_s16(__p0_621, __p1_621, __p2_621, __p3_621) __extension__ ({ \
  53133.   int32x4_t __ret_621; \
  53134.   int32x4_t __s0_621 = __p0_621; \
  53135.   int16x8_t __s1_621 = __p1_621; \
  53136.   int16x4_t __s2_621 = __p2_621; \
  53137.   __ret_621 = vqdmlsl_s16(__s0_621, vget_high_s16(__s1_621), splat_lane_s16(__s2_621, __p3_621)); \
  53138.   __ret_621; \
  53139. })
  53140. #else
  53141. #define vqdmlsl_high_lane_s16(__p0_622, __p1_622, __p2_622, __p3_622) __extension__ ({ \
  53142.   int32x4_t __ret_622; \
  53143.   int32x4_t __s0_622 = __p0_622; \
  53144.   int16x8_t __s1_622 = __p1_622; \
  53145.   int16x4_t __s2_622 = __p2_622; \
  53146.   int32x4_t __rev0_622;  __rev0_622 = __builtin_shufflevector(__s0_622, __s0_622, 3, 2, 1, 0); \
  53147.   int16x8_t __rev1_622;  __rev1_622 = __builtin_shufflevector(__s1_622, __s1_622, 7, 6, 5, 4, 3, 2, 1, 0); \
  53148.   int16x4_t __rev2_622;  __rev2_622 = __builtin_shufflevector(__s2_622, __s2_622, 3, 2, 1, 0); \
  53149.   __ret_622 = __noswap_vqdmlsl_s16(__rev0_622, __noswap_vget_high_s16(__rev1_622), __noswap_splat_lane_s16(__rev2_622, __p3_622)); \
  53150.   __ret_622 = __builtin_shufflevector(__ret_622, __ret_622, 3, 2, 1, 0); \
  53151.   __ret_622; \
  53152. })
  53153. #endif
  53154.  
  53155. #ifdef __LITTLE_ENDIAN__
  53156. #define vqdmlsl_high_laneq_s32(__p0_623, __p1_623, __p2_623, __p3_623) __extension__ ({ \
  53157.   int64x2_t __ret_623; \
  53158.   int64x2_t __s0_623 = __p0_623; \
  53159.   int32x4_t __s1_623 = __p1_623; \
  53160.   int32x4_t __s2_623 = __p2_623; \
  53161.   __ret_623 = vqdmlsl_s32(__s0_623, vget_high_s32(__s1_623), splat_laneq_s32(__s2_623, __p3_623)); \
  53162.   __ret_623; \
  53163. })
  53164. #else
  53165. #define vqdmlsl_high_laneq_s32(__p0_624, __p1_624, __p2_624, __p3_624) __extension__ ({ \
  53166.   int64x2_t __ret_624; \
  53167.   int64x2_t __s0_624 = __p0_624; \
  53168.   int32x4_t __s1_624 = __p1_624; \
  53169.   int32x4_t __s2_624 = __p2_624; \
  53170.   int64x2_t __rev0_624;  __rev0_624 = __builtin_shufflevector(__s0_624, __s0_624, 1, 0); \
  53171.   int32x4_t __rev1_624;  __rev1_624 = __builtin_shufflevector(__s1_624, __s1_624, 3, 2, 1, 0); \
  53172.   int32x4_t __rev2_624;  __rev2_624 = __builtin_shufflevector(__s2_624, __s2_624, 3, 2, 1, 0); \
  53173.   __ret_624 = __noswap_vqdmlsl_s32(__rev0_624, __noswap_vget_high_s32(__rev1_624), __noswap_splat_laneq_s32(__rev2_624, __p3_624)); \
  53174.   __ret_624 = __builtin_shufflevector(__ret_624, __ret_624, 1, 0); \
  53175.   __ret_624; \
  53176. })
  53177. #endif
  53178.  
  53179. #ifdef __LITTLE_ENDIAN__
  53180. #define vqdmlsl_high_laneq_s16(__p0_625, __p1_625, __p2_625, __p3_625) __extension__ ({ \
  53181.   int32x4_t __ret_625; \
  53182.   int32x4_t __s0_625 = __p0_625; \
  53183.   int16x8_t __s1_625 = __p1_625; \
  53184.   int16x8_t __s2_625 = __p2_625; \
  53185.   __ret_625 = vqdmlsl_s16(__s0_625, vget_high_s16(__s1_625), splat_laneq_s16(__s2_625, __p3_625)); \
  53186.   __ret_625; \
  53187. })
  53188. #else
  53189. #define vqdmlsl_high_laneq_s16(__p0_626, __p1_626, __p2_626, __p3_626) __extension__ ({ \
  53190.   int32x4_t __ret_626; \
  53191.   int32x4_t __s0_626 = __p0_626; \
  53192.   int16x8_t __s1_626 = __p1_626; \
  53193.   int16x8_t __s2_626 = __p2_626; \
  53194.   int32x4_t __rev0_626;  __rev0_626 = __builtin_shufflevector(__s0_626, __s0_626, 3, 2, 1, 0); \
  53195.   int16x8_t __rev1_626;  __rev1_626 = __builtin_shufflevector(__s1_626, __s1_626, 7, 6, 5, 4, 3, 2, 1, 0); \
  53196.   int16x8_t __rev2_626;  __rev2_626 = __builtin_shufflevector(__s2_626, __s2_626, 7, 6, 5, 4, 3, 2, 1, 0); \
  53197.   __ret_626 = __noswap_vqdmlsl_s16(__rev0_626, __noswap_vget_high_s16(__rev1_626), __noswap_splat_laneq_s16(__rev2_626, __p3_626)); \
  53198.   __ret_626 = __builtin_shufflevector(__ret_626, __ret_626, 3, 2, 1, 0); \
  53199.   __ret_626; \
  53200. })
  53201. #endif
  53202.  
  53203. #ifdef __LITTLE_ENDIAN__
  53204. __ai int64x2_t vqdmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
  53205.   int64x2_t __ret;
  53206.   __ret = vqdmlsl_n_s32(__p0, vget_high_s32(__p1), __p2);
  53207.   return __ret;
  53208. }
  53209. #else
  53210. __ai int64x2_t vqdmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
  53211.   int64x2_t __ret;
  53212.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  53213.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  53214.   __ret = __noswap_vqdmlsl_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
  53215.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  53216.   return __ret;
  53217. }
  53218. #endif
  53219.  
  53220. #ifdef __LITTLE_ENDIAN__
  53221. __ai int32x4_t vqdmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
  53222.   int32x4_t __ret;
  53223.   __ret = vqdmlsl_n_s16(__p0, vget_high_s16(__p1), __p2);
  53224.   return __ret;
  53225. }
  53226. #else
  53227. __ai int32x4_t vqdmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
  53228.   int32x4_t __ret;
  53229.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  53230.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  53231.   __ret = __noswap_vqdmlsl_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
  53232.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  53233.   return __ret;
  53234. }
  53235. #endif
  53236.  
  53237. #ifdef __LITTLE_ENDIAN__
  53238. #define vqdmlsls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
  53239.   int64_t __ret; \
  53240.   int64_t __s0 = __p0; \
  53241.   int32_t __s1 = __p1; \
  53242.   int32x2_t __s2 = __p2; \
  53243.   __ret = (int64_t) __builtin_neon_vqdmlsls_lane_s32(__s0, __s1, __s2, __p3); \
  53244.   __ret; \
  53245. })
  53246. #else
  53247. #define vqdmlsls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
  53248.   int64_t __ret; \
  53249.   int64_t __s0 = __p0; \
  53250.   int32_t __s1 = __p1; \
  53251.   int32x2_t __s2 = __p2; \
  53252.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
  53253.   __ret = (int64_t) __builtin_neon_vqdmlsls_lane_s32(__s0, __s1, __rev2, __p3); \
  53254.   __ret; \
  53255. })
  53256. #endif
  53257.  
  53258. #ifdef __LITTLE_ENDIAN__
  53259. #define vqdmlslh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
  53260.   int32_t __ret; \
  53261.   int32_t __s0 = __p0; \
  53262.   int16_t __s1 = __p1; \
  53263.   int16x4_t __s2 = __p2; \
  53264.   __ret = (int32_t) __builtin_neon_vqdmlslh_lane_s16(__s0, __s1, __s2, __p3); \
  53265.   __ret; \
  53266. })
  53267. #else
  53268. #define vqdmlslh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
  53269.   int32_t __ret; \
  53270.   int32_t __s0 = __p0; \
  53271.   int16_t __s1 = __p1; \
  53272.   int16x4_t __s2 = __p2; \
  53273.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  53274.   __ret = (int32_t) __builtin_neon_vqdmlslh_lane_s16(__s0, __s1, __rev2, __p3); \
  53275.   __ret; \
  53276. })
  53277. #endif
  53278.  
  53279. #ifdef __LITTLE_ENDIAN__
  53280. #define vqdmlsls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
  53281.   int64_t __ret; \
  53282.   int64_t __s0 = __p0; \
  53283.   int32_t __s1 = __p1; \
  53284.   int32x4_t __s2 = __p2; \
  53285.   __ret = (int64_t) __builtin_neon_vqdmlsls_laneq_s32(__s0, __s1, __s2, __p3); \
  53286.   __ret; \
  53287. })
  53288. #else
  53289. #define vqdmlsls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
  53290.   int64_t __ret; \
  53291.   int64_t __s0 = __p0; \
  53292.   int32_t __s1 = __p1; \
  53293.   int32x4_t __s2 = __p2; \
  53294.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  53295.   __ret = (int64_t) __builtin_neon_vqdmlsls_laneq_s32(__s0, __s1, __rev2, __p3); \
  53296.   __ret; \
  53297. })
  53298. #endif
  53299.  
  53300. #ifdef __LITTLE_ENDIAN__
  53301. #define vqdmlslh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
  53302.   int32_t __ret; \
  53303.   int32_t __s0 = __p0; \
  53304.   int16_t __s1 = __p1; \
  53305.   int16x8_t __s2 = __p2; \
  53306.   __ret = (int32_t) __builtin_neon_vqdmlslh_laneq_s16(__s0, __s1, __s2, __p3); \
  53307.   __ret; \
  53308. })
  53309. #else
  53310. #define vqdmlslh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
  53311.   int32_t __ret; \
  53312.   int32_t __s0 = __p0; \
  53313.   int16_t __s1 = __p1; \
  53314.   int16x8_t __s2 = __p2; \
  53315.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
  53316.   __ret = (int32_t) __builtin_neon_vqdmlslh_laneq_s16(__s0, __s1, __rev2, __p3); \
  53317.   __ret; \
  53318. })
  53319. #endif
  53320.  
  53321. #ifdef __LITTLE_ENDIAN__
  53322. #define vqdmlsl_laneq_s32(__p0_627, __p1_627, __p2_627, __p3_627) __extension__ ({ \
  53323.   int64x2_t __ret_627; \
  53324.   int64x2_t __s0_627 = __p0_627; \
  53325.   int32x2_t __s1_627 = __p1_627; \
  53326.   int32x4_t __s2_627 = __p2_627; \
  53327.   __ret_627 = vqdmlsl_s32(__s0_627, __s1_627, splat_laneq_s32(__s2_627, __p3_627)); \
  53328.   __ret_627; \
  53329. })
  53330. #else
  53331. #define vqdmlsl_laneq_s32(__p0_628, __p1_628, __p2_628, __p3_628) __extension__ ({ \
  53332.   int64x2_t __ret_628; \
  53333.   int64x2_t __s0_628 = __p0_628; \
  53334.   int32x2_t __s1_628 = __p1_628; \
  53335.   int32x4_t __s2_628 = __p2_628; \
  53336.   int64x2_t __rev0_628;  __rev0_628 = __builtin_shufflevector(__s0_628, __s0_628, 1, 0); \
  53337.   int32x2_t __rev1_628;  __rev1_628 = __builtin_shufflevector(__s1_628, __s1_628, 1, 0); \
  53338.   int32x4_t __rev2_628;  __rev2_628 = __builtin_shufflevector(__s2_628, __s2_628, 3, 2, 1, 0); \
  53339.   __ret_628 = __noswap_vqdmlsl_s32(__rev0_628, __rev1_628, __noswap_splat_laneq_s32(__rev2_628, __p3_628)); \
  53340.   __ret_628 = __builtin_shufflevector(__ret_628, __ret_628, 1, 0); \
  53341.   __ret_628; \
  53342. })
  53343. #endif
  53344.  
  53345. #ifdef __LITTLE_ENDIAN__
  53346. #define vqdmlsl_laneq_s16(__p0_629, __p1_629, __p2_629, __p3_629) __extension__ ({ \
  53347.   int32x4_t __ret_629; \
  53348.   int32x4_t __s0_629 = __p0_629; \
  53349.   int16x4_t __s1_629 = __p1_629; \
  53350.   int16x8_t __s2_629 = __p2_629; \
  53351.   __ret_629 = vqdmlsl_s16(__s0_629, __s1_629, splat_laneq_s16(__s2_629, __p3_629)); \
  53352.   __ret_629; \
  53353. })
  53354. #else
  53355. #define vqdmlsl_laneq_s16(__p0_630, __p1_630, __p2_630, __p3_630) __extension__ ({ \
  53356.   int32x4_t __ret_630; \
  53357.   int32x4_t __s0_630 = __p0_630; \
  53358.   int16x4_t __s1_630 = __p1_630; \
  53359.   int16x8_t __s2_630 = __p2_630; \
  53360.   int32x4_t __rev0_630;  __rev0_630 = __builtin_shufflevector(__s0_630, __s0_630, 3, 2, 1, 0); \
  53361.   int16x4_t __rev1_630;  __rev1_630 = __builtin_shufflevector(__s1_630, __s1_630, 3, 2, 1, 0); \
  53362.   int16x8_t __rev2_630;  __rev2_630 = __builtin_shufflevector(__s2_630, __s2_630, 7, 6, 5, 4, 3, 2, 1, 0); \
  53363.   __ret_630 = __noswap_vqdmlsl_s16(__rev0_630, __rev1_630, __noswap_splat_laneq_s16(__rev2_630, __p3_630)); \
  53364.   __ret_630 = __builtin_shufflevector(__ret_630, __ret_630, 3, 2, 1, 0); \
  53365.   __ret_630; \
  53366. })
  53367. #endif
  53368.  
  53369. __ai int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1) {
  53370.   int32_t __ret;
  53371.   __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
  53372.   return __ret;
  53373. }
  53374. __ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) {
  53375.   int16_t __ret;
  53376.   __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
  53377.   return __ret;
  53378. }
  53379. #ifdef __LITTLE_ENDIAN__
  53380. #define vqdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  53381.   int32x4_t __ret; \
  53382.   int32x4_t __s0 = __p0; \
  53383.   int32x2_t __s1 = __p1; \
  53384.   __ret = (int32x4_t) __builtin_neon_vqdmulhq_lane_v((int8x16_t)__s0, (int8x8_t)__s1, __p2, 2); \
  53385.   __ret; \
  53386. })
  53387. #else
  53388. #define vqdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  53389.   int32x4_t __ret; \
  53390.   int32x4_t __s0 = __p0; \
  53391.   int32x2_t __s1 = __p1; \
  53392.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  53393.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  53394.   __ret = (int32x4_t) __builtin_neon_vqdmulhq_lane_v((int8x16_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  53395.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  53396.   __ret; \
  53397. })
  53398. #endif
  53399.  
  53400. #ifdef __LITTLE_ENDIAN__
  53401. #define vqdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  53402.   int16x8_t __ret; \
  53403.   int16x8_t __s0 = __p0; \
  53404.   int16x4_t __s1 = __p1; \
  53405.   __ret = (int16x8_t) __builtin_neon_vqdmulhq_lane_v((int8x16_t)__s0, (int8x8_t)__s1, __p2, 1); \
  53406.   __ret; \
  53407. })
  53408. #else
  53409. #define vqdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  53410.   int16x8_t __ret; \
  53411.   int16x8_t __s0 = __p0; \
  53412.   int16x4_t __s1 = __p1; \
  53413.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  53414.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  53415.   __ret = (int16x8_t) __builtin_neon_vqdmulhq_lane_v((int8x16_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  53416.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  53417.   __ret; \
  53418. })
  53419. #endif
  53420.  
  53421. #ifdef __LITTLE_ENDIAN__
  53422. #define vqdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  53423.   int32x2_t __ret; \
  53424.   int32x2_t __s0 = __p0; \
  53425.   int32x2_t __s1 = __p1; \
  53426.   __ret = (int32x2_t) __builtin_neon_vqdmulh_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
  53427.   __ret; \
  53428. })
  53429. #else
  53430. #define vqdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  53431.   int32x2_t __ret; \
  53432.   int32x2_t __s0 = __p0; \
  53433.   int32x2_t __s1 = __p1; \
  53434.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  53435.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  53436.   __ret = (int32x2_t) __builtin_neon_vqdmulh_lane_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  53437.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  53438.   __ret; \
  53439. })
  53440. #endif
  53441.  
  53442. #ifdef __LITTLE_ENDIAN__
  53443. #define vqdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  53444.   int16x4_t __ret; \
  53445.   int16x4_t __s0 = __p0; \
  53446.   int16x4_t __s1 = __p1; \
  53447.   __ret = (int16x4_t) __builtin_neon_vqdmulh_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
  53448.   __ret; \
  53449. })
  53450. #else
  53451. #define vqdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  53452.   int16x4_t __ret; \
  53453.   int16x4_t __s0 = __p0; \
  53454.   int16x4_t __s1 = __p1; \
  53455.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  53456.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  53457.   __ret = (int16x4_t) __builtin_neon_vqdmulh_lane_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  53458.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  53459.   __ret; \
  53460. })
  53461. #endif
  53462.  
  53463. #ifdef __LITTLE_ENDIAN__
  53464. #define vqdmulhs_lane_s32(__p0_631, __p1_631, __p2_631) __extension__ ({ \
  53465.   int32_t __ret_631; \
  53466.   int32_t __s0_631 = __p0_631; \
  53467.   int32x2_t __s1_631 = __p1_631; \
  53468.   __ret_631 = vqdmulhs_s32(__s0_631, vget_lane_s32(__s1_631, __p2_631)); \
  53469.   __ret_631; \
  53470. })
  53471. #else
  53472. #define vqdmulhs_lane_s32(__p0_632, __p1_632, __p2_632) __extension__ ({ \
  53473.   int32_t __ret_632; \
  53474.   int32_t __s0_632 = __p0_632; \
  53475.   int32x2_t __s1_632 = __p1_632; \
  53476.   int32x2_t __rev1_632;  __rev1_632 = __builtin_shufflevector(__s1_632, __s1_632, 1, 0); \
  53477.   __ret_632 = vqdmulhs_s32(__s0_632, __noswap_vget_lane_s32(__rev1_632, __p2_632)); \
  53478.   __ret_632; \
  53479. })
  53480. #endif
  53481.  
  53482. #ifdef __LITTLE_ENDIAN__
  53483. #define vqdmulhh_lane_s16(__p0_633, __p1_633, __p2_633) __extension__ ({ \
  53484.   int16_t __ret_633; \
  53485.   int16_t __s0_633 = __p0_633; \
  53486.   int16x4_t __s1_633 = __p1_633; \
  53487.   __ret_633 = vqdmulhh_s16(__s0_633, vget_lane_s16(__s1_633, __p2_633)); \
  53488.   __ret_633; \
  53489. })
  53490. #else
  53491. #define vqdmulhh_lane_s16(__p0_634, __p1_634, __p2_634) __extension__ ({ \
  53492.   int16_t __ret_634; \
  53493.   int16_t __s0_634 = __p0_634; \
  53494.   int16x4_t __s1_634 = __p1_634; \
  53495.   int16x4_t __rev1_634;  __rev1_634 = __builtin_shufflevector(__s1_634, __s1_634, 3, 2, 1, 0); \
  53496.   __ret_634 = vqdmulhh_s16(__s0_634, __noswap_vget_lane_s16(__rev1_634, __p2_634)); \
  53497.   __ret_634; \
  53498. })
  53499. #endif
  53500.  
  53501. #ifdef __LITTLE_ENDIAN__
  53502. #define vqdmulhs_laneq_s32(__p0_635, __p1_635, __p2_635) __extension__ ({ \
  53503.   int32_t __ret_635; \
  53504.   int32_t __s0_635 = __p0_635; \
  53505.   int32x4_t __s1_635 = __p1_635; \
  53506.   __ret_635 = vqdmulhs_s32(__s0_635, vgetq_lane_s32(__s1_635, __p2_635)); \
  53507.   __ret_635; \
  53508. })
  53509. #else
  53510. #define vqdmulhs_laneq_s32(__p0_636, __p1_636, __p2_636) __extension__ ({ \
  53511.   int32_t __ret_636; \
  53512.   int32_t __s0_636 = __p0_636; \
  53513.   int32x4_t __s1_636 = __p1_636; \
  53514.   int32x4_t __rev1_636;  __rev1_636 = __builtin_shufflevector(__s1_636, __s1_636, 3, 2, 1, 0); \
  53515.   __ret_636 = vqdmulhs_s32(__s0_636, __noswap_vgetq_lane_s32(__rev1_636, __p2_636)); \
  53516.   __ret_636; \
  53517. })
  53518. #endif
  53519.  
  53520. #ifdef __LITTLE_ENDIAN__
  53521. #define vqdmulhh_laneq_s16(__p0_637, __p1_637, __p2_637) __extension__ ({ \
  53522.   int16_t __ret_637; \
  53523.   int16_t __s0_637 = __p0_637; \
  53524.   int16x8_t __s1_637 = __p1_637; \
  53525.   __ret_637 = vqdmulhh_s16(__s0_637, vgetq_lane_s16(__s1_637, __p2_637)); \
  53526.   __ret_637; \
  53527. })
  53528. #else
  53529. #define vqdmulhh_laneq_s16(__p0_638, __p1_638, __p2_638) __extension__ ({ \
  53530.   int16_t __ret_638; \
  53531.   int16_t __s0_638 = __p0_638; \
  53532.   int16x8_t __s1_638 = __p1_638; \
  53533.   int16x8_t __rev1_638;  __rev1_638 = __builtin_shufflevector(__s1_638, __s1_638, 7, 6, 5, 4, 3, 2, 1, 0); \
  53534.   __ret_638 = vqdmulhh_s16(__s0_638, __noswap_vgetq_lane_s16(__rev1_638, __p2_638)); \
  53535.   __ret_638; \
  53536. })
  53537. #endif
  53538.  
  53539. #ifdef __LITTLE_ENDIAN__
  53540. #define vqdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
  53541.   int32x4_t __ret; \
  53542.   int32x4_t __s0 = __p0; \
  53543.   int32x4_t __s1 = __p1; \
  53544.   __ret = (int32x4_t) __builtin_neon_vqdmulhq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
  53545.   __ret; \
  53546. })
  53547. #else
  53548. #define vqdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
  53549.   int32x4_t __ret; \
  53550.   int32x4_t __s0 = __p0; \
  53551.   int32x4_t __s1 = __p1; \
  53552.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  53553.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  53554.   __ret = (int32x4_t) __builtin_neon_vqdmulhq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
  53555.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  53556.   __ret; \
  53557. })
  53558. #endif
  53559.  
  53560. #ifdef __LITTLE_ENDIAN__
  53561. #define vqdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
  53562.   int16x8_t __ret; \
  53563.   int16x8_t __s0 = __p0; \
  53564.   int16x8_t __s1 = __p1; \
  53565.   __ret = (int16x8_t) __builtin_neon_vqdmulhq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
  53566.   __ret; \
  53567. })
  53568. #else
  53569. #define vqdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
  53570.   int16x8_t __ret; \
  53571.   int16x8_t __s0 = __p0; \
  53572.   int16x8_t __s1 = __p1; \
  53573.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  53574.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  53575.   __ret = (int16x8_t) __builtin_neon_vqdmulhq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
  53576.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  53577.   __ret; \
  53578. })
  53579. #endif
  53580.  
  53581. #ifdef __LITTLE_ENDIAN__
  53582. #define vqdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
  53583.   int32x2_t __ret; \
  53584.   int32x2_t __s0 = __p0; \
  53585.   int32x4_t __s1 = __p1; \
  53586.   __ret = (int32x2_t) __builtin_neon_vqdmulh_laneq_v((int8x8_t)__s0, (int8x16_t)__s1, __p2, 2); \
  53587.   __ret; \
  53588. })
  53589. #else
  53590. #define vqdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
  53591.   int32x2_t __ret; \
  53592.   int32x2_t __s0 = __p0; \
  53593.   int32x4_t __s1 = __p1; \
  53594.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  53595.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  53596.   __ret = (int32x2_t) __builtin_neon_vqdmulh_laneq_v((int8x8_t)__rev0, (int8x16_t)__rev1, __p2, 2); \
  53597.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  53598.   __ret; \
  53599. })
  53600. #endif
  53601.  
  53602. #ifdef __LITTLE_ENDIAN__
  53603. #define vqdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
  53604.   int16x4_t __ret; \
  53605.   int16x4_t __s0 = __p0; \
  53606.   int16x8_t __s1 = __p1; \
  53607.   __ret = (int16x4_t) __builtin_neon_vqdmulh_laneq_v((int8x8_t)__s0, (int8x16_t)__s1, __p2, 1); \
  53608.   __ret; \
  53609. })
  53610. #else
  53611. #define vqdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
  53612.   int16x4_t __ret; \
  53613.   int16x4_t __s0 = __p0; \
  53614.   int16x8_t __s1 = __p1; \
  53615.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  53616.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  53617.   __ret = (int16x4_t) __builtin_neon_vqdmulh_laneq_v((int8x8_t)__rev0, (int8x16_t)__rev1, __p2, 1); \
  53618.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  53619.   __ret; \
  53620. })
  53621. #endif
  53622.  
  53623. __ai int64_t vqdmulls_s32(int32_t __p0, int32_t __p1) {
  53624.   int64_t __ret;
  53625.   __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
  53626.   return __ret;
  53627. }
  53628. __ai int32_t vqdmullh_s16(int16_t __p0, int16_t __p1) {
  53629.   int32_t __ret;
  53630.   __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
  53631.   return __ret;
  53632. }
  53633. #ifdef __LITTLE_ENDIAN__
  53634. __ai int64x2_t vqdmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
  53635.   int64x2_t __ret;
  53636.   __ret = vqdmull_s32(vget_high_s32(__p0), vget_high_s32(__p1));
  53637.   return __ret;
  53638. }
  53639. #else
  53640. __ai int64x2_t vqdmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
  53641.   int64x2_t __ret;
  53642.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  53643.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  53644.   __ret = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0), __noswap_vget_high_s32(__rev1));
  53645.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  53646.   return __ret;
  53647. }
  53648. #endif
  53649.  
  53650. #ifdef __LITTLE_ENDIAN__
  53651. __ai int32x4_t vqdmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
  53652.   int32x4_t __ret;
  53653.   __ret = vqdmull_s16(vget_high_s16(__p0), vget_high_s16(__p1));
  53654.   return __ret;
  53655. }
  53656. #else
  53657. __ai int32x4_t vqdmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
  53658.   int32x4_t __ret;
  53659.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  53660.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  53661.   __ret = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0), __noswap_vget_high_s16(__rev1));
  53662.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  53663.   return __ret;
  53664. }
  53665. #endif
  53666.  
  53667. #ifdef __LITTLE_ENDIAN__
  53668. #define vqdmull_high_lane_s32(__p0_639, __p1_639, __p2_639) __extension__ ({ \
  53669.   int64x2_t __ret_639; \
  53670.   int32x4_t __s0_639 = __p0_639; \
  53671.   int32x2_t __s1_639 = __p1_639; \
  53672.   __ret_639 = vqdmull_s32(vget_high_s32(__s0_639), splat_lane_s32(__s1_639, __p2_639)); \
  53673.   __ret_639; \
  53674. })
  53675. #else
  53676. #define vqdmull_high_lane_s32(__p0_640, __p1_640, __p2_640) __extension__ ({ \
  53677.   int64x2_t __ret_640; \
  53678.   int32x4_t __s0_640 = __p0_640; \
  53679.   int32x2_t __s1_640 = __p1_640; \
  53680.   int32x4_t __rev0_640;  __rev0_640 = __builtin_shufflevector(__s0_640, __s0_640, 3, 2, 1, 0); \
  53681.   int32x2_t __rev1_640;  __rev1_640 = __builtin_shufflevector(__s1_640, __s1_640, 1, 0); \
  53682.   __ret_640 = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0_640), __noswap_splat_lane_s32(__rev1_640, __p2_640)); \
  53683.   __ret_640 = __builtin_shufflevector(__ret_640, __ret_640, 1, 0); \
  53684.   __ret_640; \
  53685. })
  53686. #endif
  53687.  
  53688. #ifdef __LITTLE_ENDIAN__
  53689. #define vqdmull_high_lane_s16(__p0_641, __p1_641, __p2_641) __extension__ ({ \
  53690.   int32x4_t __ret_641; \
  53691.   int16x8_t __s0_641 = __p0_641; \
  53692.   int16x4_t __s1_641 = __p1_641; \
  53693.   __ret_641 = vqdmull_s16(vget_high_s16(__s0_641), splat_lane_s16(__s1_641, __p2_641)); \
  53694.   __ret_641; \
  53695. })
  53696. #else
  53697. #define vqdmull_high_lane_s16(__p0_642, __p1_642, __p2_642) __extension__ ({ \
  53698.   int32x4_t __ret_642; \
  53699.   int16x8_t __s0_642 = __p0_642; \
  53700.   int16x4_t __s1_642 = __p1_642; \
  53701.   int16x8_t __rev0_642;  __rev0_642 = __builtin_shufflevector(__s0_642, __s0_642, 7, 6, 5, 4, 3, 2, 1, 0); \
  53702.   int16x4_t __rev1_642;  __rev1_642 = __builtin_shufflevector(__s1_642, __s1_642, 3, 2, 1, 0); \
  53703.   __ret_642 = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0_642), __noswap_splat_lane_s16(__rev1_642, __p2_642)); \
  53704.   __ret_642 = __builtin_shufflevector(__ret_642, __ret_642, 3, 2, 1, 0); \
  53705.   __ret_642; \
  53706. })
  53707. #endif
  53708.  
  53709. #ifdef __LITTLE_ENDIAN__
  53710. #define vqdmull_high_laneq_s32(__p0_643, __p1_643, __p2_643) __extension__ ({ \
  53711.   int64x2_t __ret_643; \
  53712.   int32x4_t __s0_643 = __p0_643; \
  53713.   int32x4_t __s1_643 = __p1_643; \
  53714.   __ret_643 = vqdmull_s32(vget_high_s32(__s0_643), splat_laneq_s32(__s1_643, __p2_643)); \
  53715.   __ret_643; \
  53716. })
  53717. #else
  53718. #define vqdmull_high_laneq_s32(__p0_644, __p1_644, __p2_644) __extension__ ({ \
  53719.   int64x2_t __ret_644; \
  53720.   int32x4_t __s0_644 = __p0_644; \
  53721.   int32x4_t __s1_644 = __p1_644; \
  53722.   int32x4_t __rev0_644;  __rev0_644 = __builtin_shufflevector(__s0_644, __s0_644, 3, 2, 1, 0); \
  53723.   int32x4_t __rev1_644;  __rev1_644 = __builtin_shufflevector(__s1_644, __s1_644, 3, 2, 1, 0); \
  53724.   __ret_644 = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0_644), __noswap_splat_laneq_s32(__rev1_644, __p2_644)); \
  53725.   __ret_644 = __builtin_shufflevector(__ret_644, __ret_644, 1, 0); \
  53726.   __ret_644; \
  53727. })
  53728. #endif
  53729.  
  53730. #ifdef __LITTLE_ENDIAN__
  53731. #define vqdmull_high_laneq_s16(__p0_645, __p1_645, __p2_645) __extension__ ({ \
  53732.   int32x4_t __ret_645; \
  53733.   int16x8_t __s0_645 = __p0_645; \
  53734.   int16x8_t __s1_645 = __p1_645; \
  53735.   __ret_645 = vqdmull_s16(vget_high_s16(__s0_645), splat_laneq_s16(__s1_645, __p2_645)); \
  53736.   __ret_645; \
  53737. })
  53738. #else
  53739. #define vqdmull_high_laneq_s16(__p0_646, __p1_646, __p2_646) __extension__ ({ \
  53740.   int32x4_t __ret_646; \
  53741.   int16x8_t __s0_646 = __p0_646; \
  53742.   int16x8_t __s1_646 = __p1_646; \
  53743.   int16x8_t __rev0_646;  __rev0_646 = __builtin_shufflevector(__s0_646, __s0_646, 7, 6, 5, 4, 3, 2, 1, 0); \
  53744.   int16x8_t __rev1_646;  __rev1_646 = __builtin_shufflevector(__s1_646, __s1_646, 7, 6, 5, 4, 3, 2, 1, 0); \
  53745.   __ret_646 = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0_646), __noswap_splat_laneq_s16(__rev1_646, __p2_646)); \
  53746.   __ret_646 = __builtin_shufflevector(__ret_646, __ret_646, 3, 2, 1, 0); \
  53747.   __ret_646; \
  53748. })
  53749. #endif
  53750.  
  53751. #ifdef __LITTLE_ENDIAN__
  53752. __ai int64x2_t vqdmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
  53753.   int64x2_t __ret;
  53754.   __ret = vqdmull_n_s32(vget_high_s32(__p0), __p1);
  53755.   return __ret;
  53756. }
  53757. #else
  53758. __ai int64x2_t vqdmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
  53759.   int64x2_t __ret;
  53760.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  53761.   __ret = __noswap_vqdmull_n_s32(__noswap_vget_high_s32(__rev0), __p1);
  53762.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  53763.   return __ret;
  53764. }
  53765. #endif
  53766.  
  53767. #ifdef __LITTLE_ENDIAN__
  53768. __ai int32x4_t vqdmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
  53769.   int32x4_t __ret;
  53770.   __ret = vqdmull_n_s16(vget_high_s16(__p0), __p1);
  53771.   return __ret;
  53772. }
  53773. #else
  53774. __ai int32x4_t vqdmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
  53775.   int32x4_t __ret;
  53776.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  53777.   __ret = __noswap_vqdmull_n_s16(__noswap_vget_high_s16(__rev0), __p1);
  53778.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  53779.   return __ret;
  53780. }
  53781. #endif
  53782.  
  53783. #ifdef __LITTLE_ENDIAN__
  53784. #define vqdmulls_lane_s32(__p0_647, __p1_647, __p2_647) __extension__ ({ \
  53785.   int64_t __ret_647; \
  53786.   int32_t __s0_647 = __p0_647; \
  53787.   int32x2_t __s1_647 = __p1_647; \
  53788.   __ret_647 = vqdmulls_s32(__s0_647, vget_lane_s32(__s1_647, __p2_647)); \
  53789.   __ret_647; \
  53790. })
  53791. #else
  53792. #define vqdmulls_lane_s32(__p0_648, __p1_648, __p2_648) __extension__ ({ \
  53793.   int64_t __ret_648; \
  53794.   int32_t __s0_648 = __p0_648; \
  53795.   int32x2_t __s1_648 = __p1_648; \
  53796.   int32x2_t __rev1_648;  __rev1_648 = __builtin_shufflevector(__s1_648, __s1_648, 1, 0); \
  53797.   __ret_648 = vqdmulls_s32(__s0_648, __noswap_vget_lane_s32(__rev1_648, __p2_648)); \
  53798.   __ret_648; \
  53799. })
  53800. #endif
  53801.  
  53802. #ifdef __LITTLE_ENDIAN__
  53803. #define vqdmullh_lane_s16(__p0_649, __p1_649, __p2_649) __extension__ ({ \
  53804.   int32_t __ret_649; \
  53805.   int16_t __s0_649 = __p0_649; \
  53806.   int16x4_t __s1_649 = __p1_649; \
  53807.   __ret_649 = vqdmullh_s16(__s0_649, vget_lane_s16(__s1_649, __p2_649)); \
  53808.   __ret_649; \
  53809. })
  53810. #else
  53811. #define vqdmullh_lane_s16(__p0_650, __p1_650, __p2_650) __extension__ ({ \
  53812.   int32_t __ret_650; \
  53813.   int16_t __s0_650 = __p0_650; \
  53814.   int16x4_t __s1_650 = __p1_650; \
  53815.   int16x4_t __rev1_650;  __rev1_650 = __builtin_shufflevector(__s1_650, __s1_650, 3, 2, 1, 0); \
  53816.   __ret_650 = vqdmullh_s16(__s0_650, __noswap_vget_lane_s16(__rev1_650, __p2_650)); \
  53817.   __ret_650; \
  53818. })
  53819. #endif
  53820.  
  53821. #ifdef __LITTLE_ENDIAN__
  53822. #define vqdmulls_laneq_s32(__p0_651, __p1_651, __p2_651) __extension__ ({ \
  53823.   int64_t __ret_651; \
  53824.   int32_t __s0_651 = __p0_651; \
  53825.   int32x4_t __s1_651 = __p1_651; \
  53826.   __ret_651 = vqdmulls_s32(__s0_651, vgetq_lane_s32(__s1_651, __p2_651)); \
  53827.   __ret_651; \
  53828. })
  53829. #else
  53830. #define vqdmulls_laneq_s32(__p0_652, __p1_652, __p2_652) __extension__ ({ \
  53831.   int64_t __ret_652; \
  53832.   int32_t __s0_652 = __p0_652; \
  53833.   int32x4_t __s1_652 = __p1_652; \
  53834.   int32x4_t __rev1_652;  __rev1_652 = __builtin_shufflevector(__s1_652, __s1_652, 3, 2, 1, 0); \
  53835.   __ret_652 = vqdmulls_s32(__s0_652, __noswap_vgetq_lane_s32(__rev1_652, __p2_652)); \
  53836.   __ret_652; \
  53837. })
  53838. #endif
  53839.  
  53840. #ifdef __LITTLE_ENDIAN__
  53841. #define vqdmullh_laneq_s16(__p0_653, __p1_653, __p2_653) __extension__ ({ \
  53842.   int32_t __ret_653; \
  53843.   int16_t __s0_653 = __p0_653; \
  53844.   int16x8_t __s1_653 = __p1_653; \
  53845.   __ret_653 = vqdmullh_s16(__s0_653, vgetq_lane_s16(__s1_653, __p2_653)); \
  53846.   __ret_653; \
  53847. })
  53848. #else
  53849. #define vqdmullh_laneq_s16(__p0_654, __p1_654, __p2_654) __extension__ ({ \
  53850.   int32_t __ret_654; \
  53851.   int16_t __s0_654 = __p0_654; \
  53852.   int16x8_t __s1_654 = __p1_654; \
  53853.   int16x8_t __rev1_654;  __rev1_654 = __builtin_shufflevector(__s1_654, __s1_654, 7, 6, 5, 4, 3, 2, 1, 0); \
  53854.   __ret_654 = vqdmullh_s16(__s0_654, __noswap_vgetq_lane_s16(__rev1_654, __p2_654)); \
  53855.   __ret_654; \
  53856. })
  53857. #endif
  53858.  
  53859. #ifdef __LITTLE_ENDIAN__
  53860. #define vqdmull_laneq_s32(__p0_655, __p1_655, __p2_655) __extension__ ({ \
  53861.   int64x2_t __ret_655; \
  53862.   int32x2_t __s0_655 = __p0_655; \
  53863.   int32x4_t __s1_655 = __p1_655; \
  53864.   __ret_655 = vqdmull_s32(__s0_655, splat_laneq_s32(__s1_655, __p2_655)); \
  53865.   __ret_655; \
  53866. })
  53867. #else
  53868. #define vqdmull_laneq_s32(__p0_656, __p1_656, __p2_656) __extension__ ({ \
  53869.   int64x2_t __ret_656; \
  53870.   int32x2_t __s0_656 = __p0_656; \
  53871.   int32x4_t __s1_656 = __p1_656; \
  53872.   int32x2_t __rev0_656;  __rev0_656 = __builtin_shufflevector(__s0_656, __s0_656, 1, 0); \
  53873.   int32x4_t __rev1_656;  __rev1_656 = __builtin_shufflevector(__s1_656, __s1_656, 3, 2, 1, 0); \
  53874.   __ret_656 = __noswap_vqdmull_s32(__rev0_656, __noswap_splat_laneq_s32(__rev1_656, __p2_656)); \
  53875.   __ret_656 = __builtin_shufflevector(__ret_656, __ret_656, 1, 0); \
  53876.   __ret_656; \
  53877. })
  53878. #endif
  53879.  
  53880. #ifdef __LITTLE_ENDIAN__
  53881. #define vqdmull_laneq_s16(__p0_657, __p1_657, __p2_657) __extension__ ({ \
  53882.   int32x4_t __ret_657; \
  53883.   int16x4_t __s0_657 = __p0_657; \
  53884.   int16x8_t __s1_657 = __p1_657; \
  53885.   __ret_657 = vqdmull_s16(__s0_657, splat_laneq_s16(__s1_657, __p2_657)); \
  53886.   __ret_657; \
  53887. })
  53888. #else
  53889. #define vqdmull_laneq_s16(__p0_658, __p1_658, __p2_658) __extension__ ({ \
  53890.   int32x4_t __ret_658; \
  53891.   int16x4_t __s0_658 = __p0_658; \
  53892.   int16x8_t __s1_658 = __p1_658; \
  53893.   int16x4_t __rev0_658;  __rev0_658 = __builtin_shufflevector(__s0_658, __s0_658, 3, 2, 1, 0); \
  53894.   int16x8_t __rev1_658;  __rev1_658 = __builtin_shufflevector(__s1_658, __s1_658, 7, 6, 5, 4, 3, 2, 1, 0); \
  53895.   __ret_658 = __noswap_vqdmull_s16(__rev0_658, __noswap_splat_laneq_s16(__rev1_658, __p2_658)); \
  53896.   __ret_658 = __builtin_shufflevector(__ret_658, __ret_658, 3, 2, 1, 0); \
  53897.   __ret_658; \
  53898. })
  53899. #endif
  53900.  
  53901. __ai int16_t vqmovns_s32(int32_t __p0) {
  53902.   int16_t __ret;
  53903.   __ret = (int16_t) __builtin_neon_vqmovns_s32(__p0);
  53904.   return __ret;
  53905. }
  53906. __ai int32_t vqmovnd_s64(int64_t __p0) {
  53907.   int32_t __ret;
  53908.   __ret = (int32_t) __builtin_neon_vqmovnd_s64(__p0);
  53909.   return __ret;
  53910. }
  53911. __ai int8_t vqmovnh_s16(int16_t __p0) {
  53912.   int8_t __ret;
  53913.   __ret = (int8_t) __builtin_neon_vqmovnh_s16(__p0);
  53914.   return __ret;
  53915. }
  53916. __ai uint16_t vqmovns_u32(uint32_t __p0) {
  53917.   uint16_t __ret;
  53918.   __ret = (uint16_t) __builtin_neon_vqmovns_u32(__p0);
  53919.   return __ret;
  53920. }
  53921. __ai uint32_t vqmovnd_u64(uint64_t __p0) {
  53922.   uint32_t __ret;
  53923.   __ret = (uint32_t) __builtin_neon_vqmovnd_u64(__p0);
  53924.   return __ret;
  53925. }
  53926. __ai uint8_t vqmovnh_u16(uint16_t __p0) {
  53927.   uint8_t __ret;
  53928.   __ret = (uint8_t) __builtin_neon_vqmovnh_u16(__p0);
  53929.   return __ret;
  53930. }
  53931. #ifdef __LITTLE_ENDIAN__
  53932. __ai uint16x8_t vqmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
  53933.   uint16x8_t __ret;
  53934.   __ret = vcombine_u16(__p0, vqmovn_u32(__p1));
  53935.   return __ret;
  53936. }
  53937. #else
  53938. __ai uint16x8_t vqmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
  53939.   uint16x8_t __ret;
  53940.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  53941.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  53942.   __ret = __noswap_vcombine_u16(__rev0, __noswap_vqmovn_u32(__rev1));
  53943.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  53944.   return __ret;
  53945. }
  53946. #endif
  53947.  
  53948. #ifdef __LITTLE_ENDIAN__
  53949. __ai uint32x4_t vqmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
  53950.   uint32x4_t __ret;
  53951.   __ret = vcombine_u32(__p0, vqmovn_u64(__p1));
  53952.   return __ret;
  53953. }
  53954. #else
  53955. __ai uint32x4_t vqmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
  53956.   uint32x4_t __ret;
  53957.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  53958.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  53959.   __ret = __noswap_vcombine_u32(__rev0, __noswap_vqmovn_u64(__rev1));
  53960.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  53961.   return __ret;
  53962. }
  53963. #endif
  53964.  
  53965. #ifdef __LITTLE_ENDIAN__
  53966. __ai uint8x16_t vqmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
  53967.   uint8x16_t __ret;
  53968.   __ret = vcombine_u8(__p0, vqmovn_u16(__p1));
  53969.   return __ret;
  53970. }
  53971. #else
  53972. __ai uint8x16_t vqmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
  53973.   uint8x16_t __ret;
  53974.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  53975.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  53976.   __ret = __noswap_vcombine_u8(__rev0, __noswap_vqmovn_u16(__rev1));
  53977.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  53978.   return __ret;
  53979. }
  53980. #endif
  53981.  
  53982. #ifdef __LITTLE_ENDIAN__
  53983. __ai int16x8_t vqmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
  53984.   int16x8_t __ret;
  53985.   __ret = vcombine_s16(__p0, vqmovn_s32(__p1));
  53986.   return __ret;
  53987. }
  53988. #else
  53989. __ai int16x8_t vqmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
  53990.   int16x8_t __ret;
  53991.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  53992.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  53993.   __ret = __noswap_vcombine_s16(__rev0, __noswap_vqmovn_s32(__rev1));
  53994.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  53995.   return __ret;
  53996. }
  53997. #endif
  53998.  
  53999. #ifdef __LITTLE_ENDIAN__
  54000. __ai int32x4_t vqmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
  54001.   int32x4_t __ret;
  54002.   __ret = vcombine_s32(__p0, vqmovn_s64(__p1));
  54003.   return __ret;
  54004. }
  54005. #else
  54006. __ai int32x4_t vqmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
  54007.   int32x4_t __ret;
  54008.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  54009.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  54010.   __ret = __noswap_vcombine_s32(__rev0, __noswap_vqmovn_s64(__rev1));
  54011.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  54012.   return __ret;
  54013. }
  54014. #endif
  54015.  
  54016. #ifdef __LITTLE_ENDIAN__
  54017. __ai int8x16_t vqmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
  54018.   int8x16_t __ret;
  54019.   __ret = vcombine_s8(__p0, vqmovn_s16(__p1));
  54020.   return __ret;
  54021. }
  54022. #else
  54023. __ai int8x16_t vqmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
  54024.   int8x16_t __ret;
  54025.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  54026.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  54027.   __ret = __noswap_vcombine_s8(__rev0, __noswap_vqmovn_s16(__rev1));
  54028.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  54029.   return __ret;
  54030. }
  54031. #endif
  54032.  
  54033. __ai uint16_t vqmovuns_s32(int32_t __p0) {
  54034.   uint16_t __ret;
  54035.   __ret = (uint16_t) __builtin_neon_vqmovuns_s32(__p0);
  54036.   return __ret;
  54037. }
  54038. __ai uint32_t vqmovund_s64(int64_t __p0) {
  54039.   uint32_t __ret;
  54040.   __ret = (uint32_t) __builtin_neon_vqmovund_s64(__p0);
  54041.   return __ret;
  54042. }
  54043. __ai uint8_t vqmovunh_s16(int16_t __p0) {
  54044.   uint8_t __ret;
  54045.   __ret = (uint8_t) __builtin_neon_vqmovunh_s16(__p0);
  54046.   return __ret;
  54047. }
  54048. #ifdef __LITTLE_ENDIAN__
  54049. __ai uint16x8_t vqmovun_high_s32(uint16x4_t __p0, int32x4_t __p1) {
  54050.   uint16x8_t __ret;
  54051.   __ret = vcombine_u16((uint16x4_t)(__p0), vqmovun_s32(__p1));
  54052.   return __ret;
  54053. }
  54054. #else
  54055. __ai uint16x8_t vqmovun_high_s32(uint16x4_t __p0, int32x4_t __p1) {
  54056.   uint16x8_t __ret;
  54057.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  54058.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  54059.   __ret = __noswap_vcombine_u16((uint16x4_t)(__rev0), __noswap_vqmovun_s32(__rev1));
  54060.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  54061.   return __ret;
  54062. }
  54063. #endif
  54064.  
  54065. #ifdef __LITTLE_ENDIAN__
  54066. __ai uint32x4_t vqmovun_high_s64(uint32x2_t __p0, int64x2_t __p1) {
  54067.   uint32x4_t __ret;
  54068.   __ret = vcombine_u32((uint32x2_t)(__p0), vqmovun_s64(__p1));
  54069.   return __ret;
  54070. }
  54071. #else
  54072. __ai uint32x4_t vqmovun_high_s64(uint32x2_t __p0, int64x2_t __p1) {
  54073.   uint32x4_t __ret;
  54074.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  54075.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  54076.   __ret = __noswap_vcombine_u32((uint32x2_t)(__rev0), __noswap_vqmovun_s64(__rev1));
  54077.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  54078.   return __ret;
  54079. }
  54080. #endif
  54081.  
  54082. #ifdef __LITTLE_ENDIAN__
  54083. __ai uint8x16_t vqmovun_high_s16(uint8x8_t __p0, int16x8_t __p1) {
  54084.   uint8x16_t __ret;
  54085.   __ret = vcombine_u8((uint8x8_t)(__p0), vqmovun_s16(__p1));
  54086.   return __ret;
  54087. }
  54088. #else
  54089. __ai uint8x16_t vqmovun_high_s16(uint8x8_t __p0, int16x8_t __p1) {
  54090.   uint8x16_t __ret;
  54091.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  54092.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  54093.   __ret = __noswap_vcombine_u8((uint8x8_t)(__rev0), __noswap_vqmovun_s16(__rev1));
  54094.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  54095.   return __ret;
  54096. }
  54097. #endif
  54098.  
  54099. #ifdef __LITTLE_ENDIAN__
  54100. __ai int64x2_t vqnegq_s64(int64x2_t __p0) {
  54101.   int64x2_t __ret;
  54102.   __ret = (int64x2_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 35);
  54103.   return __ret;
  54104. }
  54105. #else
  54106. __ai int64x2_t vqnegq_s64(int64x2_t __p0) {
  54107.   int64x2_t __ret;
  54108.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  54109.   __ret = (int64x2_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 35);
  54110.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  54111.   return __ret;
  54112. }
  54113. #endif
  54114.  
  54115. __ai int64x1_t vqneg_s64(int64x1_t __p0) {
  54116.   int64x1_t __ret;
  54117.   __ret = (int64x1_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 3);
  54118.   return __ret;
  54119. }
  54120. __ai int8_t vqnegb_s8(int8_t __p0) {
  54121.   int8_t __ret;
  54122.   __ret = (int8_t) __builtin_neon_vqnegb_s8(__p0);
  54123.   return __ret;
  54124. }
  54125. __ai int32_t vqnegs_s32(int32_t __p0) {
  54126.   int32_t __ret;
  54127.   __ret = (int32_t) __builtin_neon_vqnegs_s32(__p0);
  54128.   return __ret;
  54129. }
  54130. __ai int64_t vqnegd_s64(int64_t __p0) {
  54131.   int64_t __ret;
  54132.   __ret = (int64_t) __builtin_neon_vqnegd_s64(__p0);
  54133.   return __ret;
  54134. }
  54135. __ai int16_t vqnegh_s16(int16_t __p0) {
  54136.   int16_t __ret;
  54137.   __ret = (int16_t) __builtin_neon_vqnegh_s16(__p0);
  54138.   return __ret;
  54139. }
  54140. __ai int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
  54141.   int32_t __ret;
  54142.   __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
  54143.   return __ret;
  54144. }
  54145. __ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
  54146.   int16_t __ret;
  54147.   __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
  54148.   return __ret;
  54149. }
  54150. #ifdef __LITTLE_ENDIAN__
  54151. #define vqrdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  54152.   int32x4_t __ret; \
  54153.   int32x4_t __s0 = __p0; \
  54154.   int32x2_t __s1 = __p1; \
  54155.   __ret = (int32x4_t) __builtin_neon_vqrdmulhq_lane_v((int8x16_t)__s0, (int8x8_t)__s1, __p2, 2); \
  54156.   __ret; \
  54157. })
  54158. #else
  54159. #define vqrdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  54160.   int32x4_t __ret; \
  54161.   int32x4_t __s0 = __p0; \
  54162.   int32x2_t __s1 = __p1; \
  54163.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  54164.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  54165.   __ret = (int32x4_t) __builtin_neon_vqrdmulhq_lane_v((int8x16_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  54166.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  54167.   __ret; \
  54168. })
  54169. #endif
  54170.  
  54171. #ifdef __LITTLE_ENDIAN__
  54172. #define vqrdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  54173.   int16x8_t __ret; \
  54174.   int16x8_t __s0 = __p0; \
  54175.   int16x4_t __s1 = __p1; \
  54176.   __ret = (int16x8_t) __builtin_neon_vqrdmulhq_lane_v((int8x16_t)__s0, (int8x8_t)__s1, __p2, 1); \
  54177.   __ret; \
  54178. })
  54179. #else
  54180. #define vqrdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  54181.   int16x8_t __ret; \
  54182.   int16x8_t __s0 = __p0; \
  54183.   int16x4_t __s1 = __p1; \
  54184.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  54185.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  54186.   __ret = (int16x8_t) __builtin_neon_vqrdmulhq_lane_v((int8x16_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  54187.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  54188.   __ret; \
  54189. })
  54190. #endif
  54191.  
  54192. #ifdef __LITTLE_ENDIAN__
  54193. #define vqrdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  54194.   int32x2_t __ret; \
  54195.   int32x2_t __s0 = __p0; \
  54196.   int32x2_t __s1 = __p1; \
  54197.   __ret = (int32x2_t) __builtin_neon_vqrdmulh_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
  54198.   __ret; \
  54199. })
  54200. #else
  54201. #define vqrdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
  54202.   int32x2_t __ret; \
  54203.   int32x2_t __s0 = __p0; \
  54204.   int32x2_t __s1 = __p1; \
  54205.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  54206.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  54207.   __ret = (int32x2_t) __builtin_neon_vqrdmulh_lane_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
  54208.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  54209.   __ret; \
  54210. })
  54211. #endif
  54212.  
  54213. #ifdef __LITTLE_ENDIAN__
  54214. #define vqrdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  54215.   int16x4_t __ret; \
  54216.   int16x4_t __s0 = __p0; \
  54217.   int16x4_t __s1 = __p1; \
  54218.   __ret = (int16x4_t) __builtin_neon_vqrdmulh_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
  54219.   __ret; \
  54220. })
  54221. #else
  54222. #define vqrdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
  54223.   int16x4_t __ret; \
  54224.   int16x4_t __s0 = __p0; \
  54225.   int16x4_t __s1 = __p1; \
  54226.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  54227.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  54228.   __ret = (int16x4_t) __builtin_neon_vqrdmulh_lane_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
  54229.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  54230.   __ret; \
  54231. })
  54232. #endif
  54233.  
  54234. #ifdef __LITTLE_ENDIAN__
  54235. #define vqrdmulhs_lane_s32(__p0_659, __p1_659, __p2_659) __extension__ ({ \
  54236.   int32_t __ret_659; \
  54237.   int32_t __s0_659 = __p0_659; \
  54238.   int32x2_t __s1_659 = __p1_659; \
  54239.   __ret_659 = vqrdmulhs_s32(__s0_659, vget_lane_s32(__s1_659, __p2_659)); \
  54240.   __ret_659; \
  54241. })
  54242. #else
  54243. #define vqrdmulhs_lane_s32(__p0_660, __p1_660, __p2_660) __extension__ ({ \
  54244.   int32_t __ret_660; \
  54245.   int32_t __s0_660 = __p0_660; \
  54246.   int32x2_t __s1_660 = __p1_660; \
  54247.   int32x2_t __rev1_660;  __rev1_660 = __builtin_shufflevector(__s1_660, __s1_660, 1, 0); \
  54248.   __ret_660 = vqrdmulhs_s32(__s0_660, __noswap_vget_lane_s32(__rev1_660, __p2_660)); \
  54249.   __ret_660; \
  54250. })
  54251. #endif
  54252.  
  54253. #ifdef __LITTLE_ENDIAN__
  54254. #define vqrdmulhh_lane_s16(__p0_661, __p1_661, __p2_661) __extension__ ({ \
  54255.   int16_t __ret_661; \
  54256.   int16_t __s0_661 = __p0_661; \
  54257.   int16x4_t __s1_661 = __p1_661; \
  54258.   __ret_661 = vqrdmulhh_s16(__s0_661, vget_lane_s16(__s1_661, __p2_661)); \
  54259.   __ret_661; \
  54260. })
  54261. #else
  54262. #define vqrdmulhh_lane_s16(__p0_662, __p1_662, __p2_662) __extension__ ({ \
  54263.   int16_t __ret_662; \
  54264.   int16_t __s0_662 = __p0_662; \
  54265.   int16x4_t __s1_662 = __p1_662; \
  54266.   int16x4_t __rev1_662;  __rev1_662 = __builtin_shufflevector(__s1_662, __s1_662, 3, 2, 1, 0); \
  54267.   __ret_662 = vqrdmulhh_s16(__s0_662, __noswap_vget_lane_s16(__rev1_662, __p2_662)); \
  54268.   __ret_662; \
  54269. })
  54270. #endif
  54271.  
  54272. #ifdef __LITTLE_ENDIAN__
  54273. #define vqrdmulhs_laneq_s32(__p0_663, __p1_663, __p2_663) __extension__ ({ \
  54274.   int32_t __ret_663; \
  54275.   int32_t __s0_663 = __p0_663; \
  54276.   int32x4_t __s1_663 = __p1_663; \
  54277.   __ret_663 = vqrdmulhs_s32(__s0_663, vgetq_lane_s32(__s1_663, __p2_663)); \
  54278.   __ret_663; \
  54279. })
  54280. #else
  54281. #define vqrdmulhs_laneq_s32(__p0_664, __p1_664, __p2_664) __extension__ ({ \
  54282.   int32_t __ret_664; \
  54283.   int32_t __s0_664 = __p0_664; \
  54284.   int32x4_t __s1_664 = __p1_664; \
  54285.   int32x4_t __rev1_664;  __rev1_664 = __builtin_shufflevector(__s1_664, __s1_664, 3, 2, 1, 0); \
  54286.   __ret_664 = vqrdmulhs_s32(__s0_664, __noswap_vgetq_lane_s32(__rev1_664, __p2_664)); \
  54287.   __ret_664; \
  54288. })
  54289. #endif
  54290.  
  54291. #ifdef __LITTLE_ENDIAN__
  54292. #define vqrdmulhh_laneq_s16(__p0_665, __p1_665, __p2_665) __extension__ ({ \
  54293.   int16_t __ret_665; \
  54294.   int16_t __s0_665 = __p0_665; \
  54295.   int16x8_t __s1_665 = __p1_665; \
  54296.   __ret_665 = vqrdmulhh_s16(__s0_665, vgetq_lane_s16(__s1_665, __p2_665)); \
  54297.   __ret_665; \
  54298. })
  54299. #else
  54300. #define vqrdmulhh_laneq_s16(__p0_666, __p1_666, __p2_666) __extension__ ({ \
  54301.   int16_t __ret_666; \
  54302.   int16_t __s0_666 = __p0_666; \
  54303.   int16x8_t __s1_666 = __p1_666; \
  54304.   int16x8_t __rev1_666;  __rev1_666 = __builtin_shufflevector(__s1_666, __s1_666, 7, 6, 5, 4, 3, 2, 1, 0); \
  54305.   __ret_666 = vqrdmulhh_s16(__s0_666, __noswap_vgetq_lane_s16(__rev1_666, __p2_666)); \
  54306.   __ret_666; \
  54307. })
  54308. #endif
  54309.  
  54310. #ifdef __LITTLE_ENDIAN__
  54311. #define vqrdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
  54312.   int32x4_t __ret; \
  54313.   int32x4_t __s0 = __p0; \
  54314.   int32x4_t __s1 = __p1; \
  54315.   __ret = (int32x4_t) __builtin_neon_vqrdmulhq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
  54316.   __ret; \
  54317. })
  54318. #else
  54319. #define vqrdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
  54320.   int32x4_t __ret; \
  54321.   int32x4_t __s0 = __p0; \
  54322.   int32x4_t __s1 = __p1; \
  54323.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  54324.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  54325.   __ret = (int32x4_t) __builtin_neon_vqrdmulhq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
  54326.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  54327.   __ret; \
  54328. })
  54329. #endif
  54330.  
  54331. #ifdef __LITTLE_ENDIAN__
  54332. #define vqrdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
  54333.   int16x8_t __ret; \
  54334.   int16x8_t __s0 = __p0; \
  54335.   int16x8_t __s1 = __p1; \
  54336.   __ret = (int16x8_t) __builtin_neon_vqrdmulhq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
  54337.   __ret; \
  54338. })
  54339. #else
  54340. #define vqrdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
  54341.   int16x8_t __ret; \
  54342.   int16x8_t __s0 = __p0; \
  54343.   int16x8_t __s1 = __p1; \
  54344.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  54345.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  54346.   __ret = (int16x8_t) __builtin_neon_vqrdmulhq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
  54347.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  54348.   __ret; \
  54349. })
  54350. #endif
  54351.  
  54352. #ifdef __LITTLE_ENDIAN__
  54353. #define vqrdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
  54354.   int32x2_t __ret; \
  54355.   int32x2_t __s0 = __p0; \
  54356.   int32x4_t __s1 = __p1; \
  54357.   __ret = (int32x2_t) __builtin_neon_vqrdmulh_laneq_v((int8x8_t)__s0, (int8x16_t)__s1, __p2, 2); \
  54358.   __ret; \
  54359. })
  54360. #else
  54361. #define vqrdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
  54362.   int32x2_t __ret; \
  54363.   int32x2_t __s0 = __p0; \
  54364.   int32x4_t __s1 = __p1; \
  54365.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  54366.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  54367.   __ret = (int32x2_t) __builtin_neon_vqrdmulh_laneq_v((int8x8_t)__rev0, (int8x16_t)__rev1, __p2, 2); \
  54368.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  54369.   __ret; \
  54370. })
  54371. #endif
  54372.  
  54373. #ifdef __LITTLE_ENDIAN__
  54374. #define vqrdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
  54375.   int16x4_t __ret; \
  54376.   int16x4_t __s0 = __p0; \
  54377.   int16x8_t __s1 = __p1; \
  54378.   __ret = (int16x4_t) __builtin_neon_vqrdmulh_laneq_v((int8x8_t)__s0, (int8x16_t)__s1, __p2, 1); \
  54379.   __ret; \
  54380. })
  54381. #else
  54382. #define vqrdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
  54383.   int16x4_t __ret; \
  54384.   int16x4_t __s0 = __p0; \
  54385.   int16x8_t __s1 = __p1; \
  54386.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  54387.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  54388.   __ret = (int16x4_t) __builtin_neon_vqrdmulh_laneq_v((int8x8_t)__rev0, (int8x16_t)__rev1, __p2, 1); \
  54389.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  54390.   __ret; \
  54391. })
  54392. #endif
  54393.  
  54394. __ai uint8_t vqrshlb_u8(uint8_t __p0, int8_t __p1) {
  54395.   uint8_t __ret;
  54396.   __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1);
  54397.   return __ret;
  54398. }
  54399. __ai uint32_t vqrshls_u32(uint32_t __p0, int32_t __p1) {
  54400.   uint32_t __ret;
  54401.   __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1);
  54402.   return __ret;
  54403. }
  54404. __ai uint64_t vqrshld_u64(uint64_t __p0, int64_t __p1) {
  54405.   uint64_t __ret;
  54406.   __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1);
  54407.   return __ret;
  54408. }
  54409. __ai uint16_t vqrshlh_u16(uint16_t __p0, int16_t __p1) {
  54410.   uint16_t __ret;
  54411.   __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1);
  54412.   return __ret;
  54413. }
  54414. __ai int8_t vqrshlb_s8(int8_t __p0, int8_t __p1) {
  54415.   int8_t __ret;
  54416.   __ret = (int8_t) __builtin_neon_vqrshlb_s8(__p0, __p1);
  54417.   return __ret;
  54418. }
  54419. __ai int32_t vqrshls_s32(int32_t __p0, int32_t __p1) {
  54420.   int32_t __ret;
  54421.   __ret = (int32_t) __builtin_neon_vqrshls_s32(__p0, __p1);
  54422.   return __ret;
  54423. }
  54424. __ai int64_t vqrshld_s64(int64_t __p0, int64_t __p1) {
  54425.   int64_t __ret;
  54426.   __ret = (int64_t) __builtin_neon_vqrshld_s64(__p0, __p1);
  54427.   return __ret;
  54428. }
  54429. __ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) {
  54430.   int16_t __ret;
  54431.   __ret = (int16_t) __builtin_neon_vqrshlh_s16(__p0, __p1);
  54432.   return __ret;
  54433. }
  54434. #ifdef __LITTLE_ENDIAN__
  54435. #define vqrshrn_high_n_u32(__p0_667, __p1_667, __p2_667) __extension__ ({ \
  54436.   uint16x8_t __ret_667; \
  54437.   uint16x4_t __s0_667 = __p0_667; \
  54438.   uint32x4_t __s1_667 = __p1_667; \
  54439.   __ret_667 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_667), (uint16x4_t)(vqrshrn_n_u32(__s1_667, __p2_667)))); \
  54440.   __ret_667; \
  54441. })
  54442. #else
  54443. #define vqrshrn_high_n_u32(__p0_668, __p1_668, __p2_668) __extension__ ({ \
  54444.   uint16x8_t __ret_668; \
  54445.   uint16x4_t __s0_668 = __p0_668; \
  54446.   uint32x4_t __s1_668 = __p1_668; \
  54447.   uint16x4_t __rev0_668;  __rev0_668 = __builtin_shufflevector(__s0_668, __s0_668, 3, 2, 1, 0); \
  54448.   uint32x4_t __rev1_668;  __rev1_668 = __builtin_shufflevector(__s1_668, __s1_668, 3, 2, 1, 0); \
  54449.   __ret_668 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_668), (uint16x4_t)(__noswap_vqrshrn_n_u32(__rev1_668, __p2_668)))); \
  54450.   __ret_668 = __builtin_shufflevector(__ret_668, __ret_668, 7, 6, 5, 4, 3, 2, 1, 0); \
  54451.   __ret_668; \
  54452. })
  54453. #endif
  54454.  
  54455. #ifdef __LITTLE_ENDIAN__
  54456. #define vqrshrn_high_n_u64(__p0_669, __p1_669, __p2_669) __extension__ ({ \
  54457.   uint32x4_t __ret_669; \
  54458.   uint32x2_t __s0_669 = __p0_669; \
  54459.   uint64x2_t __s1_669 = __p1_669; \
  54460.   __ret_669 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_669), (uint32x2_t)(vqrshrn_n_u64(__s1_669, __p2_669)))); \
  54461.   __ret_669; \
  54462. })
  54463. #else
  54464. #define vqrshrn_high_n_u64(__p0_670, __p1_670, __p2_670) __extension__ ({ \
  54465.   uint32x4_t __ret_670; \
  54466.   uint32x2_t __s0_670 = __p0_670; \
  54467.   uint64x2_t __s1_670 = __p1_670; \
  54468.   uint32x2_t __rev0_670;  __rev0_670 = __builtin_shufflevector(__s0_670, __s0_670, 1, 0); \
  54469.   uint64x2_t __rev1_670;  __rev1_670 = __builtin_shufflevector(__s1_670, __s1_670, 1, 0); \
  54470.   __ret_670 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_670), (uint32x2_t)(__noswap_vqrshrn_n_u64(__rev1_670, __p2_670)))); \
  54471.   __ret_670 = __builtin_shufflevector(__ret_670, __ret_670, 3, 2, 1, 0); \
  54472.   __ret_670; \
  54473. })
  54474. #endif
  54475.  
  54476. #ifdef __LITTLE_ENDIAN__
  54477. #define vqrshrn_high_n_u16(__p0_671, __p1_671, __p2_671) __extension__ ({ \
  54478.   uint8x16_t __ret_671; \
  54479.   uint8x8_t __s0_671 = __p0_671; \
  54480.   uint16x8_t __s1_671 = __p1_671; \
  54481.   __ret_671 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_671), (uint8x8_t)(vqrshrn_n_u16(__s1_671, __p2_671)))); \
  54482.   __ret_671; \
  54483. })
  54484. #else
  54485. #define vqrshrn_high_n_u16(__p0_672, __p1_672, __p2_672) __extension__ ({ \
  54486.   uint8x16_t __ret_672; \
  54487.   uint8x8_t __s0_672 = __p0_672; \
  54488.   uint16x8_t __s1_672 = __p1_672; \
  54489.   uint8x8_t __rev0_672;  __rev0_672 = __builtin_shufflevector(__s0_672, __s0_672, 7, 6, 5, 4, 3, 2, 1, 0); \
  54490.   uint16x8_t __rev1_672;  __rev1_672 = __builtin_shufflevector(__s1_672, __s1_672, 7, 6, 5, 4, 3, 2, 1, 0); \
  54491.   __ret_672 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_672), (uint8x8_t)(__noswap_vqrshrn_n_u16(__rev1_672, __p2_672)))); \
  54492.   __ret_672 = __builtin_shufflevector(__ret_672, __ret_672, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  54493.   __ret_672; \
  54494. })
  54495. #endif
  54496.  
  54497. #ifdef __LITTLE_ENDIAN__
  54498. #define vqrshrn_high_n_s32(__p0_673, __p1_673, __p2_673) __extension__ ({ \
  54499.   int16x8_t __ret_673; \
  54500.   int16x4_t __s0_673 = __p0_673; \
  54501.   int32x4_t __s1_673 = __p1_673; \
  54502.   __ret_673 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_673), (int16x4_t)(vqrshrn_n_s32(__s1_673, __p2_673)))); \
  54503.   __ret_673; \
  54504. })
  54505. #else
  54506. #define vqrshrn_high_n_s32(__p0_674, __p1_674, __p2_674) __extension__ ({ \
  54507.   int16x8_t __ret_674; \
  54508.   int16x4_t __s0_674 = __p0_674; \
  54509.   int32x4_t __s1_674 = __p1_674; \
  54510.   int16x4_t __rev0_674;  __rev0_674 = __builtin_shufflevector(__s0_674, __s0_674, 3, 2, 1, 0); \
  54511.   int32x4_t __rev1_674;  __rev1_674 = __builtin_shufflevector(__s1_674, __s1_674, 3, 2, 1, 0); \
  54512.   __ret_674 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_674), (int16x4_t)(__noswap_vqrshrn_n_s32(__rev1_674, __p2_674)))); \
  54513.   __ret_674 = __builtin_shufflevector(__ret_674, __ret_674, 7, 6, 5, 4, 3, 2, 1, 0); \
  54514.   __ret_674; \
  54515. })
  54516. #endif
  54517.  
  54518. #ifdef __LITTLE_ENDIAN__
  54519. #define vqrshrn_high_n_s64(__p0_675, __p1_675, __p2_675) __extension__ ({ \
  54520.   int32x4_t __ret_675; \
  54521.   int32x2_t __s0_675 = __p0_675; \
  54522.   int64x2_t __s1_675 = __p1_675; \
  54523.   __ret_675 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_675), (int32x2_t)(vqrshrn_n_s64(__s1_675, __p2_675)))); \
  54524.   __ret_675; \
  54525. })
  54526. #else
  54527. #define vqrshrn_high_n_s64(__p0_676, __p1_676, __p2_676) __extension__ ({ \
  54528.   int32x4_t __ret_676; \
  54529.   int32x2_t __s0_676 = __p0_676; \
  54530.   int64x2_t __s1_676 = __p1_676; \
  54531.   int32x2_t __rev0_676;  __rev0_676 = __builtin_shufflevector(__s0_676, __s0_676, 1, 0); \
  54532.   int64x2_t __rev1_676;  __rev1_676 = __builtin_shufflevector(__s1_676, __s1_676, 1, 0); \
  54533.   __ret_676 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_676), (int32x2_t)(__noswap_vqrshrn_n_s64(__rev1_676, __p2_676)))); \
  54534.   __ret_676 = __builtin_shufflevector(__ret_676, __ret_676, 3, 2, 1, 0); \
  54535.   __ret_676; \
  54536. })
  54537. #endif
  54538.  
  54539. #ifdef __LITTLE_ENDIAN__
  54540. #define vqrshrn_high_n_s16(__p0_677, __p1_677, __p2_677) __extension__ ({ \
  54541.   int8x16_t __ret_677; \
  54542.   int8x8_t __s0_677 = __p0_677; \
  54543.   int16x8_t __s1_677 = __p1_677; \
  54544.   __ret_677 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_677), (int8x8_t)(vqrshrn_n_s16(__s1_677, __p2_677)))); \
  54545.   __ret_677; \
  54546. })
  54547. #else
  54548. #define vqrshrn_high_n_s16(__p0_678, __p1_678, __p2_678) __extension__ ({ \
  54549.   int8x16_t __ret_678; \
  54550.   int8x8_t __s0_678 = __p0_678; \
  54551.   int16x8_t __s1_678 = __p1_678; \
  54552.   int8x8_t __rev0_678;  __rev0_678 = __builtin_shufflevector(__s0_678, __s0_678, 7, 6, 5, 4, 3, 2, 1, 0); \
  54553.   int16x8_t __rev1_678;  __rev1_678 = __builtin_shufflevector(__s1_678, __s1_678, 7, 6, 5, 4, 3, 2, 1, 0); \
  54554.   __ret_678 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_678), (int8x8_t)(__noswap_vqrshrn_n_s16(__rev1_678, __p2_678)))); \
  54555.   __ret_678 = __builtin_shufflevector(__ret_678, __ret_678, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  54556.   __ret_678; \
  54557. })
  54558. #endif
  54559.  
  54560. #define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
  54561.   uint16_t __ret; \
  54562.   uint32_t __s0 = __p0; \
  54563.   __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
  54564.   __ret; \
  54565. })
  54566. #define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
  54567.   uint32_t __ret; \
  54568.   uint64_t __s0 = __p0; \
  54569.   __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
  54570.   __ret; \
  54571. })
  54572. #define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
  54573.   uint8_t __ret; \
  54574.   uint16_t __s0 = __p0; \
  54575.   __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
  54576.   __ret; \
  54577. })
  54578. #define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
  54579.   int16_t __ret; \
  54580.   int32_t __s0 = __p0; \
  54581.   __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
  54582.   __ret; \
  54583. })
  54584. #define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
  54585.   int32_t __ret; \
  54586.   int64_t __s0 = __p0; \
  54587.   __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
  54588.   __ret; \
  54589. })
  54590. #define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
  54591.   int8_t __ret; \
  54592.   int16_t __s0 = __p0; \
  54593.   __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
  54594.   __ret; \
  54595. })
  54596. #ifdef __LITTLE_ENDIAN__
  54597. #define vqrshrun_high_n_s32(__p0_679, __p1_679, __p2_679) __extension__ ({ \
  54598.   int16x8_t __ret_679; \
  54599.   int16x4_t __s0_679 = __p0_679; \
  54600.   int32x4_t __s1_679 = __p1_679; \
  54601.   __ret_679 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_679), (int16x4_t)(vqrshrun_n_s32(__s1_679, __p2_679)))); \
  54602.   __ret_679; \
  54603. })
  54604. #else
  54605. #define vqrshrun_high_n_s32(__p0_680, __p1_680, __p2_680) __extension__ ({ \
  54606.   int16x8_t __ret_680; \
  54607.   int16x4_t __s0_680 = __p0_680; \
  54608.   int32x4_t __s1_680 = __p1_680; \
  54609.   int16x4_t __rev0_680;  __rev0_680 = __builtin_shufflevector(__s0_680, __s0_680, 3, 2, 1, 0); \
  54610.   int32x4_t __rev1_680;  __rev1_680 = __builtin_shufflevector(__s1_680, __s1_680, 3, 2, 1, 0); \
  54611.   __ret_680 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_680), (int16x4_t)(__noswap_vqrshrun_n_s32(__rev1_680, __p2_680)))); \
  54612.   __ret_680 = __builtin_shufflevector(__ret_680, __ret_680, 7, 6, 5, 4, 3, 2, 1, 0); \
  54613.   __ret_680; \
  54614. })
  54615. #endif
  54616.  
  54617. #ifdef __LITTLE_ENDIAN__
  54618. #define vqrshrun_high_n_s64(__p0_681, __p1_681, __p2_681) __extension__ ({ \
  54619.   int32x4_t __ret_681; \
  54620.   int32x2_t __s0_681 = __p0_681; \
  54621.   int64x2_t __s1_681 = __p1_681; \
  54622.   __ret_681 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_681), (int32x2_t)(vqrshrun_n_s64(__s1_681, __p2_681)))); \
  54623.   __ret_681; \
  54624. })
  54625. #else
  54626. #define vqrshrun_high_n_s64(__p0_682, __p1_682, __p2_682) __extension__ ({ \
  54627.   int32x4_t __ret_682; \
  54628.   int32x2_t __s0_682 = __p0_682; \
  54629.   int64x2_t __s1_682 = __p1_682; \
  54630.   int32x2_t __rev0_682;  __rev0_682 = __builtin_shufflevector(__s0_682, __s0_682, 1, 0); \
  54631.   int64x2_t __rev1_682;  __rev1_682 = __builtin_shufflevector(__s1_682, __s1_682, 1, 0); \
  54632.   __ret_682 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_682), (int32x2_t)(__noswap_vqrshrun_n_s64(__rev1_682, __p2_682)))); \
  54633.   __ret_682 = __builtin_shufflevector(__ret_682, __ret_682, 3, 2, 1, 0); \
  54634.   __ret_682; \
  54635. })
  54636. #endif
  54637.  
  54638. #ifdef __LITTLE_ENDIAN__
  54639. #define vqrshrun_high_n_s16(__p0_683, __p1_683, __p2_683) __extension__ ({ \
  54640.   int8x16_t __ret_683; \
  54641.   int8x8_t __s0_683 = __p0_683; \
  54642.   int16x8_t __s1_683 = __p1_683; \
  54643.   __ret_683 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_683), (int8x8_t)(vqrshrun_n_s16(__s1_683, __p2_683)))); \
  54644.   __ret_683; \
  54645. })
  54646. #else
  54647. #define vqrshrun_high_n_s16(__p0_684, __p1_684, __p2_684) __extension__ ({ \
  54648.   int8x16_t __ret_684; \
  54649.   int8x8_t __s0_684 = __p0_684; \
  54650.   int16x8_t __s1_684 = __p1_684; \
  54651.   int8x8_t __rev0_684;  __rev0_684 = __builtin_shufflevector(__s0_684, __s0_684, 7, 6, 5, 4, 3, 2, 1, 0); \
  54652.   int16x8_t __rev1_684;  __rev1_684 = __builtin_shufflevector(__s1_684, __s1_684, 7, 6, 5, 4, 3, 2, 1, 0); \
  54653.   __ret_684 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_684), (int8x8_t)(__noswap_vqrshrun_n_s16(__rev1_684, __p2_684)))); \
  54654.   __ret_684 = __builtin_shufflevector(__ret_684, __ret_684, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  54655.   __ret_684; \
  54656. })
  54657. #endif
  54658.  
  54659. #define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
  54660.   int16_t __ret; \
  54661.   int32_t __s0 = __p0; \
  54662.   __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
  54663.   __ret; \
  54664. })
  54665. #define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
  54666.   int32_t __ret; \
  54667.   int64_t __s0 = __p0; \
  54668.   __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
  54669.   __ret; \
  54670. })
  54671. #define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
  54672.   int8_t __ret; \
  54673.   int16_t __s0 = __p0; \
  54674.   __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
  54675.   __ret; \
  54676. })
  54677. __ai uint8_t vqshlb_u8(uint8_t __p0, int8_t __p1) {
  54678.   uint8_t __ret;
  54679.   __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
  54680.   return __ret;
  54681. }
  54682. __ai uint32_t vqshls_u32(uint32_t __p0, int32_t __p1) {
  54683.   uint32_t __ret;
  54684.   __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
  54685.   return __ret;
  54686. }
  54687. __ai uint64_t vqshld_u64(uint64_t __p0, int64_t __p1) {
  54688.   uint64_t __ret;
  54689.   __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
  54690.   return __ret;
  54691. }
  54692. __ai uint16_t vqshlh_u16(uint16_t __p0, int16_t __p1) {
  54693.   uint16_t __ret;
  54694.   __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
  54695.   return __ret;
  54696. }
  54697. __ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
  54698.   int8_t __ret;
  54699.   __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
  54700.   return __ret;
  54701. }
  54702. __ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
  54703.   int32_t __ret;
  54704.   __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
  54705.   return __ret;
  54706. }
  54707. __ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
  54708.   int64_t __ret;
  54709.   __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
  54710.   return __ret;
  54711. }
  54712. __ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
  54713.   int16_t __ret;
  54714.   __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
  54715.   return __ret;
  54716. }
  54717. #define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
  54718.   uint8_t __ret; \
  54719.   uint8_t __s0 = __p0; \
  54720.   __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
  54721.   __ret; \
  54722. })
  54723. #define vqshls_n_u32(__p0, __p1) __extension__ ({ \
  54724.   uint32_t __ret; \
  54725.   uint32_t __s0 = __p0; \
  54726.   __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
  54727.   __ret; \
  54728. })
  54729. #define vqshld_n_u64(__p0, __p1) __extension__ ({ \
  54730.   uint64_t __ret; \
  54731.   uint64_t __s0 = __p0; \
  54732.   __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
  54733.   __ret; \
  54734. })
  54735. #define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
  54736.   uint16_t __ret; \
  54737.   uint16_t __s0 = __p0; \
  54738.   __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
  54739.   __ret; \
  54740. })
  54741. #define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
  54742.   int8_t __ret; \
  54743.   int8_t __s0 = __p0; \
  54744.   __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
  54745.   __ret; \
  54746. })
  54747. #define vqshls_n_s32(__p0, __p1) __extension__ ({ \
  54748.   int32_t __ret; \
  54749.   int32_t __s0 = __p0; \
  54750.   __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
  54751.   __ret; \
  54752. })
  54753. #define vqshld_n_s64(__p0, __p1) __extension__ ({ \
  54754.   int64_t __ret; \
  54755.   int64_t __s0 = __p0; \
  54756.   __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
  54757.   __ret; \
  54758. })
  54759. #define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
  54760.   int16_t __ret; \
  54761.   int16_t __s0 = __p0; \
  54762.   __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
  54763.   __ret; \
  54764. })
  54765. #define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
  54766.   int8_t __ret; \
  54767.   int8_t __s0 = __p0; \
  54768.   __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
  54769.   __ret; \
  54770. })
  54771. #define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
  54772.   int32_t __ret; \
  54773.   int32_t __s0 = __p0; \
  54774.   __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
  54775.   __ret; \
  54776. })
  54777. #define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
  54778.   int64_t __ret; \
  54779.   int64_t __s0 = __p0; \
  54780.   __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
  54781.   __ret; \
  54782. })
  54783. #define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
  54784.   int16_t __ret; \
  54785.   int16_t __s0 = __p0; \
  54786.   __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
  54787.   __ret; \
  54788. })
  54789. #ifdef __LITTLE_ENDIAN__
  54790. #define vqshrn_high_n_u32(__p0_685, __p1_685, __p2_685) __extension__ ({ \
  54791.   uint16x8_t __ret_685; \
  54792.   uint16x4_t __s0_685 = __p0_685; \
  54793.   uint32x4_t __s1_685 = __p1_685; \
  54794.   __ret_685 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_685), (uint16x4_t)(vqshrn_n_u32(__s1_685, __p2_685)))); \
  54795.   __ret_685; \
  54796. })
  54797. #else
  54798. #define vqshrn_high_n_u32(__p0_686, __p1_686, __p2_686) __extension__ ({ \
  54799.   uint16x8_t __ret_686; \
  54800.   uint16x4_t __s0_686 = __p0_686; \
  54801.   uint32x4_t __s1_686 = __p1_686; \
  54802.   uint16x4_t __rev0_686;  __rev0_686 = __builtin_shufflevector(__s0_686, __s0_686, 3, 2, 1, 0); \
  54803.   uint32x4_t __rev1_686;  __rev1_686 = __builtin_shufflevector(__s1_686, __s1_686, 3, 2, 1, 0); \
  54804.   __ret_686 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_686), (uint16x4_t)(__noswap_vqshrn_n_u32(__rev1_686, __p2_686)))); \
  54805.   __ret_686 = __builtin_shufflevector(__ret_686, __ret_686, 7, 6, 5, 4, 3, 2, 1, 0); \
  54806.   __ret_686; \
  54807. })
  54808. #endif
  54809.  
  54810. #ifdef __LITTLE_ENDIAN__
  54811. #define vqshrn_high_n_u64(__p0_687, __p1_687, __p2_687) __extension__ ({ \
  54812.   uint32x4_t __ret_687; \
  54813.   uint32x2_t __s0_687 = __p0_687; \
  54814.   uint64x2_t __s1_687 = __p1_687; \
  54815.   __ret_687 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_687), (uint32x2_t)(vqshrn_n_u64(__s1_687, __p2_687)))); \
  54816.   __ret_687; \
  54817. })
  54818. #else
  54819. #define vqshrn_high_n_u64(__p0_688, __p1_688, __p2_688) __extension__ ({ \
  54820.   uint32x4_t __ret_688; \
  54821.   uint32x2_t __s0_688 = __p0_688; \
  54822.   uint64x2_t __s1_688 = __p1_688; \
  54823.   uint32x2_t __rev0_688;  __rev0_688 = __builtin_shufflevector(__s0_688, __s0_688, 1, 0); \
  54824.   uint64x2_t __rev1_688;  __rev1_688 = __builtin_shufflevector(__s1_688, __s1_688, 1, 0); \
  54825.   __ret_688 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_688), (uint32x2_t)(__noswap_vqshrn_n_u64(__rev1_688, __p2_688)))); \
  54826.   __ret_688 = __builtin_shufflevector(__ret_688, __ret_688, 3, 2, 1, 0); \
  54827.   __ret_688; \
  54828. })
  54829. #endif
  54830.  
  54831. #ifdef __LITTLE_ENDIAN__
  54832. #define vqshrn_high_n_u16(__p0_689, __p1_689, __p2_689) __extension__ ({ \
  54833.   uint8x16_t __ret_689; \
  54834.   uint8x8_t __s0_689 = __p0_689; \
  54835.   uint16x8_t __s1_689 = __p1_689; \
  54836.   __ret_689 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_689), (uint8x8_t)(vqshrn_n_u16(__s1_689, __p2_689)))); \
  54837.   __ret_689; \
  54838. })
  54839. #else
  54840. #define vqshrn_high_n_u16(__p0_690, __p1_690, __p2_690) __extension__ ({ \
  54841.   uint8x16_t __ret_690; \
  54842.   uint8x8_t __s0_690 = __p0_690; \
  54843.   uint16x8_t __s1_690 = __p1_690; \
  54844.   uint8x8_t __rev0_690;  __rev0_690 = __builtin_shufflevector(__s0_690, __s0_690, 7, 6, 5, 4, 3, 2, 1, 0); \
  54845.   uint16x8_t __rev1_690;  __rev1_690 = __builtin_shufflevector(__s1_690, __s1_690, 7, 6, 5, 4, 3, 2, 1, 0); \
  54846.   __ret_690 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_690), (uint8x8_t)(__noswap_vqshrn_n_u16(__rev1_690, __p2_690)))); \
  54847.   __ret_690 = __builtin_shufflevector(__ret_690, __ret_690, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  54848.   __ret_690; \
  54849. })
  54850. #endif
  54851.  
  54852. #ifdef __LITTLE_ENDIAN__
  54853. #define vqshrn_high_n_s32(__p0_691, __p1_691, __p2_691) __extension__ ({ \
  54854.   int16x8_t __ret_691; \
  54855.   int16x4_t __s0_691 = __p0_691; \
  54856.   int32x4_t __s1_691 = __p1_691; \
  54857.   __ret_691 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_691), (int16x4_t)(vqshrn_n_s32(__s1_691, __p2_691)))); \
  54858.   __ret_691; \
  54859. })
  54860. #else
  54861. #define vqshrn_high_n_s32(__p0_692, __p1_692, __p2_692) __extension__ ({ \
  54862.   int16x8_t __ret_692; \
  54863.   int16x4_t __s0_692 = __p0_692; \
  54864.   int32x4_t __s1_692 = __p1_692; \
  54865.   int16x4_t __rev0_692;  __rev0_692 = __builtin_shufflevector(__s0_692, __s0_692, 3, 2, 1, 0); \
  54866.   int32x4_t __rev1_692;  __rev1_692 = __builtin_shufflevector(__s1_692, __s1_692, 3, 2, 1, 0); \
  54867.   __ret_692 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_692), (int16x4_t)(__noswap_vqshrn_n_s32(__rev1_692, __p2_692)))); \
  54868.   __ret_692 = __builtin_shufflevector(__ret_692, __ret_692, 7, 6, 5, 4, 3, 2, 1, 0); \
  54869.   __ret_692; \
  54870. })
  54871. #endif
  54872.  
  54873. #ifdef __LITTLE_ENDIAN__
  54874. #define vqshrn_high_n_s64(__p0_693, __p1_693, __p2_693) __extension__ ({ \
  54875.   int32x4_t __ret_693; \
  54876.   int32x2_t __s0_693 = __p0_693; \
  54877.   int64x2_t __s1_693 = __p1_693; \
  54878.   __ret_693 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_693), (int32x2_t)(vqshrn_n_s64(__s1_693, __p2_693)))); \
  54879.   __ret_693; \
  54880. })
  54881. #else
  54882. #define vqshrn_high_n_s64(__p0_694, __p1_694, __p2_694) __extension__ ({ \
  54883.   int32x4_t __ret_694; \
  54884.   int32x2_t __s0_694 = __p0_694; \
  54885.   int64x2_t __s1_694 = __p1_694; \
  54886.   int32x2_t __rev0_694;  __rev0_694 = __builtin_shufflevector(__s0_694, __s0_694, 1, 0); \
  54887.   int64x2_t __rev1_694;  __rev1_694 = __builtin_shufflevector(__s1_694, __s1_694, 1, 0); \
  54888.   __ret_694 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_694), (int32x2_t)(__noswap_vqshrn_n_s64(__rev1_694, __p2_694)))); \
  54889.   __ret_694 = __builtin_shufflevector(__ret_694, __ret_694, 3, 2, 1, 0); \
  54890.   __ret_694; \
  54891. })
  54892. #endif
  54893.  
  54894. #ifdef __LITTLE_ENDIAN__
  54895. #define vqshrn_high_n_s16(__p0_695, __p1_695, __p2_695) __extension__ ({ \
  54896.   int8x16_t __ret_695; \
  54897.   int8x8_t __s0_695 = __p0_695; \
  54898.   int16x8_t __s1_695 = __p1_695; \
  54899.   __ret_695 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_695), (int8x8_t)(vqshrn_n_s16(__s1_695, __p2_695)))); \
  54900.   __ret_695; \
  54901. })
  54902. #else
  54903. #define vqshrn_high_n_s16(__p0_696, __p1_696, __p2_696) __extension__ ({ \
  54904.   int8x16_t __ret_696; \
  54905.   int8x8_t __s0_696 = __p0_696; \
  54906.   int16x8_t __s1_696 = __p1_696; \
  54907.   int8x8_t __rev0_696;  __rev0_696 = __builtin_shufflevector(__s0_696, __s0_696, 7, 6, 5, 4, 3, 2, 1, 0); \
  54908.   int16x8_t __rev1_696;  __rev1_696 = __builtin_shufflevector(__s1_696, __s1_696, 7, 6, 5, 4, 3, 2, 1, 0); \
  54909.   __ret_696 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_696), (int8x8_t)(__noswap_vqshrn_n_s16(__rev1_696, __p2_696)))); \
  54910.   __ret_696 = __builtin_shufflevector(__ret_696, __ret_696, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  54911.   __ret_696; \
  54912. })
  54913. #endif
  54914.  
  54915. #define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
  54916.   uint16_t __ret; \
  54917.   uint32_t __s0 = __p0; \
  54918.   __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
  54919.   __ret; \
  54920. })
  54921. #define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
  54922.   uint32_t __ret; \
  54923.   uint64_t __s0 = __p0; \
  54924.   __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
  54925.   __ret; \
  54926. })
  54927. #define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
  54928.   uint8_t __ret; \
  54929.   uint16_t __s0 = __p0; \
  54930.   __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
  54931.   __ret; \
  54932. })
  54933. #define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
  54934.   int16_t __ret; \
  54935.   int32_t __s0 = __p0; \
  54936.   __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
  54937.   __ret; \
  54938. })
  54939. #define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
  54940.   int32_t __ret; \
  54941.   int64_t __s0 = __p0; \
  54942.   __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
  54943.   __ret; \
  54944. })
  54945. #define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
  54946.   int8_t __ret; \
  54947.   int16_t __s0 = __p0; \
  54948.   __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
  54949.   __ret; \
  54950. })
  54951. #ifdef __LITTLE_ENDIAN__
  54952. #define vqshrun_high_n_s32(__p0_697, __p1_697, __p2_697) __extension__ ({ \
  54953.   int16x8_t __ret_697; \
  54954.   int16x4_t __s0_697 = __p0_697; \
  54955.   int32x4_t __s1_697 = __p1_697; \
  54956.   __ret_697 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_697), (int16x4_t)(vqshrun_n_s32(__s1_697, __p2_697)))); \
  54957.   __ret_697; \
  54958. })
  54959. #else
  54960. #define vqshrun_high_n_s32(__p0_698, __p1_698, __p2_698) __extension__ ({ \
  54961.   int16x8_t __ret_698; \
  54962.   int16x4_t __s0_698 = __p0_698; \
  54963.   int32x4_t __s1_698 = __p1_698; \
  54964.   int16x4_t __rev0_698;  __rev0_698 = __builtin_shufflevector(__s0_698, __s0_698, 3, 2, 1, 0); \
  54965.   int32x4_t __rev1_698;  __rev1_698 = __builtin_shufflevector(__s1_698, __s1_698, 3, 2, 1, 0); \
  54966.   __ret_698 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_698), (int16x4_t)(__noswap_vqshrun_n_s32(__rev1_698, __p2_698)))); \
  54967.   __ret_698 = __builtin_shufflevector(__ret_698, __ret_698, 7, 6, 5, 4, 3, 2, 1, 0); \
  54968.   __ret_698; \
  54969. })
  54970. #endif
  54971.  
  54972. #ifdef __LITTLE_ENDIAN__
  54973. #define vqshrun_high_n_s64(__p0_699, __p1_699, __p2_699) __extension__ ({ \
  54974.   int32x4_t __ret_699; \
  54975.   int32x2_t __s0_699 = __p0_699; \
  54976.   int64x2_t __s1_699 = __p1_699; \
  54977.   __ret_699 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_699), (int32x2_t)(vqshrun_n_s64(__s1_699, __p2_699)))); \
  54978.   __ret_699; \
  54979. })
  54980. #else
  54981. #define vqshrun_high_n_s64(__p0_700, __p1_700, __p2_700) __extension__ ({ \
  54982.   int32x4_t __ret_700; \
  54983.   int32x2_t __s0_700 = __p0_700; \
  54984.   int64x2_t __s1_700 = __p1_700; \
  54985.   int32x2_t __rev0_700;  __rev0_700 = __builtin_shufflevector(__s0_700, __s0_700, 1, 0); \
  54986.   int64x2_t __rev1_700;  __rev1_700 = __builtin_shufflevector(__s1_700, __s1_700, 1, 0); \
  54987.   __ret_700 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_700), (int32x2_t)(__noswap_vqshrun_n_s64(__rev1_700, __p2_700)))); \
  54988.   __ret_700 = __builtin_shufflevector(__ret_700, __ret_700, 3, 2, 1, 0); \
  54989.   __ret_700; \
  54990. })
  54991. #endif
  54992.  
  54993. #ifdef __LITTLE_ENDIAN__
  54994. #define vqshrun_high_n_s16(__p0_701, __p1_701, __p2_701) __extension__ ({ \
  54995.   int8x16_t __ret_701; \
  54996.   int8x8_t __s0_701 = __p0_701; \
  54997.   int16x8_t __s1_701 = __p1_701; \
  54998.   __ret_701 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_701), (int8x8_t)(vqshrun_n_s16(__s1_701, __p2_701)))); \
  54999.   __ret_701; \
  55000. })
  55001. #else
  55002. #define vqshrun_high_n_s16(__p0_702, __p1_702, __p2_702) __extension__ ({ \
  55003.   int8x16_t __ret_702; \
  55004.   int8x8_t __s0_702 = __p0_702; \
  55005.   int16x8_t __s1_702 = __p1_702; \
  55006.   int8x8_t __rev0_702;  __rev0_702 = __builtin_shufflevector(__s0_702, __s0_702, 7, 6, 5, 4, 3, 2, 1, 0); \
  55007.   int16x8_t __rev1_702;  __rev1_702 = __builtin_shufflevector(__s1_702, __s1_702, 7, 6, 5, 4, 3, 2, 1, 0); \
  55008.   __ret_702 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_702), (int8x8_t)(__noswap_vqshrun_n_s16(__rev1_702, __p2_702)))); \
  55009.   __ret_702 = __builtin_shufflevector(__ret_702, __ret_702, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  55010.   __ret_702; \
  55011. })
  55012. #endif
  55013.  
  55014. #define vqshruns_n_s32(__p0, __p1) __extension__ ({ \
  55015.   int16_t __ret; \
  55016.   int32_t __s0 = __p0; \
  55017.   __ret = (int16_t) __builtin_neon_vqshruns_n_s32(__s0, __p1); \
  55018.   __ret; \
  55019. })
  55020. #define vqshrund_n_s64(__p0, __p1) __extension__ ({ \
  55021.   int32_t __ret; \
  55022.   int64_t __s0 = __p0; \
  55023.   __ret = (int32_t) __builtin_neon_vqshrund_n_s64(__s0, __p1); \
  55024.   __ret; \
  55025. })
  55026. #define vqshrunh_n_s16(__p0, __p1) __extension__ ({ \
  55027.   int8_t __ret; \
  55028.   int16_t __s0 = __p0; \
  55029.   __ret = (int8_t) __builtin_neon_vqshrunh_n_s16(__s0, __p1); \
  55030.   __ret; \
  55031. })
  55032. __ai uint8_t vqsubb_u8(uint8_t __p0, uint8_t __p1) {
  55033.   uint8_t __ret;
  55034.   __ret = (uint8_t) __builtin_neon_vqsubb_u8(__p0, __p1);
  55035.   return __ret;
  55036. }
  55037. __ai uint32_t vqsubs_u32(uint32_t __p0, uint32_t __p1) {
  55038.   uint32_t __ret;
  55039.   __ret = (uint32_t) __builtin_neon_vqsubs_u32(__p0, __p1);
  55040.   return __ret;
  55041. }
  55042. __ai uint64_t vqsubd_u64(uint64_t __p0, uint64_t __p1) {
  55043.   uint64_t __ret;
  55044.   __ret = (uint64_t) __builtin_neon_vqsubd_u64(__p0, __p1);
  55045.   return __ret;
  55046. }
  55047. __ai uint16_t vqsubh_u16(uint16_t __p0, uint16_t __p1) {
  55048.   uint16_t __ret;
  55049.   __ret = (uint16_t) __builtin_neon_vqsubh_u16(__p0, __p1);
  55050.   return __ret;
  55051. }
  55052. __ai int8_t vqsubb_s8(int8_t __p0, int8_t __p1) {
  55053.   int8_t __ret;
  55054.   __ret = (int8_t) __builtin_neon_vqsubb_s8(__p0, __p1);
  55055.   return __ret;
  55056. }
  55057. __ai int32_t vqsubs_s32(int32_t __p0, int32_t __p1) {
  55058.   int32_t __ret;
  55059.   __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
  55060.   return __ret;
  55061. }
  55062. __ai int64_t vqsubd_s64(int64_t __p0, int64_t __p1) {
  55063.   int64_t __ret;
  55064.   __ret = (int64_t) __builtin_neon_vqsubd_s64(__p0, __p1);
  55065.   return __ret;
  55066. }
  55067. __ai int16_t vqsubh_s16(int16_t __p0, int16_t __p1) {
  55068.   int16_t __ret;
  55069.   __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
  55070.   return __ret;
  55071. }
  55072. #ifdef __LITTLE_ENDIAN__
  55073. __ai poly8x8_t vqtbl1_p8(poly8x16_t __p0, uint8x8_t __p1) {
  55074.   poly8x8_t __ret;
  55075.   __ret = (poly8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__p0, (int8x8_t)__p1, 4);
  55076.   return __ret;
  55077. }
  55078. #else
  55079. __ai poly8x8_t vqtbl1_p8(poly8x16_t __p0, uint8x8_t __p1) {
  55080.   poly8x8_t __ret;
  55081.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55082.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55083.   __ret = (poly8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__rev0, (int8x8_t)__rev1, 4);
  55084.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55085.   return __ret;
  55086. }
  55087. #endif
  55088.  
  55089. #ifdef __LITTLE_ENDIAN__
  55090. __ai poly8x16_t vqtbl1q_p8(poly8x16_t __p0, uint8x16_t __p1) {
  55091.   poly8x16_t __ret;
  55092.   __ret = (poly8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__p0, (int8x16_t)__p1, 36);
  55093.   return __ret;
  55094. }
  55095. #else
  55096. __ai poly8x16_t vqtbl1q_p8(poly8x16_t __p0, uint8x16_t __p1) {
  55097.   poly8x16_t __ret;
  55098.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55099.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55100.   __ret = (poly8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 36);
  55101.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55102.   return __ret;
  55103. }
  55104. #endif
  55105.  
  55106. #ifdef __LITTLE_ENDIAN__
  55107. __ai uint8x16_t vqtbl1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  55108.   uint8x16_t __ret;
  55109.   __ret = (uint8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  55110.   return __ret;
  55111. }
  55112. #else
  55113. __ai uint8x16_t vqtbl1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  55114.   uint8x16_t __ret;
  55115.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55116.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55117.   __ret = (uint8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  55118.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55119.   return __ret;
  55120. }
  55121. #endif
  55122.  
  55123. #ifdef __LITTLE_ENDIAN__
  55124. __ai int8x16_t vqtbl1q_s8(int8x16_t __p0, uint8x16_t __p1) {
  55125.   int8x16_t __ret;
  55126.   __ret = (int8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  55127.   return __ret;
  55128. }
  55129. #else
  55130. __ai int8x16_t vqtbl1q_s8(int8x16_t __p0, uint8x16_t __p1) {
  55131.   int8x16_t __ret;
  55132.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55133.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55134.   __ret = (int8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  55135.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55136.   return __ret;
  55137. }
  55138. #endif
  55139.  
  55140. #ifdef __LITTLE_ENDIAN__
  55141. __ai uint8x8_t vqtbl1_u8(uint8x16_t __p0, uint8x8_t __p1) {
  55142.   uint8x8_t __ret;
  55143.   __ret = (uint8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__p0, (int8x8_t)__p1, 16);
  55144.   return __ret;
  55145. }
  55146. #else
  55147. __ai uint8x8_t vqtbl1_u8(uint8x16_t __p0, uint8x8_t __p1) {
  55148.   uint8x8_t __ret;
  55149.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55150.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55151.   __ret = (uint8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__rev0, (int8x8_t)__rev1, 16);
  55152.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55153.   return __ret;
  55154. }
  55155. #endif
  55156.  
  55157. #ifdef __LITTLE_ENDIAN__
  55158. __ai int8x8_t vqtbl1_s8(int8x16_t __p0, uint8x8_t __p1) {
  55159.   int8x8_t __ret;
  55160.   __ret = (int8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__p0, (int8x8_t)__p1, 0);
  55161.   return __ret;
  55162. }
  55163. #else
  55164. __ai int8x8_t vqtbl1_s8(int8x16_t __p0, uint8x8_t __p1) {
  55165.   int8x8_t __ret;
  55166.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55167.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55168.   __ret = (int8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__rev0, (int8x8_t)__rev1, 0);
  55169.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55170.   return __ret;
  55171. }
  55172. #endif
  55173.  
  55174. #ifdef __LITTLE_ENDIAN__
  55175. __ai poly8x8_t vqtbl2_p8(poly8x16x2_t __p0, uint8x8_t __p1) {
  55176.   poly8x8_t __ret;
  55177.   __ret = (poly8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x8_t)__p1, 4);
  55178.   return __ret;
  55179. }
  55180. #else
  55181. __ai poly8x8_t vqtbl2_p8(poly8x16x2_t __p0, uint8x8_t __p1) {
  55182.   poly8x8_t __ret;
  55183.   poly8x16x2_t __rev0;
  55184.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55185.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55186.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55187.   __ret = (poly8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x8_t)__rev1, 4);
  55188.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55189.   return __ret;
  55190. }
  55191. #endif
  55192.  
  55193. #ifdef __LITTLE_ENDIAN__
  55194. __ai poly8x16_t vqtbl2q_p8(poly8x16x2_t __p0, uint8x16_t __p1) {
  55195.   poly8x16_t __ret;
  55196.   __ret = (poly8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p1, 36);
  55197.   return __ret;
  55198. }
  55199. #else
  55200. __ai poly8x16_t vqtbl2q_p8(poly8x16x2_t __p0, uint8x16_t __p1) {
  55201.   poly8x16_t __ret;
  55202.   poly8x16x2_t __rev0;
  55203.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55204.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55205.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55206.   __ret = (poly8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev1, 36);
  55207.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55208.   return __ret;
  55209. }
  55210. #endif
  55211.  
  55212. #ifdef __LITTLE_ENDIAN__
  55213. __ai uint8x16_t vqtbl2q_u8(uint8x16x2_t __p0, uint8x16_t __p1) {
  55214.   uint8x16_t __ret;
  55215.   __ret = (uint8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p1, 48);
  55216.   return __ret;
  55217. }
  55218. #else
  55219. __ai uint8x16_t vqtbl2q_u8(uint8x16x2_t __p0, uint8x16_t __p1) {
  55220.   uint8x16_t __ret;
  55221.   uint8x16x2_t __rev0;
  55222.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55223.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55224.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55225.   __ret = (uint8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev1, 48);
  55226.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55227.   return __ret;
  55228. }
  55229. #endif
  55230.  
  55231. #ifdef __LITTLE_ENDIAN__
  55232. __ai int8x16_t vqtbl2q_s8(int8x16x2_t __p0, uint8x16_t __p1) {
  55233.   int8x16_t __ret;
  55234.   __ret = (int8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p1, 32);
  55235.   return __ret;
  55236. }
  55237. #else
  55238. __ai int8x16_t vqtbl2q_s8(int8x16x2_t __p0, uint8x16_t __p1) {
  55239.   int8x16_t __ret;
  55240.   int8x16x2_t __rev0;
  55241.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55242.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55243.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55244.   __ret = (int8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev1, 32);
  55245.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55246.   return __ret;
  55247. }
  55248. #endif
  55249.  
  55250. #ifdef __LITTLE_ENDIAN__
  55251. __ai uint8x8_t vqtbl2_u8(uint8x16x2_t __p0, uint8x8_t __p1) {
  55252.   uint8x8_t __ret;
  55253.   __ret = (uint8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x8_t)__p1, 16);
  55254.   return __ret;
  55255. }
  55256. #else
  55257. __ai uint8x8_t vqtbl2_u8(uint8x16x2_t __p0, uint8x8_t __p1) {
  55258.   uint8x8_t __ret;
  55259.   uint8x16x2_t __rev0;
  55260.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55261.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55262.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55263.   __ret = (uint8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x8_t)__rev1, 16);
  55264.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55265.   return __ret;
  55266. }
  55267. #endif
  55268.  
  55269. #ifdef __LITTLE_ENDIAN__
  55270. __ai int8x8_t vqtbl2_s8(int8x16x2_t __p0, uint8x8_t __p1) {
  55271.   int8x8_t __ret;
  55272.   __ret = (int8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x8_t)__p1, 0);
  55273.   return __ret;
  55274. }
  55275. #else
  55276. __ai int8x8_t vqtbl2_s8(int8x16x2_t __p0, uint8x8_t __p1) {
  55277.   int8x8_t __ret;
  55278.   int8x16x2_t __rev0;
  55279.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55280.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55281.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55282.   __ret = (int8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x8_t)__rev1, 0);
  55283.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55284.   return __ret;
  55285. }
  55286. #endif
  55287.  
  55288. #ifdef __LITTLE_ENDIAN__
  55289. __ai poly8x8_t vqtbl3_p8(poly8x16x3_t __p0, uint8x8_t __p1) {
  55290.   poly8x8_t __ret;
  55291.   __ret = (poly8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x8_t)__p1, 4);
  55292.   return __ret;
  55293. }
  55294. #else
  55295. __ai poly8x8_t vqtbl3_p8(poly8x16x3_t __p0, uint8x8_t __p1) {
  55296.   poly8x8_t __ret;
  55297.   poly8x16x3_t __rev0;
  55298.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55299.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55300.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55301.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55302.   __ret = (poly8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x8_t)__rev1, 4);
  55303.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55304.   return __ret;
  55305. }
  55306. #endif
  55307.  
  55308. #ifdef __LITTLE_ENDIAN__
  55309. __ai poly8x16_t vqtbl3q_p8(poly8x16x3_t __p0, uint8x16_t __p1) {
  55310.   poly8x16_t __ret;
  55311.   __ret = (poly8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p1, 36);
  55312.   return __ret;
  55313. }
  55314. #else
  55315. __ai poly8x16_t vqtbl3q_p8(poly8x16x3_t __p0, uint8x16_t __p1) {
  55316.   poly8x16_t __ret;
  55317.   poly8x16x3_t __rev0;
  55318.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55319.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55320.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55321.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55322.   __ret = (poly8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev1, 36);
  55323.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55324.   return __ret;
  55325. }
  55326. #endif
  55327.  
  55328. #ifdef __LITTLE_ENDIAN__
  55329. __ai uint8x16_t vqtbl3q_u8(uint8x16x3_t __p0, uint8x16_t __p1) {
  55330.   uint8x16_t __ret;
  55331.   __ret = (uint8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p1, 48);
  55332.   return __ret;
  55333. }
  55334. #else
  55335. __ai uint8x16_t vqtbl3q_u8(uint8x16x3_t __p0, uint8x16_t __p1) {
  55336.   uint8x16_t __ret;
  55337.   uint8x16x3_t __rev0;
  55338.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55339.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55340.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55341.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55342.   __ret = (uint8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev1, 48);
  55343.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55344.   return __ret;
  55345. }
  55346. #endif
  55347.  
  55348. #ifdef __LITTLE_ENDIAN__
  55349. __ai int8x16_t vqtbl3q_s8(int8x16x3_t __p0, uint8x16_t __p1) {
  55350.   int8x16_t __ret;
  55351.   __ret = (int8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p1, 32);
  55352.   return __ret;
  55353. }
  55354. #else
  55355. __ai int8x16_t vqtbl3q_s8(int8x16x3_t __p0, uint8x16_t __p1) {
  55356.   int8x16_t __ret;
  55357.   int8x16x3_t __rev0;
  55358.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55359.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55360.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55361.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55362.   __ret = (int8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev1, 32);
  55363.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55364.   return __ret;
  55365. }
  55366. #endif
  55367.  
  55368. #ifdef __LITTLE_ENDIAN__
  55369. __ai uint8x8_t vqtbl3_u8(uint8x16x3_t __p0, uint8x8_t __p1) {
  55370.   uint8x8_t __ret;
  55371.   __ret = (uint8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x8_t)__p1, 16);
  55372.   return __ret;
  55373. }
  55374. #else
  55375. __ai uint8x8_t vqtbl3_u8(uint8x16x3_t __p0, uint8x8_t __p1) {
  55376.   uint8x8_t __ret;
  55377.   uint8x16x3_t __rev0;
  55378.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55379.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55380.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55381.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55382.   __ret = (uint8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x8_t)__rev1, 16);
  55383.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55384.   return __ret;
  55385. }
  55386. #endif
  55387.  
  55388. #ifdef __LITTLE_ENDIAN__
  55389. __ai int8x8_t vqtbl3_s8(int8x16x3_t __p0, uint8x8_t __p1) {
  55390.   int8x8_t __ret;
  55391.   __ret = (int8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x8_t)__p1, 0);
  55392.   return __ret;
  55393. }
  55394. #else
  55395. __ai int8x8_t vqtbl3_s8(int8x16x3_t __p0, uint8x8_t __p1) {
  55396.   int8x8_t __ret;
  55397.   int8x16x3_t __rev0;
  55398.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55399.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55400.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55401.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55402.   __ret = (int8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x8_t)__rev1, 0);
  55403.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55404.   return __ret;
  55405. }
  55406. #endif
  55407.  
  55408. #ifdef __LITTLE_ENDIAN__
  55409. __ai poly8x8_t vqtbl4_p8(poly8x16x4_t __p0, uint8x8_t __p1) {
  55410.   poly8x8_t __ret;
  55411.   __ret = (poly8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x8_t)__p1, 4);
  55412.   return __ret;
  55413. }
  55414. #else
  55415. __ai poly8x8_t vqtbl4_p8(poly8x16x4_t __p0, uint8x8_t __p1) {
  55416.   poly8x8_t __ret;
  55417.   poly8x16x4_t __rev0;
  55418.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55419.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55420.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55421.   __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55422.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55423.   __ret = (poly8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x8_t)__rev1, 4);
  55424.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55425.   return __ret;
  55426. }
  55427. #endif
  55428.  
  55429. #ifdef __LITTLE_ENDIAN__
  55430. __ai poly8x16_t vqtbl4q_p8(poly8x16x4_t __p0, uint8x16_t __p1) {
  55431.   poly8x16_t __ret;
  55432.   __ret = (poly8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x16_t)__p1, 36);
  55433.   return __ret;
  55434. }
  55435. #else
  55436. __ai poly8x16_t vqtbl4q_p8(poly8x16x4_t __p0, uint8x16_t __p1) {
  55437.   poly8x16_t __ret;
  55438.   poly8x16x4_t __rev0;
  55439.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55440.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55441.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55442.   __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55443.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55444.   __ret = (poly8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x16_t)__rev1, 36);
  55445.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55446.   return __ret;
  55447. }
  55448. #endif
  55449.  
  55450. #ifdef __LITTLE_ENDIAN__
  55451. __ai uint8x16_t vqtbl4q_u8(uint8x16x4_t __p0, uint8x16_t __p1) {
  55452.   uint8x16_t __ret;
  55453.   __ret = (uint8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x16_t)__p1, 48);
  55454.   return __ret;
  55455. }
  55456. #else
  55457. __ai uint8x16_t vqtbl4q_u8(uint8x16x4_t __p0, uint8x16_t __p1) {
  55458.   uint8x16_t __ret;
  55459.   uint8x16x4_t __rev0;
  55460.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55461.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55462.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55463.   __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55464.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55465.   __ret = (uint8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x16_t)__rev1, 48);
  55466.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55467.   return __ret;
  55468. }
  55469. #endif
  55470.  
  55471. #ifdef __LITTLE_ENDIAN__
  55472. __ai int8x16_t vqtbl4q_s8(int8x16x4_t __p0, uint8x16_t __p1) {
  55473.   int8x16_t __ret;
  55474.   __ret = (int8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x16_t)__p1, 32);
  55475.   return __ret;
  55476. }
  55477. #else
  55478. __ai int8x16_t vqtbl4q_s8(int8x16x4_t __p0, uint8x16_t __p1) {
  55479.   int8x16_t __ret;
  55480.   int8x16x4_t __rev0;
  55481.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55482.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55483.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55484.   __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55485.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55486.   __ret = (int8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x16_t)__rev1, 32);
  55487.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55488.   return __ret;
  55489. }
  55490. #endif
  55491.  
  55492. #ifdef __LITTLE_ENDIAN__
  55493. __ai uint8x8_t vqtbl4_u8(uint8x16x4_t __p0, uint8x8_t __p1) {
  55494.   uint8x8_t __ret;
  55495.   __ret = (uint8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x8_t)__p1, 16);
  55496.   return __ret;
  55497. }
  55498. #else
  55499. __ai uint8x8_t vqtbl4_u8(uint8x16x4_t __p0, uint8x8_t __p1) {
  55500.   uint8x8_t __ret;
  55501.   uint8x16x4_t __rev0;
  55502.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55503.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55504.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55505.   __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55506.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55507.   __ret = (uint8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x8_t)__rev1, 16);
  55508.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55509.   return __ret;
  55510. }
  55511. #endif
  55512.  
  55513. #ifdef __LITTLE_ENDIAN__
  55514. __ai int8x8_t vqtbl4_s8(int8x16x4_t __p0, uint8x8_t __p1) {
  55515.   int8x8_t __ret;
  55516.   __ret = (int8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x8_t)__p1, 0);
  55517.   return __ret;
  55518. }
  55519. #else
  55520. __ai int8x8_t vqtbl4_s8(int8x16x4_t __p0, uint8x8_t __p1) {
  55521.   int8x8_t __ret;
  55522.   int8x16x4_t __rev0;
  55523.   __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55524.   __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55525.   __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55526.   __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55527.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  55528.   __ret = (int8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x8_t)__rev1, 0);
  55529.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55530.   return __ret;
  55531. }
  55532. #endif
  55533.  
  55534. #ifdef __LITTLE_ENDIAN__
  55535. __ai poly8x8_t vqtbx1_p8(poly8x8_t __p0, poly8x16_t __p1, uint8x8_t __p2) {
  55536.   poly8x8_t __ret;
  55537.   __ret = (poly8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__p0, (int8x16_t)__p1, (int8x8_t)__p2, 4);
  55538.   return __ret;
  55539. }
  55540. #else
  55541. __ai poly8x8_t vqtbx1_p8(poly8x8_t __p0, poly8x16_t __p1, uint8x8_t __p2) {
  55542.   poly8x8_t __ret;
  55543.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55544.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55545.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55546.   __ret = (poly8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, 4);
  55547.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55548.   return __ret;
  55549. }
  55550. #endif
  55551.  
  55552. #ifdef __LITTLE_ENDIAN__
  55553. __ai poly8x16_t vqtbx1q_p8(poly8x16_t __p0, poly8x16_t __p1, uint8x16_t __p2) {
  55554.   poly8x16_t __ret;
  55555.   __ret = (poly8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 36);
  55556.   return __ret;
  55557. }
  55558. #else
  55559. __ai poly8x16_t vqtbx1q_p8(poly8x16_t __p0, poly8x16_t __p1, uint8x16_t __p2) {
  55560.   poly8x16_t __ret;
  55561.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55562.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55563.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55564.   __ret = (poly8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 36);
  55565.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55566.   return __ret;
  55567. }
  55568. #endif
  55569.  
  55570. #ifdef __LITTLE_ENDIAN__
  55571. __ai uint8x16_t vqtbx1q_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  55572.   uint8x16_t __ret;
  55573.   __ret = (uint8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 48);
  55574.   return __ret;
  55575. }
  55576. #else
  55577. __ai uint8x16_t vqtbx1q_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  55578.   uint8x16_t __ret;
  55579.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55580.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55581.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55582.   __ret = (uint8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 48);
  55583.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55584.   return __ret;
  55585. }
  55586. #endif
  55587.  
  55588. #ifdef __LITTLE_ENDIAN__
  55589. __ai int8x16_t vqtbx1q_s8(int8x16_t __p0, int8x16_t __p1, uint8x16_t __p2) {
  55590.   int8x16_t __ret;
  55591.   __ret = (int8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 32);
  55592.   return __ret;
  55593. }
  55594. #else
  55595. __ai int8x16_t vqtbx1q_s8(int8x16_t __p0, int8x16_t __p1, uint8x16_t __p2) {
  55596.   int8x16_t __ret;
  55597.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55598.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55599.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55600.   __ret = (int8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 32);
  55601.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55602.   return __ret;
  55603. }
  55604. #endif
  55605.  
  55606. #ifdef __LITTLE_ENDIAN__
  55607. __ai uint8x8_t vqtbx1_u8(uint8x8_t __p0, uint8x16_t __p1, uint8x8_t __p2) {
  55608.   uint8x8_t __ret;
  55609.   __ret = (uint8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__p0, (int8x16_t)__p1, (int8x8_t)__p2, 16);
  55610.   return __ret;
  55611. }
  55612. #else
  55613. __ai uint8x8_t vqtbx1_u8(uint8x8_t __p0, uint8x16_t __p1, uint8x8_t __p2) {
  55614.   uint8x8_t __ret;
  55615.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55616.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55617.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55618.   __ret = (uint8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, 16);
  55619.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55620.   return __ret;
  55621. }
  55622. #endif
  55623.  
  55624. #ifdef __LITTLE_ENDIAN__
  55625. __ai int8x8_t vqtbx1_s8(int8x8_t __p0, int8x16_t __p1, uint8x8_t __p2) {
  55626.   int8x8_t __ret;
  55627.   __ret = (int8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__p0, (int8x16_t)__p1, (int8x8_t)__p2, 0);
  55628.   return __ret;
  55629. }
  55630. #else
  55631. __ai int8x8_t vqtbx1_s8(int8x8_t __p0, int8x16_t __p1, uint8x8_t __p2) {
  55632.   int8x8_t __ret;
  55633.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55634.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55635.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55636.   __ret = (int8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, 0);
  55637.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55638.   return __ret;
  55639. }
  55640. #endif
  55641.  
  55642. #ifdef __LITTLE_ENDIAN__
  55643. __ai poly8x8_t vqtbx2_p8(poly8x8_t __p0, poly8x16x2_t __p1, uint8x8_t __p2) {
  55644.   poly8x8_t __ret;
  55645.   __ret = (poly8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x8_t)__p2, 4);
  55646.   return __ret;
  55647. }
  55648. #else
  55649. __ai poly8x8_t vqtbx2_p8(poly8x8_t __p0, poly8x16x2_t __p1, uint8x8_t __p2) {
  55650.   poly8x8_t __ret;
  55651.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55652.   poly8x16x2_t __rev1;
  55653.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55654.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55655.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55656.   __ret = (poly8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x8_t)__rev2, 4);
  55657.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55658.   return __ret;
  55659. }
  55660. #endif
  55661.  
  55662. #ifdef __LITTLE_ENDIAN__
  55663. __ai poly8x16_t vqtbx2q_p8(poly8x16_t __p0, poly8x16x2_t __p1, uint8x16_t __p2) {
  55664.   poly8x16_t __ret;
  55665.   __ret = (poly8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p2, 36);
  55666.   return __ret;
  55667. }
  55668. #else
  55669. __ai poly8x16_t vqtbx2q_p8(poly8x16_t __p0, poly8x16x2_t __p1, uint8x16_t __p2) {
  55670.   poly8x16_t __ret;
  55671.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55672.   poly8x16x2_t __rev1;
  55673.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55674.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55675.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55676.   __ret = (poly8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev2, 36);
  55677.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55678.   return __ret;
  55679. }
  55680. #endif
  55681.  
  55682. #ifdef __LITTLE_ENDIAN__
  55683. __ai uint8x16_t vqtbx2q_u8(uint8x16_t __p0, uint8x16x2_t __p1, uint8x16_t __p2) {
  55684.   uint8x16_t __ret;
  55685.   __ret = (uint8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p2, 48);
  55686.   return __ret;
  55687. }
  55688. #else
  55689. __ai uint8x16_t vqtbx2q_u8(uint8x16_t __p0, uint8x16x2_t __p1, uint8x16_t __p2) {
  55690.   uint8x16_t __ret;
  55691.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55692.   uint8x16x2_t __rev1;
  55693.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55694.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55695.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55696.   __ret = (uint8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev2, 48);
  55697.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55698.   return __ret;
  55699. }
  55700. #endif
  55701.  
  55702. #ifdef __LITTLE_ENDIAN__
  55703. __ai int8x16_t vqtbx2q_s8(int8x16_t __p0, int8x16x2_t __p1, uint8x16_t __p2) {
  55704.   int8x16_t __ret;
  55705.   __ret = (int8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p2, 32);
  55706.   return __ret;
  55707. }
  55708. #else
  55709. __ai int8x16_t vqtbx2q_s8(int8x16_t __p0, int8x16x2_t __p1, uint8x16_t __p2) {
  55710.   int8x16_t __ret;
  55711.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55712.   int8x16x2_t __rev1;
  55713.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55714.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55715.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55716.   __ret = (int8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev2, 32);
  55717.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55718.   return __ret;
  55719. }
  55720. #endif
  55721.  
  55722. #ifdef __LITTLE_ENDIAN__
  55723. __ai uint8x8_t vqtbx2_u8(uint8x8_t __p0, uint8x16x2_t __p1, uint8x8_t __p2) {
  55724.   uint8x8_t __ret;
  55725.   __ret = (uint8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x8_t)__p2, 16);
  55726.   return __ret;
  55727. }
  55728. #else
  55729. __ai uint8x8_t vqtbx2_u8(uint8x8_t __p0, uint8x16x2_t __p1, uint8x8_t __p2) {
  55730.   uint8x8_t __ret;
  55731.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55732.   uint8x16x2_t __rev1;
  55733.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55734.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55735.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55736.   __ret = (uint8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x8_t)__rev2, 16);
  55737.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55738.   return __ret;
  55739. }
  55740. #endif
  55741.  
  55742. #ifdef __LITTLE_ENDIAN__
  55743. __ai int8x8_t vqtbx2_s8(int8x8_t __p0, int8x16x2_t __p1, uint8x8_t __p2) {
  55744.   int8x8_t __ret;
  55745.   __ret = (int8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x8_t)__p2, 0);
  55746.   return __ret;
  55747. }
  55748. #else
  55749. __ai int8x8_t vqtbx2_s8(int8x8_t __p0, int8x16x2_t __p1, uint8x8_t __p2) {
  55750.   int8x8_t __ret;
  55751.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55752.   int8x16x2_t __rev1;
  55753.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55754.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55755.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55756.   __ret = (int8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x8_t)__rev2, 0);
  55757.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55758.   return __ret;
  55759. }
  55760. #endif
  55761.  
  55762. #ifdef __LITTLE_ENDIAN__
  55763. __ai poly8x8_t vqtbx3_p8(poly8x8_t __p0, poly8x16x3_t __p1, uint8x8_t __p2) {
  55764.   poly8x8_t __ret;
  55765.   __ret = (poly8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x8_t)__p2, 4);
  55766.   return __ret;
  55767. }
  55768. #else
  55769. __ai poly8x8_t vqtbx3_p8(poly8x8_t __p0, poly8x16x3_t __p1, uint8x8_t __p2) {
  55770.   poly8x8_t __ret;
  55771.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55772.   poly8x16x3_t __rev1;
  55773.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55774.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55775.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55776.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55777.   __ret = (poly8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x8_t)__rev2, 4);
  55778.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55779.   return __ret;
  55780. }
  55781. #endif
  55782.  
  55783. #ifdef __LITTLE_ENDIAN__
  55784. __ai poly8x16_t vqtbx3q_p8(poly8x16_t __p0, poly8x16x3_t __p1, uint8x16_t __p2) {
  55785.   poly8x16_t __ret;
  55786.   __ret = (poly8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p2, 36);
  55787.   return __ret;
  55788. }
  55789. #else
  55790. __ai poly8x16_t vqtbx3q_p8(poly8x16_t __p0, poly8x16x3_t __p1, uint8x16_t __p2) {
  55791.   poly8x16_t __ret;
  55792.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55793.   poly8x16x3_t __rev1;
  55794.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55795.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55796.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55797.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55798.   __ret = (poly8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev2, 36);
  55799.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55800.   return __ret;
  55801. }
  55802. #endif
  55803.  
  55804. #ifdef __LITTLE_ENDIAN__
  55805. __ai uint8x16_t vqtbx3q_u8(uint8x16_t __p0, uint8x16x3_t __p1, uint8x16_t __p2) {
  55806.   uint8x16_t __ret;
  55807.   __ret = (uint8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p2, 48);
  55808.   return __ret;
  55809. }
  55810. #else
  55811. __ai uint8x16_t vqtbx3q_u8(uint8x16_t __p0, uint8x16x3_t __p1, uint8x16_t __p2) {
  55812.   uint8x16_t __ret;
  55813.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55814.   uint8x16x3_t __rev1;
  55815.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55816.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55817.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55818.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55819.   __ret = (uint8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev2, 48);
  55820.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55821.   return __ret;
  55822. }
  55823. #endif
  55824.  
  55825. #ifdef __LITTLE_ENDIAN__
  55826. __ai int8x16_t vqtbx3q_s8(int8x16_t __p0, int8x16x3_t __p1, uint8x16_t __p2) {
  55827.   int8x16_t __ret;
  55828.   __ret = (int8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p2, 32);
  55829.   return __ret;
  55830. }
  55831. #else
  55832. __ai int8x16_t vqtbx3q_s8(int8x16_t __p0, int8x16x3_t __p1, uint8x16_t __p2) {
  55833.   int8x16_t __ret;
  55834.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55835.   int8x16x3_t __rev1;
  55836.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55837.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55838.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55839.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55840.   __ret = (int8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev2, 32);
  55841.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55842.   return __ret;
  55843. }
  55844. #endif
  55845.  
  55846. #ifdef __LITTLE_ENDIAN__
  55847. __ai uint8x8_t vqtbx3_u8(uint8x8_t __p0, uint8x16x3_t __p1, uint8x8_t __p2) {
  55848.   uint8x8_t __ret;
  55849.   __ret = (uint8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x8_t)__p2, 16);
  55850.   return __ret;
  55851. }
  55852. #else
  55853. __ai uint8x8_t vqtbx3_u8(uint8x8_t __p0, uint8x16x3_t __p1, uint8x8_t __p2) {
  55854.   uint8x8_t __ret;
  55855.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55856.   uint8x16x3_t __rev1;
  55857.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55858.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55859.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55860.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55861.   __ret = (uint8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x8_t)__rev2, 16);
  55862.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55863.   return __ret;
  55864. }
  55865. #endif
  55866.  
  55867. #ifdef __LITTLE_ENDIAN__
  55868. __ai int8x8_t vqtbx3_s8(int8x8_t __p0, int8x16x3_t __p1, uint8x8_t __p2) {
  55869.   int8x8_t __ret;
  55870.   __ret = (int8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x8_t)__p2, 0);
  55871.   return __ret;
  55872. }
  55873. #else
  55874. __ai int8x8_t vqtbx3_s8(int8x8_t __p0, int8x16x3_t __p1, uint8x8_t __p2) {
  55875.   int8x8_t __ret;
  55876.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55877.   int8x16x3_t __rev1;
  55878.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55879.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55880.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55881.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55882.   __ret = (int8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x8_t)__rev2, 0);
  55883.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55884.   return __ret;
  55885. }
  55886. #endif
  55887.  
  55888. #ifdef __LITTLE_ENDIAN__
  55889. __ai poly8x8_t vqtbx4_p8(poly8x8_t __p0, poly8x16x4_t __p1, uint8x8_t __p2) {
  55890.   poly8x8_t __ret;
  55891.   __ret = (poly8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x8_t)__p2, 4);
  55892.   return __ret;
  55893. }
  55894. #else
  55895. __ai poly8x8_t vqtbx4_p8(poly8x8_t __p0, poly8x16x4_t __p1, uint8x8_t __p2) {
  55896.   poly8x8_t __ret;
  55897.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55898.   poly8x16x4_t __rev1;
  55899.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55900.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55901.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55902.   __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55903.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55904.   __ret = (poly8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x8_t)__rev2, 4);
  55905.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55906.   return __ret;
  55907. }
  55908. #endif
  55909.  
  55910. #ifdef __LITTLE_ENDIAN__
  55911. __ai poly8x16_t vqtbx4q_p8(poly8x16_t __p0, poly8x16x4_t __p1, uint8x16_t __p2) {
  55912.   poly8x16_t __ret;
  55913.   __ret = (poly8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x16_t)__p2, 36);
  55914.   return __ret;
  55915. }
  55916. #else
  55917. __ai poly8x16_t vqtbx4q_p8(poly8x16_t __p0, poly8x16x4_t __p1, uint8x16_t __p2) {
  55918.   poly8x16_t __ret;
  55919.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55920.   poly8x16x4_t __rev1;
  55921.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55922.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55923.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55924.   __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55925.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55926.   __ret = (poly8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x16_t)__rev2, 36);
  55927.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55928.   return __ret;
  55929. }
  55930. #endif
  55931.  
  55932. #ifdef __LITTLE_ENDIAN__
  55933. __ai uint8x16_t vqtbx4q_u8(uint8x16_t __p0, uint8x16x4_t __p1, uint8x16_t __p2) {
  55934.   uint8x16_t __ret;
  55935.   __ret = (uint8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x16_t)__p2, 48);
  55936.   return __ret;
  55937. }
  55938. #else
  55939. __ai uint8x16_t vqtbx4q_u8(uint8x16_t __p0, uint8x16x4_t __p1, uint8x16_t __p2) {
  55940.   uint8x16_t __ret;
  55941.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55942.   uint8x16x4_t __rev1;
  55943.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55944.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55945.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55946.   __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55947.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55948.   __ret = (uint8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x16_t)__rev2, 48);
  55949.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55950.   return __ret;
  55951. }
  55952. #endif
  55953.  
  55954. #ifdef __LITTLE_ENDIAN__
  55955. __ai int8x16_t vqtbx4q_s8(int8x16_t __p0, int8x16x4_t __p1, uint8x16_t __p2) {
  55956.   int8x16_t __ret;
  55957.   __ret = (int8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x16_t)__p2, 32);
  55958.   return __ret;
  55959. }
  55960. #else
  55961. __ai int8x16_t vqtbx4q_s8(int8x16_t __p0, int8x16x4_t __p1, uint8x16_t __p2) {
  55962.   int8x16_t __ret;
  55963.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55964.   int8x16x4_t __rev1;
  55965.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55966.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55967.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55968.   __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55969.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55970.   __ret = (int8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x16_t)__rev2, 32);
  55971.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55972.   return __ret;
  55973. }
  55974. #endif
  55975.  
  55976. #ifdef __LITTLE_ENDIAN__
  55977. __ai uint8x8_t vqtbx4_u8(uint8x8_t __p0, uint8x16x4_t __p1, uint8x8_t __p2) {
  55978.   uint8x8_t __ret;
  55979.   __ret = (uint8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x8_t)__p2, 16);
  55980.   return __ret;
  55981. }
  55982. #else
  55983. __ai uint8x8_t vqtbx4_u8(uint8x8_t __p0, uint8x16x4_t __p1, uint8x8_t __p2) {
  55984.   uint8x8_t __ret;
  55985.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  55986.   uint8x16x4_t __rev1;
  55987.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55988.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55989.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55990.   __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  55991.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  55992.   __ret = (uint8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x8_t)__rev2, 16);
  55993.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  55994.   return __ret;
  55995. }
  55996. #endif
  55997.  
  55998. #ifdef __LITTLE_ENDIAN__
  55999. __ai int8x8_t vqtbx4_s8(int8x8_t __p0, int8x16x4_t __p1, uint8x8_t __p2) {
  56000.   int8x8_t __ret;
  56001.   __ret = (int8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x8_t)__p2, 0);
  56002.   return __ret;
  56003. }
  56004. #else
  56005. __ai int8x8_t vqtbx4_s8(int8x8_t __p0, int8x16x4_t __p1, uint8x8_t __p2) {
  56006.   int8x8_t __ret;
  56007.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  56008.   int8x16x4_t __rev1;
  56009.   __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56010.   __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56011.   __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56012.   __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56013.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  56014.   __ret = (int8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x8_t)__rev2, 0);
  56015.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  56016.   return __ret;
  56017. }
  56018. #endif
  56019.  
  56020. #ifdef __LITTLE_ENDIAN__
  56021. __ai uint16x8_t vraddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  56022.   uint16x8_t __ret;
  56023.   __ret = vcombine_u16(__p0, vraddhn_u32(__p1, __p2));
  56024.   return __ret;
  56025. }
  56026. #else
  56027. __ai uint16x8_t vraddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  56028.   uint16x8_t __ret;
  56029.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  56030.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  56031.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  56032.   __ret = __noswap_vcombine_u16(__rev0, __noswap_vraddhn_u32(__rev1, __rev2));
  56033.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  56034.   return __ret;
  56035. }
  56036. #endif
  56037.  
  56038. #ifdef __LITTLE_ENDIAN__
  56039. __ai uint32x4_t vraddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  56040.   uint32x4_t __ret;
  56041.   __ret = vcombine_u32(__p0, vraddhn_u64(__p1, __p2));
  56042.   return __ret;
  56043. }
  56044. #else
  56045. __ai uint32x4_t vraddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  56046.   uint32x4_t __ret;
  56047.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  56048.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  56049.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  56050.   __ret = __noswap_vcombine_u32(__rev0, __noswap_vraddhn_u64(__rev1, __rev2));
  56051.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  56052.   return __ret;
  56053. }
  56054. #endif
  56055.  
  56056. #ifdef __LITTLE_ENDIAN__
  56057. __ai uint8x16_t vraddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  56058.   uint8x16_t __ret;
  56059.   __ret = vcombine_u8(__p0, vraddhn_u16(__p1, __p2));
  56060.   return __ret;
  56061. }
  56062. #else
  56063. __ai uint8x16_t vraddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  56064.   uint8x16_t __ret;
  56065.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  56066.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  56067.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  56068.   __ret = __noswap_vcombine_u8(__rev0, __noswap_vraddhn_u16(__rev1, __rev2));
  56069.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56070.   return __ret;
  56071. }
  56072. #endif
  56073.  
  56074. #ifdef __LITTLE_ENDIAN__
  56075. __ai int16x8_t vraddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  56076.   int16x8_t __ret;
  56077.   __ret = vcombine_s16(__p0, vraddhn_s32(__p1, __p2));
  56078.   return __ret;
  56079. }
  56080. #else
  56081. __ai int16x8_t vraddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  56082.   int16x8_t __ret;
  56083.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  56084.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  56085.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  56086.   __ret = __noswap_vcombine_s16(__rev0, __noswap_vraddhn_s32(__rev1, __rev2));
  56087.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  56088.   return __ret;
  56089. }
  56090. #endif
  56091.  
  56092. #ifdef __LITTLE_ENDIAN__
  56093. __ai int32x4_t vraddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  56094.   int32x4_t __ret;
  56095.   __ret = vcombine_s32(__p0, vraddhn_s64(__p1, __p2));
  56096.   return __ret;
  56097. }
  56098. #else
  56099. __ai int32x4_t vraddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  56100.   int32x4_t __ret;
  56101.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  56102.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  56103.   int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  56104.   __ret = __noswap_vcombine_s32(__rev0, __noswap_vraddhn_s64(__rev1, __rev2));
  56105.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  56106.   return __ret;
  56107. }
  56108. #endif
  56109.  
  56110. #ifdef __LITTLE_ENDIAN__
  56111. __ai int8x16_t vraddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  56112.   int8x16_t __ret;
  56113.   __ret = vcombine_s8(__p0, vraddhn_s16(__p1, __p2));
  56114.   return __ret;
  56115. }
  56116. #else
  56117. __ai int8x16_t vraddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  56118.   int8x16_t __ret;
  56119.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  56120.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  56121.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  56122.   __ret = __noswap_vcombine_s8(__rev0, __noswap_vraddhn_s16(__rev1, __rev2));
  56123.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56124.   return __ret;
  56125. }
  56126. #endif
  56127.  
  56128. #ifdef __LITTLE_ENDIAN__
  56129. __ai poly8x8_t vrbit_p8(poly8x8_t __p0) {
  56130.   poly8x8_t __ret;
  56131.   __ret = (poly8x8_t) __builtin_neon_vrbit_v((int8x8_t)__p0, 4);
  56132.   return __ret;
  56133. }
  56134. #else
  56135. __ai poly8x8_t vrbit_p8(poly8x8_t __p0) {
  56136.   poly8x8_t __ret;
  56137.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  56138.   __ret = (poly8x8_t) __builtin_neon_vrbit_v((int8x8_t)__rev0, 4);
  56139.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  56140.   return __ret;
  56141. }
  56142. #endif
  56143.  
  56144. #ifdef __LITTLE_ENDIAN__
  56145. __ai poly8x16_t vrbitq_p8(poly8x16_t __p0) {
  56146.   poly8x16_t __ret;
  56147.   __ret = (poly8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__p0, 36);
  56148.   return __ret;
  56149. }
  56150. #else
  56151. __ai poly8x16_t vrbitq_p8(poly8x16_t __p0) {
  56152.   poly8x16_t __ret;
  56153.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56154.   __ret = (poly8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__rev0, 36);
  56155.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56156.   return __ret;
  56157. }
  56158. #endif
  56159.  
  56160. #ifdef __LITTLE_ENDIAN__
  56161. __ai uint8x16_t vrbitq_u8(uint8x16_t __p0) {
  56162.   uint8x16_t __ret;
  56163.   __ret = (uint8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__p0, 48);
  56164.   return __ret;
  56165. }
  56166. #else
  56167. __ai uint8x16_t vrbitq_u8(uint8x16_t __p0) {
  56168.   uint8x16_t __ret;
  56169.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56170.   __ret = (uint8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__rev0, 48);
  56171.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56172.   return __ret;
  56173. }
  56174. #endif
  56175.  
  56176. #ifdef __LITTLE_ENDIAN__
  56177. __ai int8x16_t vrbitq_s8(int8x16_t __p0) {
  56178.   int8x16_t __ret;
  56179.   __ret = (int8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__p0, 32);
  56180.   return __ret;
  56181. }
  56182. #else
  56183. __ai int8x16_t vrbitq_s8(int8x16_t __p0) {
  56184.   int8x16_t __ret;
  56185.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56186.   __ret = (int8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__rev0, 32);
  56187.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  56188.   return __ret;
  56189. }
  56190. #endif
  56191.  
  56192. #ifdef __LITTLE_ENDIAN__
  56193. __ai uint8x8_t vrbit_u8(uint8x8_t __p0) {
  56194.   uint8x8_t __ret;
  56195.   __ret = (uint8x8_t) __builtin_neon_vrbit_v((int8x8_t)__p0, 16);
  56196.   return __ret;
  56197. }
  56198. #else
  56199. __ai uint8x8_t vrbit_u8(uint8x8_t __p0) {
  56200.   uint8x8_t __ret;
  56201.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  56202.   __ret = (uint8x8_t) __builtin_neon_vrbit_v((int8x8_t)__rev0, 16);
  56203.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  56204.   return __ret;
  56205. }
  56206. #endif
  56207.  
  56208. #ifdef __LITTLE_ENDIAN__
  56209. __ai int8x8_t vrbit_s8(int8x8_t __p0) {
  56210.   int8x8_t __ret;
  56211.   __ret = (int8x8_t) __builtin_neon_vrbit_v((int8x8_t)__p0, 0);
  56212.   return __ret;
  56213. }
  56214. #else
  56215. __ai int8x8_t vrbit_s8(int8x8_t __p0) {
  56216.   int8x8_t __ret;
  56217.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  56218.   __ret = (int8x8_t) __builtin_neon_vrbit_v((int8x8_t)__rev0, 0);
  56219.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  56220.   return __ret;
  56221. }
  56222. #endif
  56223.  
  56224. #ifdef __LITTLE_ENDIAN__
  56225. __ai float64x2_t vrecpeq_f64(float64x2_t __p0) {
  56226.   float64x2_t __ret;
  56227.   __ret = (float64x2_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 42);
  56228.   return __ret;
  56229. }
  56230. #else
  56231. __ai float64x2_t vrecpeq_f64(float64x2_t __p0) {
  56232.   float64x2_t __ret;
  56233.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  56234.   __ret = (float64x2_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 42);
  56235.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  56236.   return __ret;
  56237. }
  56238. #endif
  56239.  
  56240. __ai float64x1_t vrecpe_f64(float64x1_t __p0) {
  56241.   float64x1_t __ret;
  56242.   __ret = (float64x1_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 10);
  56243.   return __ret;
  56244. }
  56245. __ai float64_t vrecped_f64(float64_t __p0) {
  56246.   float64_t __ret;
  56247.   __ret = (float64_t) __builtin_neon_vrecped_f64(__p0);
  56248.   return __ret;
  56249. }
  56250. __ai float32_t vrecpes_f32(float32_t __p0) {
  56251.   float32_t __ret;
  56252.   __ret = (float32_t) __builtin_neon_vrecpes_f32(__p0);
  56253.   return __ret;
  56254. }
  56255. #ifdef __LITTLE_ENDIAN__
  56256. __ai float64x2_t vrecpsq_f64(float64x2_t __p0, float64x2_t __p1) {
  56257.   float64x2_t __ret;
  56258.   __ret = (float64x2_t) __builtin_neon_vrecpsq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  56259.   return __ret;
  56260. }
  56261. #else
  56262. __ai float64x2_t vrecpsq_f64(float64x2_t __p0, float64x2_t __p1) {
  56263.   float64x2_t __ret;
  56264.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  56265.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  56266.   __ret = (float64x2_t) __builtin_neon_vrecpsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  56267.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  56268.   return __ret;
  56269. }
  56270. #endif
  56271.  
  56272. __ai float64x1_t vrecps_f64(float64x1_t __p0, float64x1_t __p1) {
  56273.   float64x1_t __ret;
  56274.   __ret = (float64x1_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
  56275.   return __ret;
  56276. }
  56277. __ai float64_t vrecpsd_f64(float64_t __p0, float64_t __p1) {
  56278.   float64_t __ret;
  56279.   __ret = (float64_t) __builtin_neon_vrecpsd_f64(__p0, __p1);
  56280.   return __ret;
  56281. }
  56282. __ai float32_t vrecpss_f32(float32_t __p0, float32_t __p1) {
  56283.   float32_t __ret;
  56284.   __ret = (float32_t) __builtin_neon_vrecpss_f32(__p0, __p1);
  56285.   return __ret;
  56286. }
  56287. __ai float64_t vrecpxd_f64(float64_t __p0) {
  56288.   float64_t __ret;
  56289.   __ret = (float64_t) __builtin_neon_vrecpxd_f64(__p0);
  56290.   return __ret;
  56291. }
  56292. __ai float32_t vrecpxs_f32(float32_t __p0) {
  56293.   float32_t __ret;
  56294.   __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0);
  56295.   return __ret;
  56296. }
  56297. __ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __p0) {
  56298.   poly8x8_t __ret;
  56299.   __ret = (poly8x8_t)(__p0);
  56300.   return __ret;
  56301. }
  56302. __ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
  56303.   poly8x8_t __ret;
  56304.   __ret = (poly8x8_t)(__p0);
  56305.   return __ret;
  56306. }
  56307. __ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
  56308.   poly8x8_t __ret;
  56309.   __ret = (poly8x8_t)(__p0);
  56310.   return __ret;
  56311. }
  56312. __ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
  56313.   poly8x8_t __ret;
  56314.   __ret = (poly8x8_t)(__p0);
  56315.   return __ret;
  56316. }
  56317. __ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
  56318.   poly8x8_t __ret;
  56319.   __ret = (poly8x8_t)(__p0);
  56320.   return __ret;
  56321. }
  56322. __ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
  56323.   poly8x8_t __ret;
  56324.   __ret = (poly8x8_t)(__p0);
  56325.   return __ret;
  56326. }
  56327. __ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
  56328.   poly8x8_t __ret;
  56329.   __ret = (poly8x8_t)(__p0);
  56330.   return __ret;
  56331. }
  56332. __ai poly8x8_t vreinterpret_p8_f64(float64x1_t __p0) {
  56333.   poly8x8_t __ret;
  56334.   __ret = (poly8x8_t)(__p0);
  56335.   return __ret;
  56336. }
  56337. __ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
  56338.   poly8x8_t __ret;
  56339.   __ret = (poly8x8_t)(__p0);
  56340.   return __ret;
  56341. }
  56342. __ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
  56343.   poly8x8_t __ret;
  56344.   __ret = (poly8x8_t)(__p0);
  56345.   return __ret;
  56346. }
  56347. __ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
  56348.   poly8x8_t __ret;
  56349.   __ret = (poly8x8_t)(__p0);
  56350.   return __ret;
  56351. }
  56352. __ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
  56353.   poly8x8_t __ret;
  56354.   __ret = (poly8x8_t)(__p0);
  56355.   return __ret;
  56356. }
  56357. __ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
  56358.   poly8x8_t __ret;
  56359.   __ret = (poly8x8_t)(__p0);
  56360.   return __ret;
  56361. }
  56362. __ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __p0) {
  56363.   poly64x1_t __ret;
  56364.   __ret = (poly64x1_t)(__p0);
  56365.   return __ret;
  56366. }
  56367. __ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __p0) {
  56368.   poly64x1_t __ret;
  56369.   __ret = (poly64x1_t)(__p0);
  56370.   return __ret;
  56371. }
  56372. __ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __p0) {
  56373.   poly64x1_t __ret;
  56374.   __ret = (poly64x1_t)(__p0);
  56375.   return __ret;
  56376. }
  56377. __ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __p0) {
  56378.   poly64x1_t __ret;
  56379.   __ret = (poly64x1_t)(__p0);
  56380.   return __ret;
  56381. }
  56382. __ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __p0) {
  56383.   poly64x1_t __ret;
  56384.   __ret = (poly64x1_t)(__p0);
  56385.   return __ret;
  56386. }
  56387. __ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __p0) {
  56388.   poly64x1_t __ret;
  56389.   __ret = (poly64x1_t)(__p0);
  56390.   return __ret;
  56391. }
  56392. __ai poly64x1_t vreinterpret_p64_s8(int8x8_t __p0) {
  56393.   poly64x1_t __ret;
  56394.   __ret = (poly64x1_t)(__p0);
  56395.   return __ret;
  56396. }
  56397. __ai poly64x1_t vreinterpret_p64_f64(float64x1_t __p0) {
  56398.   poly64x1_t __ret;
  56399.   __ret = (poly64x1_t)(__p0);
  56400.   return __ret;
  56401. }
  56402. __ai poly64x1_t vreinterpret_p64_f32(float32x2_t __p0) {
  56403.   poly64x1_t __ret;
  56404.   __ret = (poly64x1_t)(__p0);
  56405.   return __ret;
  56406. }
  56407. __ai poly64x1_t vreinterpret_p64_f16(float16x4_t __p0) {
  56408.   poly64x1_t __ret;
  56409.   __ret = (poly64x1_t)(__p0);
  56410.   return __ret;
  56411. }
  56412. __ai poly64x1_t vreinterpret_p64_s32(int32x2_t __p0) {
  56413.   poly64x1_t __ret;
  56414.   __ret = (poly64x1_t)(__p0);
  56415.   return __ret;
  56416. }
  56417. __ai poly64x1_t vreinterpret_p64_s64(int64x1_t __p0) {
  56418.   poly64x1_t __ret;
  56419.   __ret = (poly64x1_t)(__p0);
  56420.   return __ret;
  56421. }
  56422. __ai poly64x1_t vreinterpret_p64_s16(int16x4_t __p0) {
  56423.   poly64x1_t __ret;
  56424.   __ret = (poly64x1_t)(__p0);
  56425.   return __ret;
  56426. }
  56427. __ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
  56428.   poly16x4_t __ret;
  56429.   __ret = (poly16x4_t)(__p0);
  56430.   return __ret;
  56431. }
  56432. __ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __p0) {
  56433.   poly16x4_t __ret;
  56434.   __ret = (poly16x4_t)(__p0);
  56435.   return __ret;
  56436. }
  56437. __ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
  56438.   poly16x4_t __ret;
  56439.   __ret = (poly16x4_t)(__p0);
  56440.   return __ret;
  56441. }
  56442. __ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
  56443.   poly16x4_t __ret;
  56444.   __ret = (poly16x4_t)(__p0);
  56445.   return __ret;
  56446. }
  56447. __ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
  56448.   poly16x4_t __ret;
  56449.   __ret = (poly16x4_t)(__p0);
  56450.   return __ret;
  56451. }
  56452. __ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
  56453.   poly16x4_t __ret;
  56454.   __ret = (poly16x4_t)(__p0);
  56455.   return __ret;
  56456. }
  56457. __ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
  56458.   poly16x4_t __ret;
  56459.   __ret = (poly16x4_t)(__p0);
  56460.   return __ret;
  56461. }
  56462. __ai poly16x4_t vreinterpret_p16_f64(float64x1_t __p0) {
  56463.   poly16x4_t __ret;
  56464.   __ret = (poly16x4_t)(__p0);
  56465.   return __ret;
  56466. }
  56467. __ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
  56468.   poly16x4_t __ret;
  56469.   __ret = (poly16x4_t)(__p0);
  56470.   return __ret;
  56471. }
  56472. __ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
  56473.   poly16x4_t __ret;
  56474.   __ret = (poly16x4_t)(__p0);
  56475.   return __ret;
  56476. }
  56477. __ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
  56478.   poly16x4_t __ret;
  56479.   __ret = (poly16x4_t)(__p0);
  56480.   return __ret;
  56481. }
  56482. __ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
  56483.   poly16x4_t __ret;
  56484.   __ret = (poly16x4_t)(__p0);
  56485.   return __ret;
  56486. }
  56487. __ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
  56488.   poly16x4_t __ret;
  56489.   __ret = (poly16x4_t)(__p0);
  56490.   return __ret;
  56491. }
  56492. __ai poly8x16_t vreinterpretq_p8_p128(poly128_t __p0) {
  56493.   poly8x16_t __ret;
  56494.   __ret = (poly8x16_t)(__p0);
  56495.   return __ret;
  56496. }
  56497. __ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __p0) {
  56498.   poly8x16_t __ret;
  56499.   __ret = (poly8x16_t)(__p0);
  56500.   return __ret;
  56501. }
  56502. __ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
  56503.   poly8x16_t __ret;
  56504.   __ret = (poly8x16_t)(__p0);
  56505.   return __ret;
  56506. }
  56507. __ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
  56508.   poly8x16_t __ret;
  56509.   __ret = (poly8x16_t)(__p0);
  56510.   return __ret;
  56511. }
  56512. __ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
  56513.   poly8x16_t __ret;
  56514.   __ret = (poly8x16_t)(__p0);
  56515.   return __ret;
  56516. }
  56517. __ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
  56518.   poly8x16_t __ret;
  56519.   __ret = (poly8x16_t)(__p0);
  56520.   return __ret;
  56521. }
  56522. __ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
  56523.   poly8x16_t __ret;
  56524.   __ret = (poly8x16_t)(__p0);
  56525.   return __ret;
  56526. }
  56527. __ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
  56528.   poly8x16_t __ret;
  56529.   __ret = (poly8x16_t)(__p0);
  56530.   return __ret;
  56531. }
  56532. __ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __p0) {
  56533.   poly8x16_t __ret;
  56534.   __ret = (poly8x16_t)(__p0);
  56535.   return __ret;
  56536. }
  56537. __ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
  56538.   poly8x16_t __ret;
  56539.   __ret = (poly8x16_t)(__p0);
  56540.   return __ret;
  56541. }
  56542. __ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
  56543.   poly8x16_t __ret;
  56544.   __ret = (poly8x16_t)(__p0);
  56545.   return __ret;
  56546. }
  56547. __ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
  56548.   poly8x16_t __ret;
  56549.   __ret = (poly8x16_t)(__p0);
  56550.   return __ret;
  56551. }
  56552. __ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
  56553.   poly8x16_t __ret;
  56554.   __ret = (poly8x16_t)(__p0);
  56555.   return __ret;
  56556. }
  56557. __ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
  56558.   poly8x16_t __ret;
  56559.   __ret = (poly8x16_t)(__p0);
  56560.   return __ret;
  56561. }
  56562. __ai poly128_t vreinterpretq_p128_p8(poly8x16_t __p0) {
  56563.   poly128_t __ret;
  56564.   __ret = (poly128_t)(__p0);
  56565.   return __ret;
  56566. }
  56567. __ai poly128_t vreinterpretq_p128_p64(poly64x2_t __p0) {
  56568.   poly128_t __ret;
  56569.   __ret = (poly128_t)(__p0);
  56570.   return __ret;
  56571. }
  56572. __ai poly128_t vreinterpretq_p128_p16(poly16x8_t __p0) {
  56573.   poly128_t __ret;
  56574.   __ret = (poly128_t)(__p0);
  56575.   return __ret;
  56576. }
  56577. __ai poly128_t vreinterpretq_p128_u8(uint8x16_t __p0) {
  56578.   poly128_t __ret;
  56579.   __ret = (poly128_t)(__p0);
  56580.   return __ret;
  56581. }
  56582. __ai poly128_t vreinterpretq_p128_u32(uint32x4_t __p0) {
  56583.   poly128_t __ret;
  56584.   __ret = (poly128_t)(__p0);
  56585.   return __ret;
  56586. }
  56587. __ai poly128_t vreinterpretq_p128_u64(uint64x2_t __p0) {
  56588.   poly128_t __ret;
  56589.   __ret = (poly128_t)(__p0);
  56590.   return __ret;
  56591. }
  56592. __ai poly128_t vreinterpretq_p128_u16(uint16x8_t __p0) {
  56593.   poly128_t __ret;
  56594.   __ret = (poly128_t)(__p0);
  56595.   return __ret;
  56596. }
  56597. __ai poly128_t vreinterpretq_p128_s8(int8x16_t __p0) {
  56598.   poly128_t __ret;
  56599.   __ret = (poly128_t)(__p0);
  56600.   return __ret;
  56601. }
  56602. __ai poly128_t vreinterpretq_p128_f64(float64x2_t __p0) {
  56603.   poly128_t __ret;
  56604.   __ret = (poly128_t)(__p0);
  56605.   return __ret;
  56606. }
  56607. __ai poly128_t vreinterpretq_p128_f32(float32x4_t __p0) {
  56608.   poly128_t __ret;
  56609.   __ret = (poly128_t)(__p0);
  56610.   return __ret;
  56611. }
  56612. __ai poly128_t vreinterpretq_p128_f16(float16x8_t __p0) {
  56613.   poly128_t __ret;
  56614.   __ret = (poly128_t)(__p0);
  56615.   return __ret;
  56616. }
  56617. __ai poly128_t vreinterpretq_p128_s32(int32x4_t __p0) {
  56618.   poly128_t __ret;
  56619.   __ret = (poly128_t)(__p0);
  56620.   return __ret;
  56621. }
  56622. __ai poly128_t vreinterpretq_p128_s64(int64x2_t __p0) {
  56623.   poly128_t __ret;
  56624.   __ret = (poly128_t)(__p0);
  56625.   return __ret;
  56626. }
  56627. __ai poly128_t vreinterpretq_p128_s16(int16x8_t __p0) {
  56628.   poly128_t __ret;
  56629.   __ret = (poly128_t)(__p0);
  56630.   return __ret;
  56631. }
  56632. __ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __p0) {
  56633.   poly64x2_t __ret;
  56634.   __ret = (poly64x2_t)(__p0);
  56635.   return __ret;
  56636. }
  56637. __ai poly64x2_t vreinterpretq_p64_p128(poly128_t __p0) {
  56638.   poly64x2_t __ret;
  56639.   __ret = (poly64x2_t)(__p0);
  56640.   return __ret;
  56641. }
  56642. __ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __p0) {
  56643.   poly64x2_t __ret;
  56644.   __ret = (poly64x2_t)(__p0);
  56645.   return __ret;
  56646. }
  56647. __ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __p0) {
  56648.   poly64x2_t __ret;
  56649.   __ret = (poly64x2_t)(__p0);
  56650.   return __ret;
  56651. }
  56652. __ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __p0) {
  56653.   poly64x2_t __ret;
  56654.   __ret = (poly64x2_t)(__p0);
  56655.   return __ret;
  56656. }
  56657. __ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __p0) {
  56658.   poly64x2_t __ret;
  56659.   __ret = (poly64x2_t)(__p0);
  56660.   return __ret;
  56661. }
  56662. __ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __p0) {
  56663.   poly64x2_t __ret;
  56664.   __ret = (poly64x2_t)(__p0);
  56665.   return __ret;
  56666. }
  56667. __ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __p0) {
  56668.   poly64x2_t __ret;
  56669.   __ret = (poly64x2_t)(__p0);
  56670.   return __ret;
  56671. }
  56672. __ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __p0) {
  56673.   poly64x2_t __ret;
  56674.   __ret = (poly64x2_t)(__p0);
  56675.   return __ret;
  56676. }
  56677. __ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __p0) {
  56678.   poly64x2_t __ret;
  56679.   __ret = (poly64x2_t)(__p0);
  56680.   return __ret;
  56681. }
  56682. __ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __p0) {
  56683.   poly64x2_t __ret;
  56684.   __ret = (poly64x2_t)(__p0);
  56685.   return __ret;
  56686. }
  56687. __ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __p0) {
  56688.   poly64x2_t __ret;
  56689.   __ret = (poly64x2_t)(__p0);
  56690.   return __ret;
  56691. }
  56692. __ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __p0) {
  56693.   poly64x2_t __ret;
  56694.   __ret = (poly64x2_t)(__p0);
  56695.   return __ret;
  56696. }
  56697. __ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __p0) {
  56698.   poly64x2_t __ret;
  56699.   __ret = (poly64x2_t)(__p0);
  56700.   return __ret;
  56701. }
  56702. __ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
  56703.   poly16x8_t __ret;
  56704.   __ret = (poly16x8_t)(__p0);
  56705.   return __ret;
  56706. }
  56707. __ai poly16x8_t vreinterpretq_p16_p128(poly128_t __p0) {
  56708.   poly16x8_t __ret;
  56709.   __ret = (poly16x8_t)(__p0);
  56710.   return __ret;
  56711. }
  56712. __ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __p0) {
  56713.   poly16x8_t __ret;
  56714.   __ret = (poly16x8_t)(__p0);
  56715.   return __ret;
  56716. }
  56717. __ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
  56718.   poly16x8_t __ret;
  56719.   __ret = (poly16x8_t)(__p0);
  56720.   return __ret;
  56721. }
  56722. __ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
  56723.   poly16x8_t __ret;
  56724.   __ret = (poly16x8_t)(__p0);
  56725.   return __ret;
  56726. }
  56727. __ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
  56728.   poly16x8_t __ret;
  56729.   __ret = (poly16x8_t)(__p0);
  56730.   return __ret;
  56731. }
  56732. __ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
  56733.   poly16x8_t __ret;
  56734.   __ret = (poly16x8_t)(__p0);
  56735.   return __ret;
  56736. }
  56737. __ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
  56738.   poly16x8_t __ret;
  56739.   __ret = (poly16x8_t)(__p0);
  56740.   return __ret;
  56741. }
  56742. __ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __p0) {
  56743.   poly16x8_t __ret;
  56744.   __ret = (poly16x8_t)(__p0);
  56745.   return __ret;
  56746. }
  56747. __ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
  56748.   poly16x8_t __ret;
  56749.   __ret = (poly16x8_t)(__p0);
  56750.   return __ret;
  56751. }
  56752. __ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
  56753.   poly16x8_t __ret;
  56754.   __ret = (poly16x8_t)(__p0);
  56755.   return __ret;
  56756. }
  56757. __ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
  56758.   poly16x8_t __ret;
  56759.   __ret = (poly16x8_t)(__p0);
  56760.   return __ret;
  56761. }
  56762. __ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
  56763.   poly16x8_t __ret;
  56764.   __ret = (poly16x8_t)(__p0);
  56765.   return __ret;
  56766. }
  56767. __ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
  56768.   poly16x8_t __ret;
  56769.   __ret = (poly16x8_t)(__p0);
  56770.   return __ret;
  56771. }
  56772. __ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
  56773.   uint8x16_t __ret;
  56774.   __ret = (uint8x16_t)(__p0);
  56775.   return __ret;
  56776. }
  56777. __ai uint8x16_t vreinterpretq_u8_p128(poly128_t __p0) {
  56778.   uint8x16_t __ret;
  56779.   __ret = (uint8x16_t)(__p0);
  56780.   return __ret;
  56781. }
  56782. __ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __p0) {
  56783.   uint8x16_t __ret;
  56784.   __ret = (uint8x16_t)(__p0);
  56785.   return __ret;
  56786. }
  56787. __ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
  56788.   uint8x16_t __ret;
  56789.   __ret = (uint8x16_t)(__p0);
  56790.   return __ret;
  56791. }
  56792. __ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
  56793.   uint8x16_t __ret;
  56794.   __ret = (uint8x16_t)(__p0);
  56795.   return __ret;
  56796. }
  56797. __ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
  56798.   uint8x16_t __ret;
  56799.   __ret = (uint8x16_t)(__p0);
  56800.   return __ret;
  56801. }
  56802. __ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
  56803.   uint8x16_t __ret;
  56804.   __ret = (uint8x16_t)(__p0);
  56805.   return __ret;
  56806. }
  56807. __ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
  56808.   uint8x16_t __ret;
  56809.   __ret = (uint8x16_t)(__p0);
  56810.   return __ret;
  56811. }
  56812. __ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __p0) {
  56813.   uint8x16_t __ret;
  56814.   __ret = (uint8x16_t)(__p0);
  56815.   return __ret;
  56816. }
  56817. __ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
  56818.   uint8x16_t __ret;
  56819.   __ret = (uint8x16_t)(__p0);
  56820.   return __ret;
  56821. }
  56822. __ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
  56823.   uint8x16_t __ret;
  56824.   __ret = (uint8x16_t)(__p0);
  56825.   return __ret;
  56826. }
  56827. __ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
  56828.   uint8x16_t __ret;
  56829.   __ret = (uint8x16_t)(__p0);
  56830.   return __ret;
  56831. }
  56832. __ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
  56833.   uint8x16_t __ret;
  56834.   __ret = (uint8x16_t)(__p0);
  56835.   return __ret;
  56836. }
  56837. __ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
  56838.   uint8x16_t __ret;
  56839.   __ret = (uint8x16_t)(__p0);
  56840.   return __ret;
  56841. }
  56842. __ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
  56843.   uint32x4_t __ret;
  56844.   __ret = (uint32x4_t)(__p0);
  56845.   return __ret;
  56846. }
  56847. __ai uint32x4_t vreinterpretq_u32_p128(poly128_t __p0) {
  56848.   uint32x4_t __ret;
  56849.   __ret = (uint32x4_t)(__p0);
  56850.   return __ret;
  56851. }
  56852. __ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __p0) {
  56853.   uint32x4_t __ret;
  56854.   __ret = (uint32x4_t)(__p0);
  56855.   return __ret;
  56856. }
  56857. __ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
  56858.   uint32x4_t __ret;
  56859.   __ret = (uint32x4_t)(__p0);
  56860.   return __ret;
  56861. }
  56862. __ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
  56863.   uint32x4_t __ret;
  56864.   __ret = (uint32x4_t)(__p0);
  56865.   return __ret;
  56866. }
  56867. __ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
  56868.   uint32x4_t __ret;
  56869.   __ret = (uint32x4_t)(__p0);
  56870.   return __ret;
  56871. }
  56872. __ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
  56873.   uint32x4_t __ret;
  56874.   __ret = (uint32x4_t)(__p0);
  56875.   return __ret;
  56876. }
  56877. __ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
  56878.   uint32x4_t __ret;
  56879.   __ret = (uint32x4_t)(__p0);
  56880.   return __ret;
  56881. }
  56882. __ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __p0) {
  56883.   uint32x4_t __ret;
  56884.   __ret = (uint32x4_t)(__p0);
  56885.   return __ret;
  56886. }
  56887. __ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
  56888.   uint32x4_t __ret;
  56889.   __ret = (uint32x4_t)(__p0);
  56890.   return __ret;
  56891. }
  56892. __ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
  56893.   uint32x4_t __ret;
  56894.   __ret = (uint32x4_t)(__p0);
  56895.   return __ret;
  56896. }
  56897. __ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
  56898.   uint32x4_t __ret;
  56899.   __ret = (uint32x4_t)(__p0);
  56900.   return __ret;
  56901. }
  56902. __ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
  56903.   uint32x4_t __ret;
  56904.   __ret = (uint32x4_t)(__p0);
  56905.   return __ret;
  56906. }
  56907. __ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
  56908.   uint32x4_t __ret;
  56909.   __ret = (uint32x4_t)(__p0);
  56910.   return __ret;
  56911. }
  56912. __ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
  56913.   uint64x2_t __ret;
  56914.   __ret = (uint64x2_t)(__p0);
  56915.   return __ret;
  56916. }
  56917. __ai uint64x2_t vreinterpretq_u64_p128(poly128_t __p0) {
  56918.   uint64x2_t __ret;
  56919.   __ret = (uint64x2_t)(__p0);
  56920.   return __ret;
  56921. }
  56922. __ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __p0) {
  56923.   uint64x2_t __ret;
  56924.   __ret = (uint64x2_t)(__p0);
  56925.   return __ret;
  56926. }
  56927. __ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
  56928.   uint64x2_t __ret;
  56929.   __ret = (uint64x2_t)(__p0);
  56930.   return __ret;
  56931. }
  56932. __ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
  56933.   uint64x2_t __ret;
  56934.   __ret = (uint64x2_t)(__p0);
  56935.   return __ret;
  56936. }
  56937. __ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
  56938.   uint64x2_t __ret;
  56939.   __ret = (uint64x2_t)(__p0);
  56940.   return __ret;
  56941. }
  56942. __ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
  56943.   uint64x2_t __ret;
  56944.   __ret = (uint64x2_t)(__p0);
  56945.   return __ret;
  56946. }
  56947. __ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
  56948.   uint64x2_t __ret;
  56949.   __ret = (uint64x2_t)(__p0);
  56950.   return __ret;
  56951. }
  56952. __ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __p0) {
  56953.   uint64x2_t __ret;
  56954.   __ret = (uint64x2_t)(__p0);
  56955.   return __ret;
  56956. }
  56957. __ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
  56958.   uint64x2_t __ret;
  56959.   __ret = (uint64x2_t)(__p0);
  56960.   return __ret;
  56961. }
  56962. __ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
  56963.   uint64x2_t __ret;
  56964.   __ret = (uint64x2_t)(__p0);
  56965.   return __ret;
  56966. }
  56967. __ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
  56968.   uint64x2_t __ret;
  56969.   __ret = (uint64x2_t)(__p0);
  56970.   return __ret;
  56971. }
  56972. __ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
  56973.   uint64x2_t __ret;
  56974.   __ret = (uint64x2_t)(__p0);
  56975.   return __ret;
  56976. }
  56977. __ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
  56978.   uint64x2_t __ret;
  56979.   __ret = (uint64x2_t)(__p0);
  56980.   return __ret;
  56981. }
  56982. __ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
  56983.   uint16x8_t __ret;
  56984.   __ret = (uint16x8_t)(__p0);
  56985.   return __ret;
  56986. }
  56987. __ai uint16x8_t vreinterpretq_u16_p128(poly128_t __p0) {
  56988.   uint16x8_t __ret;
  56989.   __ret = (uint16x8_t)(__p0);
  56990.   return __ret;
  56991. }
  56992. __ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __p0) {
  56993.   uint16x8_t __ret;
  56994.   __ret = (uint16x8_t)(__p0);
  56995.   return __ret;
  56996. }
  56997. __ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
  56998.   uint16x8_t __ret;
  56999.   __ret = (uint16x8_t)(__p0);
  57000.   return __ret;
  57001. }
  57002. __ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
  57003.   uint16x8_t __ret;
  57004.   __ret = (uint16x8_t)(__p0);
  57005.   return __ret;
  57006. }
  57007. __ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
  57008.   uint16x8_t __ret;
  57009.   __ret = (uint16x8_t)(__p0);
  57010.   return __ret;
  57011. }
  57012. __ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
  57013.   uint16x8_t __ret;
  57014.   __ret = (uint16x8_t)(__p0);
  57015.   return __ret;
  57016. }
  57017. __ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
  57018.   uint16x8_t __ret;
  57019.   __ret = (uint16x8_t)(__p0);
  57020.   return __ret;
  57021. }
  57022. __ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __p0) {
  57023.   uint16x8_t __ret;
  57024.   __ret = (uint16x8_t)(__p0);
  57025.   return __ret;
  57026. }
  57027. __ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
  57028.   uint16x8_t __ret;
  57029.   __ret = (uint16x8_t)(__p0);
  57030.   return __ret;
  57031. }
  57032. __ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
  57033.   uint16x8_t __ret;
  57034.   __ret = (uint16x8_t)(__p0);
  57035.   return __ret;
  57036. }
  57037. __ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
  57038.   uint16x8_t __ret;
  57039.   __ret = (uint16x8_t)(__p0);
  57040.   return __ret;
  57041. }
  57042. __ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
  57043.   uint16x8_t __ret;
  57044.   __ret = (uint16x8_t)(__p0);
  57045.   return __ret;
  57046. }
  57047. __ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
  57048.   uint16x8_t __ret;
  57049.   __ret = (uint16x8_t)(__p0);
  57050.   return __ret;
  57051. }
  57052. __ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
  57053.   int8x16_t __ret;
  57054.   __ret = (int8x16_t)(__p0);
  57055.   return __ret;
  57056. }
  57057. __ai int8x16_t vreinterpretq_s8_p128(poly128_t __p0) {
  57058.   int8x16_t __ret;
  57059.   __ret = (int8x16_t)(__p0);
  57060.   return __ret;
  57061. }
  57062. __ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __p0) {
  57063.   int8x16_t __ret;
  57064.   __ret = (int8x16_t)(__p0);
  57065.   return __ret;
  57066. }
  57067. __ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
  57068.   int8x16_t __ret;
  57069.   __ret = (int8x16_t)(__p0);
  57070.   return __ret;
  57071. }
  57072. __ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
  57073.   int8x16_t __ret;
  57074.   __ret = (int8x16_t)(__p0);
  57075.   return __ret;
  57076. }
  57077. __ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
  57078.   int8x16_t __ret;
  57079.   __ret = (int8x16_t)(__p0);
  57080.   return __ret;
  57081. }
  57082. __ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
  57083.   int8x16_t __ret;
  57084.   __ret = (int8x16_t)(__p0);
  57085.   return __ret;
  57086. }
  57087. __ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
  57088.   int8x16_t __ret;
  57089.   __ret = (int8x16_t)(__p0);
  57090.   return __ret;
  57091. }
  57092. __ai int8x16_t vreinterpretq_s8_f64(float64x2_t __p0) {
  57093.   int8x16_t __ret;
  57094.   __ret = (int8x16_t)(__p0);
  57095.   return __ret;
  57096. }
  57097. __ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
  57098.   int8x16_t __ret;
  57099.   __ret = (int8x16_t)(__p0);
  57100.   return __ret;
  57101. }
  57102. __ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
  57103.   int8x16_t __ret;
  57104.   __ret = (int8x16_t)(__p0);
  57105.   return __ret;
  57106. }
  57107. __ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
  57108.   int8x16_t __ret;
  57109.   __ret = (int8x16_t)(__p0);
  57110.   return __ret;
  57111. }
  57112. __ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
  57113.   int8x16_t __ret;
  57114.   __ret = (int8x16_t)(__p0);
  57115.   return __ret;
  57116. }
  57117. __ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
  57118.   int8x16_t __ret;
  57119.   __ret = (int8x16_t)(__p0);
  57120.   return __ret;
  57121. }
  57122. __ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __p0) {
  57123.   float64x2_t __ret;
  57124.   __ret = (float64x2_t)(__p0);
  57125.   return __ret;
  57126. }
  57127. __ai float64x2_t vreinterpretq_f64_p128(poly128_t __p0) {
  57128.   float64x2_t __ret;
  57129.   __ret = (float64x2_t)(__p0);
  57130.   return __ret;
  57131. }
  57132. __ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __p0) {
  57133.   float64x2_t __ret;
  57134.   __ret = (float64x2_t)(__p0);
  57135.   return __ret;
  57136. }
  57137. __ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __p0) {
  57138.   float64x2_t __ret;
  57139.   __ret = (float64x2_t)(__p0);
  57140.   return __ret;
  57141. }
  57142. __ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __p0) {
  57143.   float64x2_t __ret;
  57144.   __ret = (float64x2_t)(__p0);
  57145.   return __ret;
  57146. }
  57147. __ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __p0) {
  57148.   float64x2_t __ret;
  57149.   __ret = (float64x2_t)(__p0);
  57150.   return __ret;
  57151. }
  57152. __ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __p0) {
  57153.   float64x2_t __ret;
  57154.   __ret = (float64x2_t)(__p0);
  57155.   return __ret;
  57156. }
  57157. __ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __p0) {
  57158.   float64x2_t __ret;
  57159.   __ret = (float64x2_t)(__p0);
  57160.   return __ret;
  57161. }
  57162. __ai float64x2_t vreinterpretq_f64_s8(int8x16_t __p0) {
  57163.   float64x2_t __ret;
  57164.   __ret = (float64x2_t)(__p0);
  57165.   return __ret;
  57166. }
  57167. __ai float64x2_t vreinterpretq_f64_f32(float32x4_t __p0) {
  57168.   float64x2_t __ret;
  57169.   __ret = (float64x2_t)(__p0);
  57170.   return __ret;
  57171. }
  57172. __ai float64x2_t vreinterpretq_f64_f16(float16x8_t __p0) {
  57173.   float64x2_t __ret;
  57174.   __ret = (float64x2_t)(__p0);
  57175.   return __ret;
  57176. }
  57177. __ai float64x2_t vreinterpretq_f64_s32(int32x4_t __p0) {
  57178.   float64x2_t __ret;
  57179.   __ret = (float64x2_t)(__p0);
  57180.   return __ret;
  57181. }
  57182. __ai float64x2_t vreinterpretq_f64_s64(int64x2_t __p0) {
  57183.   float64x2_t __ret;
  57184.   __ret = (float64x2_t)(__p0);
  57185.   return __ret;
  57186. }
  57187. __ai float64x2_t vreinterpretq_f64_s16(int16x8_t __p0) {
  57188.   float64x2_t __ret;
  57189.   __ret = (float64x2_t)(__p0);
  57190.   return __ret;
  57191. }
  57192. __ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
  57193.   float32x4_t __ret;
  57194.   __ret = (float32x4_t)(__p0);
  57195.   return __ret;
  57196. }
  57197. __ai float32x4_t vreinterpretq_f32_p128(poly128_t __p0) {
  57198.   float32x4_t __ret;
  57199.   __ret = (float32x4_t)(__p0);
  57200.   return __ret;
  57201. }
  57202. __ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __p0) {
  57203.   float32x4_t __ret;
  57204.   __ret = (float32x4_t)(__p0);
  57205.   return __ret;
  57206. }
  57207. __ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
  57208.   float32x4_t __ret;
  57209.   __ret = (float32x4_t)(__p0);
  57210.   return __ret;
  57211. }
  57212. __ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
  57213.   float32x4_t __ret;
  57214.   __ret = (float32x4_t)(__p0);
  57215.   return __ret;
  57216. }
  57217. __ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
  57218.   float32x4_t __ret;
  57219.   __ret = (float32x4_t)(__p0);
  57220.   return __ret;
  57221. }
  57222. __ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
  57223.   float32x4_t __ret;
  57224.   __ret = (float32x4_t)(__p0);
  57225.   return __ret;
  57226. }
  57227. __ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
  57228.   float32x4_t __ret;
  57229.   __ret = (float32x4_t)(__p0);
  57230.   return __ret;
  57231. }
  57232. __ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
  57233.   float32x4_t __ret;
  57234.   __ret = (float32x4_t)(__p0);
  57235.   return __ret;
  57236. }
  57237. __ai float32x4_t vreinterpretq_f32_f64(float64x2_t __p0) {
  57238.   float32x4_t __ret;
  57239.   __ret = (float32x4_t)(__p0);
  57240.   return __ret;
  57241. }
  57242. __ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
  57243.   float32x4_t __ret;
  57244.   __ret = (float32x4_t)(__p0);
  57245.   return __ret;
  57246. }
  57247. __ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
  57248.   float32x4_t __ret;
  57249.   __ret = (float32x4_t)(__p0);
  57250.   return __ret;
  57251. }
  57252. __ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
  57253.   float32x4_t __ret;
  57254.   __ret = (float32x4_t)(__p0);
  57255.   return __ret;
  57256. }
  57257. __ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
  57258.   float32x4_t __ret;
  57259.   __ret = (float32x4_t)(__p0);
  57260.   return __ret;
  57261. }
  57262. __ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
  57263.   float16x8_t __ret;
  57264.   __ret = (float16x8_t)(__p0);
  57265.   return __ret;
  57266. }
  57267. __ai float16x8_t vreinterpretq_f16_p128(poly128_t __p0) {
  57268.   float16x8_t __ret;
  57269.   __ret = (float16x8_t)(__p0);
  57270.   return __ret;
  57271. }
  57272. __ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __p0) {
  57273.   float16x8_t __ret;
  57274.   __ret = (float16x8_t)(__p0);
  57275.   return __ret;
  57276. }
  57277. __ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
  57278.   float16x8_t __ret;
  57279.   __ret = (float16x8_t)(__p0);
  57280.   return __ret;
  57281. }
  57282. __ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
  57283.   float16x8_t __ret;
  57284.   __ret = (float16x8_t)(__p0);
  57285.   return __ret;
  57286. }
  57287. __ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
  57288.   float16x8_t __ret;
  57289.   __ret = (float16x8_t)(__p0);
  57290.   return __ret;
  57291. }
  57292. __ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
  57293.   float16x8_t __ret;
  57294.   __ret = (float16x8_t)(__p0);
  57295.   return __ret;
  57296. }
  57297. __ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
  57298.   float16x8_t __ret;
  57299.   __ret = (float16x8_t)(__p0);
  57300.   return __ret;
  57301. }
  57302. __ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
  57303.   float16x8_t __ret;
  57304.   __ret = (float16x8_t)(__p0);
  57305.   return __ret;
  57306. }
  57307. __ai float16x8_t vreinterpretq_f16_f64(float64x2_t __p0) {
  57308.   float16x8_t __ret;
  57309.   __ret = (float16x8_t)(__p0);
  57310.   return __ret;
  57311. }
  57312. __ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
  57313.   float16x8_t __ret;
  57314.   __ret = (float16x8_t)(__p0);
  57315.   return __ret;
  57316. }
  57317. __ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
  57318.   float16x8_t __ret;
  57319.   __ret = (float16x8_t)(__p0);
  57320.   return __ret;
  57321. }
  57322. __ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
  57323.   float16x8_t __ret;
  57324.   __ret = (float16x8_t)(__p0);
  57325.   return __ret;
  57326. }
  57327. __ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
  57328.   float16x8_t __ret;
  57329.   __ret = (float16x8_t)(__p0);
  57330.   return __ret;
  57331. }
  57332. __ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
  57333.   int32x4_t __ret;
  57334.   __ret = (int32x4_t)(__p0);
  57335.   return __ret;
  57336. }
  57337. __ai int32x4_t vreinterpretq_s32_p128(poly128_t __p0) {
  57338.   int32x4_t __ret;
  57339.   __ret = (int32x4_t)(__p0);
  57340.   return __ret;
  57341. }
  57342. __ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __p0) {
  57343.   int32x4_t __ret;
  57344.   __ret = (int32x4_t)(__p0);
  57345.   return __ret;
  57346. }
  57347. __ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
  57348.   int32x4_t __ret;
  57349.   __ret = (int32x4_t)(__p0);
  57350.   return __ret;
  57351. }
  57352. __ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
  57353.   int32x4_t __ret;
  57354.   __ret = (int32x4_t)(__p0);
  57355.   return __ret;
  57356. }
  57357. __ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
  57358.   int32x4_t __ret;
  57359.   __ret = (int32x4_t)(__p0);
  57360.   return __ret;
  57361. }
  57362. __ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
  57363.   int32x4_t __ret;
  57364.   __ret = (int32x4_t)(__p0);
  57365.   return __ret;
  57366. }
  57367. __ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
  57368.   int32x4_t __ret;
  57369.   __ret = (int32x4_t)(__p0);
  57370.   return __ret;
  57371. }
  57372. __ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
  57373.   int32x4_t __ret;
  57374.   __ret = (int32x4_t)(__p0);
  57375.   return __ret;
  57376. }
  57377. __ai int32x4_t vreinterpretq_s32_f64(float64x2_t __p0) {
  57378.   int32x4_t __ret;
  57379.   __ret = (int32x4_t)(__p0);
  57380.   return __ret;
  57381. }
  57382. __ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
  57383.   int32x4_t __ret;
  57384.   __ret = (int32x4_t)(__p0);
  57385.   return __ret;
  57386. }
  57387. __ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
  57388.   int32x4_t __ret;
  57389.   __ret = (int32x4_t)(__p0);
  57390.   return __ret;
  57391. }
  57392. __ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
  57393.   int32x4_t __ret;
  57394.   __ret = (int32x4_t)(__p0);
  57395.   return __ret;
  57396. }
  57397. __ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
  57398.   int32x4_t __ret;
  57399.   __ret = (int32x4_t)(__p0);
  57400.   return __ret;
  57401. }
  57402. __ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
  57403.   int64x2_t __ret;
  57404.   __ret = (int64x2_t)(__p0);
  57405.   return __ret;
  57406. }
  57407. __ai int64x2_t vreinterpretq_s64_p128(poly128_t __p0) {
  57408.   int64x2_t __ret;
  57409.   __ret = (int64x2_t)(__p0);
  57410.   return __ret;
  57411. }
  57412. __ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __p0) {
  57413.   int64x2_t __ret;
  57414.   __ret = (int64x2_t)(__p0);
  57415.   return __ret;
  57416. }
  57417. __ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
  57418.   int64x2_t __ret;
  57419.   __ret = (int64x2_t)(__p0);
  57420.   return __ret;
  57421. }
  57422. __ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
  57423.   int64x2_t __ret;
  57424.   __ret = (int64x2_t)(__p0);
  57425.   return __ret;
  57426. }
  57427. __ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
  57428.   int64x2_t __ret;
  57429.   __ret = (int64x2_t)(__p0);
  57430.   return __ret;
  57431. }
  57432. __ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
  57433.   int64x2_t __ret;
  57434.   __ret = (int64x2_t)(__p0);
  57435.   return __ret;
  57436. }
  57437. __ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
  57438.   int64x2_t __ret;
  57439.   __ret = (int64x2_t)(__p0);
  57440.   return __ret;
  57441. }
  57442. __ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
  57443.   int64x2_t __ret;
  57444.   __ret = (int64x2_t)(__p0);
  57445.   return __ret;
  57446. }
  57447. __ai int64x2_t vreinterpretq_s64_f64(float64x2_t __p0) {
  57448.   int64x2_t __ret;
  57449.   __ret = (int64x2_t)(__p0);
  57450.   return __ret;
  57451. }
  57452. __ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
  57453.   int64x2_t __ret;
  57454.   __ret = (int64x2_t)(__p0);
  57455.   return __ret;
  57456. }
  57457. __ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
  57458.   int64x2_t __ret;
  57459.   __ret = (int64x2_t)(__p0);
  57460.   return __ret;
  57461. }
  57462. __ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
  57463.   int64x2_t __ret;
  57464.   __ret = (int64x2_t)(__p0);
  57465.   return __ret;
  57466. }
  57467. __ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
  57468.   int64x2_t __ret;
  57469.   __ret = (int64x2_t)(__p0);
  57470.   return __ret;
  57471. }
  57472. __ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
  57473.   int16x8_t __ret;
  57474.   __ret = (int16x8_t)(__p0);
  57475.   return __ret;
  57476. }
  57477. __ai int16x8_t vreinterpretq_s16_p128(poly128_t __p0) {
  57478.   int16x8_t __ret;
  57479.   __ret = (int16x8_t)(__p0);
  57480.   return __ret;
  57481. }
  57482. __ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __p0) {
  57483.   int16x8_t __ret;
  57484.   __ret = (int16x8_t)(__p0);
  57485.   return __ret;
  57486. }
  57487. __ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
  57488.   int16x8_t __ret;
  57489.   __ret = (int16x8_t)(__p0);
  57490.   return __ret;
  57491. }
  57492. __ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
  57493.   int16x8_t __ret;
  57494.   __ret = (int16x8_t)(__p0);
  57495.   return __ret;
  57496. }
  57497. __ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
  57498.   int16x8_t __ret;
  57499.   __ret = (int16x8_t)(__p0);
  57500.   return __ret;
  57501. }
  57502. __ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
  57503.   int16x8_t __ret;
  57504.   __ret = (int16x8_t)(__p0);
  57505.   return __ret;
  57506. }
  57507. __ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
  57508.   int16x8_t __ret;
  57509.   __ret = (int16x8_t)(__p0);
  57510.   return __ret;
  57511. }
  57512. __ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
  57513.   int16x8_t __ret;
  57514.   __ret = (int16x8_t)(__p0);
  57515.   return __ret;
  57516. }
  57517. __ai int16x8_t vreinterpretq_s16_f64(float64x2_t __p0) {
  57518.   int16x8_t __ret;
  57519.   __ret = (int16x8_t)(__p0);
  57520.   return __ret;
  57521. }
  57522. __ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
  57523.   int16x8_t __ret;
  57524.   __ret = (int16x8_t)(__p0);
  57525.   return __ret;
  57526. }
  57527. __ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
  57528.   int16x8_t __ret;
  57529.   __ret = (int16x8_t)(__p0);
  57530.   return __ret;
  57531. }
  57532. __ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
  57533.   int16x8_t __ret;
  57534.   __ret = (int16x8_t)(__p0);
  57535.   return __ret;
  57536. }
  57537. __ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
  57538.   int16x8_t __ret;
  57539.   __ret = (int16x8_t)(__p0);
  57540.   return __ret;
  57541. }
  57542. __ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
  57543.   uint8x8_t __ret;
  57544.   __ret = (uint8x8_t)(__p0);
  57545.   return __ret;
  57546. }
  57547. __ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __p0) {
  57548.   uint8x8_t __ret;
  57549.   __ret = (uint8x8_t)(__p0);
  57550.   return __ret;
  57551. }
  57552. __ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
  57553.   uint8x8_t __ret;
  57554.   __ret = (uint8x8_t)(__p0);
  57555.   return __ret;
  57556. }
  57557. __ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
  57558.   uint8x8_t __ret;
  57559.   __ret = (uint8x8_t)(__p0);
  57560.   return __ret;
  57561. }
  57562. __ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
  57563.   uint8x8_t __ret;
  57564.   __ret = (uint8x8_t)(__p0);
  57565.   return __ret;
  57566. }
  57567. __ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
  57568.   uint8x8_t __ret;
  57569.   __ret = (uint8x8_t)(__p0);
  57570.   return __ret;
  57571. }
  57572. __ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
  57573.   uint8x8_t __ret;
  57574.   __ret = (uint8x8_t)(__p0);
  57575.   return __ret;
  57576. }
  57577. __ai uint8x8_t vreinterpret_u8_f64(float64x1_t __p0) {
  57578.   uint8x8_t __ret;
  57579.   __ret = (uint8x8_t)(__p0);
  57580.   return __ret;
  57581. }
  57582. __ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
  57583.   uint8x8_t __ret;
  57584.   __ret = (uint8x8_t)(__p0);
  57585.   return __ret;
  57586. }
  57587. __ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
  57588.   uint8x8_t __ret;
  57589.   __ret = (uint8x8_t)(__p0);
  57590.   return __ret;
  57591. }
  57592. __ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
  57593.   uint8x8_t __ret;
  57594.   __ret = (uint8x8_t)(__p0);
  57595.   return __ret;
  57596. }
  57597. __ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
  57598.   uint8x8_t __ret;
  57599.   __ret = (uint8x8_t)(__p0);
  57600.   return __ret;
  57601. }
  57602. __ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
  57603.   uint8x8_t __ret;
  57604.   __ret = (uint8x8_t)(__p0);
  57605.   return __ret;
  57606. }
  57607. __ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
  57608.   uint32x2_t __ret;
  57609.   __ret = (uint32x2_t)(__p0);
  57610.   return __ret;
  57611. }
  57612. __ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __p0) {
  57613.   uint32x2_t __ret;
  57614.   __ret = (uint32x2_t)(__p0);
  57615.   return __ret;
  57616. }
  57617. __ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
  57618.   uint32x2_t __ret;
  57619.   __ret = (uint32x2_t)(__p0);
  57620.   return __ret;
  57621. }
  57622. __ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
  57623.   uint32x2_t __ret;
  57624.   __ret = (uint32x2_t)(__p0);
  57625.   return __ret;
  57626. }
  57627. __ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
  57628.   uint32x2_t __ret;
  57629.   __ret = (uint32x2_t)(__p0);
  57630.   return __ret;
  57631. }
  57632. __ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
  57633.   uint32x2_t __ret;
  57634.   __ret = (uint32x2_t)(__p0);
  57635.   return __ret;
  57636. }
  57637. __ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
  57638.   uint32x2_t __ret;
  57639.   __ret = (uint32x2_t)(__p0);
  57640.   return __ret;
  57641. }
  57642. __ai uint32x2_t vreinterpret_u32_f64(float64x1_t __p0) {
  57643.   uint32x2_t __ret;
  57644.   __ret = (uint32x2_t)(__p0);
  57645.   return __ret;
  57646. }
  57647. __ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
  57648.   uint32x2_t __ret;
  57649.   __ret = (uint32x2_t)(__p0);
  57650.   return __ret;
  57651. }
  57652. __ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
  57653.   uint32x2_t __ret;
  57654.   __ret = (uint32x2_t)(__p0);
  57655.   return __ret;
  57656. }
  57657. __ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
  57658.   uint32x2_t __ret;
  57659.   __ret = (uint32x2_t)(__p0);
  57660.   return __ret;
  57661. }
  57662. __ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
  57663.   uint32x2_t __ret;
  57664.   __ret = (uint32x2_t)(__p0);
  57665.   return __ret;
  57666. }
  57667. __ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
  57668.   uint32x2_t __ret;
  57669.   __ret = (uint32x2_t)(__p0);
  57670.   return __ret;
  57671. }
  57672. __ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
  57673.   uint64x1_t __ret;
  57674.   __ret = (uint64x1_t)(__p0);
  57675.   return __ret;
  57676. }
  57677. __ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __p0) {
  57678.   uint64x1_t __ret;
  57679.   __ret = (uint64x1_t)(__p0);
  57680.   return __ret;
  57681. }
  57682. __ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
  57683.   uint64x1_t __ret;
  57684.   __ret = (uint64x1_t)(__p0);
  57685.   return __ret;
  57686. }
  57687. __ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
  57688.   uint64x1_t __ret;
  57689.   __ret = (uint64x1_t)(__p0);
  57690.   return __ret;
  57691. }
  57692. __ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
  57693.   uint64x1_t __ret;
  57694.   __ret = (uint64x1_t)(__p0);
  57695.   return __ret;
  57696. }
  57697. __ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
  57698.   uint64x1_t __ret;
  57699.   __ret = (uint64x1_t)(__p0);
  57700.   return __ret;
  57701. }
  57702. __ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
  57703.   uint64x1_t __ret;
  57704.   __ret = (uint64x1_t)(__p0);
  57705.   return __ret;
  57706. }
  57707. __ai uint64x1_t vreinterpret_u64_f64(float64x1_t __p0) {
  57708.   uint64x1_t __ret;
  57709.   __ret = (uint64x1_t)(__p0);
  57710.   return __ret;
  57711. }
  57712. __ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
  57713.   uint64x1_t __ret;
  57714.   __ret = (uint64x1_t)(__p0);
  57715.   return __ret;
  57716. }
  57717. __ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
  57718.   uint64x1_t __ret;
  57719.   __ret = (uint64x1_t)(__p0);
  57720.   return __ret;
  57721. }
  57722. __ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
  57723.   uint64x1_t __ret;
  57724.   __ret = (uint64x1_t)(__p0);
  57725.   return __ret;
  57726. }
  57727. __ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
  57728.   uint64x1_t __ret;
  57729.   __ret = (uint64x1_t)(__p0);
  57730.   return __ret;
  57731. }
  57732. __ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
  57733.   uint64x1_t __ret;
  57734.   __ret = (uint64x1_t)(__p0);
  57735.   return __ret;
  57736. }
  57737. __ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
  57738.   uint16x4_t __ret;
  57739.   __ret = (uint16x4_t)(__p0);
  57740.   return __ret;
  57741. }
  57742. __ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __p0) {
  57743.   uint16x4_t __ret;
  57744.   __ret = (uint16x4_t)(__p0);
  57745.   return __ret;
  57746. }
  57747. __ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
  57748.   uint16x4_t __ret;
  57749.   __ret = (uint16x4_t)(__p0);
  57750.   return __ret;
  57751. }
  57752. __ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
  57753.   uint16x4_t __ret;
  57754.   __ret = (uint16x4_t)(__p0);
  57755.   return __ret;
  57756. }
  57757. __ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
  57758.   uint16x4_t __ret;
  57759.   __ret = (uint16x4_t)(__p0);
  57760.   return __ret;
  57761. }
  57762. __ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
  57763.   uint16x4_t __ret;
  57764.   __ret = (uint16x4_t)(__p0);
  57765.   return __ret;
  57766. }
  57767. __ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
  57768.   uint16x4_t __ret;
  57769.   __ret = (uint16x4_t)(__p0);
  57770.   return __ret;
  57771. }
  57772. __ai uint16x4_t vreinterpret_u16_f64(float64x1_t __p0) {
  57773.   uint16x4_t __ret;
  57774.   __ret = (uint16x4_t)(__p0);
  57775.   return __ret;
  57776. }
  57777. __ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
  57778.   uint16x4_t __ret;
  57779.   __ret = (uint16x4_t)(__p0);
  57780.   return __ret;
  57781. }
  57782. __ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
  57783.   uint16x4_t __ret;
  57784.   __ret = (uint16x4_t)(__p0);
  57785.   return __ret;
  57786. }
  57787. __ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
  57788.   uint16x4_t __ret;
  57789.   __ret = (uint16x4_t)(__p0);
  57790.   return __ret;
  57791. }
  57792. __ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
  57793.   uint16x4_t __ret;
  57794.   __ret = (uint16x4_t)(__p0);
  57795.   return __ret;
  57796. }
  57797. __ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
  57798.   uint16x4_t __ret;
  57799.   __ret = (uint16x4_t)(__p0);
  57800.   return __ret;
  57801. }
  57802. __ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
  57803.   int8x8_t __ret;
  57804.   __ret = (int8x8_t)(__p0);
  57805.   return __ret;
  57806. }
  57807. __ai int8x8_t vreinterpret_s8_p64(poly64x1_t __p0) {
  57808.   int8x8_t __ret;
  57809.   __ret = (int8x8_t)(__p0);
  57810.   return __ret;
  57811. }
  57812. __ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
  57813.   int8x8_t __ret;
  57814.   __ret = (int8x8_t)(__p0);
  57815.   return __ret;
  57816. }
  57817. __ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
  57818.   int8x8_t __ret;
  57819.   __ret = (int8x8_t)(__p0);
  57820.   return __ret;
  57821. }
  57822. __ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
  57823.   int8x8_t __ret;
  57824.   __ret = (int8x8_t)(__p0);
  57825.   return __ret;
  57826. }
  57827. __ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
  57828.   int8x8_t __ret;
  57829.   __ret = (int8x8_t)(__p0);
  57830.   return __ret;
  57831. }
  57832. __ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
  57833.   int8x8_t __ret;
  57834.   __ret = (int8x8_t)(__p0);
  57835.   return __ret;
  57836. }
  57837. __ai int8x8_t vreinterpret_s8_f64(float64x1_t __p0) {
  57838.   int8x8_t __ret;
  57839.   __ret = (int8x8_t)(__p0);
  57840.   return __ret;
  57841. }
  57842. __ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
  57843.   int8x8_t __ret;
  57844.   __ret = (int8x8_t)(__p0);
  57845.   return __ret;
  57846. }
  57847. __ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
  57848.   int8x8_t __ret;
  57849.   __ret = (int8x8_t)(__p0);
  57850.   return __ret;
  57851. }
  57852. __ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
  57853.   int8x8_t __ret;
  57854.   __ret = (int8x8_t)(__p0);
  57855.   return __ret;
  57856. }
  57857. __ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
  57858.   int8x8_t __ret;
  57859.   __ret = (int8x8_t)(__p0);
  57860.   return __ret;
  57861. }
  57862. __ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
  57863.   int8x8_t __ret;
  57864.   __ret = (int8x8_t)(__p0);
  57865.   return __ret;
  57866. }
  57867. __ai float64x1_t vreinterpret_f64_p8(poly8x8_t __p0) {
  57868.   float64x1_t __ret;
  57869.   __ret = (float64x1_t)(__p0);
  57870.   return __ret;
  57871. }
  57872. __ai float64x1_t vreinterpret_f64_p64(poly64x1_t __p0) {
  57873.   float64x1_t __ret;
  57874.   __ret = (float64x1_t)(__p0);
  57875.   return __ret;
  57876. }
  57877. __ai float64x1_t vreinterpret_f64_p16(poly16x4_t __p0) {
  57878.   float64x1_t __ret;
  57879.   __ret = (float64x1_t)(__p0);
  57880.   return __ret;
  57881. }
  57882. __ai float64x1_t vreinterpret_f64_u8(uint8x8_t __p0) {
  57883.   float64x1_t __ret;
  57884.   __ret = (float64x1_t)(__p0);
  57885.   return __ret;
  57886. }
  57887. __ai float64x1_t vreinterpret_f64_u32(uint32x2_t __p0) {
  57888.   float64x1_t __ret;
  57889.   __ret = (float64x1_t)(__p0);
  57890.   return __ret;
  57891. }
  57892. __ai float64x1_t vreinterpret_f64_u64(uint64x1_t __p0) {
  57893.   float64x1_t __ret;
  57894.   __ret = (float64x1_t)(__p0);
  57895.   return __ret;
  57896. }
  57897. __ai float64x1_t vreinterpret_f64_u16(uint16x4_t __p0) {
  57898.   float64x1_t __ret;
  57899.   __ret = (float64x1_t)(__p0);
  57900.   return __ret;
  57901. }
  57902. __ai float64x1_t vreinterpret_f64_s8(int8x8_t __p0) {
  57903.   float64x1_t __ret;
  57904.   __ret = (float64x1_t)(__p0);
  57905.   return __ret;
  57906. }
  57907. __ai float64x1_t vreinterpret_f64_f32(float32x2_t __p0) {
  57908.   float64x1_t __ret;
  57909.   __ret = (float64x1_t)(__p0);
  57910.   return __ret;
  57911. }
  57912. __ai float64x1_t vreinterpret_f64_f16(float16x4_t __p0) {
  57913.   float64x1_t __ret;
  57914.   __ret = (float64x1_t)(__p0);
  57915.   return __ret;
  57916. }
  57917. __ai float64x1_t vreinterpret_f64_s32(int32x2_t __p0) {
  57918.   float64x1_t __ret;
  57919.   __ret = (float64x1_t)(__p0);
  57920.   return __ret;
  57921. }
  57922. __ai float64x1_t vreinterpret_f64_s64(int64x1_t __p0) {
  57923.   float64x1_t __ret;
  57924.   __ret = (float64x1_t)(__p0);
  57925.   return __ret;
  57926. }
  57927. __ai float64x1_t vreinterpret_f64_s16(int16x4_t __p0) {
  57928.   float64x1_t __ret;
  57929.   __ret = (float64x1_t)(__p0);
  57930.   return __ret;
  57931. }
  57932. __ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
  57933.   float32x2_t __ret;
  57934.   __ret = (float32x2_t)(__p0);
  57935.   return __ret;
  57936. }
  57937. __ai float32x2_t vreinterpret_f32_p64(poly64x1_t __p0) {
  57938.   float32x2_t __ret;
  57939.   __ret = (float32x2_t)(__p0);
  57940.   return __ret;
  57941. }
  57942. __ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
  57943.   float32x2_t __ret;
  57944.   __ret = (float32x2_t)(__p0);
  57945.   return __ret;
  57946. }
  57947. __ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
  57948.   float32x2_t __ret;
  57949.   __ret = (float32x2_t)(__p0);
  57950.   return __ret;
  57951. }
  57952. __ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
  57953.   float32x2_t __ret;
  57954.   __ret = (float32x2_t)(__p0);
  57955.   return __ret;
  57956. }
  57957. __ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
  57958.   float32x2_t __ret;
  57959.   __ret = (float32x2_t)(__p0);
  57960.   return __ret;
  57961. }
  57962. __ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
  57963.   float32x2_t __ret;
  57964.   __ret = (float32x2_t)(__p0);
  57965.   return __ret;
  57966. }
  57967. __ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
  57968.   float32x2_t __ret;
  57969.   __ret = (float32x2_t)(__p0);
  57970.   return __ret;
  57971. }
  57972. __ai float32x2_t vreinterpret_f32_f64(float64x1_t __p0) {
  57973.   float32x2_t __ret;
  57974.   __ret = (float32x2_t)(__p0);
  57975.   return __ret;
  57976. }
  57977. __ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
  57978.   float32x2_t __ret;
  57979.   __ret = (float32x2_t)(__p0);
  57980.   return __ret;
  57981. }
  57982. __ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
  57983.   float32x2_t __ret;
  57984.   __ret = (float32x2_t)(__p0);
  57985.   return __ret;
  57986. }
  57987. __ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
  57988.   float32x2_t __ret;
  57989.   __ret = (float32x2_t)(__p0);
  57990.   return __ret;
  57991. }
  57992. __ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
  57993.   float32x2_t __ret;
  57994.   __ret = (float32x2_t)(__p0);
  57995.   return __ret;
  57996. }
  57997. __ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
  57998.   float16x4_t __ret;
  57999.   __ret = (float16x4_t)(__p0);
  58000.   return __ret;
  58001. }
  58002. __ai float16x4_t vreinterpret_f16_p64(poly64x1_t __p0) {
  58003.   float16x4_t __ret;
  58004.   __ret = (float16x4_t)(__p0);
  58005.   return __ret;
  58006. }
  58007. __ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
  58008.   float16x4_t __ret;
  58009.   __ret = (float16x4_t)(__p0);
  58010.   return __ret;
  58011. }
  58012. __ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
  58013.   float16x4_t __ret;
  58014.   __ret = (float16x4_t)(__p0);
  58015.   return __ret;
  58016. }
  58017. __ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
  58018.   float16x4_t __ret;
  58019.   __ret = (float16x4_t)(__p0);
  58020.   return __ret;
  58021. }
  58022. __ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
  58023.   float16x4_t __ret;
  58024.   __ret = (float16x4_t)(__p0);
  58025.   return __ret;
  58026. }
  58027. __ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
  58028.   float16x4_t __ret;
  58029.   __ret = (float16x4_t)(__p0);
  58030.   return __ret;
  58031. }
  58032. __ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
  58033.   float16x4_t __ret;
  58034.   __ret = (float16x4_t)(__p0);
  58035.   return __ret;
  58036. }
  58037. __ai float16x4_t vreinterpret_f16_f64(float64x1_t __p0) {
  58038.   float16x4_t __ret;
  58039.   __ret = (float16x4_t)(__p0);
  58040.   return __ret;
  58041. }
  58042. __ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
  58043.   float16x4_t __ret;
  58044.   __ret = (float16x4_t)(__p0);
  58045.   return __ret;
  58046. }
  58047. __ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
  58048.   float16x4_t __ret;
  58049.   __ret = (float16x4_t)(__p0);
  58050.   return __ret;
  58051. }
  58052. __ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
  58053.   float16x4_t __ret;
  58054.   __ret = (float16x4_t)(__p0);
  58055.   return __ret;
  58056. }
  58057. __ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
  58058.   float16x4_t __ret;
  58059.   __ret = (float16x4_t)(__p0);
  58060.   return __ret;
  58061. }
  58062. __ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
  58063.   int32x2_t __ret;
  58064.   __ret = (int32x2_t)(__p0);
  58065.   return __ret;
  58066. }
  58067. __ai int32x2_t vreinterpret_s32_p64(poly64x1_t __p0) {
  58068.   int32x2_t __ret;
  58069.   __ret = (int32x2_t)(__p0);
  58070.   return __ret;
  58071. }
  58072. __ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
  58073.   int32x2_t __ret;
  58074.   __ret = (int32x2_t)(__p0);
  58075.   return __ret;
  58076. }
  58077. __ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
  58078.   int32x2_t __ret;
  58079.   __ret = (int32x2_t)(__p0);
  58080.   return __ret;
  58081. }
  58082. __ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
  58083.   int32x2_t __ret;
  58084.   __ret = (int32x2_t)(__p0);
  58085.   return __ret;
  58086. }
  58087. __ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
  58088.   int32x2_t __ret;
  58089.   __ret = (int32x2_t)(__p0);
  58090.   return __ret;
  58091. }
  58092. __ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
  58093.   int32x2_t __ret;
  58094.   __ret = (int32x2_t)(__p0);
  58095.   return __ret;
  58096. }
  58097. __ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
  58098.   int32x2_t __ret;
  58099.   __ret = (int32x2_t)(__p0);
  58100.   return __ret;
  58101. }
  58102. __ai int32x2_t vreinterpret_s32_f64(float64x1_t __p0) {
  58103.   int32x2_t __ret;
  58104.   __ret = (int32x2_t)(__p0);
  58105.   return __ret;
  58106. }
  58107. __ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
  58108.   int32x2_t __ret;
  58109.   __ret = (int32x2_t)(__p0);
  58110.   return __ret;
  58111. }
  58112. __ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
  58113.   int32x2_t __ret;
  58114.   __ret = (int32x2_t)(__p0);
  58115.   return __ret;
  58116. }
  58117. __ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
  58118.   int32x2_t __ret;
  58119.   __ret = (int32x2_t)(__p0);
  58120.   return __ret;
  58121. }
  58122. __ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
  58123.   int32x2_t __ret;
  58124.   __ret = (int32x2_t)(__p0);
  58125.   return __ret;
  58126. }
  58127. __ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
  58128.   int64x1_t __ret;
  58129.   __ret = (int64x1_t)(__p0);
  58130.   return __ret;
  58131. }
  58132. __ai int64x1_t vreinterpret_s64_p64(poly64x1_t __p0) {
  58133.   int64x1_t __ret;
  58134.   __ret = (int64x1_t)(__p0);
  58135.   return __ret;
  58136. }
  58137. __ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
  58138.   int64x1_t __ret;
  58139.   __ret = (int64x1_t)(__p0);
  58140.   return __ret;
  58141. }
  58142. __ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
  58143.   int64x1_t __ret;
  58144.   __ret = (int64x1_t)(__p0);
  58145.   return __ret;
  58146. }
  58147. __ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
  58148.   int64x1_t __ret;
  58149.   __ret = (int64x1_t)(__p0);
  58150.   return __ret;
  58151. }
  58152. __ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
  58153.   int64x1_t __ret;
  58154.   __ret = (int64x1_t)(__p0);
  58155.   return __ret;
  58156. }
  58157. __ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
  58158.   int64x1_t __ret;
  58159.   __ret = (int64x1_t)(__p0);
  58160.   return __ret;
  58161. }
  58162. __ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
  58163.   int64x1_t __ret;
  58164.   __ret = (int64x1_t)(__p0);
  58165.   return __ret;
  58166. }
  58167. __ai int64x1_t vreinterpret_s64_f64(float64x1_t __p0) {
  58168.   int64x1_t __ret;
  58169.   __ret = (int64x1_t)(__p0);
  58170.   return __ret;
  58171. }
  58172. __ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
  58173.   int64x1_t __ret;
  58174.   __ret = (int64x1_t)(__p0);
  58175.   return __ret;
  58176. }
  58177. __ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
  58178.   int64x1_t __ret;
  58179.   __ret = (int64x1_t)(__p0);
  58180.   return __ret;
  58181. }
  58182. __ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
  58183.   int64x1_t __ret;
  58184.   __ret = (int64x1_t)(__p0);
  58185.   return __ret;
  58186. }
  58187. __ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
  58188.   int64x1_t __ret;
  58189.   __ret = (int64x1_t)(__p0);
  58190.   return __ret;
  58191. }
  58192. __ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
  58193.   int16x4_t __ret;
  58194.   __ret = (int16x4_t)(__p0);
  58195.   return __ret;
  58196. }
  58197. __ai int16x4_t vreinterpret_s16_p64(poly64x1_t __p0) {
  58198.   int16x4_t __ret;
  58199.   __ret = (int16x4_t)(__p0);
  58200.   return __ret;
  58201. }
  58202. __ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
  58203.   int16x4_t __ret;
  58204.   __ret = (int16x4_t)(__p0);
  58205.   return __ret;
  58206. }
  58207. __ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
  58208.   int16x4_t __ret;
  58209.   __ret = (int16x4_t)(__p0);
  58210.   return __ret;
  58211. }
  58212. __ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
  58213.   int16x4_t __ret;
  58214.   __ret = (int16x4_t)(__p0);
  58215.   return __ret;
  58216. }
  58217. __ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
  58218.   int16x4_t __ret;
  58219.   __ret = (int16x4_t)(__p0);
  58220.   return __ret;
  58221. }
  58222. __ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
  58223.   int16x4_t __ret;
  58224.   __ret = (int16x4_t)(__p0);
  58225.   return __ret;
  58226. }
  58227. __ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
  58228.   int16x4_t __ret;
  58229.   __ret = (int16x4_t)(__p0);
  58230.   return __ret;
  58231. }
  58232. __ai int16x4_t vreinterpret_s16_f64(float64x1_t __p0) {
  58233.   int16x4_t __ret;
  58234.   __ret = (int16x4_t)(__p0);
  58235.   return __ret;
  58236. }
  58237. __ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
  58238.   int16x4_t __ret;
  58239.   __ret = (int16x4_t)(__p0);
  58240.   return __ret;
  58241. }
  58242. __ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
  58243.   int16x4_t __ret;
  58244.   __ret = (int16x4_t)(__p0);
  58245.   return __ret;
  58246. }
  58247. __ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
  58248.   int16x4_t __ret;
  58249.   __ret = (int16x4_t)(__p0);
  58250.   return __ret;
  58251. }
  58252. __ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
  58253.   int16x4_t __ret;
  58254.   __ret = (int16x4_t)(__p0);
  58255.   return __ret;
  58256. }
  58257. __ai uint64_t vrshld_u64(uint64_t __p0, int64_t __p1) {
  58258.   uint64_t __ret;
  58259.   __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1);
  58260.   return __ret;
  58261. }
  58262. __ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) {
  58263.   int64_t __ret;
  58264.   __ret = (int64_t) __builtin_neon_vrshld_s64(__p0, __p1);
  58265.   return __ret;
  58266. }
  58267. #define vrshrd_n_u64(__p0, __p1) __extension__ ({ \
  58268.   uint64_t __ret; \
  58269.   uint64_t __s0 = __p0; \
  58270.   __ret = (uint64_t) __builtin_neon_vrshrd_n_u64(__s0, __p1); \
  58271.   __ret; \
  58272. })
  58273. #define vrshrd_n_s64(__p0, __p1) __extension__ ({ \
  58274.   int64_t __ret; \
  58275.   int64_t __s0 = __p0; \
  58276.   __ret = (int64_t) __builtin_neon_vrshrd_n_s64(__s0, __p1); \
  58277.   __ret; \
  58278. })
  58279. #ifdef __LITTLE_ENDIAN__
  58280. #define vrshrn_high_n_u32(__p0_703, __p1_703, __p2_703) __extension__ ({ \
  58281.   uint16x8_t __ret_703; \
  58282.   uint16x4_t __s0_703 = __p0_703; \
  58283.   uint32x4_t __s1_703 = __p1_703; \
  58284.   __ret_703 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_703), (uint16x4_t)(vrshrn_n_u32(__s1_703, __p2_703)))); \
  58285.   __ret_703; \
  58286. })
  58287. #else
  58288. #define vrshrn_high_n_u32(__p0_704, __p1_704, __p2_704) __extension__ ({ \
  58289.   uint16x8_t __ret_704; \
  58290.   uint16x4_t __s0_704 = __p0_704; \
  58291.   uint32x4_t __s1_704 = __p1_704; \
  58292.   uint16x4_t __rev0_704;  __rev0_704 = __builtin_shufflevector(__s0_704, __s0_704, 3, 2, 1, 0); \
  58293.   uint32x4_t __rev1_704;  __rev1_704 = __builtin_shufflevector(__s1_704, __s1_704, 3, 2, 1, 0); \
  58294.   __ret_704 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_704), (uint16x4_t)(__noswap_vrshrn_n_u32(__rev1_704, __p2_704)))); \
  58295.   __ret_704 = __builtin_shufflevector(__ret_704, __ret_704, 7, 6, 5, 4, 3, 2, 1, 0); \
  58296.   __ret_704; \
  58297. })
  58298. #endif
  58299.  
  58300. #ifdef __LITTLE_ENDIAN__
  58301. #define vrshrn_high_n_u64(__p0_705, __p1_705, __p2_705) __extension__ ({ \
  58302.   uint32x4_t __ret_705; \
  58303.   uint32x2_t __s0_705 = __p0_705; \
  58304.   uint64x2_t __s1_705 = __p1_705; \
  58305.   __ret_705 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_705), (uint32x2_t)(vrshrn_n_u64(__s1_705, __p2_705)))); \
  58306.   __ret_705; \
  58307. })
  58308. #else
  58309. #define vrshrn_high_n_u64(__p0_706, __p1_706, __p2_706) __extension__ ({ \
  58310.   uint32x4_t __ret_706; \
  58311.   uint32x2_t __s0_706 = __p0_706; \
  58312.   uint64x2_t __s1_706 = __p1_706; \
  58313.   uint32x2_t __rev0_706;  __rev0_706 = __builtin_shufflevector(__s0_706, __s0_706, 1, 0); \
  58314.   uint64x2_t __rev1_706;  __rev1_706 = __builtin_shufflevector(__s1_706, __s1_706, 1, 0); \
  58315.   __ret_706 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_706), (uint32x2_t)(__noswap_vrshrn_n_u64(__rev1_706, __p2_706)))); \
  58316.   __ret_706 = __builtin_shufflevector(__ret_706, __ret_706, 3, 2, 1, 0); \
  58317.   __ret_706; \
  58318. })
  58319. #endif
  58320.  
  58321. #ifdef __LITTLE_ENDIAN__
  58322. #define vrshrn_high_n_u16(__p0_707, __p1_707, __p2_707) __extension__ ({ \
  58323.   uint8x16_t __ret_707; \
  58324.   uint8x8_t __s0_707 = __p0_707; \
  58325.   uint16x8_t __s1_707 = __p1_707; \
  58326.   __ret_707 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_707), (uint8x8_t)(vrshrn_n_u16(__s1_707, __p2_707)))); \
  58327.   __ret_707; \
  58328. })
  58329. #else
  58330. #define vrshrn_high_n_u16(__p0_708, __p1_708, __p2_708) __extension__ ({ \
  58331.   uint8x16_t __ret_708; \
  58332.   uint8x8_t __s0_708 = __p0_708; \
  58333.   uint16x8_t __s1_708 = __p1_708; \
  58334.   uint8x8_t __rev0_708;  __rev0_708 = __builtin_shufflevector(__s0_708, __s0_708, 7, 6, 5, 4, 3, 2, 1, 0); \
  58335.   uint16x8_t __rev1_708;  __rev1_708 = __builtin_shufflevector(__s1_708, __s1_708, 7, 6, 5, 4, 3, 2, 1, 0); \
  58336.   __ret_708 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_708), (uint8x8_t)(__noswap_vrshrn_n_u16(__rev1_708, __p2_708)))); \
  58337.   __ret_708 = __builtin_shufflevector(__ret_708, __ret_708, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  58338.   __ret_708; \
  58339. })
  58340. #endif
  58341.  
  58342. #ifdef __LITTLE_ENDIAN__
  58343. #define vrshrn_high_n_s32(__p0_709, __p1_709, __p2_709) __extension__ ({ \
  58344.   int16x8_t __ret_709; \
  58345.   int16x4_t __s0_709 = __p0_709; \
  58346.   int32x4_t __s1_709 = __p1_709; \
  58347.   __ret_709 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_709), (int16x4_t)(vrshrn_n_s32(__s1_709, __p2_709)))); \
  58348.   __ret_709; \
  58349. })
  58350. #else
  58351. #define vrshrn_high_n_s32(__p0_710, __p1_710, __p2_710) __extension__ ({ \
  58352.   int16x8_t __ret_710; \
  58353.   int16x4_t __s0_710 = __p0_710; \
  58354.   int32x4_t __s1_710 = __p1_710; \
  58355.   int16x4_t __rev0_710;  __rev0_710 = __builtin_shufflevector(__s0_710, __s0_710, 3, 2, 1, 0); \
  58356.   int32x4_t __rev1_710;  __rev1_710 = __builtin_shufflevector(__s1_710, __s1_710, 3, 2, 1, 0); \
  58357.   __ret_710 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_710), (int16x4_t)(__noswap_vrshrn_n_s32(__rev1_710, __p2_710)))); \
  58358.   __ret_710 = __builtin_shufflevector(__ret_710, __ret_710, 7, 6, 5, 4, 3, 2, 1, 0); \
  58359.   __ret_710; \
  58360. })
  58361. #endif
  58362.  
  58363. #ifdef __LITTLE_ENDIAN__
  58364. #define vrshrn_high_n_s64(__p0_711, __p1_711, __p2_711) __extension__ ({ \
  58365.   int32x4_t __ret_711; \
  58366.   int32x2_t __s0_711 = __p0_711; \
  58367.   int64x2_t __s1_711 = __p1_711; \
  58368.   __ret_711 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_711), (int32x2_t)(vrshrn_n_s64(__s1_711, __p2_711)))); \
  58369.   __ret_711; \
  58370. })
  58371. #else
  58372. #define vrshrn_high_n_s64(__p0_712, __p1_712, __p2_712) __extension__ ({ \
  58373.   int32x4_t __ret_712; \
  58374.   int32x2_t __s0_712 = __p0_712; \
  58375.   int64x2_t __s1_712 = __p1_712; \
  58376.   int32x2_t __rev0_712;  __rev0_712 = __builtin_shufflevector(__s0_712, __s0_712, 1, 0); \
  58377.   int64x2_t __rev1_712;  __rev1_712 = __builtin_shufflevector(__s1_712, __s1_712, 1, 0); \
  58378.   __ret_712 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_712), (int32x2_t)(__noswap_vrshrn_n_s64(__rev1_712, __p2_712)))); \
  58379.   __ret_712 = __builtin_shufflevector(__ret_712, __ret_712, 3, 2, 1, 0); \
  58380.   __ret_712; \
  58381. })
  58382. #endif
  58383.  
  58384. #ifdef __LITTLE_ENDIAN__
  58385. #define vrshrn_high_n_s16(__p0_713, __p1_713, __p2_713) __extension__ ({ \
  58386.   int8x16_t __ret_713; \
  58387.   int8x8_t __s0_713 = __p0_713; \
  58388.   int16x8_t __s1_713 = __p1_713; \
  58389.   __ret_713 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_713), (int8x8_t)(vrshrn_n_s16(__s1_713, __p2_713)))); \
  58390.   __ret_713; \
  58391. })
  58392. #else
  58393. #define vrshrn_high_n_s16(__p0_714, __p1_714, __p2_714) __extension__ ({ \
  58394.   int8x16_t __ret_714; \
  58395.   int8x8_t __s0_714 = __p0_714; \
  58396.   int16x8_t __s1_714 = __p1_714; \
  58397.   int8x8_t __rev0_714;  __rev0_714 = __builtin_shufflevector(__s0_714, __s0_714, 7, 6, 5, 4, 3, 2, 1, 0); \
  58398.   int16x8_t __rev1_714;  __rev1_714 = __builtin_shufflevector(__s1_714, __s1_714, 7, 6, 5, 4, 3, 2, 1, 0); \
  58399.   __ret_714 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_714), (int8x8_t)(__noswap_vrshrn_n_s16(__rev1_714, __p2_714)))); \
  58400.   __ret_714 = __builtin_shufflevector(__ret_714, __ret_714, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  58401.   __ret_714; \
  58402. })
  58403. #endif
  58404.  
  58405. #ifdef __LITTLE_ENDIAN__
  58406. __ai float64x2_t vrsqrteq_f64(float64x2_t __p0) {
  58407.   float64x2_t __ret;
  58408.   __ret = (float64x2_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 42);
  58409.   return __ret;
  58410. }
  58411. #else
  58412. __ai float64x2_t vrsqrteq_f64(float64x2_t __p0) {
  58413.   float64x2_t __ret;
  58414.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  58415.   __ret = (float64x2_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 42);
  58416.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  58417.   return __ret;
  58418. }
  58419. #endif
  58420.  
  58421. __ai float64x1_t vrsqrte_f64(float64x1_t __p0) {
  58422.   float64x1_t __ret;
  58423.   __ret = (float64x1_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 10);
  58424.   return __ret;
  58425. }
  58426. __ai float64_t vrsqrted_f64(float64_t __p0) {
  58427.   float64_t __ret;
  58428.   __ret = (float64_t) __builtin_neon_vrsqrted_f64(__p0);
  58429.   return __ret;
  58430. }
  58431. __ai float32_t vrsqrtes_f32(float32_t __p0) {
  58432.   float32_t __ret;
  58433.   __ret = (float32_t) __builtin_neon_vrsqrtes_f32(__p0);
  58434.   return __ret;
  58435. }
  58436. #ifdef __LITTLE_ENDIAN__
  58437. __ai float64x2_t vrsqrtsq_f64(float64x2_t __p0, float64x2_t __p1) {
  58438.   float64x2_t __ret;
  58439.   __ret = (float64x2_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  58440.   return __ret;
  58441. }
  58442. #else
  58443. __ai float64x2_t vrsqrtsq_f64(float64x2_t __p0, float64x2_t __p1) {
  58444.   float64x2_t __ret;
  58445.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  58446.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  58447.   __ret = (float64x2_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  58448.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  58449.   return __ret;
  58450. }
  58451. #endif
  58452.  
  58453. __ai float64x1_t vrsqrts_f64(float64x1_t __p0, float64x1_t __p1) {
  58454.   float64x1_t __ret;
  58455.   __ret = (float64x1_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
  58456.   return __ret;
  58457. }
  58458. __ai float64_t vrsqrtsd_f64(float64_t __p0, float64_t __p1) {
  58459.   float64_t __ret;
  58460.   __ret = (float64_t) __builtin_neon_vrsqrtsd_f64(__p0, __p1);
  58461.   return __ret;
  58462. }
  58463. __ai float32_t vrsqrtss_f32(float32_t __p0, float32_t __p1) {
  58464.   float32_t __ret;
  58465.   __ret = (float32_t) __builtin_neon_vrsqrtss_f32(__p0, __p1);
  58466.   return __ret;
  58467. }
  58468. #define vrsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
  58469.   uint64_t __ret; \
  58470.   uint64_t __s0 = __p0; \
  58471.   uint64_t __s1 = __p1; \
  58472.   __ret = (uint64_t) __builtin_neon_vrsrad_n_u64(__s0, __s1, __p2); \
  58473.   __ret; \
  58474. })
  58475. #define vrsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
  58476.   int64_t __ret; \
  58477.   int64_t __s0 = __p0; \
  58478.   int64_t __s1 = __p1; \
  58479.   __ret = (int64_t) __builtin_neon_vrsrad_n_s64(__s0, __s1, __p2); \
  58480.   __ret; \
  58481. })
  58482. #ifdef __LITTLE_ENDIAN__
  58483. __ai uint16x8_t vrsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  58484.   uint16x8_t __ret;
  58485.   __ret = vcombine_u16(__p0, vrsubhn_u32(__p1, __p2));
  58486.   return __ret;
  58487. }
  58488. #else
  58489. __ai uint16x8_t vrsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  58490.   uint16x8_t __ret;
  58491.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  58492.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  58493.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  58494.   __ret = __noswap_vcombine_u16(__rev0, __noswap_vrsubhn_u32(__rev1, __rev2));
  58495.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  58496.   return __ret;
  58497. }
  58498. #endif
  58499.  
  58500. #ifdef __LITTLE_ENDIAN__
  58501. __ai uint32x4_t vrsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  58502.   uint32x4_t __ret;
  58503.   __ret = vcombine_u32(__p0, vrsubhn_u64(__p1, __p2));
  58504.   return __ret;
  58505. }
  58506. #else
  58507. __ai uint32x4_t vrsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  58508.   uint32x4_t __ret;
  58509.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  58510.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  58511.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  58512.   __ret = __noswap_vcombine_u32(__rev0, __noswap_vrsubhn_u64(__rev1, __rev2));
  58513.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  58514.   return __ret;
  58515. }
  58516. #endif
  58517.  
  58518. #ifdef __LITTLE_ENDIAN__
  58519. __ai uint8x16_t vrsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  58520.   uint8x16_t __ret;
  58521.   __ret = vcombine_u8(__p0, vrsubhn_u16(__p1, __p2));
  58522.   return __ret;
  58523. }
  58524. #else
  58525. __ai uint8x16_t vrsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  58526.   uint8x16_t __ret;
  58527.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  58528.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  58529.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  58530.   __ret = __noswap_vcombine_u8(__rev0, __noswap_vrsubhn_u16(__rev1, __rev2));
  58531.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  58532.   return __ret;
  58533. }
  58534. #endif
  58535.  
  58536. #ifdef __LITTLE_ENDIAN__
  58537. __ai int16x8_t vrsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  58538.   int16x8_t __ret;
  58539.   __ret = vcombine_s16(__p0, vrsubhn_s32(__p1, __p2));
  58540.   return __ret;
  58541. }
  58542. #else
  58543. __ai int16x8_t vrsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  58544.   int16x8_t __ret;
  58545.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  58546.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  58547.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  58548.   __ret = __noswap_vcombine_s16(__rev0, __noswap_vrsubhn_s32(__rev1, __rev2));
  58549.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  58550.   return __ret;
  58551. }
  58552. #endif
  58553.  
  58554. #ifdef __LITTLE_ENDIAN__
  58555. __ai int32x4_t vrsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  58556.   int32x4_t __ret;
  58557.   __ret = vcombine_s32(__p0, vrsubhn_s64(__p1, __p2));
  58558.   return __ret;
  58559. }
  58560. #else
  58561. __ai int32x4_t vrsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  58562.   int32x4_t __ret;
  58563.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  58564.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  58565.   int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  58566.   __ret = __noswap_vcombine_s32(__rev0, __noswap_vrsubhn_s64(__rev1, __rev2));
  58567.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  58568.   return __ret;
  58569. }
  58570. #endif
  58571.  
  58572. #ifdef __LITTLE_ENDIAN__
  58573. __ai int8x16_t vrsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  58574.   int8x16_t __ret;
  58575.   __ret = vcombine_s8(__p0, vrsubhn_s16(__p1, __p2));
  58576.   return __ret;
  58577. }
  58578. #else
  58579. __ai int8x16_t vrsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  58580.   int8x16_t __ret;
  58581.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  58582.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  58583.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  58584.   __ret = __noswap_vcombine_s8(__rev0, __noswap_vrsubhn_s16(__rev1, __rev2));
  58585.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  58586.   return __ret;
  58587. }
  58588. #endif
  58589.  
  58590. #define vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  58591.   poly64x1_t __ret; \
  58592.   poly64_t __s0 = __p0; \
  58593.   poly64x1_t __s1 = __p1; \
  58594.   __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (poly64x1_t)__s1, __p2); \
  58595.   __ret; \
  58596. })
  58597. #ifdef __LITTLE_ENDIAN__
  58598. #define vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  58599.   poly64x2_t __ret; \
  58600.   poly64_t __s0 = __p0; \
  58601.   poly64x2_t __s1 = __p1; \
  58602.   __ret = (poly64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (poly64x2_t)__s1, __p2); \
  58603.   __ret; \
  58604. })
  58605. #else
  58606. #define vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  58607.   poly64x2_t __ret; \
  58608.   poly64_t __s0 = __p0; \
  58609.   poly64x2_t __s1 = __p1; \
  58610.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  58611.   __ret = (poly64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (poly64x2_t)__rev1, __p2); \
  58612.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  58613.   __ret; \
  58614. })
  58615. #define __noswap_vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  58616.   poly64x2_t __ret; \
  58617.   poly64_t __s0 = __p0; \
  58618.   poly64x2_t __s1 = __p1; \
  58619.   __ret = (poly64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (poly64x2_t)__s1, __p2); \
  58620.   __ret; \
  58621. })
  58622. #endif
  58623.  
  58624. #ifdef __LITTLE_ENDIAN__
  58625. #define vsetq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  58626.   float64x2_t __ret; \
  58627.   float64_t __s0 = __p0; \
  58628.   float64x2_t __s1 = __p1; \
  58629.   __ret = (float64x2_t) __builtin_neon_vsetq_lane_f64(__s0, (float64x2_t)__s1, __p2); \
  58630.   __ret; \
  58631. })
  58632. #else
  58633. #define vsetq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  58634.   float64x2_t __ret; \
  58635.   float64_t __s0 = __p0; \
  58636.   float64x2_t __s1 = __p1; \
  58637.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  58638.   __ret = (float64x2_t) __builtin_neon_vsetq_lane_f64(__s0, (float64x2_t)__rev1, __p2); \
  58639.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  58640.   __ret; \
  58641. })
  58642. #define __noswap_vsetq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  58643.   float64x2_t __ret; \
  58644.   float64_t __s0 = __p0; \
  58645.   float64x2_t __s1 = __p1; \
  58646.   __ret = (float64x2_t) __builtin_neon_vsetq_lane_f64(__s0, (float64x2_t)__s1, __p2); \
  58647.   __ret; \
  58648. })
  58649. #endif
  58650.  
  58651. #define vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  58652.   float64x1_t __ret; \
  58653.   float64_t __s0 = __p0; \
  58654.   float64x1_t __s1 = __p1; \
  58655.   __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (float64x1_t)__s1, __p2); \
  58656.   __ret; \
  58657. })
  58658. __ai uint64_t vshld_u64(uint64_t __p0, int64_t __p1) {
  58659.   uint64_t __ret;
  58660.   __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1);
  58661.   return __ret;
  58662. }
  58663. __ai int64_t vshld_s64(int64_t __p0, int64_t __p1) {
  58664.   int64_t __ret;
  58665.   __ret = (int64_t) __builtin_neon_vshld_s64(__p0, __p1);
  58666.   return __ret;
  58667. }
  58668. #define vshld_n_u64(__p0, __p1) __extension__ ({ \
  58669.   uint64_t __ret; \
  58670.   uint64_t __s0 = __p0; \
  58671.   __ret = (uint64_t) __builtin_neon_vshld_n_u64(__s0, __p1); \
  58672.   __ret; \
  58673. })
  58674. #define vshld_n_s64(__p0, __p1) __extension__ ({ \
  58675.   int64_t __ret; \
  58676.   int64_t __s0 = __p0; \
  58677.   __ret = (int64_t) __builtin_neon_vshld_n_s64(__s0, __p1); \
  58678.   __ret; \
  58679. })
  58680. #ifdef __LITTLE_ENDIAN__
  58681. #define vshll_high_n_u8(__p0_715, __p1_715) __extension__ ({ \
  58682.   uint16x8_t __ret_715; \
  58683.   uint8x16_t __s0_715 = __p0_715; \
  58684.   __ret_715 = (uint16x8_t)(vshll_n_u8(vget_high_u8(__s0_715), __p1_715)); \
  58685.   __ret_715; \
  58686. })
  58687. #else
  58688. #define vshll_high_n_u8(__p0_716, __p1_716) __extension__ ({ \
  58689.   uint16x8_t __ret_716; \
  58690.   uint8x16_t __s0_716 = __p0_716; \
  58691.   uint8x16_t __rev0_716;  __rev0_716 = __builtin_shufflevector(__s0_716, __s0_716, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  58692.   __ret_716 = (uint16x8_t)(__noswap_vshll_n_u8(__noswap_vget_high_u8(__rev0_716), __p1_716)); \
  58693.   __ret_716 = __builtin_shufflevector(__ret_716, __ret_716, 7, 6, 5, 4, 3, 2, 1, 0); \
  58694.   __ret_716; \
  58695. })
  58696. #endif
  58697.  
  58698. #ifdef __LITTLE_ENDIAN__
  58699. #define vshll_high_n_u32(__p0_717, __p1_717) __extension__ ({ \
  58700.   uint64x2_t __ret_717; \
  58701.   uint32x4_t __s0_717 = __p0_717; \
  58702.   __ret_717 = (uint64x2_t)(vshll_n_u32(vget_high_u32(__s0_717), __p1_717)); \
  58703.   __ret_717; \
  58704. })
  58705. #else
  58706. #define vshll_high_n_u32(__p0_718, __p1_718) __extension__ ({ \
  58707.   uint64x2_t __ret_718; \
  58708.   uint32x4_t __s0_718 = __p0_718; \
  58709.   uint32x4_t __rev0_718;  __rev0_718 = __builtin_shufflevector(__s0_718, __s0_718, 3, 2, 1, 0); \
  58710.   __ret_718 = (uint64x2_t)(__noswap_vshll_n_u32(__noswap_vget_high_u32(__rev0_718), __p1_718)); \
  58711.   __ret_718 = __builtin_shufflevector(__ret_718, __ret_718, 1, 0); \
  58712.   __ret_718; \
  58713. })
  58714. #endif
  58715.  
  58716. #ifdef __LITTLE_ENDIAN__
  58717. #define vshll_high_n_u16(__p0_719, __p1_719) __extension__ ({ \
  58718.   uint32x4_t __ret_719; \
  58719.   uint16x8_t __s0_719 = __p0_719; \
  58720.   __ret_719 = (uint32x4_t)(vshll_n_u16(vget_high_u16(__s0_719), __p1_719)); \
  58721.   __ret_719; \
  58722. })
  58723. #else
  58724. #define vshll_high_n_u16(__p0_720, __p1_720) __extension__ ({ \
  58725.   uint32x4_t __ret_720; \
  58726.   uint16x8_t __s0_720 = __p0_720; \
  58727.   uint16x8_t __rev0_720;  __rev0_720 = __builtin_shufflevector(__s0_720, __s0_720, 7, 6, 5, 4, 3, 2, 1, 0); \
  58728.   __ret_720 = (uint32x4_t)(__noswap_vshll_n_u16(__noswap_vget_high_u16(__rev0_720), __p1_720)); \
  58729.   __ret_720 = __builtin_shufflevector(__ret_720, __ret_720, 3, 2, 1, 0); \
  58730.   __ret_720; \
  58731. })
  58732. #endif
  58733.  
  58734. #ifdef __LITTLE_ENDIAN__
  58735. #define vshll_high_n_s8(__p0_721, __p1_721) __extension__ ({ \
  58736.   int16x8_t __ret_721; \
  58737.   int8x16_t __s0_721 = __p0_721; \
  58738.   __ret_721 = (int16x8_t)(vshll_n_s8(vget_high_s8(__s0_721), __p1_721)); \
  58739.   __ret_721; \
  58740. })
  58741. #else
  58742. #define vshll_high_n_s8(__p0_722, __p1_722) __extension__ ({ \
  58743.   int16x8_t __ret_722; \
  58744.   int8x16_t __s0_722 = __p0_722; \
  58745.   int8x16_t __rev0_722;  __rev0_722 = __builtin_shufflevector(__s0_722, __s0_722, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  58746.   __ret_722 = (int16x8_t)(__noswap_vshll_n_s8(__noswap_vget_high_s8(__rev0_722), __p1_722)); \
  58747.   __ret_722 = __builtin_shufflevector(__ret_722, __ret_722, 7, 6, 5, 4, 3, 2, 1, 0); \
  58748.   __ret_722; \
  58749. })
  58750. #endif
  58751.  
  58752. #ifdef __LITTLE_ENDIAN__
  58753. #define vshll_high_n_s32(__p0_723, __p1_723) __extension__ ({ \
  58754.   int64x2_t __ret_723; \
  58755.   int32x4_t __s0_723 = __p0_723; \
  58756.   __ret_723 = (int64x2_t)(vshll_n_s32(vget_high_s32(__s0_723), __p1_723)); \
  58757.   __ret_723; \
  58758. })
  58759. #else
  58760. #define vshll_high_n_s32(__p0_724, __p1_724) __extension__ ({ \
  58761.   int64x2_t __ret_724; \
  58762.   int32x4_t __s0_724 = __p0_724; \
  58763.   int32x4_t __rev0_724;  __rev0_724 = __builtin_shufflevector(__s0_724, __s0_724, 3, 2, 1, 0); \
  58764.   __ret_724 = (int64x2_t)(__noswap_vshll_n_s32(__noswap_vget_high_s32(__rev0_724), __p1_724)); \
  58765.   __ret_724 = __builtin_shufflevector(__ret_724, __ret_724, 1, 0); \
  58766.   __ret_724; \
  58767. })
  58768. #endif
  58769.  
  58770. #ifdef __LITTLE_ENDIAN__
  58771. #define vshll_high_n_s16(__p0_725, __p1_725) __extension__ ({ \
  58772.   int32x4_t __ret_725; \
  58773.   int16x8_t __s0_725 = __p0_725; \
  58774.   __ret_725 = (int32x4_t)(vshll_n_s16(vget_high_s16(__s0_725), __p1_725)); \
  58775.   __ret_725; \
  58776. })
  58777. #else
  58778. #define vshll_high_n_s16(__p0_726, __p1_726) __extension__ ({ \
  58779.   int32x4_t __ret_726; \
  58780.   int16x8_t __s0_726 = __p0_726; \
  58781.   int16x8_t __rev0_726;  __rev0_726 = __builtin_shufflevector(__s0_726, __s0_726, 7, 6, 5, 4, 3, 2, 1, 0); \
  58782.   __ret_726 = (int32x4_t)(__noswap_vshll_n_s16(__noswap_vget_high_s16(__rev0_726), __p1_726)); \
  58783.   __ret_726 = __builtin_shufflevector(__ret_726, __ret_726, 3, 2, 1, 0); \
  58784.   __ret_726; \
  58785. })
  58786. #endif
  58787.  
  58788. #define vshrd_n_u64(__p0, __p1) __extension__ ({ \
  58789.   uint64_t __ret; \
  58790.   uint64_t __s0 = __p0; \
  58791.   __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
  58792.   __ret; \
  58793. })
  58794. #define vshrd_n_s64(__p0, __p1) __extension__ ({ \
  58795.   int64_t __ret; \
  58796.   int64_t __s0 = __p0; \
  58797.   __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
  58798.   __ret; \
  58799. })
  58800. #ifdef __LITTLE_ENDIAN__
  58801. #define vshrn_high_n_u32(__p0_727, __p1_727, __p2_727) __extension__ ({ \
  58802.   uint16x8_t __ret_727; \
  58803.   uint16x4_t __s0_727 = __p0_727; \
  58804.   uint32x4_t __s1_727 = __p1_727; \
  58805.   __ret_727 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_727), (uint16x4_t)(vshrn_n_u32(__s1_727, __p2_727)))); \
  58806.   __ret_727; \
  58807. })
  58808. #else
  58809. #define vshrn_high_n_u32(__p0_728, __p1_728, __p2_728) __extension__ ({ \
  58810.   uint16x8_t __ret_728; \
  58811.   uint16x4_t __s0_728 = __p0_728; \
  58812.   uint32x4_t __s1_728 = __p1_728; \
  58813.   uint16x4_t __rev0_728;  __rev0_728 = __builtin_shufflevector(__s0_728, __s0_728, 3, 2, 1, 0); \
  58814.   uint32x4_t __rev1_728;  __rev1_728 = __builtin_shufflevector(__s1_728, __s1_728, 3, 2, 1, 0); \
  58815.   __ret_728 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_728), (uint16x4_t)(__noswap_vshrn_n_u32(__rev1_728, __p2_728)))); \
  58816.   __ret_728 = __builtin_shufflevector(__ret_728, __ret_728, 7, 6, 5, 4, 3, 2, 1, 0); \
  58817.   __ret_728; \
  58818. })
  58819. #endif
  58820.  
  58821. #ifdef __LITTLE_ENDIAN__
  58822. #define vshrn_high_n_u64(__p0_729, __p1_729, __p2_729) __extension__ ({ \
  58823.   uint32x4_t __ret_729; \
  58824.   uint32x2_t __s0_729 = __p0_729; \
  58825.   uint64x2_t __s1_729 = __p1_729; \
  58826.   __ret_729 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_729), (uint32x2_t)(vshrn_n_u64(__s1_729, __p2_729)))); \
  58827.   __ret_729; \
  58828. })
  58829. #else
  58830. #define vshrn_high_n_u64(__p0_730, __p1_730, __p2_730) __extension__ ({ \
  58831.   uint32x4_t __ret_730; \
  58832.   uint32x2_t __s0_730 = __p0_730; \
  58833.   uint64x2_t __s1_730 = __p1_730; \
  58834.   uint32x2_t __rev0_730;  __rev0_730 = __builtin_shufflevector(__s0_730, __s0_730, 1, 0); \
  58835.   uint64x2_t __rev1_730;  __rev1_730 = __builtin_shufflevector(__s1_730, __s1_730, 1, 0); \
  58836.   __ret_730 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_730), (uint32x2_t)(__noswap_vshrn_n_u64(__rev1_730, __p2_730)))); \
  58837.   __ret_730 = __builtin_shufflevector(__ret_730, __ret_730, 3, 2, 1, 0); \
  58838.   __ret_730; \
  58839. })
  58840. #endif
  58841.  
  58842. #ifdef __LITTLE_ENDIAN__
  58843. #define vshrn_high_n_u16(__p0_731, __p1_731, __p2_731) __extension__ ({ \
  58844.   uint8x16_t __ret_731; \
  58845.   uint8x8_t __s0_731 = __p0_731; \
  58846.   uint16x8_t __s1_731 = __p1_731; \
  58847.   __ret_731 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_731), (uint8x8_t)(vshrn_n_u16(__s1_731, __p2_731)))); \
  58848.   __ret_731; \
  58849. })
  58850. #else
  58851. #define vshrn_high_n_u16(__p0_732, __p1_732, __p2_732) __extension__ ({ \
  58852.   uint8x16_t __ret_732; \
  58853.   uint8x8_t __s0_732 = __p0_732; \
  58854.   uint16x8_t __s1_732 = __p1_732; \
  58855.   uint8x8_t __rev0_732;  __rev0_732 = __builtin_shufflevector(__s0_732, __s0_732, 7, 6, 5, 4, 3, 2, 1, 0); \
  58856.   uint16x8_t __rev1_732;  __rev1_732 = __builtin_shufflevector(__s1_732, __s1_732, 7, 6, 5, 4, 3, 2, 1, 0); \
  58857.   __ret_732 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_732), (uint8x8_t)(__noswap_vshrn_n_u16(__rev1_732, __p2_732)))); \
  58858.   __ret_732 = __builtin_shufflevector(__ret_732, __ret_732, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  58859.   __ret_732; \
  58860. })
  58861. #endif
  58862.  
  58863. #ifdef __LITTLE_ENDIAN__
  58864. #define vshrn_high_n_s32(__p0_733, __p1_733, __p2_733) __extension__ ({ \
  58865.   int16x8_t __ret_733; \
  58866.   int16x4_t __s0_733 = __p0_733; \
  58867.   int32x4_t __s1_733 = __p1_733; \
  58868.   __ret_733 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_733), (int16x4_t)(vshrn_n_s32(__s1_733, __p2_733)))); \
  58869.   __ret_733; \
  58870. })
  58871. #else
  58872. #define vshrn_high_n_s32(__p0_734, __p1_734, __p2_734) __extension__ ({ \
  58873.   int16x8_t __ret_734; \
  58874.   int16x4_t __s0_734 = __p0_734; \
  58875.   int32x4_t __s1_734 = __p1_734; \
  58876.   int16x4_t __rev0_734;  __rev0_734 = __builtin_shufflevector(__s0_734, __s0_734, 3, 2, 1, 0); \
  58877.   int32x4_t __rev1_734;  __rev1_734 = __builtin_shufflevector(__s1_734, __s1_734, 3, 2, 1, 0); \
  58878.   __ret_734 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_734), (int16x4_t)(__noswap_vshrn_n_s32(__rev1_734, __p2_734)))); \
  58879.   __ret_734 = __builtin_shufflevector(__ret_734, __ret_734, 7, 6, 5, 4, 3, 2, 1, 0); \
  58880.   __ret_734; \
  58881. })
  58882. #endif
  58883.  
  58884. #ifdef __LITTLE_ENDIAN__
  58885. #define vshrn_high_n_s64(__p0_735, __p1_735, __p2_735) __extension__ ({ \
  58886.   int32x4_t __ret_735; \
  58887.   int32x2_t __s0_735 = __p0_735; \
  58888.   int64x2_t __s1_735 = __p1_735; \
  58889.   __ret_735 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_735), (int32x2_t)(vshrn_n_s64(__s1_735, __p2_735)))); \
  58890.   __ret_735; \
  58891. })
  58892. #else
  58893. #define vshrn_high_n_s64(__p0_736, __p1_736, __p2_736) __extension__ ({ \
  58894.   int32x4_t __ret_736; \
  58895.   int32x2_t __s0_736 = __p0_736; \
  58896.   int64x2_t __s1_736 = __p1_736; \
  58897.   int32x2_t __rev0_736;  __rev0_736 = __builtin_shufflevector(__s0_736, __s0_736, 1, 0); \
  58898.   int64x2_t __rev1_736;  __rev1_736 = __builtin_shufflevector(__s1_736, __s1_736, 1, 0); \
  58899.   __ret_736 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_736), (int32x2_t)(__noswap_vshrn_n_s64(__rev1_736, __p2_736)))); \
  58900.   __ret_736 = __builtin_shufflevector(__ret_736, __ret_736, 3, 2, 1, 0); \
  58901.   __ret_736; \
  58902. })
  58903. #endif
  58904.  
  58905. #ifdef __LITTLE_ENDIAN__
  58906. #define vshrn_high_n_s16(__p0_737, __p1_737, __p2_737) __extension__ ({ \
  58907.   int8x16_t __ret_737; \
  58908.   int8x8_t __s0_737 = __p0_737; \
  58909.   int16x8_t __s1_737 = __p1_737; \
  58910.   __ret_737 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_737), (int8x8_t)(vshrn_n_s16(__s1_737, __p2_737)))); \
  58911.   __ret_737; \
  58912. })
  58913. #else
  58914. #define vshrn_high_n_s16(__p0_738, __p1_738, __p2_738) __extension__ ({ \
  58915.   int8x16_t __ret_738; \
  58916.   int8x8_t __s0_738 = __p0_738; \
  58917.   int16x8_t __s1_738 = __p1_738; \
  58918.   int8x8_t __rev0_738;  __rev0_738 = __builtin_shufflevector(__s0_738, __s0_738, 7, 6, 5, 4, 3, 2, 1, 0); \
  58919.   int16x8_t __rev1_738;  __rev1_738 = __builtin_shufflevector(__s1_738, __s1_738, 7, 6, 5, 4, 3, 2, 1, 0); \
  58920.   __ret_738 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_738), (int8x8_t)(__noswap_vshrn_n_s16(__rev1_738, __p2_738)))); \
  58921.   __ret_738 = __builtin_shufflevector(__ret_738, __ret_738, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  58922.   __ret_738; \
  58923. })
  58924. #endif
  58925.  
  58926. #define vslid_n_u64(__p0, __p1, __p2) __extension__ ({ \
  58927.   uint64_t __ret; \
  58928.   uint64_t __s0 = __p0; \
  58929.   uint64_t __s1 = __p1; \
  58930.   __ret = (uint64_t) __builtin_neon_vslid_n_u64(__s0, __s1, __p2); \
  58931.   __ret; \
  58932. })
  58933. #define vslid_n_s64(__p0, __p1, __p2) __extension__ ({ \
  58934.   int64_t __ret; \
  58935.   int64_t __s0 = __p0; \
  58936.   int64_t __s1 = __p1; \
  58937.   __ret = (int64_t) __builtin_neon_vslid_n_s64(__s0, __s1, __p2); \
  58938.   __ret; \
  58939. })
  58940. #define vsli_n_p64(__p0, __p1, __p2) __extension__ ({ \
  58941.   poly64x1_t __ret; \
  58942.   poly64x1_t __s0 = __p0; \
  58943.   poly64x1_t __s1 = __p1; \
  58944.   __ret = (poly64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
  58945.   __ret; \
  58946. })
  58947. #ifdef __LITTLE_ENDIAN__
  58948. #define vsliq_n_p64(__p0, __p1, __p2) __extension__ ({ \
  58949.   poly64x2_t __ret; \
  58950.   poly64x2_t __s0 = __p0; \
  58951.   poly64x2_t __s1 = __p1; \
  58952.   __ret = (poly64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 38); \
  58953.   __ret; \
  58954. })
  58955. #else
  58956. #define vsliq_n_p64(__p0, __p1, __p2) __extension__ ({ \
  58957.   poly64x2_t __ret; \
  58958.   poly64x2_t __s0 = __p0; \
  58959.   poly64x2_t __s1 = __p1; \
  58960.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  58961.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  58962.   __ret = (poly64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 38); \
  58963.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  58964.   __ret; \
  58965. })
  58966. #endif
  58967.  
  58968. __ai uint8_t vsqaddb_u8(uint8_t __p0, int8_t __p1) {
  58969.   uint8_t __ret;
  58970.   __ret = (uint8_t) __builtin_neon_vsqaddb_u8(__p0, __p1);
  58971.   return __ret;
  58972. }
  58973. __ai uint32_t vsqadds_u32(uint32_t __p0, int32_t __p1) {
  58974.   uint32_t __ret;
  58975.   __ret = (uint32_t) __builtin_neon_vsqadds_u32(__p0, __p1);
  58976.   return __ret;
  58977. }
  58978. __ai uint64_t vsqaddd_u64(uint64_t __p0, int64_t __p1) {
  58979.   uint64_t __ret;
  58980.   __ret = (uint64_t) __builtin_neon_vsqaddd_u64(__p0, __p1);
  58981.   return __ret;
  58982. }
  58983. __ai uint16_t vsqaddh_u16(uint16_t __p0, int16_t __p1) {
  58984.   uint16_t __ret;
  58985.   __ret = (uint16_t) __builtin_neon_vsqaddh_u16(__p0, __p1);
  58986.   return __ret;
  58987. }
  58988. #ifdef __LITTLE_ENDIAN__
  58989. __ai uint8x16_t vsqaddq_u8(uint8x16_t __p0, int8x16_t __p1) {
  58990.   uint8x16_t __ret;
  58991.   __ret = (uint8x16_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
  58992.   return __ret;
  58993. }
  58994. #else
  58995. __ai uint8x16_t vsqaddq_u8(uint8x16_t __p0, int8x16_t __p1) {
  58996.   uint8x16_t __ret;
  58997.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  58998.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  58999.   __ret = (uint8x16_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
  59000.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  59001.   return __ret;
  59002. }
  59003. #endif
  59004.  
  59005. #ifdef __LITTLE_ENDIAN__
  59006. __ai uint32x4_t vsqaddq_u32(uint32x4_t __p0, int32x4_t __p1) {
  59007.   uint32x4_t __ret;
  59008.   __ret = (uint32x4_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
  59009.   return __ret;
  59010. }
  59011. #else
  59012. __ai uint32x4_t vsqaddq_u32(uint32x4_t __p0, int32x4_t __p1) {
  59013.   uint32x4_t __ret;
  59014.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  59015.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  59016.   __ret = (uint32x4_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  59017.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  59018.   return __ret;
  59019. }
  59020. #endif
  59021.  
  59022. #ifdef __LITTLE_ENDIAN__
  59023. __ai uint64x2_t vsqaddq_u64(uint64x2_t __p0, int64x2_t __p1) {
  59024.   uint64x2_t __ret;
  59025.   __ret = (uint64x2_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  59026.   return __ret;
  59027. }
  59028. #else
  59029. __ai uint64x2_t vsqaddq_u64(uint64x2_t __p0, int64x2_t __p1) {
  59030.   uint64x2_t __ret;
  59031.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  59032.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  59033.   __ret = (uint64x2_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  59034.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  59035.   return __ret;
  59036. }
  59037. #endif
  59038.  
  59039. #ifdef __LITTLE_ENDIAN__
  59040. __ai uint16x8_t vsqaddq_u16(uint16x8_t __p0, int16x8_t __p1) {
  59041.   uint16x8_t __ret;
  59042.   __ret = (uint16x8_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
  59043.   return __ret;
  59044. }
  59045. #else
  59046. __ai uint16x8_t vsqaddq_u16(uint16x8_t __p0, int16x8_t __p1) {
  59047.   uint16x8_t __ret;
  59048.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  59049.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  59050.   __ret = (uint16x8_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
  59051.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  59052.   return __ret;
  59053. }
  59054. #endif
  59055.  
  59056. #ifdef __LITTLE_ENDIAN__
  59057. __ai uint8x8_t vsqadd_u8(uint8x8_t __p0, int8x8_t __p1) {
  59058.   uint8x8_t __ret;
  59059.   __ret = (uint8x8_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
  59060.   return __ret;
  59061. }
  59062. #else
  59063. __ai uint8x8_t vsqadd_u8(uint8x8_t __p0, int8x8_t __p1) {
  59064.   uint8x8_t __ret;
  59065.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  59066.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  59067.   __ret = (uint8x8_t) __builtin_neon_vsqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
  59068.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  59069.   return __ret;
  59070. }
  59071. #endif
  59072.  
  59073. #ifdef __LITTLE_ENDIAN__
  59074. __ai uint32x2_t vsqadd_u32(uint32x2_t __p0, int32x2_t __p1) {
  59075.   uint32x2_t __ret;
  59076.   __ret = (uint32x2_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
  59077.   return __ret;
  59078. }
  59079. #else
  59080. __ai uint32x2_t vsqadd_u32(uint32x2_t __p0, int32x2_t __p1) {
  59081.   uint32x2_t __ret;
  59082.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  59083.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  59084.   __ret = (uint32x2_t) __builtin_neon_vsqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
  59085.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  59086.   return __ret;
  59087. }
  59088. #endif
  59089.  
  59090. __ai uint64x1_t vsqadd_u64(uint64x1_t __p0, int64x1_t __p1) {
  59091.   uint64x1_t __ret;
  59092.   __ret = (uint64x1_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  59093.   return __ret;
  59094. }
  59095. #ifdef __LITTLE_ENDIAN__
  59096. __ai uint16x4_t vsqadd_u16(uint16x4_t __p0, int16x4_t __p1) {
  59097.   uint16x4_t __ret;
  59098.   __ret = (uint16x4_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
  59099.   return __ret;
  59100. }
  59101. #else
  59102. __ai uint16x4_t vsqadd_u16(uint16x4_t __p0, int16x4_t __p1) {
  59103.   uint16x4_t __ret;
  59104.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  59105.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  59106.   __ret = (uint16x4_t) __builtin_neon_vsqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
  59107.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  59108.   return __ret;
  59109. }
  59110. #endif
  59111.  
  59112. #ifdef __LITTLE_ENDIAN__
  59113. __ai float64x2_t vsqrtq_f64(float64x2_t __p0) {
  59114.   float64x2_t __ret;
  59115.   __ret = (float64x2_t) __builtin_neon_vsqrtq_v((int8x16_t)__p0, 42);
  59116.   return __ret;
  59117. }
  59118. #else
  59119. __ai float64x2_t vsqrtq_f64(float64x2_t __p0) {
  59120.   float64x2_t __ret;
  59121.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  59122.   __ret = (float64x2_t) __builtin_neon_vsqrtq_v((int8x16_t)__rev0, 42);
  59123.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  59124.   return __ret;
  59125. }
  59126. #endif
  59127.  
  59128. #ifdef __LITTLE_ENDIAN__
  59129. __ai float32x4_t vsqrtq_f32(float32x4_t __p0) {
  59130.   float32x4_t __ret;
  59131.   __ret = (float32x4_t) __builtin_neon_vsqrtq_v((int8x16_t)__p0, 41);
  59132.   return __ret;
  59133. }
  59134. #else
  59135. __ai float32x4_t vsqrtq_f32(float32x4_t __p0) {
  59136.   float32x4_t __ret;
  59137.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  59138.   __ret = (float32x4_t) __builtin_neon_vsqrtq_v((int8x16_t)__rev0, 41);
  59139.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  59140.   return __ret;
  59141. }
  59142. #endif
  59143.  
  59144. __ai float64x1_t vsqrt_f64(float64x1_t __p0) {
  59145.   float64x1_t __ret;
  59146.   __ret = (float64x1_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 10);
  59147.   return __ret;
  59148. }
  59149. #ifdef __LITTLE_ENDIAN__
  59150. __ai float32x2_t vsqrt_f32(float32x2_t __p0) {
  59151.   float32x2_t __ret;
  59152.   __ret = (float32x2_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 9);
  59153.   return __ret;
  59154. }
  59155. #else
  59156. __ai float32x2_t vsqrt_f32(float32x2_t __p0) {
  59157.   float32x2_t __ret;
  59158.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  59159.   __ret = (float32x2_t) __builtin_neon_vsqrt_v((int8x8_t)__rev0, 9);
  59160.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  59161.   return __ret;
  59162. }
  59163. #endif
  59164.  
  59165. #define vsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
  59166.   uint64_t __ret; \
  59167.   uint64_t __s0 = __p0; \
  59168.   uint64_t __s1 = __p1; \
  59169.   __ret = (uint64_t) __builtin_neon_vsrad_n_u64(__s0, __s1, __p2); \
  59170.   __ret; \
  59171. })
  59172. #define vsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
  59173.   int64_t __ret; \
  59174.   int64_t __s0 = __p0; \
  59175.   int64_t __s1 = __p1; \
  59176.   __ret = (int64_t) __builtin_neon_vsrad_n_s64(__s0, __s1, __p2); \
  59177.   __ret; \
  59178. })
  59179. #define vsrid_n_u64(__p0, __p1, __p2) __extension__ ({ \
  59180.   uint64_t __ret; \
  59181.   uint64_t __s0 = __p0; \
  59182.   uint64_t __s1 = __p1; \
  59183.   __ret = (uint64_t) __builtin_neon_vsrid_n_u64(__s0, __s1, __p2); \
  59184.   __ret; \
  59185. })
  59186. #define vsrid_n_s64(__p0, __p1, __p2) __extension__ ({ \
  59187.   int64_t __ret; \
  59188.   int64_t __s0 = __p0; \
  59189.   int64_t __s1 = __p1; \
  59190.   __ret = (int64_t) __builtin_neon_vsrid_n_s64(__s0, __s1, __p2); \
  59191.   __ret; \
  59192. })
  59193. #define vsri_n_p64(__p0, __p1, __p2) __extension__ ({ \
  59194.   poly64x1_t __ret; \
  59195.   poly64x1_t __s0 = __p0; \
  59196.   poly64x1_t __s1 = __p1; \
  59197.   __ret = (poly64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
  59198.   __ret; \
  59199. })
  59200. #ifdef __LITTLE_ENDIAN__
  59201. #define vsriq_n_p64(__p0, __p1, __p2) __extension__ ({ \
  59202.   poly64x2_t __ret; \
  59203.   poly64x2_t __s0 = __p0; \
  59204.   poly64x2_t __s1 = __p1; \
  59205.   __ret = (poly64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 38); \
  59206.   __ret; \
  59207. })
  59208. #else
  59209. #define vsriq_n_p64(__p0, __p1, __p2) __extension__ ({ \
  59210.   poly64x2_t __ret; \
  59211.   poly64x2_t __s0 = __p0; \
  59212.   poly64x2_t __s1 = __p1; \
  59213.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  59214.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  59215.   __ret = (poly64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 38); \
  59216.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  59217.   __ret; \
  59218. })
  59219. #endif
  59220.  
  59221. #define vst1_p64(__p0, __p1) __extension__ ({ \
  59222.   poly64x1_t __s1 = __p1; \
  59223.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 6); \
  59224. })
  59225. #ifdef __LITTLE_ENDIAN__
  59226. #define vst1q_p64(__p0, __p1) __extension__ ({ \
  59227.   poly64x2_t __s1 = __p1; \
  59228.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 38); \
  59229. })
  59230. #else
  59231. #define vst1q_p64(__p0, __p1) __extension__ ({ \
  59232.   poly64x2_t __s1 = __p1; \
  59233.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  59234.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 38); \
  59235. })
  59236. #endif
  59237.  
  59238. #ifdef __LITTLE_ENDIAN__
  59239. #define vst1q_f64(__p0, __p1) __extension__ ({ \
  59240.   float64x2_t __s1 = __p1; \
  59241.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 42); \
  59242. })
  59243. #else
  59244. #define vst1q_f64(__p0, __p1) __extension__ ({ \
  59245.   float64x2_t __s1 = __p1; \
  59246.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  59247.   __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 42); \
  59248. })
  59249. #endif
  59250.  
  59251. #define vst1_f64(__p0, __p1) __extension__ ({ \
  59252.   float64x1_t __s1 = __p1; \
  59253.   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 10); \
  59254. })
  59255. #define vst1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59256.   poly64x1_t __s1 = __p1; \
  59257.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
  59258. })
  59259. #ifdef __LITTLE_ENDIAN__
  59260. #define vst1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59261.   poly64x2_t __s1 = __p1; \
  59262.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 38); \
  59263. })
  59264. #else
  59265. #define vst1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59266.   poly64x2_t __s1 = __p1; \
  59267.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  59268.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 38); \
  59269. })
  59270. #endif
  59271.  
  59272. #ifdef __LITTLE_ENDIAN__
  59273. #define vst1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59274.   float64x2_t __s1 = __p1; \
  59275.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 42); \
  59276. })
  59277. #else
  59278. #define vst1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59279.   float64x2_t __s1 = __p1; \
  59280.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  59281.   __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 42); \
  59282. })
  59283. #endif
  59284.  
  59285. #define vst1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59286.   float64x1_t __s1 = __p1; \
  59287.   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
  59288. })
  59289. #define vst1_p64_x2(__p0, __p1) __extension__ ({ \
  59290.   poly64x1x2_t __s1 = __p1; \
  59291.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
  59292. })
  59293. #ifdef __LITTLE_ENDIAN__
  59294. #define vst1q_p64_x2(__p0, __p1) __extension__ ({ \
  59295.   poly64x2x2_t __s1 = __p1; \
  59296.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 38); \
  59297. })
  59298. #else
  59299. #define vst1q_p64_x2(__p0, __p1) __extension__ ({ \
  59300.   poly64x2x2_t __s1 = __p1; \
  59301.   poly64x2x2_t __rev1; \
  59302.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59303.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59304.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 38); \
  59305. })
  59306. #endif
  59307.  
  59308. #ifdef __LITTLE_ENDIAN__
  59309. #define vst1q_f64_x2(__p0, __p1) __extension__ ({ \
  59310.   float64x2x2_t __s1 = __p1; \
  59311.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 42); \
  59312. })
  59313. #else
  59314. #define vst1q_f64_x2(__p0, __p1) __extension__ ({ \
  59315.   float64x2x2_t __s1 = __p1; \
  59316.   float64x2x2_t __rev1; \
  59317.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59318.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59319.   __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 42); \
  59320. })
  59321. #endif
  59322.  
  59323. #define vst1_f64_x2(__p0, __p1) __extension__ ({ \
  59324.   float64x1x2_t __s1 = __p1; \
  59325.   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 10); \
  59326. })
  59327. #define vst1_p64_x3(__p0, __p1) __extension__ ({ \
  59328.   poly64x1x3_t __s1 = __p1; \
  59329.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
  59330. })
  59331. #ifdef __LITTLE_ENDIAN__
  59332. #define vst1q_p64_x3(__p0, __p1) __extension__ ({ \
  59333.   poly64x2x3_t __s1 = __p1; \
  59334.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 38); \
  59335. })
  59336. #else
  59337. #define vst1q_p64_x3(__p0, __p1) __extension__ ({ \
  59338.   poly64x2x3_t __s1 = __p1; \
  59339.   poly64x2x3_t __rev1; \
  59340.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59341.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59342.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59343.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 38); \
  59344. })
  59345. #endif
  59346.  
  59347. #ifdef __LITTLE_ENDIAN__
  59348. #define vst1q_f64_x3(__p0, __p1) __extension__ ({ \
  59349.   float64x2x3_t __s1 = __p1; \
  59350.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 42); \
  59351. })
  59352. #else
  59353. #define vst1q_f64_x3(__p0, __p1) __extension__ ({ \
  59354.   float64x2x3_t __s1 = __p1; \
  59355.   float64x2x3_t __rev1; \
  59356.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59357.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59358.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59359.   __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 42); \
  59360. })
  59361. #endif
  59362.  
  59363. #define vst1_f64_x3(__p0, __p1) __extension__ ({ \
  59364.   float64x1x3_t __s1 = __p1; \
  59365.   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 10); \
  59366. })
  59367. #define vst1_p64_x4(__p0, __p1) __extension__ ({ \
  59368.   poly64x1x4_t __s1 = __p1; \
  59369.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
  59370. })
  59371. #ifdef __LITTLE_ENDIAN__
  59372. #define vst1q_p64_x4(__p0, __p1) __extension__ ({ \
  59373.   poly64x2x4_t __s1 = __p1; \
  59374.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 38); \
  59375. })
  59376. #else
  59377. #define vst1q_p64_x4(__p0, __p1) __extension__ ({ \
  59378.   poly64x2x4_t __s1 = __p1; \
  59379.   poly64x2x4_t __rev1; \
  59380.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59381.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59382.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59383.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59384.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 38); \
  59385. })
  59386. #endif
  59387.  
  59388. #ifdef __LITTLE_ENDIAN__
  59389. #define vst1q_f64_x4(__p0, __p1) __extension__ ({ \
  59390.   float64x2x4_t __s1 = __p1; \
  59391.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 42); \
  59392. })
  59393. #else
  59394. #define vst1q_f64_x4(__p0, __p1) __extension__ ({ \
  59395.   float64x2x4_t __s1 = __p1; \
  59396.   float64x2x4_t __rev1; \
  59397.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59398.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59399.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59400.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59401.   __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 42); \
  59402. })
  59403. #endif
  59404.  
  59405. #define vst1_f64_x4(__p0, __p1) __extension__ ({ \
  59406.   float64x1x4_t __s1 = __p1; \
  59407.   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 10); \
  59408. })
  59409. #define vst2_p64(__p0, __p1) __extension__ ({ \
  59410.   poly64x1x2_t __s1 = __p1; \
  59411.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
  59412. })
  59413. #ifdef __LITTLE_ENDIAN__
  59414. #define vst2q_p64(__p0, __p1) __extension__ ({ \
  59415.   poly64x2x2_t __s1 = __p1; \
  59416.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 38); \
  59417. })
  59418. #else
  59419. #define vst2q_p64(__p0, __p1) __extension__ ({ \
  59420.   poly64x2x2_t __s1 = __p1; \
  59421.   poly64x2x2_t __rev1; \
  59422.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59423.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59424.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 38); \
  59425. })
  59426. #endif
  59427.  
  59428. #ifdef __LITTLE_ENDIAN__
  59429. #define vst2q_u64(__p0, __p1) __extension__ ({ \
  59430.   uint64x2x2_t __s1 = __p1; \
  59431.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 51); \
  59432. })
  59433. #else
  59434. #define vst2q_u64(__p0, __p1) __extension__ ({ \
  59435.   uint64x2x2_t __s1 = __p1; \
  59436.   uint64x2x2_t __rev1; \
  59437.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59438.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59439.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 51); \
  59440. })
  59441. #endif
  59442.  
  59443. #ifdef __LITTLE_ENDIAN__
  59444. #define vst2q_f64(__p0, __p1) __extension__ ({ \
  59445.   float64x2x2_t __s1 = __p1; \
  59446.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 42); \
  59447. })
  59448. #else
  59449. #define vst2q_f64(__p0, __p1) __extension__ ({ \
  59450.   float64x2x2_t __s1 = __p1; \
  59451.   float64x2x2_t __rev1; \
  59452.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59453.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59454.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 42); \
  59455. })
  59456. #endif
  59457.  
  59458. #ifdef __LITTLE_ENDIAN__
  59459. #define vst2q_s64(__p0, __p1) __extension__ ({ \
  59460.   int64x2x2_t __s1 = __p1; \
  59461.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 35); \
  59462. })
  59463. #else
  59464. #define vst2q_s64(__p0, __p1) __extension__ ({ \
  59465.   int64x2x2_t __s1 = __p1; \
  59466.   int64x2x2_t __rev1; \
  59467.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59468.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59469.   __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 35); \
  59470. })
  59471. #endif
  59472.  
  59473. #define vst2_f64(__p0, __p1) __extension__ ({ \
  59474.   float64x1x2_t __s1 = __p1; \
  59475.   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 10); \
  59476. })
  59477. #define vst2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59478.   poly64x1x2_t __s1 = __p1; \
  59479.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
  59480. })
  59481. #ifdef __LITTLE_ENDIAN__
  59482. #define vst2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  59483.   poly8x16x2_t __s1 = __p1; \
  59484.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 36); \
  59485. })
  59486. #else
  59487. #define vst2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  59488.   poly8x16x2_t __s1 = __p1; \
  59489.   poly8x16x2_t __rev1; \
  59490.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59491.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59492.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 36); \
  59493. })
  59494. #endif
  59495.  
  59496. #ifdef __LITTLE_ENDIAN__
  59497. #define vst2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59498.   poly64x2x2_t __s1 = __p1; \
  59499.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 38); \
  59500. })
  59501. #else
  59502. #define vst2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59503.   poly64x2x2_t __s1 = __p1; \
  59504.   poly64x2x2_t __rev1; \
  59505.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59506.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59507.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 38); \
  59508. })
  59509. #endif
  59510.  
  59511. #ifdef __LITTLE_ENDIAN__
  59512. #define vst2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  59513.   uint8x16x2_t __s1 = __p1; \
  59514.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 48); \
  59515. })
  59516. #else
  59517. #define vst2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  59518.   uint8x16x2_t __s1 = __p1; \
  59519.   uint8x16x2_t __rev1; \
  59520.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59521.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59522.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 48); \
  59523. })
  59524. #endif
  59525.  
  59526. #ifdef __LITTLE_ENDIAN__
  59527. #define vst2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  59528.   uint64x2x2_t __s1 = __p1; \
  59529.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 51); \
  59530. })
  59531. #else
  59532. #define vst2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  59533.   uint64x2x2_t __s1 = __p1; \
  59534.   uint64x2x2_t __rev1; \
  59535.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59536.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59537.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 51); \
  59538. })
  59539. #endif
  59540.  
  59541. #ifdef __LITTLE_ENDIAN__
  59542. #define vst2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  59543.   int8x16x2_t __s1 = __p1; \
  59544.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 32); \
  59545. })
  59546. #else
  59547. #define vst2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  59548.   int8x16x2_t __s1 = __p1; \
  59549.   int8x16x2_t __rev1; \
  59550.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59551.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59552.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 32); \
  59553. })
  59554. #endif
  59555.  
  59556. #ifdef __LITTLE_ENDIAN__
  59557. #define vst2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59558.   float64x2x2_t __s1 = __p1; \
  59559.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 42); \
  59560. })
  59561. #else
  59562. #define vst2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59563.   float64x2x2_t __s1 = __p1; \
  59564.   float64x2x2_t __rev1; \
  59565.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59566.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59567.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 42); \
  59568. })
  59569. #endif
  59570.  
  59571. #ifdef __LITTLE_ENDIAN__
  59572. #define vst2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  59573.   int64x2x2_t __s1 = __p1; \
  59574.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 35); \
  59575. })
  59576. #else
  59577. #define vst2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  59578.   int64x2x2_t __s1 = __p1; \
  59579.   int64x2x2_t __rev1; \
  59580.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59581.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59582.   __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 35); \
  59583. })
  59584. #endif
  59585.  
  59586. #define vst2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  59587.   uint64x1x2_t __s1 = __p1; \
  59588.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
  59589. })
  59590. #define vst2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59591.   float64x1x2_t __s1 = __p1; \
  59592.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 10); \
  59593. })
  59594. #define vst2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  59595.   int64x1x2_t __s1 = __p1; \
  59596.   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 3); \
  59597. })
  59598. #define vst3_p64(__p0, __p1) __extension__ ({ \
  59599.   poly64x1x3_t __s1 = __p1; \
  59600.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
  59601. })
  59602. #ifdef __LITTLE_ENDIAN__
  59603. #define vst3q_p64(__p0, __p1) __extension__ ({ \
  59604.   poly64x2x3_t __s1 = __p1; \
  59605.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 38); \
  59606. })
  59607. #else
  59608. #define vst3q_p64(__p0, __p1) __extension__ ({ \
  59609.   poly64x2x3_t __s1 = __p1; \
  59610.   poly64x2x3_t __rev1; \
  59611.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59612.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59613.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59614.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 38); \
  59615. })
  59616. #endif
  59617.  
  59618. #ifdef __LITTLE_ENDIAN__
  59619. #define vst3q_u64(__p0, __p1) __extension__ ({ \
  59620.   uint64x2x3_t __s1 = __p1; \
  59621.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 51); \
  59622. })
  59623. #else
  59624. #define vst3q_u64(__p0, __p1) __extension__ ({ \
  59625.   uint64x2x3_t __s1 = __p1; \
  59626.   uint64x2x3_t __rev1; \
  59627.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59628.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59629.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59630.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 51); \
  59631. })
  59632. #endif
  59633.  
  59634. #ifdef __LITTLE_ENDIAN__
  59635. #define vst3q_f64(__p0, __p1) __extension__ ({ \
  59636.   float64x2x3_t __s1 = __p1; \
  59637.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 42); \
  59638. })
  59639. #else
  59640. #define vst3q_f64(__p0, __p1) __extension__ ({ \
  59641.   float64x2x3_t __s1 = __p1; \
  59642.   float64x2x3_t __rev1; \
  59643.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59644.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59645.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59646.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 42); \
  59647. })
  59648. #endif
  59649.  
  59650. #ifdef __LITTLE_ENDIAN__
  59651. #define vst3q_s64(__p0, __p1) __extension__ ({ \
  59652.   int64x2x3_t __s1 = __p1; \
  59653.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 35); \
  59654. })
  59655. #else
  59656. #define vst3q_s64(__p0, __p1) __extension__ ({ \
  59657.   int64x2x3_t __s1 = __p1; \
  59658.   int64x2x3_t __rev1; \
  59659.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59660.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59661.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59662.   __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 35); \
  59663. })
  59664. #endif
  59665.  
  59666. #define vst3_f64(__p0, __p1) __extension__ ({ \
  59667.   float64x1x3_t __s1 = __p1; \
  59668.   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 10); \
  59669. })
  59670. #define vst3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59671.   poly64x1x3_t __s1 = __p1; \
  59672.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
  59673. })
  59674. #ifdef __LITTLE_ENDIAN__
  59675. #define vst3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  59676.   poly8x16x3_t __s1 = __p1; \
  59677.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 36); \
  59678. })
  59679. #else
  59680. #define vst3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  59681.   poly8x16x3_t __s1 = __p1; \
  59682.   poly8x16x3_t __rev1; \
  59683.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59684.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59685.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59686.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 36); \
  59687. })
  59688. #endif
  59689.  
  59690. #ifdef __LITTLE_ENDIAN__
  59691. #define vst3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59692.   poly64x2x3_t __s1 = __p1; \
  59693.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 38); \
  59694. })
  59695. #else
  59696. #define vst3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59697.   poly64x2x3_t __s1 = __p1; \
  59698.   poly64x2x3_t __rev1; \
  59699.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59700.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59701.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59702.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 38); \
  59703. })
  59704. #endif
  59705.  
  59706. #ifdef __LITTLE_ENDIAN__
  59707. #define vst3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  59708.   uint8x16x3_t __s1 = __p1; \
  59709.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 48); \
  59710. })
  59711. #else
  59712. #define vst3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  59713.   uint8x16x3_t __s1 = __p1; \
  59714.   uint8x16x3_t __rev1; \
  59715.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59716.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59717.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59718.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 48); \
  59719. })
  59720. #endif
  59721.  
  59722. #ifdef __LITTLE_ENDIAN__
  59723. #define vst3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  59724.   uint64x2x3_t __s1 = __p1; \
  59725.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 51); \
  59726. })
  59727. #else
  59728. #define vst3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  59729.   uint64x2x3_t __s1 = __p1; \
  59730.   uint64x2x3_t __rev1; \
  59731.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59732.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59733.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59734.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 51); \
  59735. })
  59736. #endif
  59737.  
  59738. #ifdef __LITTLE_ENDIAN__
  59739. #define vst3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  59740.   int8x16x3_t __s1 = __p1; \
  59741.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 32); \
  59742. })
  59743. #else
  59744. #define vst3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  59745.   int8x16x3_t __s1 = __p1; \
  59746.   int8x16x3_t __rev1; \
  59747.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59748.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59749.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59750.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 32); \
  59751. })
  59752. #endif
  59753.  
  59754. #ifdef __LITTLE_ENDIAN__
  59755. #define vst3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59756.   float64x2x3_t __s1 = __p1; \
  59757.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 42); \
  59758. })
  59759. #else
  59760. #define vst3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59761.   float64x2x3_t __s1 = __p1; \
  59762.   float64x2x3_t __rev1; \
  59763.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59764.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59765.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59766.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 42); \
  59767. })
  59768. #endif
  59769.  
  59770. #ifdef __LITTLE_ENDIAN__
  59771. #define vst3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  59772.   int64x2x3_t __s1 = __p1; \
  59773.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 35); \
  59774. })
  59775. #else
  59776. #define vst3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  59777.   int64x2x3_t __s1 = __p1; \
  59778.   int64x2x3_t __rev1; \
  59779.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59780.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59781.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59782.   __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 35); \
  59783. })
  59784. #endif
  59785.  
  59786. #define vst3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  59787.   uint64x1x3_t __s1 = __p1; \
  59788.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
  59789. })
  59790. #define vst3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59791.   float64x1x3_t __s1 = __p1; \
  59792.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 10); \
  59793. })
  59794. #define vst3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  59795.   int64x1x3_t __s1 = __p1; \
  59796.   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 3); \
  59797. })
  59798. #define vst4_p64(__p0, __p1) __extension__ ({ \
  59799.   poly64x1x4_t __s1 = __p1; \
  59800.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
  59801. })
  59802. #ifdef __LITTLE_ENDIAN__
  59803. #define vst4q_p64(__p0, __p1) __extension__ ({ \
  59804.   poly64x2x4_t __s1 = __p1; \
  59805.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 38); \
  59806. })
  59807. #else
  59808. #define vst4q_p64(__p0, __p1) __extension__ ({ \
  59809.   poly64x2x4_t __s1 = __p1; \
  59810.   poly64x2x4_t __rev1; \
  59811.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59812.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59813.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59814.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59815.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 38); \
  59816. })
  59817. #endif
  59818.  
  59819. #ifdef __LITTLE_ENDIAN__
  59820. #define vst4q_u64(__p0, __p1) __extension__ ({ \
  59821.   uint64x2x4_t __s1 = __p1; \
  59822.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 51); \
  59823. })
  59824. #else
  59825. #define vst4q_u64(__p0, __p1) __extension__ ({ \
  59826.   uint64x2x4_t __s1 = __p1; \
  59827.   uint64x2x4_t __rev1; \
  59828.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59829.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59830.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59831.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59832.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 51); \
  59833. })
  59834. #endif
  59835.  
  59836. #ifdef __LITTLE_ENDIAN__
  59837. #define vst4q_f64(__p0, __p1) __extension__ ({ \
  59838.   float64x2x4_t __s1 = __p1; \
  59839.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 42); \
  59840. })
  59841. #else
  59842. #define vst4q_f64(__p0, __p1) __extension__ ({ \
  59843.   float64x2x4_t __s1 = __p1; \
  59844.   float64x2x4_t __rev1; \
  59845.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59846.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59847.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59848.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59849.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 42); \
  59850. })
  59851. #endif
  59852.  
  59853. #ifdef __LITTLE_ENDIAN__
  59854. #define vst4q_s64(__p0, __p1) __extension__ ({ \
  59855.   int64x2x4_t __s1 = __p1; \
  59856.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 35); \
  59857. })
  59858. #else
  59859. #define vst4q_s64(__p0, __p1) __extension__ ({ \
  59860.   int64x2x4_t __s1 = __p1; \
  59861.   int64x2x4_t __rev1; \
  59862.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59863.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59864.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59865.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59866.   __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 35); \
  59867. })
  59868. #endif
  59869.  
  59870. #define vst4_f64(__p0, __p1) __extension__ ({ \
  59871.   float64x1x4_t __s1 = __p1; \
  59872.   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 10); \
  59873. })
  59874. #define vst4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59875.   poly64x1x4_t __s1 = __p1; \
  59876.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
  59877. })
  59878. #ifdef __LITTLE_ENDIAN__
  59879. #define vst4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  59880.   poly8x16x4_t __s1 = __p1; \
  59881.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 36); \
  59882. })
  59883. #else
  59884. #define vst4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
  59885.   poly8x16x4_t __s1 = __p1; \
  59886.   poly8x16x4_t __rev1; \
  59887.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59888.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59889.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59890.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59891.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 36); \
  59892. })
  59893. #endif
  59894.  
  59895. #ifdef __LITTLE_ENDIAN__
  59896. #define vst4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59897.   poly64x2x4_t __s1 = __p1; \
  59898.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 38); \
  59899. })
  59900. #else
  59901. #define vst4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
  59902.   poly64x2x4_t __s1 = __p1; \
  59903.   poly64x2x4_t __rev1; \
  59904.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59905.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59906.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59907.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59908.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 38); \
  59909. })
  59910. #endif
  59911.  
  59912. #ifdef __LITTLE_ENDIAN__
  59913. #define vst4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  59914.   uint8x16x4_t __s1 = __p1; \
  59915.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 48); \
  59916. })
  59917. #else
  59918. #define vst4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
  59919.   uint8x16x4_t __s1 = __p1; \
  59920.   uint8x16x4_t __rev1; \
  59921.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59922.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59923.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59924.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59925.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 48); \
  59926. })
  59927. #endif
  59928.  
  59929. #ifdef __LITTLE_ENDIAN__
  59930. #define vst4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  59931.   uint64x2x4_t __s1 = __p1; \
  59932.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 51); \
  59933. })
  59934. #else
  59935. #define vst4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  59936.   uint64x2x4_t __s1 = __p1; \
  59937.   uint64x2x4_t __rev1; \
  59938.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59939.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59940.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59941.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59942.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 51); \
  59943. })
  59944. #endif
  59945.  
  59946. #ifdef __LITTLE_ENDIAN__
  59947. #define vst4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  59948.   int8x16x4_t __s1 = __p1; \
  59949.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 32); \
  59950. })
  59951. #else
  59952. #define vst4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
  59953.   int8x16x4_t __s1 = __p1; \
  59954.   int8x16x4_t __rev1; \
  59955.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59956.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59957.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59958.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  59959.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 32); \
  59960. })
  59961. #endif
  59962.  
  59963. #ifdef __LITTLE_ENDIAN__
  59964. #define vst4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59965.   float64x2x4_t __s1 = __p1; \
  59966.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 42); \
  59967. })
  59968. #else
  59969. #define vst4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  59970.   float64x2x4_t __s1 = __p1; \
  59971.   float64x2x4_t __rev1; \
  59972.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59973.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59974.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59975.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59976.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 42); \
  59977. })
  59978. #endif
  59979.  
  59980. #ifdef __LITTLE_ENDIAN__
  59981. #define vst4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  59982.   int64x2x4_t __s1 = __p1; \
  59983.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 35); \
  59984. })
  59985. #else
  59986. #define vst4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  59987.   int64x2x4_t __s1 = __p1; \
  59988.   int64x2x4_t __rev1; \
  59989.   __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
  59990.   __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
  59991.   __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
  59992.   __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
  59993.   __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 35); \
  59994. })
  59995. #endif
  59996.  
  59997. #define vst4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
  59998.   uint64x1x4_t __s1 = __p1; \
  59999.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
  60000. })
  60001. #define vst4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
  60002.   float64x1x4_t __s1 = __p1; \
  60003.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 10); \
  60004. })
  60005. #define vst4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
  60006.   int64x1x4_t __s1 = __p1; \
  60007.   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 3); \
  60008. })
  60009. #define vstrq_p128(__p0, __p1) __extension__ ({ \
  60010.   poly128_t __s1 = __p1; \
  60011.   __builtin_neon_vstrq_p128(__p0, __s1); \
  60012. })
  60013. __ai uint64_t vsubd_u64(uint64_t __p0, uint64_t __p1) {
  60014.   uint64_t __ret;
  60015.   __ret = (uint64_t) __builtin_neon_vsubd_u64(__p0, __p1);
  60016.   return __ret;
  60017. }
  60018. __ai int64_t vsubd_s64(int64_t __p0, int64_t __p1) {
  60019.   int64_t __ret;
  60020.   __ret = (int64_t) __builtin_neon_vsubd_s64(__p0, __p1);
  60021.   return __ret;
  60022. }
  60023. #ifdef __LITTLE_ENDIAN__
  60024. __ai float64x2_t vsubq_f64(float64x2_t __p0, float64x2_t __p1) {
  60025.   float64x2_t __ret;
  60026.   __ret = __p0 - __p1;
  60027.   return __ret;
  60028. }
  60029. #else
  60030. __ai float64x2_t vsubq_f64(float64x2_t __p0, float64x2_t __p1) {
  60031.   float64x2_t __ret;
  60032.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60033.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60034.   __ret = __rev0 - __rev1;
  60035.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60036.   return __ret;
  60037. }
  60038. #endif
  60039.  
  60040. __ai float64x1_t vsub_f64(float64x1_t __p0, float64x1_t __p1) {
  60041.   float64x1_t __ret;
  60042.   __ret = __p0 - __p1;
  60043.   return __ret;
  60044. }
  60045. #ifdef __LITTLE_ENDIAN__
  60046. __ai uint16x8_t vsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  60047.   uint16x8_t __ret;
  60048.   __ret = vcombine_u16(__p0, vsubhn_u32(__p1, __p2));
  60049.   return __ret;
  60050. }
  60051. #else
  60052. __ai uint16x8_t vsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  60053.   uint16x8_t __ret;
  60054.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60055.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60056.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  60057.   __ret = __noswap_vcombine_u16(__rev0, __noswap_vsubhn_u32(__rev1, __rev2));
  60058.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60059.   return __ret;
  60060. }
  60061. #endif
  60062.  
  60063. #ifdef __LITTLE_ENDIAN__
  60064. __ai uint32x4_t vsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  60065.   uint32x4_t __ret;
  60066.   __ret = vcombine_u32(__p0, vsubhn_u64(__p1, __p2));
  60067.   return __ret;
  60068. }
  60069. #else
  60070. __ai uint32x4_t vsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  60071.   uint32x4_t __ret;
  60072.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60073.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60074.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  60075.   __ret = __noswap_vcombine_u32(__rev0, __noswap_vsubhn_u64(__rev1, __rev2));
  60076.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60077.   return __ret;
  60078. }
  60079. #endif
  60080.  
  60081. #ifdef __LITTLE_ENDIAN__
  60082. __ai uint8x16_t vsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  60083.   uint8x16_t __ret;
  60084.   __ret = vcombine_u8(__p0, vsubhn_u16(__p1, __p2));
  60085.   return __ret;
  60086. }
  60087. #else
  60088. __ai uint8x16_t vsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  60089.   uint8x16_t __ret;
  60090.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60091.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60092.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  60093.   __ret = __noswap_vcombine_u8(__rev0, __noswap_vsubhn_u16(__rev1, __rev2));
  60094.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60095.   return __ret;
  60096. }
  60097. #endif
  60098.  
  60099. #ifdef __LITTLE_ENDIAN__
  60100. __ai int16x8_t vsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  60101.   int16x8_t __ret;
  60102.   __ret = vcombine_s16(__p0, vsubhn_s32(__p1, __p2));
  60103.   return __ret;
  60104. }
  60105. #else
  60106. __ai int16x8_t vsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  60107.   int16x8_t __ret;
  60108.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60109.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60110.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  60111.   __ret = __noswap_vcombine_s16(__rev0, __noswap_vsubhn_s32(__rev1, __rev2));
  60112.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60113.   return __ret;
  60114. }
  60115. #endif
  60116.  
  60117. #ifdef __LITTLE_ENDIAN__
  60118. __ai int32x4_t vsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  60119.   int32x4_t __ret;
  60120.   __ret = vcombine_s32(__p0, vsubhn_s64(__p1, __p2));
  60121.   return __ret;
  60122. }
  60123. #else
  60124. __ai int32x4_t vsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  60125.   int32x4_t __ret;
  60126.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60127.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60128.   int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  60129.   __ret = __noswap_vcombine_s32(__rev0, __noswap_vsubhn_s64(__rev1, __rev2));
  60130.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60131.   return __ret;
  60132. }
  60133. #endif
  60134.  
  60135. #ifdef __LITTLE_ENDIAN__
  60136. __ai int8x16_t vsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  60137.   int8x16_t __ret;
  60138.   __ret = vcombine_s8(__p0, vsubhn_s16(__p1, __p2));
  60139.   return __ret;
  60140. }
  60141. #else
  60142. __ai int8x16_t vsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  60143.   int8x16_t __ret;
  60144.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60145.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60146.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  60147.   __ret = __noswap_vcombine_s8(__rev0, __noswap_vsubhn_s16(__rev1, __rev2));
  60148.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60149.   return __ret;
  60150. }
  60151. #endif
  60152.  
  60153. #ifdef __LITTLE_ENDIAN__
  60154. __ai uint16x8_t vsubl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
  60155.   uint16x8_t __ret;
  60156.   __ret = vmovl_high_u8(__p0) - vmovl_high_u8(__p1);
  60157.   return __ret;
  60158. }
  60159. #else
  60160. __ai uint16x8_t vsubl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
  60161.   uint16x8_t __ret;
  60162.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60163.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60164.   __ret = __noswap_vmovl_high_u8(__rev0) - __noswap_vmovl_high_u8(__rev1);
  60165.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60166.   return __ret;
  60167. }
  60168. #endif
  60169.  
  60170. #ifdef __LITTLE_ENDIAN__
  60171. __ai uint64x2_t vsubl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
  60172.   uint64x2_t __ret;
  60173.   __ret = vmovl_high_u32(__p0) - vmovl_high_u32(__p1);
  60174.   return __ret;
  60175. }
  60176. #else
  60177. __ai uint64x2_t vsubl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
  60178.   uint64x2_t __ret;
  60179.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60180.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60181.   __ret = __noswap_vmovl_high_u32(__rev0) - __noswap_vmovl_high_u32(__rev1);
  60182.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60183.   return __ret;
  60184. }
  60185. #endif
  60186.  
  60187. #ifdef __LITTLE_ENDIAN__
  60188. __ai uint32x4_t vsubl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
  60189.   uint32x4_t __ret;
  60190.   __ret = vmovl_high_u16(__p0) - vmovl_high_u16(__p1);
  60191.   return __ret;
  60192. }
  60193. #else
  60194. __ai uint32x4_t vsubl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
  60195.   uint32x4_t __ret;
  60196.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60197.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60198.   __ret = __noswap_vmovl_high_u16(__rev0) - __noswap_vmovl_high_u16(__rev1);
  60199.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60200.   return __ret;
  60201. }
  60202. #endif
  60203.  
  60204. #ifdef __LITTLE_ENDIAN__
  60205. __ai int16x8_t vsubl_high_s8(int8x16_t __p0, int8x16_t __p1) {
  60206.   int16x8_t __ret;
  60207.   __ret = vmovl_high_s8(__p0) - vmovl_high_s8(__p1);
  60208.   return __ret;
  60209. }
  60210. #else
  60211. __ai int16x8_t vsubl_high_s8(int8x16_t __p0, int8x16_t __p1) {
  60212.   int16x8_t __ret;
  60213.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60214.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60215.   __ret = __noswap_vmovl_high_s8(__rev0) - __noswap_vmovl_high_s8(__rev1);
  60216.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60217.   return __ret;
  60218. }
  60219. #endif
  60220.  
  60221. #ifdef __LITTLE_ENDIAN__
  60222. __ai int64x2_t vsubl_high_s32(int32x4_t __p0, int32x4_t __p1) {
  60223.   int64x2_t __ret;
  60224.   __ret = vmovl_high_s32(__p0) - vmovl_high_s32(__p1);
  60225.   return __ret;
  60226. }
  60227. #else
  60228. __ai int64x2_t vsubl_high_s32(int32x4_t __p0, int32x4_t __p1) {
  60229.   int64x2_t __ret;
  60230.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60231.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60232.   __ret = __noswap_vmovl_high_s32(__rev0) - __noswap_vmovl_high_s32(__rev1);
  60233.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60234.   return __ret;
  60235. }
  60236. #endif
  60237.  
  60238. #ifdef __LITTLE_ENDIAN__
  60239. __ai int32x4_t vsubl_high_s16(int16x8_t __p0, int16x8_t __p1) {
  60240.   int32x4_t __ret;
  60241.   __ret = vmovl_high_s16(__p0) - vmovl_high_s16(__p1);
  60242.   return __ret;
  60243. }
  60244. #else
  60245. __ai int32x4_t vsubl_high_s16(int16x8_t __p0, int16x8_t __p1) {
  60246.   int32x4_t __ret;
  60247.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60248.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60249.   __ret = __noswap_vmovl_high_s16(__rev0) - __noswap_vmovl_high_s16(__rev1);
  60250.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60251.   return __ret;
  60252. }
  60253. #endif
  60254.  
  60255. #ifdef __LITTLE_ENDIAN__
  60256. __ai uint16x8_t vsubw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
  60257.   uint16x8_t __ret;
  60258.   __ret = __p0 - vmovl_high_u8(__p1);
  60259.   return __ret;
  60260. }
  60261. #else
  60262. __ai uint16x8_t vsubw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
  60263.   uint16x8_t __ret;
  60264.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60265.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60266.   __ret = __rev0 - __noswap_vmovl_high_u8(__rev1);
  60267.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60268.   return __ret;
  60269. }
  60270. #endif
  60271.  
  60272. #ifdef __LITTLE_ENDIAN__
  60273. __ai uint64x2_t vsubw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
  60274.   uint64x2_t __ret;
  60275.   __ret = __p0 - vmovl_high_u32(__p1);
  60276.   return __ret;
  60277. }
  60278. #else
  60279. __ai uint64x2_t vsubw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
  60280.   uint64x2_t __ret;
  60281.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60282.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60283.   __ret = __rev0 - __noswap_vmovl_high_u32(__rev1);
  60284.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60285.   return __ret;
  60286. }
  60287. #endif
  60288.  
  60289. #ifdef __LITTLE_ENDIAN__
  60290. __ai uint32x4_t vsubw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
  60291.   uint32x4_t __ret;
  60292.   __ret = __p0 - vmovl_high_u16(__p1);
  60293.   return __ret;
  60294. }
  60295. #else
  60296. __ai uint32x4_t vsubw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
  60297.   uint32x4_t __ret;
  60298.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60299.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60300.   __ret = __rev0 - __noswap_vmovl_high_u16(__rev1);
  60301.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60302.   return __ret;
  60303. }
  60304. #endif
  60305.  
  60306. #ifdef __LITTLE_ENDIAN__
  60307. __ai int16x8_t vsubw_high_s8(int16x8_t __p0, int8x16_t __p1) {
  60308.   int16x8_t __ret;
  60309.   __ret = __p0 - vmovl_high_s8(__p1);
  60310.   return __ret;
  60311. }
  60312. #else
  60313. __ai int16x8_t vsubw_high_s8(int16x8_t __p0, int8x16_t __p1) {
  60314.   int16x8_t __ret;
  60315.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60316.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60317.   __ret = __rev0 - __noswap_vmovl_high_s8(__rev1);
  60318.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60319.   return __ret;
  60320. }
  60321. #endif
  60322.  
  60323. #ifdef __LITTLE_ENDIAN__
  60324. __ai int64x2_t vsubw_high_s32(int64x2_t __p0, int32x4_t __p1) {
  60325.   int64x2_t __ret;
  60326.   __ret = __p0 - vmovl_high_s32(__p1);
  60327.   return __ret;
  60328. }
  60329. #else
  60330. __ai int64x2_t vsubw_high_s32(int64x2_t __p0, int32x4_t __p1) {
  60331.   int64x2_t __ret;
  60332.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60333.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60334.   __ret = __rev0 - __noswap_vmovl_high_s32(__rev1);
  60335.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60336.   return __ret;
  60337. }
  60338. #endif
  60339.  
  60340. #ifdef __LITTLE_ENDIAN__
  60341. __ai int32x4_t vsubw_high_s16(int32x4_t __p0, int16x8_t __p1) {
  60342.   int32x4_t __ret;
  60343.   __ret = __p0 - vmovl_high_s16(__p1);
  60344.   return __ret;
  60345. }
  60346. #else
  60347. __ai int32x4_t vsubw_high_s16(int32x4_t __p0, int16x8_t __p1) {
  60348.   int32x4_t __ret;
  60349.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60350.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60351.   __ret = __rev0 - __noswap_vmovl_high_s16(__rev1);
  60352.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60353.   return __ret;
  60354. }
  60355. #endif
  60356.  
  60357. #ifdef __LITTLE_ENDIAN__
  60358. __ai poly8x8_t vtrn1_p8(poly8x8_t __p0, poly8x8_t __p1) {
  60359.   poly8x8_t __ret;
  60360.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
  60361.   return __ret;
  60362. }
  60363. #else
  60364. __ai poly8x8_t vtrn1_p8(poly8x8_t __p0, poly8x8_t __p1) {
  60365.   poly8x8_t __ret;
  60366.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60367.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60368.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
  60369.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60370.   return __ret;
  60371. }
  60372. #endif
  60373.  
  60374. #ifdef __LITTLE_ENDIAN__
  60375. __ai poly16x4_t vtrn1_p16(poly16x4_t __p0, poly16x4_t __p1) {
  60376.   poly16x4_t __ret;
  60377.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
  60378.   return __ret;
  60379. }
  60380. #else
  60381. __ai poly16x4_t vtrn1_p16(poly16x4_t __p0, poly16x4_t __p1) {
  60382.   poly16x4_t __ret;
  60383.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60384.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60385.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
  60386.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60387.   return __ret;
  60388. }
  60389. #endif
  60390.  
  60391. #ifdef __LITTLE_ENDIAN__
  60392. __ai poly8x16_t vtrn1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  60393.   poly8x16_t __ret;
  60394.   __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
  60395.   return __ret;
  60396. }
  60397. #else
  60398. __ai poly8x16_t vtrn1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  60399.   poly8x16_t __ret;
  60400.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60401.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60402.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
  60403.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60404.   return __ret;
  60405. }
  60406. #endif
  60407.  
  60408. #ifdef __LITTLE_ENDIAN__
  60409. __ai poly64x2_t vtrn1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  60410.   poly64x2_t __ret;
  60411.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  60412.   return __ret;
  60413. }
  60414. #else
  60415. __ai poly64x2_t vtrn1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  60416.   poly64x2_t __ret;
  60417.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60418.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60419.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  60420.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60421.   return __ret;
  60422. }
  60423. #endif
  60424.  
  60425. #ifdef __LITTLE_ENDIAN__
  60426. __ai poly16x8_t vtrn1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  60427.   poly16x8_t __ret;
  60428.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
  60429.   return __ret;
  60430. }
  60431. #else
  60432. __ai poly16x8_t vtrn1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  60433.   poly16x8_t __ret;
  60434.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60435.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60436.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
  60437.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60438.   return __ret;
  60439. }
  60440. #endif
  60441.  
  60442. #ifdef __LITTLE_ENDIAN__
  60443. __ai uint8x16_t vtrn1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  60444.   uint8x16_t __ret;
  60445.   __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
  60446.   return __ret;
  60447. }
  60448. #else
  60449. __ai uint8x16_t vtrn1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  60450.   uint8x16_t __ret;
  60451.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60452.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60453.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
  60454.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60455.   return __ret;
  60456. }
  60457. #endif
  60458.  
  60459. #ifdef __LITTLE_ENDIAN__
  60460. __ai uint32x4_t vtrn1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  60461.   uint32x4_t __ret;
  60462.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
  60463.   return __ret;
  60464. }
  60465. #else
  60466. __ai uint32x4_t vtrn1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  60467.   uint32x4_t __ret;
  60468.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60469.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60470.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
  60471.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60472.   return __ret;
  60473. }
  60474. #endif
  60475.  
  60476. #ifdef __LITTLE_ENDIAN__
  60477. __ai uint64x2_t vtrn1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  60478.   uint64x2_t __ret;
  60479.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  60480.   return __ret;
  60481. }
  60482. #else
  60483. __ai uint64x2_t vtrn1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  60484.   uint64x2_t __ret;
  60485.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60486.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60487.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  60488.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60489.   return __ret;
  60490. }
  60491. #endif
  60492.  
  60493. #ifdef __LITTLE_ENDIAN__
  60494. __ai uint16x8_t vtrn1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  60495.   uint16x8_t __ret;
  60496.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
  60497.   return __ret;
  60498. }
  60499. #else
  60500. __ai uint16x8_t vtrn1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  60501.   uint16x8_t __ret;
  60502.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60503.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60504.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
  60505.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60506.   return __ret;
  60507. }
  60508. #endif
  60509.  
  60510. #ifdef __LITTLE_ENDIAN__
  60511. __ai int8x16_t vtrn1q_s8(int8x16_t __p0, int8x16_t __p1) {
  60512.   int8x16_t __ret;
  60513.   __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
  60514.   return __ret;
  60515. }
  60516. #else
  60517. __ai int8x16_t vtrn1q_s8(int8x16_t __p0, int8x16_t __p1) {
  60518.   int8x16_t __ret;
  60519.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60520.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60521.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
  60522.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60523.   return __ret;
  60524. }
  60525. #endif
  60526.  
  60527. #ifdef __LITTLE_ENDIAN__
  60528. __ai float64x2_t vtrn1q_f64(float64x2_t __p0, float64x2_t __p1) {
  60529.   float64x2_t __ret;
  60530.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  60531.   return __ret;
  60532. }
  60533. #else
  60534. __ai float64x2_t vtrn1q_f64(float64x2_t __p0, float64x2_t __p1) {
  60535.   float64x2_t __ret;
  60536.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60537.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60538.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  60539.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60540.   return __ret;
  60541. }
  60542. #endif
  60543.  
  60544. #ifdef __LITTLE_ENDIAN__
  60545. __ai float32x4_t vtrn1q_f32(float32x4_t __p0, float32x4_t __p1) {
  60546.   float32x4_t __ret;
  60547.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
  60548.   return __ret;
  60549. }
  60550. #else
  60551. __ai float32x4_t vtrn1q_f32(float32x4_t __p0, float32x4_t __p1) {
  60552.   float32x4_t __ret;
  60553.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60554.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60555.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
  60556.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60557.   return __ret;
  60558. }
  60559. #endif
  60560.  
  60561. #ifdef __LITTLE_ENDIAN__
  60562. __ai int32x4_t vtrn1q_s32(int32x4_t __p0, int32x4_t __p1) {
  60563.   int32x4_t __ret;
  60564.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
  60565.   return __ret;
  60566. }
  60567. #else
  60568. __ai int32x4_t vtrn1q_s32(int32x4_t __p0, int32x4_t __p1) {
  60569.   int32x4_t __ret;
  60570.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60571.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60572.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
  60573.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60574.   return __ret;
  60575. }
  60576. #endif
  60577.  
  60578. #ifdef __LITTLE_ENDIAN__
  60579. __ai int64x2_t vtrn1q_s64(int64x2_t __p0, int64x2_t __p1) {
  60580.   int64x2_t __ret;
  60581.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  60582.   return __ret;
  60583. }
  60584. #else
  60585. __ai int64x2_t vtrn1q_s64(int64x2_t __p0, int64x2_t __p1) {
  60586.   int64x2_t __ret;
  60587.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60588.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60589.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  60590.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60591.   return __ret;
  60592. }
  60593. #endif
  60594.  
  60595. #ifdef __LITTLE_ENDIAN__
  60596. __ai int16x8_t vtrn1q_s16(int16x8_t __p0, int16x8_t __p1) {
  60597.   int16x8_t __ret;
  60598.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
  60599.   return __ret;
  60600. }
  60601. #else
  60602. __ai int16x8_t vtrn1q_s16(int16x8_t __p0, int16x8_t __p1) {
  60603.   int16x8_t __ret;
  60604.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60605.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60606.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
  60607.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60608.   return __ret;
  60609. }
  60610. #endif
  60611.  
  60612. #ifdef __LITTLE_ENDIAN__
  60613. __ai uint8x8_t vtrn1_u8(uint8x8_t __p0, uint8x8_t __p1) {
  60614.   uint8x8_t __ret;
  60615.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
  60616.   return __ret;
  60617. }
  60618. #else
  60619. __ai uint8x8_t vtrn1_u8(uint8x8_t __p0, uint8x8_t __p1) {
  60620.   uint8x8_t __ret;
  60621.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60622.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60623.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
  60624.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60625.   return __ret;
  60626. }
  60627. #endif
  60628.  
  60629. #ifdef __LITTLE_ENDIAN__
  60630. __ai uint32x2_t vtrn1_u32(uint32x2_t __p0, uint32x2_t __p1) {
  60631.   uint32x2_t __ret;
  60632.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  60633.   return __ret;
  60634. }
  60635. #else
  60636. __ai uint32x2_t vtrn1_u32(uint32x2_t __p0, uint32x2_t __p1) {
  60637.   uint32x2_t __ret;
  60638.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60639.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60640.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  60641.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60642.   return __ret;
  60643. }
  60644. #endif
  60645.  
  60646. #ifdef __LITTLE_ENDIAN__
  60647. __ai uint16x4_t vtrn1_u16(uint16x4_t __p0, uint16x4_t __p1) {
  60648.   uint16x4_t __ret;
  60649.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
  60650.   return __ret;
  60651. }
  60652. #else
  60653. __ai uint16x4_t vtrn1_u16(uint16x4_t __p0, uint16x4_t __p1) {
  60654.   uint16x4_t __ret;
  60655.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60656.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60657.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
  60658.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60659.   return __ret;
  60660. }
  60661. #endif
  60662.  
  60663. #ifdef __LITTLE_ENDIAN__
  60664. __ai int8x8_t vtrn1_s8(int8x8_t __p0, int8x8_t __p1) {
  60665.   int8x8_t __ret;
  60666.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
  60667.   return __ret;
  60668. }
  60669. #else
  60670. __ai int8x8_t vtrn1_s8(int8x8_t __p0, int8x8_t __p1) {
  60671.   int8x8_t __ret;
  60672.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60673.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60674.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
  60675.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60676.   return __ret;
  60677. }
  60678. #endif
  60679.  
  60680. #ifdef __LITTLE_ENDIAN__
  60681. __ai float32x2_t vtrn1_f32(float32x2_t __p0, float32x2_t __p1) {
  60682.   float32x2_t __ret;
  60683.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  60684.   return __ret;
  60685. }
  60686. #else
  60687. __ai float32x2_t vtrn1_f32(float32x2_t __p0, float32x2_t __p1) {
  60688.   float32x2_t __ret;
  60689.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60690.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60691.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  60692.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60693.   return __ret;
  60694. }
  60695. #endif
  60696.  
  60697. #ifdef __LITTLE_ENDIAN__
  60698. __ai int32x2_t vtrn1_s32(int32x2_t __p0, int32x2_t __p1) {
  60699.   int32x2_t __ret;
  60700.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  60701.   return __ret;
  60702. }
  60703. #else
  60704. __ai int32x2_t vtrn1_s32(int32x2_t __p0, int32x2_t __p1) {
  60705.   int32x2_t __ret;
  60706.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60707.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60708.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  60709.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60710.   return __ret;
  60711. }
  60712. #endif
  60713.  
  60714. #ifdef __LITTLE_ENDIAN__
  60715. __ai int16x4_t vtrn1_s16(int16x4_t __p0, int16x4_t __p1) {
  60716.   int16x4_t __ret;
  60717.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
  60718.   return __ret;
  60719. }
  60720. #else
  60721. __ai int16x4_t vtrn1_s16(int16x4_t __p0, int16x4_t __p1) {
  60722.   int16x4_t __ret;
  60723.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60724.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60725.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
  60726.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60727.   return __ret;
  60728. }
  60729. #endif
  60730.  
  60731. #ifdef __LITTLE_ENDIAN__
  60732. __ai poly8x8_t vtrn2_p8(poly8x8_t __p0, poly8x8_t __p1) {
  60733.   poly8x8_t __ret;
  60734.   __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
  60735.   return __ret;
  60736. }
  60737. #else
  60738. __ai poly8x8_t vtrn2_p8(poly8x8_t __p0, poly8x8_t __p1) {
  60739.   poly8x8_t __ret;
  60740.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60741.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60742.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
  60743.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60744.   return __ret;
  60745. }
  60746. #endif
  60747.  
  60748. #ifdef __LITTLE_ENDIAN__
  60749. __ai poly16x4_t vtrn2_p16(poly16x4_t __p0, poly16x4_t __p1) {
  60750.   poly16x4_t __ret;
  60751.   __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
  60752.   return __ret;
  60753. }
  60754. #else
  60755. __ai poly16x4_t vtrn2_p16(poly16x4_t __p0, poly16x4_t __p1) {
  60756.   poly16x4_t __ret;
  60757.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60758.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60759.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
  60760.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60761.   return __ret;
  60762. }
  60763. #endif
  60764.  
  60765. #ifdef __LITTLE_ENDIAN__
  60766. __ai poly8x16_t vtrn2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  60767.   poly8x16_t __ret;
  60768.   __ret = __builtin_shufflevector(__p0, __p1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
  60769.   return __ret;
  60770. }
  60771. #else
  60772. __ai poly8x16_t vtrn2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  60773.   poly8x16_t __ret;
  60774.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60775.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60776.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
  60777.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60778.   return __ret;
  60779. }
  60780. #endif
  60781.  
  60782. #ifdef __LITTLE_ENDIAN__
  60783. __ai poly64x2_t vtrn2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  60784.   poly64x2_t __ret;
  60785.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  60786.   return __ret;
  60787. }
  60788. #else
  60789. __ai poly64x2_t vtrn2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  60790.   poly64x2_t __ret;
  60791.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60792.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60793.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  60794.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60795.   return __ret;
  60796. }
  60797. #endif
  60798.  
  60799. #ifdef __LITTLE_ENDIAN__
  60800. __ai poly16x8_t vtrn2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  60801.   poly16x8_t __ret;
  60802.   __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
  60803.   return __ret;
  60804. }
  60805. #else
  60806. __ai poly16x8_t vtrn2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  60807.   poly16x8_t __ret;
  60808.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60809.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60810.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
  60811.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60812.   return __ret;
  60813. }
  60814. #endif
  60815.  
  60816. #ifdef __LITTLE_ENDIAN__
  60817. __ai uint8x16_t vtrn2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  60818.   uint8x16_t __ret;
  60819.   __ret = __builtin_shufflevector(__p0, __p1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
  60820.   return __ret;
  60821. }
  60822. #else
  60823. __ai uint8x16_t vtrn2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  60824.   uint8x16_t __ret;
  60825.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60826.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60827.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
  60828.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60829.   return __ret;
  60830. }
  60831. #endif
  60832.  
  60833. #ifdef __LITTLE_ENDIAN__
  60834. __ai uint32x4_t vtrn2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  60835.   uint32x4_t __ret;
  60836.   __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
  60837.   return __ret;
  60838. }
  60839. #else
  60840. __ai uint32x4_t vtrn2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  60841.   uint32x4_t __ret;
  60842.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60843.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60844.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
  60845.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60846.   return __ret;
  60847. }
  60848. #endif
  60849.  
  60850. #ifdef __LITTLE_ENDIAN__
  60851. __ai uint64x2_t vtrn2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  60852.   uint64x2_t __ret;
  60853.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  60854.   return __ret;
  60855. }
  60856. #else
  60857. __ai uint64x2_t vtrn2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  60858.   uint64x2_t __ret;
  60859.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60860.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60861.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  60862.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60863.   return __ret;
  60864. }
  60865. #endif
  60866.  
  60867. #ifdef __LITTLE_ENDIAN__
  60868. __ai uint16x8_t vtrn2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  60869.   uint16x8_t __ret;
  60870.   __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
  60871.   return __ret;
  60872. }
  60873. #else
  60874. __ai uint16x8_t vtrn2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  60875.   uint16x8_t __ret;
  60876.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60877.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60878.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
  60879.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60880.   return __ret;
  60881. }
  60882. #endif
  60883.  
  60884. #ifdef __LITTLE_ENDIAN__
  60885. __ai int8x16_t vtrn2q_s8(int8x16_t __p0, int8x16_t __p1) {
  60886.   int8x16_t __ret;
  60887.   __ret = __builtin_shufflevector(__p0, __p1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
  60888.   return __ret;
  60889. }
  60890. #else
  60891. __ai int8x16_t vtrn2q_s8(int8x16_t __p0, int8x16_t __p1) {
  60892.   int8x16_t __ret;
  60893.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60894.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60895.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
  60896.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  60897.   return __ret;
  60898. }
  60899. #endif
  60900.  
  60901. #ifdef __LITTLE_ENDIAN__
  60902. __ai float64x2_t vtrn2q_f64(float64x2_t __p0, float64x2_t __p1) {
  60903.   float64x2_t __ret;
  60904.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  60905.   return __ret;
  60906. }
  60907. #else
  60908. __ai float64x2_t vtrn2q_f64(float64x2_t __p0, float64x2_t __p1) {
  60909.   float64x2_t __ret;
  60910.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60911.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60912.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  60913.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60914.   return __ret;
  60915. }
  60916. #endif
  60917.  
  60918. #ifdef __LITTLE_ENDIAN__
  60919. __ai float32x4_t vtrn2q_f32(float32x4_t __p0, float32x4_t __p1) {
  60920.   float32x4_t __ret;
  60921.   __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
  60922.   return __ret;
  60923. }
  60924. #else
  60925. __ai float32x4_t vtrn2q_f32(float32x4_t __p0, float32x4_t __p1) {
  60926.   float32x4_t __ret;
  60927.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60928.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60929.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
  60930.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60931.   return __ret;
  60932. }
  60933. #endif
  60934.  
  60935. #ifdef __LITTLE_ENDIAN__
  60936. __ai int32x4_t vtrn2q_s32(int32x4_t __p0, int32x4_t __p1) {
  60937.   int32x4_t __ret;
  60938.   __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
  60939.   return __ret;
  60940. }
  60941. #else
  60942. __ai int32x4_t vtrn2q_s32(int32x4_t __p0, int32x4_t __p1) {
  60943.   int32x4_t __ret;
  60944.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  60945.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  60946.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
  60947.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  60948.   return __ret;
  60949. }
  60950. #endif
  60951.  
  60952. #ifdef __LITTLE_ENDIAN__
  60953. __ai int64x2_t vtrn2q_s64(int64x2_t __p0, int64x2_t __p1) {
  60954.   int64x2_t __ret;
  60955.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  60956.   return __ret;
  60957. }
  60958. #else
  60959. __ai int64x2_t vtrn2q_s64(int64x2_t __p0, int64x2_t __p1) {
  60960.   int64x2_t __ret;
  60961.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  60962.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  60963.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  60964.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  60965.   return __ret;
  60966. }
  60967. #endif
  60968.  
  60969. #ifdef __LITTLE_ENDIAN__
  60970. __ai int16x8_t vtrn2q_s16(int16x8_t __p0, int16x8_t __p1) {
  60971.   int16x8_t __ret;
  60972.   __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
  60973.   return __ret;
  60974. }
  60975. #else
  60976. __ai int16x8_t vtrn2q_s16(int16x8_t __p0, int16x8_t __p1) {
  60977.   int16x8_t __ret;
  60978.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60979.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60980.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
  60981.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60982.   return __ret;
  60983. }
  60984. #endif
  60985.  
  60986. #ifdef __LITTLE_ENDIAN__
  60987. __ai uint8x8_t vtrn2_u8(uint8x8_t __p0, uint8x8_t __p1) {
  60988.   uint8x8_t __ret;
  60989.   __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
  60990.   return __ret;
  60991. }
  60992. #else
  60993. __ai uint8x8_t vtrn2_u8(uint8x8_t __p0, uint8x8_t __p1) {
  60994.   uint8x8_t __ret;
  60995.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  60996.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  60997.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
  60998.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  60999.   return __ret;
  61000. }
  61001. #endif
  61002.  
  61003. #ifdef __LITTLE_ENDIAN__
  61004. __ai uint32x2_t vtrn2_u32(uint32x2_t __p0, uint32x2_t __p1) {
  61005.   uint32x2_t __ret;
  61006.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  61007.   return __ret;
  61008. }
  61009. #else
  61010. __ai uint32x2_t vtrn2_u32(uint32x2_t __p0, uint32x2_t __p1) {
  61011.   uint32x2_t __ret;
  61012.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61013.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61014.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  61015.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61016.   return __ret;
  61017. }
  61018. #endif
  61019.  
  61020. #ifdef __LITTLE_ENDIAN__
  61021. __ai uint16x4_t vtrn2_u16(uint16x4_t __p0, uint16x4_t __p1) {
  61022.   uint16x4_t __ret;
  61023.   __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
  61024.   return __ret;
  61025. }
  61026. #else
  61027. __ai uint16x4_t vtrn2_u16(uint16x4_t __p0, uint16x4_t __p1) {
  61028.   uint16x4_t __ret;
  61029.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61030.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61031.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
  61032.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61033.   return __ret;
  61034. }
  61035. #endif
  61036.  
  61037. #ifdef __LITTLE_ENDIAN__
  61038. __ai int8x8_t vtrn2_s8(int8x8_t __p0, int8x8_t __p1) {
  61039.   int8x8_t __ret;
  61040.   __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
  61041.   return __ret;
  61042. }
  61043. #else
  61044. __ai int8x8_t vtrn2_s8(int8x8_t __p0, int8x8_t __p1) {
  61045.   int8x8_t __ret;
  61046.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61047.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61048.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
  61049.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61050.   return __ret;
  61051. }
  61052. #endif
  61053.  
  61054. #ifdef __LITTLE_ENDIAN__
  61055. __ai float32x2_t vtrn2_f32(float32x2_t __p0, float32x2_t __p1) {
  61056.   float32x2_t __ret;
  61057.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  61058.   return __ret;
  61059. }
  61060. #else
  61061. __ai float32x2_t vtrn2_f32(float32x2_t __p0, float32x2_t __p1) {
  61062.   float32x2_t __ret;
  61063.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61064.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61065.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  61066.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61067.   return __ret;
  61068. }
  61069. #endif
  61070.  
  61071. #ifdef __LITTLE_ENDIAN__
  61072. __ai int32x2_t vtrn2_s32(int32x2_t __p0, int32x2_t __p1) {
  61073.   int32x2_t __ret;
  61074.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  61075.   return __ret;
  61076. }
  61077. #else
  61078. __ai int32x2_t vtrn2_s32(int32x2_t __p0, int32x2_t __p1) {
  61079.   int32x2_t __ret;
  61080.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61081.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61082.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  61083.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61084.   return __ret;
  61085. }
  61086. #endif
  61087.  
  61088. #ifdef __LITTLE_ENDIAN__
  61089. __ai int16x4_t vtrn2_s16(int16x4_t __p0, int16x4_t __p1) {
  61090.   int16x4_t __ret;
  61091.   __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
  61092.   return __ret;
  61093. }
  61094. #else
  61095. __ai int16x4_t vtrn2_s16(int16x4_t __p0, int16x4_t __p1) {
  61096.   int16x4_t __ret;
  61097.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61098.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61099.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
  61100.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61101.   return __ret;
  61102. }
  61103. #endif
  61104.  
  61105. __ai uint64x1_t vtst_p64(poly64x1_t __p0, poly64x1_t __p1) {
  61106.   uint64x1_t __ret;
  61107.   __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  61108.   return __ret;
  61109. }
  61110. #ifdef __LITTLE_ENDIAN__
  61111. __ai uint64x2_t vtstq_p64(poly64x2_t __p0, poly64x2_t __p1) {
  61112.   uint64x2_t __ret;
  61113.   __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  61114.   return __ret;
  61115. }
  61116. #else
  61117. __ai uint64x2_t vtstq_p64(poly64x2_t __p0, poly64x2_t __p1) {
  61118.   uint64x2_t __ret;
  61119.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61120.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61121.   __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  61122.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61123.   return __ret;
  61124. }
  61125. #endif
  61126.  
  61127. #ifdef __LITTLE_ENDIAN__
  61128. __ai uint64x2_t vtstq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  61129.   uint64x2_t __ret;
  61130.   __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  61131.   return __ret;
  61132. }
  61133. #else
  61134. __ai uint64x2_t vtstq_u64(uint64x2_t __p0, uint64x2_t __p1) {
  61135.   uint64x2_t __ret;
  61136.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61137.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61138.   __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  61139.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61140.   return __ret;
  61141. }
  61142. #endif
  61143.  
  61144. #ifdef __LITTLE_ENDIAN__
  61145. __ai uint64x2_t vtstq_s64(int64x2_t __p0, int64x2_t __p1) {
  61146.   uint64x2_t __ret;
  61147.   __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
  61148.   return __ret;
  61149. }
  61150. #else
  61151. __ai uint64x2_t vtstq_s64(int64x2_t __p0, int64x2_t __p1) {
  61152.   uint64x2_t __ret;
  61153.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61154.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61155.   __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  61156.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61157.   return __ret;
  61158. }
  61159. #endif
  61160.  
  61161. __ai uint64x1_t vtst_u64(uint64x1_t __p0, uint64x1_t __p1) {
  61162.   uint64x1_t __ret;
  61163.   __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  61164.   return __ret;
  61165. }
  61166. __ai uint64x1_t vtst_s64(int64x1_t __p0, int64x1_t __p1) {
  61167.   uint64x1_t __ret;
  61168.   __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
  61169.   return __ret;
  61170. }
  61171. __ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) {
  61172.   uint64_t __ret;
  61173.   __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1);
  61174.   return __ret;
  61175. }
  61176. __ai uint64_t vtstd_s64(int64_t __p0, int64_t __p1) {
  61177.   uint64_t __ret;
  61178.   __ret = (uint64_t) __builtin_neon_vtstd_s64(__p0, __p1);
  61179.   return __ret;
  61180. }
  61181. __ai int8_t vuqaddb_s8(int8_t __p0, uint8_t __p1) {
  61182.   int8_t __ret;
  61183.   __ret = (int8_t) __builtin_neon_vuqaddb_s8(__p0, __p1);
  61184.   return __ret;
  61185. }
  61186. __ai int32_t vuqadds_s32(int32_t __p0, uint32_t __p1) {
  61187.   int32_t __ret;
  61188.   __ret = (int32_t) __builtin_neon_vuqadds_s32(__p0, __p1);
  61189.   return __ret;
  61190. }
  61191. __ai int64_t vuqaddd_s64(int64_t __p0, uint64_t __p1) {
  61192.   int64_t __ret;
  61193.   __ret = (int64_t) __builtin_neon_vuqaddd_s64(__p0, __p1);
  61194.   return __ret;
  61195. }
  61196. __ai int16_t vuqaddh_s16(int16_t __p0, uint16_t __p1) {
  61197.   int16_t __ret;
  61198.   __ret = (int16_t) __builtin_neon_vuqaddh_s16(__p0, __p1);
  61199.   return __ret;
  61200. }
  61201. #ifdef __LITTLE_ENDIAN__
  61202. __ai int8x16_t vuqaddq_s8(int8x16_t __p0, uint8x16_t __p1) {
  61203.   int8x16_t __ret;
  61204.   __ret = (int8x16_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
  61205.   return __ret;
  61206. }
  61207. #else
  61208. __ai int8x16_t vuqaddq_s8(int8x16_t __p0, uint8x16_t __p1) {
  61209.   int8x16_t __ret;
  61210.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61211.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61212.   __ret = (int8x16_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
  61213.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61214.   return __ret;
  61215. }
  61216. #endif
  61217.  
  61218. #ifdef __LITTLE_ENDIAN__
  61219. __ai int32x4_t vuqaddq_s32(int32x4_t __p0, uint32x4_t __p1) {
  61220.   int32x4_t __ret;
  61221.   __ret = (int32x4_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
  61222.   return __ret;
  61223. }
  61224. #else
  61225. __ai int32x4_t vuqaddq_s32(int32x4_t __p0, uint32x4_t __p1) {
  61226.   int32x4_t __ret;
  61227.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61228.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61229.   __ret = (int32x4_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
  61230.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61231.   return __ret;
  61232. }
  61233. #endif
  61234.  
  61235. #ifdef __LITTLE_ENDIAN__
  61236. __ai int64x2_t vuqaddq_s64(int64x2_t __p0, uint64x2_t __p1) {
  61237.   int64x2_t __ret;
  61238.   __ret = (int64x2_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
  61239.   return __ret;
  61240. }
  61241. #else
  61242. __ai int64x2_t vuqaddq_s64(int64x2_t __p0, uint64x2_t __p1) {
  61243.   int64x2_t __ret;
  61244.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61245.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61246.   __ret = (int64x2_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
  61247.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61248.   return __ret;
  61249. }
  61250. #endif
  61251.  
  61252. #ifdef __LITTLE_ENDIAN__
  61253. __ai int16x8_t vuqaddq_s16(int16x8_t __p0, uint16x8_t __p1) {
  61254.   int16x8_t __ret;
  61255.   __ret = (int16x8_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
  61256.   return __ret;
  61257. }
  61258. #else
  61259. __ai int16x8_t vuqaddq_s16(int16x8_t __p0, uint16x8_t __p1) {
  61260.   int16x8_t __ret;
  61261.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61262.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61263.   __ret = (int16x8_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
  61264.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61265.   return __ret;
  61266. }
  61267. #endif
  61268.  
  61269. #ifdef __LITTLE_ENDIAN__
  61270. __ai int8x8_t vuqadd_s8(int8x8_t __p0, uint8x8_t __p1) {
  61271.   int8x8_t __ret;
  61272.   __ret = (int8x8_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
  61273.   return __ret;
  61274. }
  61275. #else
  61276. __ai int8x8_t vuqadd_s8(int8x8_t __p0, uint8x8_t __p1) {
  61277.   int8x8_t __ret;
  61278.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61279.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61280.   __ret = (int8x8_t) __builtin_neon_vuqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
  61281.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61282.   return __ret;
  61283. }
  61284. #endif
  61285.  
  61286. #ifdef __LITTLE_ENDIAN__
  61287. __ai int32x2_t vuqadd_s32(int32x2_t __p0, uint32x2_t __p1) {
  61288.   int32x2_t __ret;
  61289.   __ret = (int32x2_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
  61290.   return __ret;
  61291. }
  61292. #else
  61293. __ai int32x2_t vuqadd_s32(int32x2_t __p0, uint32x2_t __p1) {
  61294.   int32x2_t __ret;
  61295.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61296.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61297.   __ret = (int32x2_t) __builtin_neon_vuqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
  61298.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61299.   return __ret;
  61300. }
  61301. #endif
  61302.  
  61303. __ai int64x1_t vuqadd_s64(int64x1_t __p0, uint64x1_t __p1) {
  61304.   int64x1_t __ret;
  61305.   __ret = (int64x1_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
  61306.   return __ret;
  61307. }
  61308. #ifdef __LITTLE_ENDIAN__
  61309. __ai int16x4_t vuqadd_s16(int16x4_t __p0, uint16x4_t __p1) {
  61310.   int16x4_t __ret;
  61311.   __ret = (int16x4_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
  61312.   return __ret;
  61313. }
  61314. #else
  61315. __ai int16x4_t vuqadd_s16(int16x4_t __p0, uint16x4_t __p1) {
  61316.   int16x4_t __ret;
  61317.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61318.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61319.   __ret = (int16x4_t) __builtin_neon_vuqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
  61320.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61321.   return __ret;
  61322. }
  61323. #endif
  61324.  
  61325. #ifdef __LITTLE_ENDIAN__
  61326. __ai poly8x8_t vuzp1_p8(poly8x8_t __p0, poly8x8_t __p1) {
  61327.   poly8x8_t __ret;
  61328.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
  61329.   return __ret;
  61330. }
  61331. #else
  61332. __ai poly8x8_t vuzp1_p8(poly8x8_t __p0, poly8x8_t __p1) {
  61333.   poly8x8_t __ret;
  61334.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61335.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61336.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
  61337.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61338.   return __ret;
  61339. }
  61340. #endif
  61341.  
  61342. #ifdef __LITTLE_ENDIAN__
  61343. __ai poly16x4_t vuzp1_p16(poly16x4_t __p0, poly16x4_t __p1) {
  61344.   poly16x4_t __ret;
  61345.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
  61346.   return __ret;
  61347. }
  61348. #else
  61349. __ai poly16x4_t vuzp1_p16(poly16x4_t __p0, poly16x4_t __p1) {
  61350.   poly16x4_t __ret;
  61351.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61352.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61353.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
  61354.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61355.   return __ret;
  61356. }
  61357. #endif
  61358.  
  61359. #ifdef __LITTLE_ENDIAN__
  61360. __ai poly8x16_t vuzp1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  61361.   poly8x16_t __ret;
  61362.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
  61363.   return __ret;
  61364. }
  61365. #else
  61366. __ai poly8x16_t vuzp1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  61367.   poly8x16_t __ret;
  61368.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61369.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61370.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
  61371.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61372.   return __ret;
  61373. }
  61374. #endif
  61375.  
  61376. #ifdef __LITTLE_ENDIAN__
  61377. __ai poly64x2_t vuzp1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  61378.   poly64x2_t __ret;
  61379.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  61380.   return __ret;
  61381. }
  61382. #else
  61383. __ai poly64x2_t vuzp1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  61384.   poly64x2_t __ret;
  61385.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61386.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61387.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  61388.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61389.   return __ret;
  61390. }
  61391. #endif
  61392.  
  61393. #ifdef __LITTLE_ENDIAN__
  61394. __ai poly16x8_t vuzp1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  61395.   poly16x8_t __ret;
  61396.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
  61397.   return __ret;
  61398. }
  61399. #else
  61400. __ai poly16x8_t vuzp1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  61401.   poly16x8_t __ret;
  61402.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61403.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61404.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
  61405.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61406.   return __ret;
  61407. }
  61408. #endif
  61409.  
  61410. #ifdef __LITTLE_ENDIAN__
  61411. __ai uint8x16_t vuzp1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  61412.   uint8x16_t __ret;
  61413.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
  61414.   return __ret;
  61415. }
  61416. #else
  61417. __ai uint8x16_t vuzp1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  61418.   uint8x16_t __ret;
  61419.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61420.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61421.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
  61422.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61423.   return __ret;
  61424. }
  61425. #endif
  61426.  
  61427. #ifdef __LITTLE_ENDIAN__
  61428. __ai uint32x4_t vuzp1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  61429.   uint32x4_t __ret;
  61430.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
  61431.   return __ret;
  61432. }
  61433. #else
  61434. __ai uint32x4_t vuzp1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  61435.   uint32x4_t __ret;
  61436.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61437.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61438.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
  61439.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61440.   return __ret;
  61441. }
  61442. #endif
  61443.  
  61444. #ifdef __LITTLE_ENDIAN__
  61445. __ai uint64x2_t vuzp1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  61446.   uint64x2_t __ret;
  61447.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  61448.   return __ret;
  61449. }
  61450. #else
  61451. __ai uint64x2_t vuzp1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  61452.   uint64x2_t __ret;
  61453.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61454.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61455.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  61456.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61457.   return __ret;
  61458. }
  61459. #endif
  61460.  
  61461. #ifdef __LITTLE_ENDIAN__
  61462. __ai uint16x8_t vuzp1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  61463.   uint16x8_t __ret;
  61464.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
  61465.   return __ret;
  61466. }
  61467. #else
  61468. __ai uint16x8_t vuzp1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  61469.   uint16x8_t __ret;
  61470.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61471.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61472.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
  61473.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61474.   return __ret;
  61475. }
  61476. #endif
  61477.  
  61478. #ifdef __LITTLE_ENDIAN__
  61479. __ai int8x16_t vuzp1q_s8(int8x16_t __p0, int8x16_t __p1) {
  61480.   int8x16_t __ret;
  61481.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
  61482.   return __ret;
  61483. }
  61484. #else
  61485. __ai int8x16_t vuzp1q_s8(int8x16_t __p0, int8x16_t __p1) {
  61486.   int8x16_t __ret;
  61487.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61488.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61489.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
  61490.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61491.   return __ret;
  61492. }
  61493. #endif
  61494.  
  61495. #ifdef __LITTLE_ENDIAN__
  61496. __ai float64x2_t vuzp1q_f64(float64x2_t __p0, float64x2_t __p1) {
  61497.   float64x2_t __ret;
  61498.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  61499.   return __ret;
  61500. }
  61501. #else
  61502. __ai float64x2_t vuzp1q_f64(float64x2_t __p0, float64x2_t __p1) {
  61503.   float64x2_t __ret;
  61504.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61505.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61506.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  61507.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61508.   return __ret;
  61509. }
  61510. #endif
  61511.  
  61512. #ifdef __LITTLE_ENDIAN__
  61513. __ai float32x4_t vuzp1q_f32(float32x4_t __p0, float32x4_t __p1) {
  61514.   float32x4_t __ret;
  61515.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
  61516.   return __ret;
  61517. }
  61518. #else
  61519. __ai float32x4_t vuzp1q_f32(float32x4_t __p0, float32x4_t __p1) {
  61520.   float32x4_t __ret;
  61521.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61522.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61523.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
  61524.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61525.   return __ret;
  61526. }
  61527. #endif
  61528.  
  61529. #ifdef __LITTLE_ENDIAN__
  61530. __ai int32x4_t vuzp1q_s32(int32x4_t __p0, int32x4_t __p1) {
  61531.   int32x4_t __ret;
  61532.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
  61533.   return __ret;
  61534. }
  61535. #else
  61536. __ai int32x4_t vuzp1q_s32(int32x4_t __p0, int32x4_t __p1) {
  61537.   int32x4_t __ret;
  61538.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61539.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61540.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
  61541.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61542.   return __ret;
  61543. }
  61544. #endif
  61545.  
  61546. #ifdef __LITTLE_ENDIAN__
  61547. __ai int64x2_t vuzp1q_s64(int64x2_t __p0, int64x2_t __p1) {
  61548.   int64x2_t __ret;
  61549.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  61550.   return __ret;
  61551. }
  61552. #else
  61553. __ai int64x2_t vuzp1q_s64(int64x2_t __p0, int64x2_t __p1) {
  61554.   int64x2_t __ret;
  61555.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61556.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61557.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  61558.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61559.   return __ret;
  61560. }
  61561. #endif
  61562.  
  61563. #ifdef __LITTLE_ENDIAN__
  61564. __ai int16x8_t vuzp1q_s16(int16x8_t __p0, int16x8_t __p1) {
  61565.   int16x8_t __ret;
  61566.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
  61567.   return __ret;
  61568. }
  61569. #else
  61570. __ai int16x8_t vuzp1q_s16(int16x8_t __p0, int16x8_t __p1) {
  61571.   int16x8_t __ret;
  61572.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61573.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61574.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
  61575.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61576.   return __ret;
  61577. }
  61578. #endif
  61579.  
  61580. #ifdef __LITTLE_ENDIAN__
  61581. __ai uint8x8_t vuzp1_u8(uint8x8_t __p0, uint8x8_t __p1) {
  61582.   uint8x8_t __ret;
  61583.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
  61584.   return __ret;
  61585. }
  61586. #else
  61587. __ai uint8x8_t vuzp1_u8(uint8x8_t __p0, uint8x8_t __p1) {
  61588.   uint8x8_t __ret;
  61589.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61590.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61591.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
  61592.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61593.   return __ret;
  61594. }
  61595. #endif
  61596.  
  61597. #ifdef __LITTLE_ENDIAN__
  61598. __ai uint32x2_t vuzp1_u32(uint32x2_t __p0, uint32x2_t __p1) {
  61599.   uint32x2_t __ret;
  61600.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  61601.   return __ret;
  61602. }
  61603. #else
  61604. __ai uint32x2_t vuzp1_u32(uint32x2_t __p0, uint32x2_t __p1) {
  61605.   uint32x2_t __ret;
  61606.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61607.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61608.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  61609.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61610.   return __ret;
  61611. }
  61612. #endif
  61613.  
  61614. #ifdef __LITTLE_ENDIAN__
  61615. __ai uint16x4_t vuzp1_u16(uint16x4_t __p0, uint16x4_t __p1) {
  61616.   uint16x4_t __ret;
  61617.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
  61618.   return __ret;
  61619. }
  61620. #else
  61621. __ai uint16x4_t vuzp1_u16(uint16x4_t __p0, uint16x4_t __p1) {
  61622.   uint16x4_t __ret;
  61623.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61624.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61625.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
  61626.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61627.   return __ret;
  61628. }
  61629. #endif
  61630.  
  61631. #ifdef __LITTLE_ENDIAN__
  61632. __ai int8x8_t vuzp1_s8(int8x8_t __p0, int8x8_t __p1) {
  61633.   int8x8_t __ret;
  61634.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
  61635.   return __ret;
  61636. }
  61637. #else
  61638. __ai int8x8_t vuzp1_s8(int8x8_t __p0, int8x8_t __p1) {
  61639.   int8x8_t __ret;
  61640.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61641.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61642.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
  61643.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61644.   return __ret;
  61645. }
  61646. #endif
  61647.  
  61648. #ifdef __LITTLE_ENDIAN__
  61649. __ai float32x2_t vuzp1_f32(float32x2_t __p0, float32x2_t __p1) {
  61650.   float32x2_t __ret;
  61651.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  61652.   return __ret;
  61653. }
  61654. #else
  61655. __ai float32x2_t vuzp1_f32(float32x2_t __p0, float32x2_t __p1) {
  61656.   float32x2_t __ret;
  61657.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61658.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61659.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  61660.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61661.   return __ret;
  61662. }
  61663. #endif
  61664.  
  61665. #ifdef __LITTLE_ENDIAN__
  61666. __ai int32x2_t vuzp1_s32(int32x2_t __p0, int32x2_t __p1) {
  61667.   int32x2_t __ret;
  61668.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  61669.   return __ret;
  61670. }
  61671. #else
  61672. __ai int32x2_t vuzp1_s32(int32x2_t __p0, int32x2_t __p1) {
  61673.   int32x2_t __ret;
  61674.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61675.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61676.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  61677.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61678.   return __ret;
  61679. }
  61680. #endif
  61681.  
  61682. #ifdef __LITTLE_ENDIAN__
  61683. __ai int16x4_t vuzp1_s16(int16x4_t __p0, int16x4_t __p1) {
  61684.   int16x4_t __ret;
  61685.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
  61686.   return __ret;
  61687. }
  61688. #else
  61689. __ai int16x4_t vuzp1_s16(int16x4_t __p0, int16x4_t __p1) {
  61690.   int16x4_t __ret;
  61691.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61692.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61693.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
  61694.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61695.   return __ret;
  61696. }
  61697. #endif
  61698.  
  61699. #ifdef __LITTLE_ENDIAN__
  61700. __ai poly8x8_t vuzp2_p8(poly8x8_t __p0, poly8x8_t __p1) {
  61701.   poly8x8_t __ret;
  61702.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
  61703.   return __ret;
  61704. }
  61705. #else
  61706. __ai poly8x8_t vuzp2_p8(poly8x8_t __p0, poly8x8_t __p1) {
  61707.   poly8x8_t __ret;
  61708.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61709.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61710.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
  61711.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61712.   return __ret;
  61713. }
  61714. #endif
  61715.  
  61716. #ifdef __LITTLE_ENDIAN__
  61717. __ai poly16x4_t vuzp2_p16(poly16x4_t __p0, poly16x4_t __p1) {
  61718.   poly16x4_t __ret;
  61719.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
  61720.   return __ret;
  61721. }
  61722. #else
  61723. __ai poly16x4_t vuzp2_p16(poly16x4_t __p0, poly16x4_t __p1) {
  61724.   poly16x4_t __ret;
  61725.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61726.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61727.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
  61728.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61729.   return __ret;
  61730. }
  61731. #endif
  61732.  
  61733. #ifdef __LITTLE_ENDIAN__
  61734. __ai poly8x16_t vuzp2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  61735.   poly8x16_t __ret;
  61736.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
  61737.   return __ret;
  61738. }
  61739. #else
  61740. __ai poly8x16_t vuzp2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  61741.   poly8x16_t __ret;
  61742.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61743.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61744.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
  61745.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61746.   return __ret;
  61747. }
  61748. #endif
  61749.  
  61750. #ifdef __LITTLE_ENDIAN__
  61751. __ai poly64x2_t vuzp2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  61752.   poly64x2_t __ret;
  61753.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  61754.   return __ret;
  61755. }
  61756. #else
  61757. __ai poly64x2_t vuzp2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  61758.   poly64x2_t __ret;
  61759.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61760.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61761.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  61762.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61763.   return __ret;
  61764. }
  61765. #endif
  61766.  
  61767. #ifdef __LITTLE_ENDIAN__
  61768. __ai poly16x8_t vuzp2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  61769.   poly16x8_t __ret;
  61770.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
  61771.   return __ret;
  61772. }
  61773. #else
  61774. __ai poly16x8_t vuzp2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  61775.   poly16x8_t __ret;
  61776.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61777.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61778.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
  61779.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61780.   return __ret;
  61781. }
  61782. #endif
  61783.  
  61784. #ifdef __LITTLE_ENDIAN__
  61785. __ai uint8x16_t vuzp2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  61786.   uint8x16_t __ret;
  61787.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
  61788.   return __ret;
  61789. }
  61790. #else
  61791. __ai uint8x16_t vuzp2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  61792.   uint8x16_t __ret;
  61793.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61794.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61795.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
  61796.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61797.   return __ret;
  61798. }
  61799. #endif
  61800.  
  61801. #ifdef __LITTLE_ENDIAN__
  61802. __ai uint32x4_t vuzp2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  61803.   uint32x4_t __ret;
  61804.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
  61805.   return __ret;
  61806. }
  61807. #else
  61808. __ai uint32x4_t vuzp2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  61809.   uint32x4_t __ret;
  61810.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61811.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61812.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
  61813.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61814.   return __ret;
  61815. }
  61816. #endif
  61817.  
  61818. #ifdef __LITTLE_ENDIAN__
  61819. __ai uint64x2_t vuzp2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  61820.   uint64x2_t __ret;
  61821.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  61822.   return __ret;
  61823. }
  61824. #else
  61825. __ai uint64x2_t vuzp2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  61826.   uint64x2_t __ret;
  61827.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61828.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61829.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  61830.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61831.   return __ret;
  61832. }
  61833. #endif
  61834.  
  61835. #ifdef __LITTLE_ENDIAN__
  61836. __ai uint16x8_t vuzp2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  61837.   uint16x8_t __ret;
  61838.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
  61839.   return __ret;
  61840. }
  61841. #else
  61842. __ai uint16x8_t vuzp2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  61843.   uint16x8_t __ret;
  61844.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61845.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61846.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
  61847.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61848.   return __ret;
  61849. }
  61850. #endif
  61851.  
  61852. #ifdef __LITTLE_ENDIAN__
  61853. __ai int8x16_t vuzp2q_s8(int8x16_t __p0, int8x16_t __p1) {
  61854.   int8x16_t __ret;
  61855.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
  61856.   return __ret;
  61857. }
  61858. #else
  61859. __ai int8x16_t vuzp2q_s8(int8x16_t __p0, int8x16_t __p1) {
  61860.   int8x16_t __ret;
  61861.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61862.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61863.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
  61864.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  61865.   return __ret;
  61866. }
  61867. #endif
  61868.  
  61869. #ifdef __LITTLE_ENDIAN__
  61870. __ai float64x2_t vuzp2q_f64(float64x2_t __p0, float64x2_t __p1) {
  61871.   float64x2_t __ret;
  61872.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  61873.   return __ret;
  61874. }
  61875. #else
  61876. __ai float64x2_t vuzp2q_f64(float64x2_t __p0, float64x2_t __p1) {
  61877.   float64x2_t __ret;
  61878.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61879.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61880.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  61881.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61882.   return __ret;
  61883. }
  61884. #endif
  61885.  
  61886. #ifdef __LITTLE_ENDIAN__
  61887. __ai float32x4_t vuzp2q_f32(float32x4_t __p0, float32x4_t __p1) {
  61888.   float32x4_t __ret;
  61889.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
  61890.   return __ret;
  61891. }
  61892. #else
  61893. __ai float32x4_t vuzp2q_f32(float32x4_t __p0, float32x4_t __p1) {
  61894.   float32x4_t __ret;
  61895.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61896.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61897.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
  61898.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61899.   return __ret;
  61900. }
  61901. #endif
  61902.  
  61903. #ifdef __LITTLE_ENDIAN__
  61904. __ai int32x4_t vuzp2q_s32(int32x4_t __p0, int32x4_t __p1) {
  61905.   int32x4_t __ret;
  61906.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
  61907.   return __ret;
  61908. }
  61909. #else
  61910. __ai int32x4_t vuzp2q_s32(int32x4_t __p0, int32x4_t __p1) {
  61911.   int32x4_t __ret;
  61912.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61913.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61914.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
  61915.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  61916.   return __ret;
  61917. }
  61918. #endif
  61919.  
  61920. #ifdef __LITTLE_ENDIAN__
  61921. __ai int64x2_t vuzp2q_s64(int64x2_t __p0, int64x2_t __p1) {
  61922.   int64x2_t __ret;
  61923.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  61924.   return __ret;
  61925. }
  61926. #else
  61927. __ai int64x2_t vuzp2q_s64(int64x2_t __p0, int64x2_t __p1) {
  61928.   int64x2_t __ret;
  61929.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61930.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61931.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  61932.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61933.   return __ret;
  61934. }
  61935. #endif
  61936.  
  61937. #ifdef __LITTLE_ENDIAN__
  61938. __ai int16x8_t vuzp2q_s16(int16x8_t __p0, int16x8_t __p1) {
  61939.   int16x8_t __ret;
  61940.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
  61941.   return __ret;
  61942. }
  61943. #else
  61944. __ai int16x8_t vuzp2q_s16(int16x8_t __p0, int16x8_t __p1) {
  61945.   int16x8_t __ret;
  61946.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61947.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61948.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
  61949.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61950.   return __ret;
  61951. }
  61952. #endif
  61953.  
  61954. #ifdef __LITTLE_ENDIAN__
  61955. __ai uint8x8_t vuzp2_u8(uint8x8_t __p0, uint8x8_t __p1) {
  61956.   uint8x8_t __ret;
  61957.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
  61958.   return __ret;
  61959. }
  61960. #else
  61961. __ai uint8x8_t vuzp2_u8(uint8x8_t __p0, uint8x8_t __p1) {
  61962.   uint8x8_t __ret;
  61963.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  61964.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  61965.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
  61966.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  61967.   return __ret;
  61968. }
  61969. #endif
  61970.  
  61971. #ifdef __LITTLE_ENDIAN__
  61972. __ai uint32x2_t vuzp2_u32(uint32x2_t __p0, uint32x2_t __p1) {
  61973.   uint32x2_t __ret;
  61974.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  61975.   return __ret;
  61976. }
  61977. #else
  61978. __ai uint32x2_t vuzp2_u32(uint32x2_t __p0, uint32x2_t __p1) {
  61979.   uint32x2_t __ret;
  61980.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  61981.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  61982.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  61983.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  61984.   return __ret;
  61985. }
  61986. #endif
  61987.  
  61988. #ifdef __LITTLE_ENDIAN__
  61989. __ai uint16x4_t vuzp2_u16(uint16x4_t __p0, uint16x4_t __p1) {
  61990.   uint16x4_t __ret;
  61991.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
  61992.   return __ret;
  61993. }
  61994. #else
  61995. __ai uint16x4_t vuzp2_u16(uint16x4_t __p0, uint16x4_t __p1) {
  61996.   uint16x4_t __ret;
  61997.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  61998.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  61999.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
  62000.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62001.   return __ret;
  62002. }
  62003. #endif
  62004.  
  62005. #ifdef __LITTLE_ENDIAN__
  62006. __ai int8x8_t vuzp2_s8(int8x8_t __p0, int8x8_t __p1) {
  62007.   int8x8_t __ret;
  62008.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
  62009.   return __ret;
  62010. }
  62011. #else
  62012. __ai int8x8_t vuzp2_s8(int8x8_t __p0, int8x8_t __p1) {
  62013.   int8x8_t __ret;
  62014.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62015.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62016.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
  62017.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62018.   return __ret;
  62019. }
  62020. #endif
  62021.  
  62022. #ifdef __LITTLE_ENDIAN__
  62023. __ai float32x2_t vuzp2_f32(float32x2_t __p0, float32x2_t __p1) {
  62024.   float32x2_t __ret;
  62025.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  62026.   return __ret;
  62027. }
  62028. #else
  62029. __ai float32x2_t vuzp2_f32(float32x2_t __p0, float32x2_t __p1) {
  62030.   float32x2_t __ret;
  62031.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62032.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62033.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  62034.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62035.   return __ret;
  62036. }
  62037. #endif
  62038.  
  62039. #ifdef __LITTLE_ENDIAN__
  62040. __ai int32x2_t vuzp2_s32(int32x2_t __p0, int32x2_t __p1) {
  62041.   int32x2_t __ret;
  62042.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  62043.   return __ret;
  62044. }
  62045. #else
  62046. __ai int32x2_t vuzp2_s32(int32x2_t __p0, int32x2_t __p1) {
  62047.   int32x2_t __ret;
  62048.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62049.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62050.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  62051.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62052.   return __ret;
  62053. }
  62054. #endif
  62055.  
  62056. #ifdef __LITTLE_ENDIAN__
  62057. __ai int16x4_t vuzp2_s16(int16x4_t __p0, int16x4_t __p1) {
  62058.   int16x4_t __ret;
  62059.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
  62060.   return __ret;
  62061. }
  62062. #else
  62063. __ai int16x4_t vuzp2_s16(int16x4_t __p0, int16x4_t __p1) {
  62064.   int16x4_t __ret;
  62065.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62066.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62067.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
  62068.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62069.   return __ret;
  62070. }
  62071. #endif
  62072.  
  62073. #ifdef __LITTLE_ENDIAN__
  62074. __ai poly8x8_t vzip1_p8(poly8x8_t __p0, poly8x8_t __p1) {
  62075.   poly8x8_t __ret;
  62076.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
  62077.   return __ret;
  62078. }
  62079. #else
  62080. __ai poly8x8_t vzip1_p8(poly8x8_t __p0, poly8x8_t __p1) {
  62081.   poly8x8_t __ret;
  62082.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62083.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62084.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
  62085.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62086.   return __ret;
  62087. }
  62088. #endif
  62089.  
  62090. #ifdef __LITTLE_ENDIAN__
  62091. __ai poly16x4_t vzip1_p16(poly16x4_t __p0, poly16x4_t __p1) {
  62092.   poly16x4_t __ret;
  62093.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
  62094.   return __ret;
  62095. }
  62096. #else
  62097. __ai poly16x4_t vzip1_p16(poly16x4_t __p0, poly16x4_t __p1) {
  62098.   poly16x4_t __ret;
  62099.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62100.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62101.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
  62102.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62103.   return __ret;
  62104. }
  62105. #endif
  62106.  
  62107. #ifdef __LITTLE_ENDIAN__
  62108. __ai poly8x16_t vzip1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  62109.   poly8x16_t __ret;
  62110.   __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
  62111.   return __ret;
  62112. }
  62113. #else
  62114. __ai poly8x16_t vzip1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  62115.   poly8x16_t __ret;
  62116.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62117.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62118.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
  62119.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62120.   return __ret;
  62121. }
  62122. #endif
  62123.  
  62124. #ifdef __LITTLE_ENDIAN__
  62125. __ai poly64x2_t vzip1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  62126.   poly64x2_t __ret;
  62127.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  62128.   return __ret;
  62129. }
  62130. #else
  62131. __ai poly64x2_t vzip1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  62132.   poly64x2_t __ret;
  62133.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62134.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62135.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  62136.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62137.   return __ret;
  62138. }
  62139. #endif
  62140.  
  62141. #ifdef __LITTLE_ENDIAN__
  62142. __ai poly16x8_t vzip1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  62143.   poly16x8_t __ret;
  62144.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
  62145.   return __ret;
  62146. }
  62147. #else
  62148. __ai poly16x8_t vzip1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  62149.   poly16x8_t __ret;
  62150.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62151.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62152.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
  62153.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62154.   return __ret;
  62155. }
  62156. #endif
  62157.  
  62158. #ifdef __LITTLE_ENDIAN__
  62159. __ai uint8x16_t vzip1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  62160.   uint8x16_t __ret;
  62161.   __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
  62162.   return __ret;
  62163. }
  62164. #else
  62165. __ai uint8x16_t vzip1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  62166.   uint8x16_t __ret;
  62167.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62168.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62169.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
  62170.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62171.   return __ret;
  62172. }
  62173. #endif
  62174.  
  62175. #ifdef __LITTLE_ENDIAN__
  62176. __ai uint32x4_t vzip1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  62177.   uint32x4_t __ret;
  62178.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
  62179.   return __ret;
  62180. }
  62181. #else
  62182. __ai uint32x4_t vzip1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  62183.   uint32x4_t __ret;
  62184.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62185.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62186.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
  62187.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62188.   return __ret;
  62189. }
  62190. #endif
  62191.  
  62192. #ifdef __LITTLE_ENDIAN__
  62193. __ai uint64x2_t vzip1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  62194.   uint64x2_t __ret;
  62195.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  62196.   return __ret;
  62197. }
  62198. #else
  62199. __ai uint64x2_t vzip1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  62200.   uint64x2_t __ret;
  62201.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62202.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62203.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  62204.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62205.   return __ret;
  62206. }
  62207. #endif
  62208.  
  62209. #ifdef __LITTLE_ENDIAN__
  62210. __ai uint16x8_t vzip1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  62211.   uint16x8_t __ret;
  62212.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
  62213.   return __ret;
  62214. }
  62215. #else
  62216. __ai uint16x8_t vzip1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  62217.   uint16x8_t __ret;
  62218.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62219.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62220.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
  62221.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62222.   return __ret;
  62223. }
  62224. #endif
  62225.  
  62226. #ifdef __LITTLE_ENDIAN__
  62227. __ai int8x16_t vzip1q_s8(int8x16_t __p0, int8x16_t __p1) {
  62228.   int8x16_t __ret;
  62229.   __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
  62230.   return __ret;
  62231. }
  62232. #else
  62233. __ai int8x16_t vzip1q_s8(int8x16_t __p0, int8x16_t __p1) {
  62234.   int8x16_t __ret;
  62235.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62236.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62237.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
  62238.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62239.   return __ret;
  62240. }
  62241. #endif
  62242.  
  62243. #ifdef __LITTLE_ENDIAN__
  62244. __ai float64x2_t vzip1q_f64(float64x2_t __p0, float64x2_t __p1) {
  62245.   float64x2_t __ret;
  62246.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  62247.   return __ret;
  62248. }
  62249. #else
  62250. __ai float64x2_t vzip1q_f64(float64x2_t __p0, float64x2_t __p1) {
  62251.   float64x2_t __ret;
  62252.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62253.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62254.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  62255.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62256.   return __ret;
  62257. }
  62258. #endif
  62259.  
  62260. #ifdef __LITTLE_ENDIAN__
  62261. __ai float32x4_t vzip1q_f32(float32x4_t __p0, float32x4_t __p1) {
  62262.   float32x4_t __ret;
  62263.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
  62264.   return __ret;
  62265. }
  62266. #else
  62267. __ai float32x4_t vzip1q_f32(float32x4_t __p0, float32x4_t __p1) {
  62268.   float32x4_t __ret;
  62269.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62270.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62271.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
  62272.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62273.   return __ret;
  62274. }
  62275. #endif
  62276.  
  62277. #ifdef __LITTLE_ENDIAN__
  62278. __ai int32x4_t vzip1q_s32(int32x4_t __p0, int32x4_t __p1) {
  62279.   int32x4_t __ret;
  62280.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
  62281.   return __ret;
  62282. }
  62283. #else
  62284. __ai int32x4_t vzip1q_s32(int32x4_t __p0, int32x4_t __p1) {
  62285.   int32x4_t __ret;
  62286.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62287.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62288.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
  62289.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62290.   return __ret;
  62291. }
  62292. #endif
  62293.  
  62294. #ifdef __LITTLE_ENDIAN__
  62295. __ai int64x2_t vzip1q_s64(int64x2_t __p0, int64x2_t __p1) {
  62296.   int64x2_t __ret;
  62297.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  62298.   return __ret;
  62299. }
  62300. #else
  62301. __ai int64x2_t vzip1q_s64(int64x2_t __p0, int64x2_t __p1) {
  62302.   int64x2_t __ret;
  62303.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62304.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62305.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  62306.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62307.   return __ret;
  62308. }
  62309. #endif
  62310.  
  62311. #ifdef __LITTLE_ENDIAN__
  62312. __ai int16x8_t vzip1q_s16(int16x8_t __p0, int16x8_t __p1) {
  62313.   int16x8_t __ret;
  62314.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
  62315.   return __ret;
  62316. }
  62317. #else
  62318. __ai int16x8_t vzip1q_s16(int16x8_t __p0, int16x8_t __p1) {
  62319.   int16x8_t __ret;
  62320.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62321.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62322.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
  62323.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62324.   return __ret;
  62325. }
  62326. #endif
  62327.  
  62328. #ifdef __LITTLE_ENDIAN__
  62329. __ai uint8x8_t vzip1_u8(uint8x8_t __p0, uint8x8_t __p1) {
  62330.   uint8x8_t __ret;
  62331.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
  62332.   return __ret;
  62333. }
  62334. #else
  62335. __ai uint8x8_t vzip1_u8(uint8x8_t __p0, uint8x8_t __p1) {
  62336.   uint8x8_t __ret;
  62337.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62338.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62339.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
  62340.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62341.   return __ret;
  62342. }
  62343. #endif
  62344.  
  62345. #ifdef __LITTLE_ENDIAN__
  62346. __ai uint32x2_t vzip1_u32(uint32x2_t __p0, uint32x2_t __p1) {
  62347.   uint32x2_t __ret;
  62348.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  62349.   return __ret;
  62350. }
  62351. #else
  62352. __ai uint32x2_t vzip1_u32(uint32x2_t __p0, uint32x2_t __p1) {
  62353.   uint32x2_t __ret;
  62354.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62355.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62356.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  62357.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62358.   return __ret;
  62359. }
  62360. #endif
  62361.  
  62362. #ifdef __LITTLE_ENDIAN__
  62363. __ai uint16x4_t vzip1_u16(uint16x4_t __p0, uint16x4_t __p1) {
  62364.   uint16x4_t __ret;
  62365.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
  62366.   return __ret;
  62367. }
  62368. #else
  62369. __ai uint16x4_t vzip1_u16(uint16x4_t __p0, uint16x4_t __p1) {
  62370.   uint16x4_t __ret;
  62371.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62372.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62373.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
  62374.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62375.   return __ret;
  62376. }
  62377. #endif
  62378.  
  62379. #ifdef __LITTLE_ENDIAN__
  62380. __ai int8x8_t vzip1_s8(int8x8_t __p0, int8x8_t __p1) {
  62381.   int8x8_t __ret;
  62382.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
  62383.   return __ret;
  62384. }
  62385. #else
  62386. __ai int8x8_t vzip1_s8(int8x8_t __p0, int8x8_t __p1) {
  62387.   int8x8_t __ret;
  62388.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62389.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62390.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
  62391.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62392.   return __ret;
  62393. }
  62394. #endif
  62395.  
  62396. #ifdef __LITTLE_ENDIAN__
  62397. __ai float32x2_t vzip1_f32(float32x2_t __p0, float32x2_t __p1) {
  62398.   float32x2_t __ret;
  62399.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  62400.   return __ret;
  62401. }
  62402. #else
  62403. __ai float32x2_t vzip1_f32(float32x2_t __p0, float32x2_t __p1) {
  62404.   float32x2_t __ret;
  62405.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62406.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62407.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  62408.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62409.   return __ret;
  62410. }
  62411. #endif
  62412.  
  62413. #ifdef __LITTLE_ENDIAN__
  62414. __ai int32x2_t vzip1_s32(int32x2_t __p0, int32x2_t __p1) {
  62415.   int32x2_t __ret;
  62416.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
  62417.   return __ret;
  62418. }
  62419. #else
  62420. __ai int32x2_t vzip1_s32(int32x2_t __p0, int32x2_t __p1) {
  62421.   int32x2_t __ret;
  62422.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62423.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62424.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
  62425.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62426.   return __ret;
  62427. }
  62428. #endif
  62429.  
  62430. #ifdef __LITTLE_ENDIAN__
  62431. __ai int16x4_t vzip1_s16(int16x4_t __p0, int16x4_t __p1) {
  62432.   int16x4_t __ret;
  62433.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
  62434.   return __ret;
  62435. }
  62436. #else
  62437. __ai int16x4_t vzip1_s16(int16x4_t __p0, int16x4_t __p1) {
  62438.   int16x4_t __ret;
  62439.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62440.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62441.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
  62442.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62443.   return __ret;
  62444. }
  62445. #endif
  62446.  
  62447. #ifdef __LITTLE_ENDIAN__
  62448. __ai poly8x8_t vzip2_p8(poly8x8_t __p0, poly8x8_t __p1) {
  62449.   poly8x8_t __ret;
  62450.   __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
  62451.   return __ret;
  62452. }
  62453. #else
  62454. __ai poly8x8_t vzip2_p8(poly8x8_t __p0, poly8x8_t __p1) {
  62455.   poly8x8_t __ret;
  62456.   poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62457.   poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62458.   __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
  62459.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62460.   return __ret;
  62461. }
  62462. #endif
  62463.  
  62464. #ifdef __LITTLE_ENDIAN__
  62465. __ai poly16x4_t vzip2_p16(poly16x4_t __p0, poly16x4_t __p1) {
  62466.   poly16x4_t __ret;
  62467.   __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
  62468.   return __ret;
  62469. }
  62470. #else
  62471. __ai poly16x4_t vzip2_p16(poly16x4_t __p0, poly16x4_t __p1) {
  62472.   poly16x4_t __ret;
  62473.   poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62474.   poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62475.   __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
  62476.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62477.   return __ret;
  62478. }
  62479. #endif
  62480.  
  62481. #ifdef __LITTLE_ENDIAN__
  62482. __ai poly8x16_t vzip2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  62483.   poly8x16_t __ret;
  62484.   __ret = __builtin_shufflevector(__p0, __p1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
  62485.   return __ret;
  62486. }
  62487. #else
  62488. __ai poly8x16_t vzip2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
  62489.   poly8x16_t __ret;
  62490.   poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62491.   poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62492.   __ret = __builtin_shufflevector(__rev0, __rev1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
  62493.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62494.   return __ret;
  62495. }
  62496. #endif
  62497.  
  62498. #ifdef __LITTLE_ENDIAN__
  62499. __ai poly64x2_t vzip2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  62500.   poly64x2_t __ret;
  62501.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  62502.   return __ret;
  62503. }
  62504. #else
  62505. __ai poly64x2_t vzip2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
  62506.   poly64x2_t __ret;
  62507.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62508.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62509.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  62510.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62511.   return __ret;
  62512. }
  62513. #endif
  62514.  
  62515. #ifdef __LITTLE_ENDIAN__
  62516. __ai poly16x8_t vzip2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  62517.   poly16x8_t __ret;
  62518.   __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
  62519.   return __ret;
  62520. }
  62521. #else
  62522. __ai poly16x8_t vzip2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
  62523.   poly16x8_t __ret;
  62524.   poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62525.   poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62526.   __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
  62527.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62528.   return __ret;
  62529. }
  62530. #endif
  62531.  
  62532. #ifdef __LITTLE_ENDIAN__
  62533. __ai uint8x16_t vzip2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  62534.   uint8x16_t __ret;
  62535.   __ret = __builtin_shufflevector(__p0, __p1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
  62536.   return __ret;
  62537. }
  62538. #else
  62539. __ai uint8x16_t vzip2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
  62540.   uint8x16_t __ret;
  62541.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62542.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62543.   __ret = __builtin_shufflevector(__rev0, __rev1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
  62544.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62545.   return __ret;
  62546. }
  62547. #endif
  62548.  
  62549. #ifdef __LITTLE_ENDIAN__
  62550. __ai uint32x4_t vzip2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  62551.   uint32x4_t __ret;
  62552.   __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
  62553.   return __ret;
  62554. }
  62555. #else
  62556. __ai uint32x4_t vzip2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
  62557.   uint32x4_t __ret;
  62558.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62559.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62560.   __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
  62561.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62562.   return __ret;
  62563. }
  62564. #endif
  62565.  
  62566. #ifdef __LITTLE_ENDIAN__
  62567. __ai uint64x2_t vzip2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  62568.   uint64x2_t __ret;
  62569.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  62570.   return __ret;
  62571. }
  62572. #else
  62573. __ai uint64x2_t vzip2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  62574.   uint64x2_t __ret;
  62575.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62576.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62577.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  62578.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62579.   return __ret;
  62580. }
  62581. #endif
  62582.  
  62583. #ifdef __LITTLE_ENDIAN__
  62584. __ai uint16x8_t vzip2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  62585.   uint16x8_t __ret;
  62586.   __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
  62587.   return __ret;
  62588. }
  62589. #else
  62590. __ai uint16x8_t vzip2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
  62591.   uint16x8_t __ret;
  62592.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62593.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62594.   __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
  62595.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62596.   return __ret;
  62597. }
  62598. #endif
  62599.  
  62600. #ifdef __LITTLE_ENDIAN__
  62601. __ai int8x16_t vzip2q_s8(int8x16_t __p0, int8x16_t __p1) {
  62602.   int8x16_t __ret;
  62603.   __ret = __builtin_shufflevector(__p0, __p1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
  62604.   return __ret;
  62605. }
  62606. #else
  62607. __ai int8x16_t vzip2q_s8(int8x16_t __p0, int8x16_t __p1) {
  62608.   int8x16_t __ret;
  62609.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62610.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62611.   __ret = __builtin_shufflevector(__rev0, __rev1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
  62612.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  62613.   return __ret;
  62614. }
  62615. #endif
  62616.  
  62617. #ifdef __LITTLE_ENDIAN__
  62618. __ai float64x2_t vzip2q_f64(float64x2_t __p0, float64x2_t __p1) {
  62619.   float64x2_t __ret;
  62620.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  62621.   return __ret;
  62622. }
  62623. #else
  62624. __ai float64x2_t vzip2q_f64(float64x2_t __p0, float64x2_t __p1) {
  62625.   float64x2_t __ret;
  62626.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62627.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62628.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  62629.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62630.   return __ret;
  62631. }
  62632. #endif
  62633.  
  62634. #ifdef __LITTLE_ENDIAN__
  62635. __ai float32x4_t vzip2q_f32(float32x4_t __p0, float32x4_t __p1) {
  62636.   float32x4_t __ret;
  62637.   __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
  62638.   return __ret;
  62639. }
  62640. #else
  62641. __ai float32x4_t vzip2q_f32(float32x4_t __p0, float32x4_t __p1) {
  62642.   float32x4_t __ret;
  62643.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62644.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62645.   __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
  62646.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62647.   return __ret;
  62648. }
  62649. #endif
  62650.  
  62651. #ifdef __LITTLE_ENDIAN__
  62652. __ai int32x4_t vzip2q_s32(int32x4_t __p0, int32x4_t __p1) {
  62653.   int32x4_t __ret;
  62654.   __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
  62655.   return __ret;
  62656. }
  62657. #else
  62658. __ai int32x4_t vzip2q_s32(int32x4_t __p0, int32x4_t __p1) {
  62659.   int32x4_t __ret;
  62660.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62661.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62662.   __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
  62663.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62664.   return __ret;
  62665. }
  62666. #endif
  62667.  
  62668. #ifdef __LITTLE_ENDIAN__
  62669. __ai int64x2_t vzip2q_s64(int64x2_t __p0, int64x2_t __p1) {
  62670.   int64x2_t __ret;
  62671.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  62672.   return __ret;
  62673. }
  62674. #else
  62675. __ai int64x2_t vzip2q_s64(int64x2_t __p0, int64x2_t __p1) {
  62676.   int64x2_t __ret;
  62677.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62678.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62679.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  62680.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62681.   return __ret;
  62682. }
  62683. #endif
  62684.  
  62685. #ifdef __LITTLE_ENDIAN__
  62686. __ai int16x8_t vzip2q_s16(int16x8_t __p0, int16x8_t __p1) {
  62687.   int16x8_t __ret;
  62688.   __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
  62689.   return __ret;
  62690. }
  62691. #else
  62692. __ai int16x8_t vzip2q_s16(int16x8_t __p0, int16x8_t __p1) {
  62693.   int16x8_t __ret;
  62694.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62695.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62696.   __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
  62697.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62698.   return __ret;
  62699. }
  62700. #endif
  62701.  
  62702. #ifdef __LITTLE_ENDIAN__
  62703. __ai uint8x8_t vzip2_u8(uint8x8_t __p0, uint8x8_t __p1) {
  62704.   uint8x8_t __ret;
  62705.   __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
  62706.   return __ret;
  62707. }
  62708. #else
  62709. __ai uint8x8_t vzip2_u8(uint8x8_t __p0, uint8x8_t __p1) {
  62710.   uint8x8_t __ret;
  62711.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62712.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62713.   __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
  62714.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62715.   return __ret;
  62716. }
  62717. #endif
  62718.  
  62719. #ifdef __LITTLE_ENDIAN__
  62720. __ai uint32x2_t vzip2_u32(uint32x2_t __p0, uint32x2_t __p1) {
  62721.   uint32x2_t __ret;
  62722.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  62723.   return __ret;
  62724. }
  62725. #else
  62726. __ai uint32x2_t vzip2_u32(uint32x2_t __p0, uint32x2_t __p1) {
  62727.   uint32x2_t __ret;
  62728.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62729.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62730.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  62731.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62732.   return __ret;
  62733. }
  62734. #endif
  62735.  
  62736. #ifdef __LITTLE_ENDIAN__
  62737. __ai uint16x4_t vzip2_u16(uint16x4_t __p0, uint16x4_t __p1) {
  62738.   uint16x4_t __ret;
  62739.   __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
  62740.   return __ret;
  62741. }
  62742. #else
  62743. __ai uint16x4_t vzip2_u16(uint16x4_t __p0, uint16x4_t __p1) {
  62744.   uint16x4_t __ret;
  62745.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62746.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62747.   __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
  62748.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62749.   return __ret;
  62750. }
  62751. #endif
  62752.  
  62753. #ifdef __LITTLE_ENDIAN__
  62754. __ai int8x8_t vzip2_s8(int8x8_t __p0, int8x8_t __p1) {
  62755.   int8x8_t __ret;
  62756.   __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
  62757.   return __ret;
  62758. }
  62759. #else
  62760. __ai int8x8_t vzip2_s8(int8x8_t __p0, int8x8_t __p1) {
  62761.   int8x8_t __ret;
  62762.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62763.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  62764.   __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
  62765.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62766.   return __ret;
  62767. }
  62768. #endif
  62769.  
  62770. #ifdef __LITTLE_ENDIAN__
  62771. __ai float32x2_t vzip2_f32(float32x2_t __p0, float32x2_t __p1) {
  62772.   float32x2_t __ret;
  62773.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  62774.   return __ret;
  62775. }
  62776. #else
  62777. __ai float32x2_t vzip2_f32(float32x2_t __p0, float32x2_t __p1) {
  62778.   float32x2_t __ret;
  62779.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62780.   float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62781.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  62782.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62783.   return __ret;
  62784. }
  62785. #endif
  62786.  
  62787. #ifdef __LITTLE_ENDIAN__
  62788. __ai int32x2_t vzip2_s32(int32x2_t __p0, int32x2_t __p1) {
  62789.   int32x2_t __ret;
  62790.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
  62791.   return __ret;
  62792. }
  62793. #else
  62794. __ai int32x2_t vzip2_s32(int32x2_t __p0, int32x2_t __p1) {
  62795.   int32x2_t __ret;
  62796.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62797.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62798.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
  62799.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  62800.   return __ret;
  62801. }
  62802. #endif
  62803.  
  62804. #ifdef __LITTLE_ENDIAN__
  62805. __ai int16x4_t vzip2_s16(int16x4_t __p0, int16x4_t __p1) {
  62806.   int16x4_t __ret;
  62807.   __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
  62808.   return __ret;
  62809. }
  62810. #else
  62811. __ai int16x4_t vzip2_s16(int16x4_t __p0, int16x4_t __p1) {
  62812.   int16x4_t __ret;
  62813.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62814.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62815.   __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
  62816.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62817.   return __ret;
  62818. }
  62819. #endif
  62820.  
  62821. __ai __attribute__((target("aes"))) poly128_t vmull_p64(poly64_t __p0, poly64_t __p1) {
  62822.   poly128_t __ret;
  62823.   __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
  62824.   return __ret;
  62825. }
  62826. #ifdef __LITTLE_ENDIAN__
  62827. __ai __attribute__((target("aes"))) poly128_t vmull_high_p64(poly64x2_t __p0, poly64x2_t __p1) {
  62828.   poly128_t __ret;
  62829.   __ret = vmull_p64((poly64_t)(vget_high_p64(__p0)), (poly64_t)(vget_high_p64(__p1)));
  62830.   return __ret;
  62831. }
  62832. #else
  62833. __ai __attribute__((target("aes"))) poly128_t vmull_high_p64(poly64x2_t __p0, poly64x2_t __p1) {
  62834.   poly128_t __ret;
  62835.   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  62836.   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  62837.   __ret = vmull_p64((poly64_t)(__noswap_vget_high_p64(__rev0)), (poly64_t)(__noswap_vget_high_p64(__rev1)));
  62838.   return __ret;
  62839. }
  62840. #endif
  62841.  
  62842. #ifdef __LITTLE_ENDIAN__
  62843. __ai __attribute__((target("bf16"))) bfloat16x8_t __a64_vcvtq_low_bf16_f32(float32x4_t __p0) {
  62844.   bfloat16x8_t __ret;
  62845.   __ret = (bfloat16x8_t) __builtin_neon___a64_vcvtq_low_bf16_f32((int8x16_t)__p0, 43);
  62846.   return __ret;
  62847. }
  62848. #else
  62849. __ai __attribute__((target("bf16"))) bfloat16x8_t __a64_vcvtq_low_bf16_f32(float32x4_t __p0) {
  62850.   bfloat16x8_t __ret;
  62851.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62852.   __ret = (bfloat16x8_t) __builtin_neon___a64_vcvtq_low_bf16_f32((int8x16_t)__rev0, 43);
  62853.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62854.   return __ret;
  62855. }
  62856. __ai __attribute__((target("bf16"))) bfloat16x8_t __noswap___a64_vcvtq_low_bf16_f32(float32x4_t __p0) {
  62857.   bfloat16x8_t __ret;
  62858.   __ret = (bfloat16x8_t) __builtin_neon___a64_vcvtq_low_bf16_f32((int8x16_t)__p0, 43);
  62859.   return __ret;
  62860. }
  62861. #endif
  62862.  
  62863. #ifdef __LITTLE_ENDIAN__
  62864. #define vcopyq_lane_bf16(__p0_739, __p1_739, __p2_739, __p3_739) __extension__ ({ \
  62865.   bfloat16x8_t __ret_739; \
  62866.   bfloat16x8_t __s0_739 = __p0_739; \
  62867.   bfloat16x4_t __s2_739 = __p2_739; \
  62868.   __ret_739 = vsetq_lane_bf16(vget_lane_bf16(__s2_739, __p3_739), __s0_739, __p1_739); \
  62869.   __ret_739; \
  62870. })
  62871. #else
  62872. #define vcopyq_lane_bf16(__p0_740, __p1_740, __p2_740, __p3_740) __extension__ ({ \
  62873.   bfloat16x8_t __ret_740; \
  62874.   bfloat16x8_t __s0_740 = __p0_740; \
  62875.   bfloat16x4_t __s2_740 = __p2_740; \
  62876.   bfloat16x8_t __rev0_740;  __rev0_740 = __builtin_shufflevector(__s0_740, __s0_740, 7, 6, 5, 4, 3, 2, 1, 0); \
  62877.   bfloat16x4_t __rev2_740;  __rev2_740 = __builtin_shufflevector(__s2_740, __s2_740, 3, 2, 1, 0); \
  62878.   __ret_740 = __noswap_vsetq_lane_bf16(__noswap_vget_lane_bf16(__rev2_740, __p3_740), __rev0_740, __p1_740); \
  62879.   __ret_740 = __builtin_shufflevector(__ret_740, __ret_740, 7, 6, 5, 4, 3, 2, 1, 0); \
  62880.   __ret_740; \
  62881. })
  62882. #endif
  62883.  
  62884. #ifdef __LITTLE_ENDIAN__
  62885. #define vcopy_lane_bf16(__p0_741, __p1_741, __p2_741, __p3_741) __extension__ ({ \
  62886.   bfloat16x4_t __ret_741; \
  62887.   bfloat16x4_t __s0_741 = __p0_741; \
  62888.   bfloat16x4_t __s2_741 = __p2_741; \
  62889.   __ret_741 = vset_lane_bf16(vget_lane_bf16(__s2_741, __p3_741), __s0_741, __p1_741); \
  62890.   __ret_741; \
  62891. })
  62892. #else
  62893. #define vcopy_lane_bf16(__p0_742, __p1_742, __p2_742, __p3_742) __extension__ ({ \
  62894.   bfloat16x4_t __ret_742; \
  62895.   bfloat16x4_t __s0_742 = __p0_742; \
  62896.   bfloat16x4_t __s2_742 = __p2_742; \
  62897.   bfloat16x4_t __rev0_742;  __rev0_742 = __builtin_shufflevector(__s0_742, __s0_742, 3, 2, 1, 0); \
  62898.   bfloat16x4_t __rev2_742;  __rev2_742 = __builtin_shufflevector(__s2_742, __s2_742, 3, 2, 1, 0); \
  62899.   __ret_742 = __noswap_vset_lane_bf16(__noswap_vget_lane_bf16(__rev2_742, __p3_742), __rev0_742, __p1_742); \
  62900.   __ret_742 = __builtin_shufflevector(__ret_742, __ret_742, 3, 2, 1, 0); \
  62901.   __ret_742; \
  62902. })
  62903. #endif
  62904.  
  62905. #ifdef __LITTLE_ENDIAN__
  62906. #define vcopyq_laneq_bf16(__p0_743, __p1_743, __p2_743, __p3_743) __extension__ ({ \
  62907.   bfloat16x8_t __ret_743; \
  62908.   bfloat16x8_t __s0_743 = __p0_743; \
  62909.   bfloat16x8_t __s2_743 = __p2_743; \
  62910.   __ret_743 = vsetq_lane_bf16(vgetq_lane_bf16(__s2_743, __p3_743), __s0_743, __p1_743); \
  62911.   __ret_743; \
  62912. })
  62913. #else
  62914. #define vcopyq_laneq_bf16(__p0_744, __p1_744, __p2_744, __p3_744) __extension__ ({ \
  62915.   bfloat16x8_t __ret_744; \
  62916.   bfloat16x8_t __s0_744 = __p0_744; \
  62917.   bfloat16x8_t __s2_744 = __p2_744; \
  62918.   bfloat16x8_t __rev0_744;  __rev0_744 = __builtin_shufflevector(__s0_744, __s0_744, 7, 6, 5, 4, 3, 2, 1, 0); \
  62919.   bfloat16x8_t __rev2_744;  __rev2_744 = __builtin_shufflevector(__s2_744, __s2_744, 7, 6, 5, 4, 3, 2, 1, 0); \
  62920.   __ret_744 = __noswap_vsetq_lane_bf16(__noswap_vgetq_lane_bf16(__rev2_744, __p3_744), __rev0_744, __p1_744); \
  62921.   __ret_744 = __builtin_shufflevector(__ret_744, __ret_744, 7, 6, 5, 4, 3, 2, 1, 0); \
  62922.   __ret_744; \
  62923. })
  62924. #endif
  62925.  
  62926. #ifdef __LITTLE_ENDIAN__
  62927. #define vcopy_laneq_bf16(__p0_745, __p1_745, __p2_745, __p3_745) __extension__ ({ \
  62928.   bfloat16x4_t __ret_745; \
  62929.   bfloat16x4_t __s0_745 = __p0_745; \
  62930.   bfloat16x8_t __s2_745 = __p2_745; \
  62931.   __ret_745 = vset_lane_bf16(vgetq_lane_bf16(__s2_745, __p3_745), __s0_745, __p1_745); \
  62932.   __ret_745; \
  62933. })
  62934. #else
  62935. #define vcopy_laneq_bf16(__p0_746, __p1_746, __p2_746, __p3_746) __extension__ ({ \
  62936.   bfloat16x4_t __ret_746; \
  62937.   bfloat16x4_t __s0_746 = __p0_746; \
  62938.   bfloat16x8_t __s2_746 = __p2_746; \
  62939.   bfloat16x4_t __rev0_746;  __rev0_746 = __builtin_shufflevector(__s0_746, __s0_746, 3, 2, 1, 0); \
  62940.   bfloat16x8_t __rev2_746;  __rev2_746 = __builtin_shufflevector(__s2_746, __s2_746, 7, 6, 5, 4, 3, 2, 1, 0); \
  62941.   __ret_746 = __noswap_vset_lane_bf16(__noswap_vgetq_lane_bf16(__rev2_746, __p3_746), __rev0_746, __p1_746); \
  62942.   __ret_746 = __builtin_shufflevector(__ret_746, __ret_746, 3, 2, 1, 0); \
  62943.   __ret_746; \
  62944. })
  62945. #endif
  62946.  
  62947. #ifdef __LITTLE_ENDIAN__
  62948. __ai __attribute__((target("bf16"))) bfloat16x4_t vcvt_bf16_f32(float32x4_t __p0) {
  62949.   bfloat16x4_t __ret;
  62950.   __ret = vget_low_bf16(__a64_vcvtq_low_bf16_f32(__p0));
  62951.   return __ret;
  62952. }
  62953. #else
  62954. __ai __attribute__((target("bf16"))) bfloat16x4_t vcvt_bf16_f32(float32x4_t __p0) {
  62955.   bfloat16x4_t __ret;
  62956.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62957.   __ret = __noswap_vget_low_bf16(__noswap___a64_vcvtq_low_bf16_f32(__rev0));
  62958.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  62959.   return __ret;
  62960. }
  62961. #endif
  62962.  
  62963. #ifdef __LITTLE_ENDIAN__
  62964. __ai __attribute__((target("bf16"))) bfloat16x8_t vcvtq_high_bf16_f32(bfloat16x8_t __p0, float32x4_t __p1) {
  62965.   bfloat16x8_t __ret;
  62966.   __ret = (bfloat16x8_t) __builtin_neon_vcvtq_high_bf16_f32((int8x16_t)__p0, (int8x16_t)__p1, 43);
  62967.   return __ret;
  62968. }
  62969. #else
  62970. __ai __attribute__((target("bf16"))) bfloat16x8_t vcvtq_high_bf16_f32(bfloat16x8_t __p0, float32x4_t __p1) {
  62971.   bfloat16x8_t __ret;
  62972.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  62973.   float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  62974.   __ret = (bfloat16x8_t) __builtin_neon_vcvtq_high_bf16_f32((int8x16_t)__rev0, (int8x16_t)__rev1, 43);
  62975.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62976.   return __ret;
  62977. }
  62978. #endif
  62979.  
  62980. #ifdef __LITTLE_ENDIAN__
  62981. __ai __attribute__((target("bf16"))) bfloat16x8_t vcvtq_low_bf16_f32(float32x4_t __p0) {
  62982.   bfloat16x8_t __ret;
  62983.   __ret = __a64_vcvtq_low_bf16_f32(__p0);
  62984.   return __ret;
  62985. }
  62986. #else
  62987. __ai __attribute__((target("bf16"))) bfloat16x8_t vcvtq_low_bf16_f32(float32x4_t __p0) {
  62988.   bfloat16x8_t __ret;
  62989.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  62990.   __ret = __noswap___a64_vcvtq_low_bf16_f32(__rev0);
  62991.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  62992.   return __ret;
  62993. }
  62994. #endif
  62995.  
  62996. __ai __attribute__((target("bf16"))) poly8x8_t vreinterpret_p8_bf16(bfloat16x4_t __p0) {
  62997.   poly8x8_t __ret;
  62998.   __ret = (poly8x8_t)(__p0);
  62999.   return __ret;
  63000. }
  63001. __ai __attribute__((target("bf16"))) poly64x1_t vreinterpret_p64_bf16(bfloat16x4_t __p0) {
  63002.   poly64x1_t __ret;
  63003.   __ret = (poly64x1_t)(__p0);
  63004.   return __ret;
  63005. }
  63006. __ai __attribute__((target("bf16"))) poly16x4_t vreinterpret_p16_bf16(bfloat16x4_t __p0) {
  63007.   poly16x4_t __ret;
  63008.   __ret = (poly16x4_t)(__p0);
  63009.   return __ret;
  63010. }
  63011. __ai __attribute__((target("bf16"))) poly8x16_t vreinterpretq_p8_bf16(bfloat16x8_t __p0) {
  63012.   poly8x16_t __ret;
  63013.   __ret = (poly8x16_t)(__p0);
  63014.   return __ret;
  63015. }
  63016. __ai __attribute__((target("bf16"))) poly128_t vreinterpretq_p128_bf16(bfloat16x8_t __p0) {
  63017.   poly128_t __ret;
  63018.   __ret = (poly128_t)(__p0);
  63019.   return __ret;
  63020. }
  63021. __ai __attribute__((target("bf16"))) poly64x2_t vreinterpretq_p64_bf16(bfloat16x8_t __p0) {
  63022.   poly64x2_t __ret;
  63023.   __ret = (poly64x2_t)(__p0);
  63024.   return __ret;
  63025. }
  63026. __ai __attribute__((target("bf16"))) poly16x8_t vreinterpretq_p16_bf16(bfloat16x8_t __p0) {
  63027.   poly16x8_t __ret;
  63028.   __ret = (poly16x8_t)(__p0);
  63029.   return __ret;
  63030. }
  63031. __ai __attribute__((target("bf16"))) uint8x16_t vreinterpretq_u8_bf16(bfloat16x8_t __p0) {
  63032.   uint8x16_t __ret;
  63033.   __ret = (uint8x16_t)(__p0);
  63034.   return __ret;
  63035. }
  63036. __ai __attribute__((target("bf16"))) uint32x4_t vreinterpretq_u32_bf16(bfloat16x8_t __p0) {
  63037.   uint32x4_t __ret;
  63038.   __ret = (uint32x4_t)(__p0);
  63039.   return __ret;
  63040. }
  63041. __ai __attribute__((target("bf16"))) uint64x2_t vreinterpretq_u64_bf16(bfloat16x8_t __p0) {
  63042.   uint64x2_t __ret;
  63043.   __ret = (uint64x2_t)(__p0);
  63044.   return __ret;
  63045. }
  63046. __ai __attribute__((target("bf16"))) uint16x8_t vreinterpretq_u16_bf16(bfloat16x8_t __p0) {
  63047.   uint16x8_t __ret;
  63048.   __ret = (uint16x8_t)(__p0);
  63049.   return __ret;
  63050. }
  63051. __ai __attribute__((target("bf16"))) int8x16_t vreinterpretq_s8_bf16(bfloat16x8_t __p0) {
  63052.   int8x16_t __ret;
  63053.   __ret = (int8x16_t)(__p0);
  63054.   return __ret;
  63055. }
  63056. __ai __attribute__((target("bf16"))) float64x2_t vreinterpretq_f64_bf16(bfloat16x8_t __p0) {
  63057.   float64x2_t __ret;
  63058.   __ret = (float64x2_t)(__p0);
  63059.   return __ret;
  63060. }
  63061. __ai __attribute__((target("bf16"))) float32x4_t vreinterpretq_f32_bf16(bfloat16x8_t __p0) {
  63062.   float32x4_t __ret;
  63063.   __ret = (float32x4_t)(__p0);
  63064.   return __ret;
  63065. }
  63066. __ai __attribute__((target("bf16"))) float16x8_t vreinterpretq_f16_bf16(bfloat16x8_t __p0) {
  63067.   float16x8_t __ret;
  63068.   __ret = (float16x8_t)(__p0);
  63069.   return __ret;
  63070. }
  63071. __ai __attribute__((target("bf16"))) int32x4_t vreinterpretq_s32_bf16(bfloat16x8_t __p0) {
  63072.   int32x4_t __ret;
  63073.   __ret = (int32x4_t)(__p0);
  63074.   return __ret;
  63075. }
  63076. __ai __attribute__((target("bf16"))) int64x2_t vreinterpretq_s64_bf16(bfloat16x8_t __p0) {
  63077.   int64x2_t __ret;
  63078.   __ret = (int64x2_t)(__p0);
  63079.   return __ret;
  63080. }
  63081. __ai __attribute__((target("bf16"))) int16x8_t vreinterpretq_s16_bf16(bfloat16x8_t __p0) {
  63082.   int16x8_t __ret;
  63083.   __ret = (int16x8_t)(__p0);
  63084.   return __ret;
  63085. }
  63086. __ai __attribute__((target("bf16"))) uint8x8_t vreinterpret_u8_bf16(bfloat16x4_t __p0) {
  63087.   uint8x8_t __ret;
  63088.   __ret = (uint8x8_t)(__p0);
  63089.   return __ret;
  63090. }
  63091. __ai __attribute__((target("bf16"))) uint32x2_t vreinterpret_u32_bf16(bfloat16x4_t __p0) {
  63092.   uint32x2_t __ret;
  63093.   __ret = (uint32x2_t)(__p0);
  63094.   return __ret;
  63095. }
  63096. __ai __attribute__((target("bf16"))) uint64x1_t vreinterpret_u64_bf16(bfloat16x4_t __p0) {
  63097.   uint64x1_t __ret;
  63098.   __ret = (uint64x1_t)(__p0);
  63099.   return __ret;
  63100. }
  63101. __ai __attribute__((target("bf16"))) uint16x4_t vreinterpret_u16_bf16(bfloat16x4_t __p0) {
  63102.   uint16x4_t __ret;
  63103.   __ret = (uint16x4_t)(__p0);
  63104.   return __ret;
  63105. }
  63106. __ai __attribute__((target("bf16"))) int8x8_t vreinterpret_s8_bf16(bfloat16x4_t __p0) {
  63107.   int8x8_t __ret;
  63108.   __ret = (int8x8_t)(__p0);
  63109.   return __ret;
  63110. }
  63111. __ai __attribute__((target("bf16"))) float64x1_t vreinterpret_f64_bf16(bfloat16x4_t __p0) {
  63112.   float64x1_t __ret;
  63113.   __ret = (float64x1_t)(__p0);
  63114.   return __ret;
  63115. }
  63116. __ai __attribute__((target("bf16"))) float32x2_t vreinterpret_f32_bf16(bfloat16x4_t __p0) {
  63117.   float32x2_t __ret;
  63118.   __ret = (float32x2_t)(__p0);
  63119.   return __ret;
  63120. }
  63121. __ai __attribute__((target("bf16"))) float16x4_t vreinterpret_f16_bf16(bfloat16x4_t __p0) {
  63122.   float16x4_t __ret;
  63123.   __ret = (float16x4_t)(__p0);
  63124.   return __ret;
  63125. }
  63126. __ai __attribute__((target("bf16"))) int32x2_t vreinterpret_s32_bf16(bfloat16x4_t __p0) {
  63127.   int32x2_t __ret;
  63128.   __ret = (int32x2_t)(__p0);
  63129.   return __ret;
  63130. }
  63131. __ai __attribute__((target("bf16"))) int64x1_t vreinterpret_s64_bf16(bfloat16x4_t __p0) {
  63132.   int64x1_t __ret;
  63133.   __ret = (int64x1_t)(__p0);
  63134.   return __ret;
  63135. }
  63136. __ai __attribute__((target("bf16"))) int16x4_t vreinterpret_s16_bf16(bfloat16x4_t __p0) {
  63137.   int16x4_t __ret;
  63138.   __ret = (int16x4_t)(__p0);
  63139.   return __ret;
  63140. }
  63141. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_p8(poly8x16_t __p0) {
  63142.   bfloat16x8_t __ret;
  63143.   __ret = (bfloat16x8_t)(__p0);
  63144.   return __ret;
  63145. }
  63146. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_p128(poly128_t __p0) {
  63147.   bfloat16x8_t __ret;
  63148.   __ret = (bfloat16x8_t)(__p0);
  63149.   return __ret;
  63150. }
  63151. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_p64(poly64x2_t __p0) {
  63152.   bfloat16x8_t __ret;
  63153.   __ret = (bfloat16x8_t)(__p0);
  63154.   return __ret;
  63155. }
  63156. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_p16(poly16x8_t __p0) {
  63157.   bfloat16x8_t __ret;
  63158.   __ret = (bfloat16x8_t)(__p0);
  63159.   return __ret;
  63160. }
  63161. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_u8(uint8x16_t __p0) {
  63162.   bfloat16x8_t __ret;
  63163.   __ret = (bfloat16x8_t)(__p0);
  63164.   return __ret;
  63165. }
  63166. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_u32(uint32x4_t __p0) {
  63167.   bfloat16x8_t __ret;
  63168.   __ret = (bfloat16x8_t)(__p0);
  63169.   return __ret;
  63170. }
  63171. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_u64(uint64x2_t __p0) {
  63172.   bfloat16x8_t __ret;
  63173.   __ret = (bfloat16x8_t)(__p0);
  63174.   return __ret;
  63175. }
  63176. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_u16(uint16x8_t __p0) {
  63177.   bfloat16x8_t __ret;
  63178.   __ret = (bfloat16x8_t)(__p0);
  63179.   return __ret;
  63180. }
  63181. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_s8(int8x16_t __p0) {
  63182.   bfloat16x8_t __ret;
  63183.   __ret = (bfloat16x8_t)(__p0);
  63184.   return __ret;
  63185. }
  63186. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_f64(float64x2_t __p0) {
  63187.   bfloat16x8_t __ret;
  63188.   __ret = (bfloat16x8_t)(__p0);
  63189.   return __ret;
  63190. }
  63191. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_f32(float32x4_t __p0) {
  63192.   bfloat16x8_t __ret;
  63193.   __ret = (bfloat16x8_t)(__p0);
  63194.   return __ret;
  63195. }
  63196. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_f16(float16x8_t __p0) {
  63197.   bfloat16x8_t __ret;
  63198.   __ret = (bfloat16x8_t)(__p0);
  63199.   return __ret;
  63200. }
  63201. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_s32(int32x4_t __p0) {
  63202.   bfloat16x8_t __ret;
  63203.   __ret = (bfloat16x8_t)(__p0);
  63204.   return __ret;
  63205. }
  63206. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_s64(int64x2_t __p0) {
  63207.   bfloat16x8_t __ret;
  63208.   __ret = (bfloat16x8_t)(__p0);
  63209.   return __ret;
  63210. }
  63211. __ai __attribute__((target("bf16"))) bfloat16x8_t vreinterpretq_bf16_s16(int16x8_t __p0) {
  63212.   bfloat16x8_t __ret;
  63213.   __ret = (bfloat16x8_t)(__p0);
  63214.   return __ret;
  63215. }
  63216. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_p8(poly8x8_t __p0) {
  63217.   bfloat16x4_t __ret;
  63218.   __ret = (bfloat16x4_t)(__p0);
  63219.   return __ret;
  63220. }
  63221. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_p64(poly64x1_t __p0) {
  63222.   bfloat16x4_t __ret;
  63223.   __ret = (bfloat16x4_t)(__p0);
  63224.   return __ret;
  63225. }
  63226. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_p16(poly16x4_t __p0) {
  63227.   bfloat16x4_t __ret;
  63228.   __ret = (bfloat16x4_t)(__p0);
  63229.   return __ret;
  63230. }
  63231. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_u8(uint8x8_t __p0) {
  63232.   bfloat16x4_t __ret;
  63233.   __ret = (bfloat16x4_t)(__p0);
  63234.   return __ret;
  63235. }
  63236. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_u32(uint32x2_t __p0) {
  63237.   bfloat16x4_t __ret;
  63238.   __ret = (bfloat16x4_t)(__p0);
  63239.   return __ret;
  63240. }
  63241. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_u64(uint64x1_t __p0) {
  63242.   bfloat16x4_t __ret;
  63243.   __ret = (bfloat16x4_t)(__p0);
  63244.   return __ret;
  63245. }
  63246. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_u16(uint16x4_t __p0) {
  63247.   bfloat16x4_t __ret;
  63248.   __ret = (bfloat16x4_t)(__p0);
  63249.   return __ret;
  63250. }
  63251. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_s8(int8x8_t __p0) {
  63252.   bfloat16x4_t __ret;
  63253.   __ret = (bfloat16x4_t)(__p0);
  63254.   return __ret;
  63255. }
  63256. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_f64(float64x1_t __p0) {
  63257.   bfloat16x4_t __ret;
  63258.   __ret = (bfloat16x4_t)(__p0);
  63259.   return __ret;
  63260. }
  63261. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_f32(float32x2_t __p0) {
  63262.   bfloat16x4_t __ret;
  63263.   __ret = (bfloat16x4_t)(__p0);
  63264.   return __ret;
  63265. }
  63266. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_f16(float16x4_t __p0) {
  63267.   bfloat16x4_t __ret;
  63268.   __ret = (bfloat16x4_t)(__p0);
  63269.   return __ret;
  63270. }
  63271. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_s32(int32x2_t __p0) {
  63272.   bfloat16x4_t __ret;
  63273.   __ret = (bfloat16x4_t)(__p0);
  63274.   return __ret;
  63275. }
  63276. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_s64(int64x1_t __p0) {
  63277.   bfloat16x4_t __ret;
  63278.   __ret = (bfloat16x4_t)(__p0);
  63279.   return __ret;
  63280. }
  63281. __ai __attribute__((target("bf16"))) bfloat16x4_t vreinterpret_bf16_s16(int16x4_t __p0) {
  63282.   bfloat16x4_t __ret;
  63283.   __ret = (bfloat16x4_t)(__p0);
  63284.   return __ret;
  63285. }
  63286. #ifdef __LITTLE_ENDIAN__
  63287. #define vdotq_laneq_u32(__p0_747, __p1_747, __p2_747, __p3_747) __extension__ ({ \
  63288.   uint32x4_t __ret_747; \
  63289.   uint32x4_t __s0_747 = __p0_747; \
  63290.   uint8x16_t __s1_747 = __p1_747; \
  63291.   uint8x16_t __s2_747 = __p2_747; \
  63292. uint8x16_t __reint_747 = __s2_747; \
  63293. uint32x4_t __reint1_747 = splatq_laneq_u32(*(uint32x4_t *) &__reint_747, __p3_747); \
  63294.   __ret_747 = vdotq_u32(__s0_747, __s1_747, *(uint8x16_t *) &__reint1_747); \
  63295.   __ret_747; \
  63296. })
  63297. #else
  63298. #define vdotq_laneq_u32(__p0_748, __p1_748, __p2_748, __p3_748) __extension__ ({ \
  63299.   uint32x4_t __ret_748; \
  63300.   uint32x4_t __s0_748 = __p0_748; \
  63301.   uint8x16_t __s1_748 = __p1_748; \
  63302.   uint8x16_t __s2_748 = __p2_748; \
  63303.   uint32x4_t __rev0_748;  __rev0_748 = __builtin_shufflevector(__s0_748, __s0_748, 3, 2, 1, 0); \
  63304.   uint8x16_t __rev1_748;  __rev1_748 = __builtin_shufflevector(__s1_748, __s1_748, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  63305.   uint8x16_t __rev2_748;  __rev2_748 = __builtin_shufflevector(__s2_748, __s2_748, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  63306. uint8x16_t __reint_748 = __rev2_748; \
  63307. uint32x4_t __reint1_748 = __noswap_splatq_laneq_u32(*(uint32x4_t *) &__reint_748, __p3_748); \
  63308.   __ret_748 = __noswap_vdotq_u32(__rev0_748, __rev1_748, *(uint8x16_t *) &__reint1_748); \
  63309.   __ret_748 = __builtin_shufflevector(__ret_748, __ret_748, 3, 2, 1, 0); \
  63310.   __ret_748; \
  63311. })
  63312. #endif
  63313.  
  63314. #ifdef __LITTLE_ENDIAN__
  63315. #define vdotq_laneq_s32(__p0_749, __p1_749, __p2_749, __p3_749) __extension__ ({ \
  63316.   int32x4_t __ret_749; \
  63317.   int32x4_t __s0_749 = __p0_749; \
  63318.   int8x16_t __s1_749 = __p1_749; \
  63319.   int8x16_t __s2_749 = __p2_749; \
  63320. int8x16_t __reint_749 = __s2_749; \
  63321. int32x4_t __reint1_749 = splatq_laneq_s32(*(int32x4_t *) &__reint_749, __p3_749); \
  63322.   __ret_749 = vdotq_s32(__s0_749, __s1_749, *(int8x16_t *) &__reint1_749); \
  63323.   __ret_749; \
  63324. })
  63325. #else
  63326. #define vdotq_laneq_s32(__p0_750, __p1_750, __p2_750, __p3_750) __extension__ ({ \
  63327.   int32x4_t __ret_750; \
  63328.   int32x4_t __s0_750 = __p0_750; \
  63329.   int8x16_t __s1_750 = __p1_750; \
  63330.   int8x16_t __s2_750 = __p2_750; \
  63331.   int32x4_t __rev0_750;  __rev0_750 = __builtin_shufflevector(__s0_750, __s0_750, 3, 2, 1, 0); \
  63332.   int8x16_t __rev1_750;  __rev1_750 = __builtin_shufflevector(__s1_750, __s1_750, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  63333.   int8x16_t __rev2_750;  __rev2_750 = __builtin_shufflevector(__s2_750, __s2_750, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  63334. int8x16_t __reint_750 = __rev2_750; \
  63335. int32x4_t __reint1_750 = __noswap_splatq_laneq_s32(*(int32x4_t *) &__reint_750, __p3_750); \
  63336.   __ret_750 = __noswap_vdotq_s32(__rev0_750, __rev1_750, *(int8x16_t *) &__reint1_750); \
  63337.   __ret_750 = __builtin_shufflevector(__ret_750, __ret_750, 3, 2, 1, 0); \
  63338.   __ret_750; \
  63339. })
  63340. #endif
  63341.  
  63342. #ifdef __LITTLE_ENDIAN__
  63343. #define vdot_laneq_u32(__p0_751, __p1_751, __p2_751, __p3_751) __extension__ ({ \
  63344.   uint32x2_t __ret_751; \
  63345.   uint32x2_t __s0_751 = __p0_751; \
  63346.   uint8x8_t __s1_751 = __p1_751; \
  63347.   uint8x16_t __s2_751 = __p2_751; \
  63348. uint8x16_t __reint_751 = __s2_751; \
  63349. uint32x2_t __reint1_751 = splat_laneq_u32(*(uint32x4_t *) &__reint_751, __p3_751); \
  63350.   __ret_751 = vdot_u32(__s0_751, __s1_751, *(uint8x8_t *) &__reint1_751); \
  63351.   __ret_751; \
  63352. })
  63353. #else
  63354. #define vdot_laneq_u32(__p0_752, __p1_752, __p2_752, __p3_752) __extension__ ({ \
  63355.   uint32x2_t __ret_752; \
  63356.   uint32x2_t __s0_752 = __p0_752; \
  63357.   uint8x8_t __s1_752 = __p1_752; \
  63358.   uint8x16_t __s2_752 = __p2_752; \
  63359.   uint32x2_t __rev0_752;  __rev0_752 = __builtin_shufflevector(__s0_752, __s0_752, 1, 0); \
  63360.   uint8x8_t __rev1_752;  __rev1_752 = __builtin_shufflevector(__s1_752, __s1_752, 7, 6, 5, 4, 3, 2, 1, 0); \
  63361.   uint8x16_t __rev2_752;  __rev2_752 = __builtin_shufflevector(__s2_752, __s2_752, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  63362. uint8x16_t __reint_752 = __rev2_752; \
  63363. uint32x2_t __reint1_752 = __noswap_splat_laneq_u32(*(uint32x4_t *) &__reint_752, __p3_752); \
  63364.   __ret_752 = __noswap_vdot_u32(__rev0_752, __rev1_752, *(uint8x8_t *) &__reint1_752); \
  63365.   __ret_752 = __builtin_shufflevector(__ret_752, __ret_752, 1, 0); \
  63366.   __ret_752; \
  63367. })
  63368. #endif
  63369.  
  63370. #ifdef __LITTLE_ENDIAN__
  63371. #define vdot_laneq_s32(__p0_753, __p1_753, __p2_753, __p3_753) __extension__ ({ \
  63372.   int32x2_t __ret_753; \
  63373.   int32x2_t __s0_753 = __p0_753; \
  63374.   int8x8_t __s1_753 = __p1_753; \
  63375.   int8x16_t __s2_753 = __p2_753; \
  63376. int8x16_t __reint_753 = __s2_753; \
  63377. int32x2_t __reint1_753 = splat_laneq_s32(*(int32x4_t *) &__reint_753, __p3_753); \
  63378.   __ret_753 = vdot_s32(__s0_753, __s1_753, *(int8x8_t *) &__reint1_753); \
  63379.   __ret_753; \
  63380. })
  63381. #else
  63382. #define vdot_laneq_s32(__p0_754, __p1_754, __p2_754, __p3_754) __extension__ ({ \
  63383.   int32x2_t __ret_754; \
  63384.   int32x2_t __s0_754 = __p0_754; \
  63385.   int8x8_t __s1_754 = __p1_754; \
  63386.   int8x16_t __s2_754 = __p2_754; \
  63387.   int32x2_t __rev0_754;  __rev0_754 = __builtin_shufflevector(__s0_754, __s0_754, 1, 0); \
  63388.   int8x8_t __rev1_754;  __rev1_754 = __builtin_shufflevector(__s1_754, __s1_754, 7, 6, 5, 4, 3, 2, 1, 0); \
  63389.   int8x16_t __rev2_754;  __rev2_754 = __builtin_shufflevector(__s2_754, __s2_754, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  63390. int8x16_t __reint_754 = __rev2_754; \
  63391. int32x2_t __reint1_754 = __noswap_splat_laneq_s32(*(int32x4_t *) &__reint_754, __p3_754); \
  63392.   __ret_754 = __noswap_vdot_s32(__rev0_754, __rev1_754, *(int8x8_t *) &__reint1_754); \
  63393.   __ret_754 = __builtin_shufflevector(__ret_754, __ret_754, 1, 0); \
  63394.   __ret_754; \
  63395. })
  63396. #endif
  63397.  
  63398. #ifdef __LITTLE_ENDIAN__
  63399. __ai __attribute__((target("fp16fml"))) float32x4_t vfmlalq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63400.   float32x4_t __ret;
  63401.   __ret = (float32x4_t) __builtin_neon_vfmlalq_high_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  63402.   return __ret;
  63403. }
  63404. #else
  63405. __ai __attribute__((target("fp16fml"))) float32x4_t vfmlalq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63406.   float32x4_t __ret;
  63407.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  63408.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  63409.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  63410.   __ret = (float32x4_t) __builtin_neon_vfmlalq_high_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  63411.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  63412.   return __ret;
  63413. }
  63414. __ai __attribute__((target("fp16fml"))) float32x4_t __noswap_vfmlalq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63415.   float32x4_t __ret;
  63416.   __ret = (float32x4_t) __builtin_neon_vfmlalq_high_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  63417.   return __ret;
  63418. }
  63419. #endif
  63420.  
  63421. #ifdef __LITTLE_ENDIAN__
  63422. __ai __attribute__((target("fp16fml"))) float32x2_t vfmlal_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63423.   float32x2_t __ret;
  63424.   __ret = (float32x2_t) __builtin_neon_vfmlal_high_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  63425.   return __ret;
  63426. }
  63427. #else
  63428. __ai __attribute__((target("fp16fml"))) float32x2_t vfmlal_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63429.   float32x2_t __ret;
  63430.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  63431.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  63432.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  63433.   __ret = (float32x2_t) __builtin_neon_vfmlal_high_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  63434.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  63435.   return __ret;
  63436. }
  63437. __ai __attribute__((target("fp16fml"))) float32x2_t __noswap_vfmlal_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63438.   float32x2_t __ret;
  63439.   __ret = (float32x2_t) __builtin_neon_vfmlal_high_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  63440.   return __ret;
  63441. }
  63442. #endif
  63443.  
  63444. #ifdef __LITTLE_ENDIAN__
  63445. __ai __attribute__((target("fp16fml"))) float32x4_t vfmlalq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63446.   float32x4_t __ret;
  63447.   __ret = (float32x4_t) __builtin_neon_vfmlalq_low_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  63448.   return __ret;
  63449. }
  63450. #else
  63451. __ai __attribute__((target("fp16fml"))) float32x4_t vfmlalq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63452.   float32x4_t __ret;
  63453.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  63454.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  63455.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  63456.   __ret = (float32x4_t) __builtin_neon_vfmlalq_low_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  63457.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  63458.   return __ret;
  63459. }
  63460. __ai __attribute__((target("fp16fml"))) float32x4_t __noswap_vfmlalq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63461.   float32x4_t __ret;
  63462.   __ret = (float32x4_t) __builtin_neon_vfmlalq_low_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  63463.   return __ret;
  63464. }
  63465. #endif
  63466.  
  63467. #ifdef __LITTLE_ENDIAN__
  63468. __ai __attribute__((target("fp16fml"))) float32x2_t vfmlal_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63469.   float32x2_t __ret;
  63470.   __ret = (float32x2_t) __builtin_neon_vfmlal_low_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  63471.   return __ret;
  63472. }
  63473. #else
  63474. __ai __attribute__((target("fp16fml"))) float32x2_t vfmlal_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63475.   float32x2_t __ret;
  63476.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  63477.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  63478.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  63479.   __ret = (float32x2_t) __builtin_neon_vfmlal_low_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  63480.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  63481.   return __ret;
  63482. }
  63483. __ai __attribute__((target("fp16fml"))) float32x2_t __noswap_vfmlal_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63484.   float32x2_t __ret;
  63485.   __ret = (float32x2_t) __builtin_neon_vfmlal_low_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  63486.   return __ret;
  63487. }
  63488. #endif
  63489.  
  63490. #ifdef __LITTLE_ENDIAN__
  63491. __ai __attribute__((target("fp16fml"))) float32x4_t vfmlslq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63492.   float32x4_t __ret;
  63493.   __ret = (float32x4_t) __builtin_neon_vfmlslq_high_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  63494.   return __ret;
  63495. }
  63496. #else
  63497. __ai __attribute__((target("fp16fml"))) float32x4_t vfmlslq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63498.   float32x4_t __ret;
  63499.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  63500.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  63501.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  63502.   __ret = (float32x4_t) __builtin_neon_vfmlslq_high_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  63503.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  63504.   return __ret;
  63505. }
  63506. __ai __attribute__((target("fp16fml"))) float32x4_t __noswap_vfmlslq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63507.   float32x4_t __ret;
  63508.   __ret = (float32x4_t) __builtin_neon_vfmlslq_high_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  63509.   return __ret;
  63510. }
  63511. #endif
  63512.  
  63513. #ifdef __LITTLE_ENDIAN__
  63514. __ai __attribute__((target("fp16fml"))) float32x2_t vfmlsl_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63515.   float32x2_t __ret;
  63516.   __ret = (float32x2_t) __builtin_neon_vfmlsl_high_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  63517.   return __ret;
  63518. }
  63519. #else
  63520. __ai __attribute__((target("fp16fml"))) float32x2_t vfmlsl_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63521.   float32x2_t __ret;
  63522.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  63523.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  63524.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  63525.   __ret = (float32x2_t) __builtin_neon_vfmlsl_high_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  63526.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  63527.   return __ret;
  63528. }
  63529. __ai __attribute__((target("fp16fml"))) float32x2_t __noswap_vfmlsl_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63530.   float32x2_t __ret;
  63531.   __ret = (float32x2_t) __builtin_neon_vfmlsl_high_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  63532.   return __ret;
  63533. }
  63534. #endif
  63535.  
  63536. #ifdef __LITTLE_ENDIAN__
  63537. __ai __attribute__((target("fp16fml"))) float32x4_t vfmlslq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63538.   float32x4_t __ret;
  63539.   __ret = (float32x4_t) __builtin_neon_vfmlslq_low_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  63540.   return __ret;
  63541. }
  63542. #else
  63543. __ai __attribute__((target("fp16fml"))) float32x4_t vfmlslq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63544.   float32x4_t __ret;
  63545.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  63546.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  63547.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  63548.   __ret = (float32x4_t) __builtin_neon_vfmlslq_low_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
  63549.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  63550.   return __ret;
  63551. }
  63552. __ai __attribute__((target("fp16fml"))) float32x4_t __noswap_vfmlslq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
  63553.   float32x4_t __ret;
  63554.   __ret = (float32x4_t) __builtin_neon_vfmlslq_low_f16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
  63555.   return __ret;
  63556. }
  63557. #endif
  63558.  
  63559. #ifdef __LITTLE_ENDIAN__
  63560. __ai __attribute__((target("fp16fml"))) float32x2_t vfmlsl_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63561.   float32x2_t __ret;
  63562.   __ret = (float32x2_t) __builtin_neon_vfmlsl_low_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  63563.   return __ret;
  63564. }
  63565. #else
  63566. __ai __attribute__((target("fp16fml"))) float32x2_t vfmlsl_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63567.   float32x2_t __ret;
  63568.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  63569.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  63570.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  63571.   __ret = (float32x2_t) __builtin_neon_vfmlsl_low_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
  63572.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  63573.   return __ret;
  63574. }
  63575. __ai __attribute__((target("fp16fml"))) float32x2_t __noswap_vfmlsl_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
  63576.   float32x2_t __ret;
  63577.   __ret = (float32x2_t) __builtin_neon_vfmlsl_low_f16((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
  63578.   return __ret;
  63579. }
  63580. #endif
  63581.  
  63582. #ifdef __LITTLE_ENDIAN__
  63583. __ai __attribute__((target("fullfp16"))) float16x8_t vdivq_f16(float16x8_t __p0, float16x8_t __p1) {
  63584.   float16x8_t __ret;
  63585.   __ret = __p0 / __p1;
  63586.   return __ret;
  63587. }
  63588. #else
  63589. __ai __attribute__((target("fullfp16"))) float16x8_t vdivq_f16(float16x8_t __p0, float16x8_t __p1) {
  63590.   float16x8_t __ret;
  63591.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  63592.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  63593.   __ret = __rev0 / __rev1;
  63594.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  63595.   return __ret;
  63596. }
  63597. #endif
  63598.  
  63599. #ifdef __LITTLE_ENDIAN__
  63600. __ai __attribute__((target("fullfp16"))) float16x4_t vdiv_f16(float16x4_t __p0, float16x4_t __p1) {
  63601.   float16x4_t __ret;
  63602.   __ret = __p0 / __p1;
  63603.   return __ret;
  63604. }
  63605. #else
  63606. __ai __attribute__((target("fullfp16"))) float16x4_t vdiv_f16(float16x4_t __p0, float16x4_t __p1) {
  63607.   float16x4_t __ret;
  63608.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  63609.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  63610.   __ret = __rev0 / __rev1;
  63611.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  63612.   return __ret;
  63613. }
  63614. #endif
  63615.  
  63616. #ifdef __LITTLE_ENDIAN__
  63617. #define vduph_lane_f16(__p0, __p1) __extension__ ({ \
  63618.   float16_t __ret; \
  63619.   float16x4_t __s0 = __p0; \
  63620.   __ret = (float16_t) __builtin_neon_vduph_lane_f16((float16x4_t)__s0, __p1); \
  63621.   __ret; \
  63622. })
  63623. #else
  63624. #define vduph_lane_f16(__p0, __p1) __extension__ ({ \
  63625.   float16_t __ret; \
  63626.   float16x4_t __s0 = __p0; \
  63627.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  63628.   __ret = (float16_t) __builtin_neon_vduph_lane_f16((float16x4_t)__rev0, __p1); \
  63629.   __ret; \
  63630. })
  63631. #endif
  63632.  
  63633. #ifdef __LITTLE_ENDIAN__
  63634. #define vduph_laneq_f16(__p0, __p1) __extension__ ({ \
  63635.   float16_t __ret; \
  63636.   float16x8_t __s0 = __p0; \
  63637.   __ret = (float16_t) __builtin_neon_vduph_laneq_f16((float16x8_t)__s0, __p1); \
  63638.   __ret; \
  63639. })
  63640. #else
  63641. #define vduph_laneq_f16(__p0, __p1) __extension__ ({ \
  63642.   float16_t __ret; \
  63643.   float16x8_t __s0 = __p0; \
  63644.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  63645.   __ret = (float16_t) __builtin_neon_vduph_laneq_f16((float16x8_t)__rev0, __p1); \
  63646.   __ret; \
  63647. })
  63648. #endif
  63649.  
  63650. #ifdef __LITTLE_ENDIAN__
  63651. #define vfmah_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63652.   float16_t __ret; \
  63653.   float16_t __s0 = __p0; \
  63654.   float16_t __s1 = __p1; \
  63655.   float16x4_t __s2 = __p2; \
  63656.   __ret = (float16_t) __builtin_neon_vfmah_lane_f16(__s0, __s1, (float16x4_t)__s2, __p3); \
  63657.   __ret; \
  63658. })
  63659. #else
  63660. #define vfmah_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63661.   float16_t __ret; \
  63662.   float16_t __s0 = __p0; \
  63663.   float16_t __s1 = __p1; \
  63664.   float16x4_t __s2 = __p2; \
  63665.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  63666.   __ret = (float16_t) __builtin_neon_vfmah_lane_f16(__s0, __s1, (float16x4_t)__rev2, __p3); \
  63667.   __ret; \
  63668. })
  63669. #define __noswap_vfmah_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63670.   float16_t __ret; \
  63671.   float16_t __s0 = __p0; \
  63672.   float16_t __s1 = __p1; \
  63673.   float16x4_t __s2 = __p2; \
  63674.   __ret = (float16_t) __builtin_neon_vfmah_lane_f16(__s0, __s1, (float16x4_t)__s2, __p3); \
  63675.   __ret; \
  63676. })
  63677. #endif
  63678.  
  63679. #ifdef __LITTLE_ENDIAN__
  63680. #define vfmaq_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63681.   float16x8_t __ret; \
  63682.   float16x8_t __s0 = __p0; \
  63683.   float16x8_t __s1 = __p1; \
  63684.   float16x4_t __s2 = __p2; \
  63685.   __ret = (float16x8_t) __builtin_neon_vfmaq_lane_f16((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 40); \
  63686.   __ret; \
  63687. })
  63688. #else
  63689. #define vfmaq_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63690.   float16x8_t __ret; \
  63691.   float16x8_t __s0 = __p0; \
  63692.   float16x8_t __s1 = __p1; \
  63693.   float16x4_t __s2 = __p2; \
  63694.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  63695.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  63696.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  63697.   __ret = (float16x8_t) __builtin_neon_vfmaq_lane_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, __p3, 40); \
  63698.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  63699.   __ret; \
  63700. })
  63701. #define __noswap_vfmaq_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63702.   float16x8_t __ret; \
  63703.   float16x8_t __s0 = __p0; \
  63704.   float16x8_t __s1 = __p1; \
  63705.   float16x4_t __s2 = __p2; \
  63706.   __ret = (float16x8_t) __builtin_neon_vfmaq_lane_f16((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 40); \
  63707.   __ret; \
  63708. })
  63709. #endif
  63710.  
  63711. #ifdef __LITTLE_ENDIAN__
  63712. #define vfma_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63713.   float16x4_t __ret; \
  63714.   float16x4_t __s0 = __p0; \
  63715.   float16x4_t __s1 = __p1; \
  63716.   float16x4_t __s2 = __p2; \
  63717.   __ret = (float16x4_t) __builtin_neon_vfma_lane_f16((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 8); \
  63718.   __ret; \
  63719. })
  63720. #else
  63721. #define vfma_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63722.   float16x4_t __ret; \
  63723.   float16x4_t __s0 = __p0; \
  63724.   float16x4_t __s1 = __p1; \
  63725.   float16x4_t __s2 = __p2; \
  63726.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  63727.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  63728.   float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  63729.   __ret = (float16x4_t) __builtin_neon_vfma_lane_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, __p3, 8); \
  63730.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  63731.   __ret; \
  63732. })
  63733. #define __noswap_vfma_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63734.   float16x4_t __ret; \
  63735.   float16x4_t __s0 = __p0; \
  63736.   float16x4_t __s1 = __p1; \
  63737.   float16x4_t __s2 = __p2; \
  63738.   __ret = (float16x4_t) __builtin_neon_vfma_lane_f16((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 8); \
  63739.   __ret; \
  63740. })
  63741. #endif
  63742.  
  63743. #ifdef __LITTLE_ENDIAN__
  63744. #define vfmah_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63745.   float16_t __ret; \
  63746.   float16_t __s0 = __p0; \
  63747.   float16_t __s1 = __p1; \
  63748.   float16x8_t __s2 = __p2; \
  63749.   __ret = (float16_t) __builtin_neon_vfmah_laneq_f16(__s0, __s1, (float16x8_t)__s2, __p3); \
  63750.   __ret; \
  63751. })
  63752. #else
  63753. #define vfmah_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63754.   float16_t __ret; \
  63755.   float16_t __s0 = __p0; \
  63756.   float16_t __s1 = __p1; \
  63757.   float16x8_t __s2 = __p2; \
  63758.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
  63759.   __ret = (float16_t) __builtin_neon_vfmah_laneq_f16(__s0, __s1, (float16x8_t)__rev2, __p3); \
  63760.   __ret; \
  63761. })
  63762. #define __noswap_vfmah_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63763.   float16_t __ret; \
  63764.   float16_t __s0 = __p0; \
  63765.   float16_t __s1 = __p1; \
  63766.   float16x8_t __s2 = __p2; \
  63767.   __ret = (float16_t) __builtin_neon_vfmah_laneq_f16(__s0, __s1, (float16x8_t)__s2, __p3); \
  63768.   __ret; \
  63769. })
  63770. #endif
  63771.  
  63772. #ifdef __LITTLE_ENDIAN__
  63773. #define vfmaq_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63774.   float16x8_t __ret; \
  63775.   float16x8_t __s0 = __p0; \
  63776.   float16x8_t __s1 = __p1; \
  63777.   float16x8_t __s2 = __p2; \
  63778.   __ret = (float16x8_t) __builtin_neon_vfmaq_laneq_f16((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 40); \
  63779.   __ret; \
  63780. })
  63781. #else
  63782. #define vfmaq_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63783.   float16x8_t __ret; \
  63784.   float16x8_t __s0 = __p0; \
  63785.   float16x8_t __s1 = __p1; \
  63786.   float16x8_t __s2 = __p2; \
  63787.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  63788.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  63789.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
  63790.   __ret = (float16x8_t) __builtin_neon_vfmaq_laneq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 40); \
  63791.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  63792.   __ret; \
  63793. })
  63794. #define __noswap_vfmaq_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63795.   float16x8_t __ret; \
  63796.   float16x8_t __s0 = __p0; \
  63797.   float16x8_t __s1 = __p1; \
  63798.   float16x8_t __s2 = __p2; \
  63799.   __ret = (float16x8_t) __builtin_neon_vfmaq_laneq_f16((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 40); \
  63800.   __ret; \
  63801. })
  63802. #endif
  63803.  
  63804. #ifdef __LITTLE_ENDIAN__
  63805. #define vfma_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63806.   float16x4_t __ret; \
  63807.   float16x4_t __s0 = __p0; \
  63808.   float16x4_t __s1 = __p1; \
  63809.   float16x8_t __s2 = __p2; \
  63810.   __ret = (float16x4_t) __builtin_neon_vfma_laneq_f16((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 8); \
  63811.   __ret; \
  63812. })
  63813. #else
  63814. #define vfma_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63815.   float16x4_t __ret; \
  63816.   float16x4_t __s0 = __p0; \
  63817.   float16x4_t __s1 = __p1; \
  63818.   float16x8_t __s2 = __p2; \
  63819.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  63820.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  63821.   float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
  63822.   __ret = (float16x4_t) __builtin_neon_vfma_laneq_f16((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x16_t)__rev2, __p3, 8); \
  63823.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  63824.   __ret; \
  63825. })
  63826. #define __noswap_vfma_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
  63827.   float16x4_t __ret; \
  63828.   float16x4_t __s0 = __p0; \
  63829.   float16x4_t __s1 = __p1; \
  63830.   float16x8_t __s2 = __p2; \
  63831.   __ret = (float16x4_t) __builtin_neon_vfma_laneq_f16((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 8); \
  63832.   __ret; \
  63833. })
  63834. #endif
  63835.  
  63836. #ifdef __LITTLE_ENDIAN__
  63837. #define vfmaq_n_f16(__p0, __p1, __p2) __extension__ ({ \
  63838.   float16x8_t __ret; \
  63839.   float16x8_t __s0 = __p0; \
  63840.   float16x8_t __s1 = __p1; \
  63841.   float16_t __s2 = __p2; \
  63842.   __ret = vfmaq_f16(__s0, __s1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2}); \
  63843.   __ret; \
  63844. })
  63845. #else
  63846. #define vfmaq_n_f16(__p0, __p1, __p2) __extension__ ({ \
  63847.   float16x8_t __ret; \
  63848.   float16x8_t __s0 = __p0; \
  63849.   float16x8_t __s1 = __p1; \
  63850.   float16_t __s2 = __p2; \
  63851.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  63852.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  63853.   __ret = __noswap_vfmaq_f16(__rev0, __rev1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2}); \
  63854.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  63855.   __ret; \
  63856. })
  63857. #endif
  63858.  
  63859. #ifdef __LITTLE_ENDIAN__
  63860. #define vfma_n_f16(__p0, __p1, __p2) __extension__ ({ \
  63861.   float16x4_t __ret; \
  63862.   float16x4_t __s0 = __p0; \
  63863.   float16x4_t __s1 = __p1; \
  63864.   float16_t __s2 = __p2; \
  63865.   __ret = vfma_f16(__s0, __s1, (float16x4_t) {__s2, __s2, __s2, __s2}); \
  63866.   __ret; \
  63867. })
  63868. #else
  63869. #define vfma_n_f16(__p0, __p1, __p2) __extension__ ({ \
  63870.   float16x4_t __ret; \
  63871.   float16x4_t __s0 = __p0; \
  63872.   float16x4_t __s1 = __p1; \
  63873.   float16_t __s2 = __p2; \
  63874.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  63875.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  63876.   __ret = __noswap_vfma_f16(__rev0, __rev1, (float16x4_t) {__s2, __s2, __s2, __s2}); \
  63877.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  63878.   __ret; \
  63879. })
  63880. #endif
  63881.  
  63882. #ifdef __LITTLE_ENDIAN__
  63883. #define vfmsh_lane_f16(__p0_755, __p1_755, __p2_755, __p3_755) __extension__ ({ \
  63884.   float16_t __ret_755; \
  63885.   float16_t __s0_755 = __p0_755; \
  63886.   float16_t __s1_755 = __p1_755; \
  63887.   float16x4_t __s2_755 = __p2_755; \
  63888.   __ret_755 = vfmah_lane_f16(__s0_755, -__s1_755, __s2_755, __p3_755); \
  63889.   __ret_755; \
  63890. })
  63891. #else
  63892. #define vfmsh_lane_f16(__p0_756, __p1_756, __p2_756, __p3_756) __extension__ ({ \
  63893.   float16_t __ret_756; \
  63894.   float16_t __s0_756 = __p0_756; \
  63895.   float16_t __s1_756 = __p1_756; \
  63896.   float16x4_t __s2_756 = __p2_756; \
  63897.   float16x4_t __rev2_756;  __rev2_756 = __builtin_shufflevector(__s2_756, __s2_756, 3, 2, 1, 0); \
  63898.   __ret_756 = __noswap_vfmah_lane_f16(__s0_756, -__s1_756, __rev2_756, __p3_756); \
  63899.   __ret_756; \
  63900. })
  63901. #endif
  63902.  
  63903. #ifdef __LITTLE_ENDIAN__
  63904. #define vfmsq_lane_f16(__p0_757, __p1_757, __p2_757, __p3_757) __extension__ ({ \
  63905.   float16x8_t __ret_757; \
  63906.   float16x8_t __s0_757 = __p0_757; \
  63907.   float16x8_t __s1_757 = __p1_757; \
  63908.   float16x4_t __s2_757 = __p2_757; \
  63909.   __ret_757 = vfmaq_lane_f16(__s0_757, -__s1_757, __s2_757, __p3_757); \
  63910.   __ret_757; \
  63911. })
  63912. #else
  63913. #define vfmsq_lane_f16(__p0_758, __p1_758, __p2_758, __p3_758) __extension__ ({ \
  63914.   float16x8_t __ret_758; \
  63915.   float16x8_t __s0_758 = __p0_758; \
  63916.   float16x8_t __s1_758 = __p1_758; \
  63917.   float16x4_t __s2_758 = __p2_758; \
  63918.   float16x8_t __rev0_758;  __rev0_758 = __builtin_shufflevector(__s0_758, __s0_758, 7, 6, 5, 4, 3, 2, 1, 0); \
  63919.   float16x8_t __rev1_758;  __rev1_758 = __builtin_shufflevector(__s1_758, __s1_758, 7, 6, 5, 4, 3, 2, 1, 0); \
  63920.   float16x4_t __rev2_758;  __rev2_758 = __builtin_shufflevector(__s2_758, __s2_758, 3, 2, 1, 0); \
  63921.   __ret_758 = __noswap_vfmaq_lane_f16(__rev0_758, -__rev1_758, __rev2_758, __p3_758); \
  63922.   __ret_758 = __builtin_shufflevector(__ret_758, __ret_758, 7, 6, 5, 4, 3, 2, 1, 0); \
  63923.   __ret_758; \
  63924. })
  63925. #endif
  63926.  
  63927. #ifdef __LITTLE_ENDIAN__
  63928. #define vfms_lane_f16(__p0_759, __p1_759, __p2_759, __p3_759) __extension__ ({ \
  63929.   float16x4_t __ret_759; \
  63930.   float16x4_t __s0_759 = __p0_759; \
  63931.   float16x4_t __s1_759 = __p1_759; \
  63932.   float16x4_t __s2_759 = __p2_759; \
  63933.   __ret_759 = vfma_lane_f16(__s0_759, -__s1_759, __s2_759, __p3_759); \
  63934.   __ret_759; \
  63935. })
  63936. #else
  63937. #define vfms_lane_f16(__p0_760, __p1_760, __p2_760, __p3_760) __extension__ ({ \
  63938.   float16x4_t __ret_760; \
  63939.   float16x4_t __s0_760 = __p0_760; \
  63940.   float16x4_t __s1_760 = __p1_760; \
  63941.   float16x4_t __s2_760 = __p2_760; \
  63942.   float16x4_t __rev0_760;  __rev0_760 = __builtin_shufflevector(__s0_760, __s0_760, 3, 2, 1, 0); \
  63943.   float16x4_t __rev1_760;  __rev1_760 = __builtin_shufflevector(__s1_760, __s1_760, 3, 2, 1, 0); \
  63944.   float16x4_t __rev2_760;  __rev2_760 = __builtin_shufflevector(__s2_760, __s2_760, 3, 2, 1, 0); \
  63945.   __ret_760 = __noswap_vfma_lane_f16(__rev0_760, -__rev1_760, __rev2_760, __p3_760); \
  63946.   __ret_760 = __builtin_shufflevector(__ret_760, __ret_760, 3, 2, 1, 0); \
  63947.   __ret_760; \
  63948. })
  63949. #endif
  63950.  
  63951. #ifdef __LITTLE_ENDIAN__
  63952. #define vfmsh_laneq_f16(__p0_761, __p1_761, __p2_761, __p3_761) __extension__ ({ \
  63953.   float16_t __ret_761; \
  63954.   float16_t __s0_761 = __p0_761; \
  63955.   float16_t __s1_761 = __p1_761; \
  63956.   float16x8_t __s2_761 = __p2_761; \
  63957.   __ret_761 = vfmah_laneq_f16(__s0_761, -__s1_761, __s2_761, __p3_761); \
  63958.   __ret_761; \
  63959. })
  63960. #else
  63961. #define vfmsh_laneq_f16(__p0_762, __p1_762, __p2_762, __p3_762) __extension__ ({ \
  63962.   float16_t __ret_762; \
  63963.   float16_t __s0_762 = __p0_762; \
  63964.   float16_t __s1_762 = __p1_762; \
  63965.   float16x8_t __s2_762 = __p2_762; \
  63966.   float16x8_t __rev2_762;  __rev2_762 = __builtin_shufflevector(__s2_762, __s2_762, 7, 6, 5, 4, 3, 2, 1, 0); \
  63967.   __ret_762 = __noswap_vfmah_laneq_f16(__s0_762, -__s1_762, __rev2_762, __p3_762); \
  63968.   __ret_762; \
  63969. })
  63970. #endif
  63971.  
  63972. #ifdef __LITTLE_ENDIAN__
  63973. #define vfmsq_laneq_f16(__p0_763, __p1_763, __p2_763, __p3_763) __extension__ ({ \
  63974.   float16x8_t __ret_763; \
  63975.   float16x8_t __s0_763 = __p0_763; \
  63976.   float16x8_t __s1_763 = __p1_763; \
  63977.   float16x8_t __s2_763 = __p2_763; \
  63978.   __ret_763 = vfmaq_laneq_f16(__s0_763, -__s1_763, __s2_763, __p3_763); \
  63979.   __ret_763; \
  63980. })
  63981. #else
  63982. #define vfmsq_laneq_f16(__p0_764, __p1_764, __p2_764, __p3_764) __extension__ ({ \
  63983.   float16x8_t __ret_764; \
  63984.   float16x8_t __s0_764 = __p0_764; \
  63985.   float16x8_t __s1_764 = __p1_764; \
  63986.   float16x8_t __s2_764 = __p2_764; \
  63987.   float16x8_t __rev0_764;  __rev0_764 = __builtin_shufflevector(__s0_764, __s0_764, 7, 6, 5, 4, 3, 2, 1, 0); \
  63988.   float16x8_t __rev1_764;  __rev1_764 = __builtin_shufflevector(__s1_764, __s1_764, 7, 6, 5, 4, 3, 2, 1, 0); \
  63989.   float16x8_t __rev2_764;  __rev2_764 = __builtin_shufflevector(__s2_764, __s2_764, 7, 6, 5, 4, 3, 2, 1, 0); \
  63990.   __ret_764 = __noswap_vfmaq_laneq_f16(__rev0_764, -__rev1_764, __rev2_764, __p3_764); \
  63991.   __ret_764 = __builtin_shufflevector(__ret_764, __ret_764, 7, 6, 5, 4, 3, 2, 1, 0); \
  63992.   __ret_764; \
  63993. })
  63994. #endif
  63995.  
  63996. #ifdef __LITTLE_ENDIAN__
  63997. #define vfms_laneq_f16(__p0_765, __p1_765, __p2_765, __p3_765) __extension__ ({ \
  63998.   float16x4_t __ret_765; \
  63999.   float16x4_t __s0_765 = __p0_765; \
  64000.   float16x4_t __s1_765 = __p1_765; \
  64001.   float16x8_t __s2_765 = __p2_765; \
  64002.   __ret_765 = vfma_laneq_f16(__s0_765, -__s1_765, __s2_765, __p3_765); \
  64003.   __ret_765; \
  64004. })
  64005. #else
  64006. #define vfms_laneq_f16(__p0_766, __p1_766, __p2_766, __p3_766) __extension__ ({ \
  64007.   float16x4_t __ret_766; \
  64008.   float16x4_t __s0_766 = __p0_766; \
  64009.   float16x4_t __s1_766 = __p1_766; \
  64010.   float16x8_t __s2_766 = __p2_766; \
  64011.   float16x4_t __rev0_766;  __rev0_766 = __builtin_shufflevector(__s0_766, __s0_766, 3, 2, 1, 0); \
  64012.   float16x4_t __rev1_766;  __rev1_766 = __builtin_shufflevector(__s1_766, __s1_766, 3, 2, 1, 0); \
  64013.   float16x8_t __rev2_766;  __rev2_766 = __builtin_shufflevector(__s2_766, __s2_766, 7, 6, 5, 4, 3, 2, 1, 0); \
  64014.   __ret_766 = __noswap_vfma_laneq_f16(__rev0_766, -__rev1_766, __rev2_766, __p3_766); \
  64015.   __ret_766 = __builtin_shufflevector(__ret_766, __ret_766, 3, 2, 1, 0); \
  64016.   __ret_766; \
  64017. })
  64018. #endif
  64019.  
  64020. #ifdef __LITTLE_ENDIAN__
  64021. #define vfmsq_n_f16(__p0, __p1, __p2) __extension__ ({ \
  64022.   float16x8_t __ret; \
  64023.   float16x8_t __s0 = __p0; \
  64024.   float16x8_t __s1 = __p1; \
  64025.   float16_t __s2 = __p2; \
  64026.   __ret = vfmaq_f16(__s0, -__s1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2}); \
  64027.   __ret; \
  64028. })
  64029. #else
  64030. #define vfmsq_n_f16(__p0, __p1, __p2) __extension__ ({ \
  64031.   float16x8_t __ret; \
  64032.   float16x8_t __s0 = __p0; \
  64033.   float16x8_t __s1 = __p1; \
  64034.   float16_t __s2 = __p2; \
  64035.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  64036.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  64037.   __ret = __noswap_vfmaq_f16(__rev0, -__rev1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2}); \
  64038.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  64039.   __ret; \
  64040. })
  64041. #endif
  64042.  
  64043. #ifdef __LITTLE_ENDIAN__
  64044. #define vfms_n_f16(__p0, __p1, __p2) __extension__ ({ \
  64045.   float16x4_t __ret; \
  64046.   float16x4_t __s0 = __p0; \
  64047.   float16x4_t __s1 = __p1; \
  64048.   float16_t __s2 = __p2; \
  64049.   __ret = vfma_f16(__s0, -__s1, (float16x4_t) {__s2, __s2, __s2, __s2}); \
  64050.   __ret; \
  64051. })
  64052. #else
  64053. #define vfms_n_f16(__p0, __p1, __p2) __extension__ ({ \
  64054.   float16x4_t __ret; \
  64055.   float16x4_t __s0 = __p0; \
  64056.   float16x4_t __s1 = __p1; \
  64057.   float16_t __s2 = __p2; \
  64058.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  64059.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  64060.   __ret = __noswap_vfma_f16(__rev0, -__rev1, (float16x4_t) {__s2, __s2, __s2, __s2}); \
  64061.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  64062.   __ret; \
  64063. })
  64064. #endif
  64065.  
  64066. #ifdef __LITTLE_ENDIAN__
  64067. #define vmaxnmvq_f16(__p0) __extension__ ({ \
  64068.   float16_t __ret; \
  64069.   float16x8_t __s0 = __p0; \
  64070.   __ret = (float16_t) __builtin_neon_vmaxnmvq_f16((int8x16_t)__s0); \
  64071.   __ret; \
  64072. })
  64073. #else
  64074. #define vmaxnmvq_f16(__p0) __extension__ ({ \
  64075.   float16_t __ret; \
  64076.   float16x8_t __s0 = __p0; \
  64077.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  64078.   __ret = (float16_t) __builtin_neon_vmaxnmvq_f16((int8x16_t)__rev0); \
  64079.   __ret; \
  64080. })
  64081. #endif
  64082.  
  64083. #ifdef __LITTLE_ENDIAN__
  64084. #define vmaxnmv_f16(__p0) __extension__ ({ \
  64085.   float16_t __ret; \
  64086.   float16x4_t __s0 = __p0; \
  64087.   __ret = (float16_t) __builtin_neon_vmaxnmv_f16((int8x8_t)__s0); \
  64088.   __ret; \
  64089. })
  64090. #else
  64091. #define vmaxnmv_f16(__p0) __extension__ ({ \
  64092.   float16_t __ret; \
  64093.   float16x4_t __s0 = __p0; \
  64094.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  64095.   __ret = (float16_t) __builtin_neon_vmaxnmv_f16((int8x8_t)__rev0); \
  64096.   __ret; \
  64097. })
  64098. #endif
  64099.  
  64100. #ifdef __LITTLE_ENDIAN__
  64101. #define vmaxvq_f16(__p0) __extension__ ({ \
  64102.   float16_t __ret; \
  64103.   float16x8_t __s0 = __p0; \
  64104.   __ret = (float16_t) __builtin_neon_vmaxvq_f16((int8x16_t)__s0); \
  64105.   __ret; \
  64106. })
  64107. #else
  64108. #define vmaxvq_f16(__p0) __extension__ ({ \
  64109.   float16_t __ret; \
  64110.   float16x8_t __s0 = __p0; \
  64111.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  64112.   __ret = (float16_t) __builtin_neon_vmaxvq_f16((int8x16_t)__rev0); \
  64113.   __ret; \
  64114. })
  64115. #endif
  64116.  
  64117. #ifdef __LITTLE_ENDIAN__
  64118. #define vmaxv_f16(__p0) __extension__ ({ \
  64119.   float16_t __ret; \
  64120.   float16x4_t __s0 = __p0; \
  64121.   __ret = (float16_t) __builtin_neon_vmaxv_f16((int8x8_t)__s0); \
  64122.   __ret; \
  64123. })
  64124. #else
  64125. #define vmaxv_f16(__p0) __extension__ ({ \
  64126.   float16_t __ret; \
  64127.   float16x4_t __s0 = __p0; \
  64128.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  64129.   __ret = (float16_t) __builtin_neon_vmaxv_f16((int8x8_t)__rev0); \
  64130.   __ret; \
  64131. })
  64132. #endif
  64133.  
  64134. #ifdef __LITTLE_ENDIAN__
  64135. #define vminnmvq_f16(__p0) __extension__ ({ \
  64136.   float16_t __ret; \
  64137.   float16x8_t __s0 = __p0; \
  64138.   __ret = (float16_t) __builtin_neon_vminnmvq_f16((int8x16_t)__s0); \
  64139.   __ret; \
  64140. })
  64141. #else
  64142. #define vminnmvq_f16(__p0) __extension__ ({ \
  64143.   float16_t __ret; \
  64144.   float16x8_t __s0 = __p0; \
  64145.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  64146.   __ret = (float16_t) __builtin_neon_vminnmvq_f16((int8x16_t)__rev0); \
  64147.   __ret; \
  64148. })
  64149. #endif
  64150.  
  64151. #ifdef __LITTLE_ENDIAN__
  64152. #define vminnmv_f16(__p0) __extension__ ({ \
  64153.   float16_t __ret; \
  64154.   float16x4_t __s0 = __p0; \
  64155.   __ret = (float16_t) __builtin_neon_vminnmv_f16((int8x8_t)__s0); \
  64156.   __ret; \
  64157. })
  64158. #else
  64159. #define vminnmv_f16(__p0) __extension__ ({ \
  64160.   float16_t __ret; \
  64161.   float16x4_t __s0 = __p0; \
  64162.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  64163.   __ret = (float16_t) __builtin_neon_vminnmv_f16((int8x8_t)__rev0); \
  64164.   __ret; \
  64165. })
  64166. #endif
  64167.  
  64168. #ifdef __LITTLE_ENDIAN__
  64169. #define vminvq_f16(__p0) __extension__ ({ \
  64170.   float16_t __ret; \
  64171.   float16x8_t __s0 = __p0; \
  64172.   __ret = (float16_t) __builtin_neon_vminvq_f16((int8x16_t)__s0); \
  64173.   __ret; \
  64174. })
  64175. #else
  64176. #define vminvq_f16(__p0) __extension__ ({ \
  64177.   float16_t __ret; \
  64178.   float16x8_t __s0 = __p0; \
  64179.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  64180.   __ret = (float16_t) __builtin_neon_vminvq_f16((int8x16_t)__rev0); \
  64181.   __ret; \
  64182. })
  64183. #endif
  64184.  
  64185. #ifdef __LITTLE_ENDIAN__
  64186. #define vminv_f16(__p0) __extension__ ({ \
  64187.   float16_t __ret; \
  64188.   float16x4_t __s0 = __p0; \
  64189.   __ret = (float16_t) __builtin_neon_vminv_f16((int8x8_t)__s0); \
  64190.   __ret; \
  64191. })
  64192. #else
  64193. #define vminv_f16(__p0) __extension__ ({ \
  64194.   float16_t __ret; \
  64195.   float16x4_t __s0 = __p0; \
  64196.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  64197.   __ret = (float16_t) __builtin_neon_vminv_f16((int8x8_t)__rev0); \
  64198.   __ret; \
  64199. })
  64200. #endif
  64201.  
  64202. #ifdef __LITTLE_ENDIAN__
  64203. #define vmulq_laneq_f16(__p0_767, __p1_767, __p2_767) __extension__ ({ \
  64204.   float16x8_t __ret_767; \
  64205.   float16x8_t __s0_767 = __p0_767; \
  64206.   float16x8_t __s1_767 = __p1_767; \
  64207.   __ret_767 = __s0_767 * splatq_laneq_f16(__s1_767, __p2_767); \
  64208.   __ret_767; \
  64209. })
  64210. #else
  64211. #define vmulq_laneq_f16(__p0_768, __p1_768, __p2_768) __extension__ ({ \
  64212.   float16x8_t __ret_768; \
  64213.   float16x8_t __s0_768 = __p0_768; \
  64214.   float16x8_t __s1_768 = __p1_768; \
  64215.   float16x8_t __rev0_768;  __rev0_768 = __builtin_shufflevector(__s0_768, __s0_768, 7, 6, 5, 4, 3, 2, 1, 0); \
  64216.   float16x8_t __rev1_768;  __rev1_768 = __builtin_shufflevector(__s1_768, __s1_768, 7, 6, 5, 4, 3, 2, 1, 0); \
  64217.   __ret_768 = __rev0_768 * __noswap_splatq_laneq_f16(__rev1_768, __p2_768); \
  64218.   __ret_768 = __builtin_shufflevector(__ret_768, __ret_768, 7, 6, 5, 4, 3, 2, 1, 0); \
  64219.   __ret_768; \
  64220. })
  64221. #endif
  64222.  
  64223. #ifdef __LITTLE_ENDIAN__
  64224. #define vmul_laneq_f16(__p0_769, __p1_769, __p2_769) __extension__ ({ \
  64225.   float16x4_t __ret_769; \
  64226.   float16x4_t __s0_769 = __p0_769; \
  64227.   float16x8_t __s1_769 = __p1_769; \
  64228.   __ret_769 = __s0_769 * splat_laneq_f16(__s1_769, __p2_769); \
  64229.   __ret_769; \
  64230. })
  64231. #else
  64232. #define vmul_laneq_f16(__p0_770, __p1_770, __p2_770) __extension__ ({ \
  64233.   float16x4_t __ret_770; \
  64234.   float16x4_t __s0_770 = __p0_770; \
  64235.   float16x8_t __s1_770 = __p1_770; \
  64236.   float16x4_t __rev0_770;  __rev0_770 = __builtin_shufflevector(__s0_770, __s0_770, 3, 2, 1, 0); \
  64237.   float16x8_t __rev1_770;  __rev1_770 = __builtin_shufflevector(__s1_770, __s1_770, 7, 6, 5, 4, 3, 2, 1, 0); \
  64238.   __ret_770 = __rev0_770 * __noswap_splat_laneq_f16(__rev1_770, __p2_770); \
  64239.   __ret_770 = __builtin_shufflevector(__ret_770, __ret_770, 3, 2, 1, 0); \
  64240.   __ret_770; \
  64241. })
  64242. #endif
  64243.  
  64244. #ifdef __LITTLE_ENDIAN__
  64245. __ai __attribute__((target("fullfp16"))) float16x8_t vmulxq_f16(float16x8_t __p0, float16x8_t __p1) {
  64246.   float16x8_t __ret;
  64247.   __ret = (float16x8_t) __builtin_neon_vmulxq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  64248.   return __ret;
  64249. }
  64250. #else
  64251. __ai __attribute__((target("fullfp16"))) float16x8_t vmulxq_f16(float16x8_t __p0, float16x8_t __p1) {
  64252.   float16x8_t __ret;
  64253.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64254.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64255.   __ret = (float16x8_t) __builtin_neon_vmulxq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  64256.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64257.   return __ret;
  64258. }
  64259. __ai __attribute__((target("fullfp16"))) float16x8_t __noswap_vmulxq_f16(float16x8_t __p0, float16x8_t __p1) {
  64260.   float16x8_t __ret;
  64261.   __ret = (float16x8_t) __builtin_neon_vmulxq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  64262.   return __ret;
  64263. }
  64264. #endif
  64265.  
  64266. #ifdef __LITTLE_ENDIAN__
  64267. __ai __attribute__((target("fullfp16"))) float16x4_t vmulx_f16(float16x4_t __p0, float16x4_t __p1) {
  64268.   float16x4_t __ret;
  64269.   __ret = (float16x4_t) __builtin_neon_vmulx_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  64270.   return __ret;
  64271. }
  64272. #else
  64273. __ai __attribute__((target("fullfp16"))) float16x4_t vmulx_f16(float16x4_t __p0, float16x4_t __p1) {
  64274.   float16x4_t __ret;
  64275.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64276.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64277.   __ret = (float16x4_t) __builtin_neon_vmulx_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  64278.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64279.   return __ret;
  64280. }
  64281. __ai __attribute__((target("fullfp16"))) float16x4_t __noswap_vmulx_f16(float16x4_t __p0, float16x4_t __p1) {
  64282.   float16x4_t __ret;
  64283.   __ret = (float16x4_t) __builtin_neon_vmulx_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  64284.   return __ret;
  64285. }
  64286. #endif
  64287.  
  64288. #ifdef __LITTLE_ENDIAN__
  64289. #define vmulxh_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  64290.   float16_t __ret; \
  64291.   float16_t __s0 = __p0; \
  64292.   float16x4_t __s1 = __p1; \
  64293.   __ret = (float16_t) __builtin_neon_vmulxh_lane_f16(__s0, (float16x4_t)__s1, __p2); \
  64294.   __ret; \
  64295. })
  64296. #else
  64297. #define vmulxh_lane_f16(__p0, __p1, __p2) __extension__ ({ \
  64298.   float16_t __ret; \
  64299.   float16_t __s0 = __p0; \
  64300.   float16x4_t __s1 = __p1; \
  64301.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  64302.   __ret = (float16_t) __builtin_neon_vmulxh_lane_f16(__s0, (float16x4_t)__rev1, __p2); \
  64303.   __ret; \
  64304. })
  64305. #endif
  64306.  
  64307. #ifdef __LITTLE_ENDIAN__
  64308. #define vmulxq_lane_f16(__p0_771, __p1_771, __p2_771) __extension__ ({ \
  64309.   float16x8_t __ret_771; \
  64310.   float16x8_t __s0_771 = __p0_771; \
  64311.   float16x4_t __s1_771 = __p1_771; \
  64312.   __ret_771 = vmulxq_f16(__s0_771, splatq_lane_f16(__s1_771, __p2_771)); \
  64313.   __ret_771; \
  64314. })
  64315. #else
  64316. #define vmulxq_lane_f16(__p0_772, __p1_772, __p2_772) __extension__ ({ \
  64317.   float16x8_t __ret_772; \
  64318.   float16x8_t __s0_772 = __p0_772; \
  64319.   float16x4_t __s1_772 = __p1_772; \
  64320.   float16x8_t __rev0_772;  __rev0_772 = __builtin_shufflevector(__s0_772, __s0_772, 7, 6, 5, 4, 3, 2, 1, 0); \
  64321.   float16x4_t __rev1_772;  __rev1_772 = __builtin_shufflevector(__s1_772, __s1_772, 3, 2, 1, 0); \
  64322.   __ret_772 = __noswap_vmulxq_f16(__rev0_772, __noswap_splatq_lane_f16(__rev1_772, __p2_772)); \
  64323.   __ret_772 = __builtin_shufflevector(__ret_772, __ret_772, 7, 6, 5, 4, 3, 2, 1, 0); \
  64324.   __ret_772; \
  64325. })
  64326. #endif
  64327.  
  64328. #ifdef __LITTLE_ENDIAN__
  64329. #define vmulx_lane_f16(__p0_773, __p1_773, __p2_773) __extension__ ({ \
  64330.   float16x4_t __ret_773; \
  64331.   float16x4_t __s0_773 = __p0_773; \
  64332.   float16x4_t __s1_773 = __p1_773; \
  64333.   __ret_773 = vmulx_f16(__s0_773, splat_lane_f16(__s1_773, __p2_773)); \
  64334.   __ret_773; \
  64335. })
  64336. #else
  64337. #define vmulx_lane_f16(__p0_774, __p1_774, __p2_774) __extension__ ({ \
  64338.   float16x4_t __ret_774; \
  64339.   float16x4_t __s0_774 = __p0_774; \
  64340.   float16x4_t __s1_774 = __p1_774; \
  64341.   float16x4_t __rev0_774;  __rev0_774 = __builtin_shufflevector(__s0_774, __s0_774, 3, 2, 1, 0); \
  64342.   float16x4_t __rev1_774;  __rev1_774 = __builtin_shufflevector(__s1_774, __s1_774, 3, 2, 1, 0); \
  64343.   __ret_774 = __noswap_vmulx_f16(__rev0_774, __noswap_splat_lane_f16(__rev1_774, __p2_774)); \
  64344.   __ret_774 = __builtin_shufflevector(__ret_774, __ret_774, 3, 2, 1, 0); \
  64345.   __ret_774; \
  64346. })
  64347. #endif
  64348.  
  64349. #ifdef __LITTLE_ENDIAN__
  64350. #define vmulxh_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
  64351.   float16_t __ret; \
  64352.   float16_t __s0 = __p0; \
  64353.   float16x8_t __s1 = __p1; \
  64354.   __ret = (float16_t) __builtin_neon_vmulxh_laneq_f16(__s0, (float16x8_t)__s1, __p2); \
  64355.   __ret; \
  64356. })
  64357. #else
  64358. #define vmulxh_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
  64359.   float16_t __ret; \
  64360.   float16_t __s0 = __p0; \
  64361.   float16x8_t __s1 = __p1; \
  64362.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
  64363.   __ret = (float16_t) __builtin_neon_vmulxh_laneq_f16(__s0, (float16x8_t)__rev1, __p2); \
  64364.   __ret; \
  64365. })
  64366. #endif
  64367.  
  64368. #ifdef __LITTLE_ENDIAN__
  64369. #define vmulxq_laneq_f16(__p0_775, __p1_775, __p2_775) __extension__ ({ \
  64370.   float16x8_t __ret_775; \
  64371.   float16x8_t __s0_775 = __p0_775; \
  64372.   float16x8_t __s1_775 = __p1_775; \
  64373.   __ret_775 = vmulxq_f16(__s0_775, splatq_laneq_f16(__s1_775, __p2_775)); \
  64374.   __ret_775; \
  64375. })
  64376. #else
  64377. #define vmulxq_laneq_f16(__p0_776, __p1_776, __p2_776) __extension__ ({ \
  64378.   float16x8_t __ret_776; \
  64379.   float16x8_t __s0_776 = __p0_776; \
  64380.   float16x8_t __s1_776 = __p1_776; \
  64381.   float16x8_t __rev0_776;  __rev0_776 = __builtin_shufflevector(__s0_776, __s0_776, 7, 6, 5, 4, 3, 2, 1, 0); \
  64382.   float16x8_t __rev1_776;  __rev1_776 = __builtin_shufflevector(__s1_776, __s1_776, 7, 6, 5, 4, 3, 2, 1, 0); \
  64383.   __ret_776 = __noswap_vmulxq_f16(__rev0_776, __noswap_splatq_laneq_f16(__rev1_776, __p2_776)); \
  64384.   __ret_776 = __builtin_shufflevector(__ret_776, __ret_776, 7, 6, 5, 4, 3, 2, 1, 0); \
  64385.   __ret_776; \
  64386. })
  64387. #endif
  64388.  
  64389. #ifdef __LITTLE_ENDIAN__
  64390. #define vmulx_laneq_f16(__p0_777, __p1_777, __p2_777) __extension__ ({ \
  64391.   float16x4_t __ret_777; \
  64392.   float16x4_t __s0_777 = __p0_777; \
  64393.   float16x8_t __s1_777 = __p1_777; \
  64394.   __ret_777 = vmulx_f16(__s0_777, splat_laneq_f16(__s1_777, __p2_777)); \
  64395.   __ret_777; \
  64396. })
  64397. #else
  64398. #define vmulx_laneq_f16(__p0_778, __p1_778, __p2_778) __extension__ ({ \
  64399.   float16x4_t __ret_778; \
  64400.   float16x4_t __s0_778 = __p0_778; \
  64401.   float16x8_t __s1_778 = __p1_778; \
  64402.   float16x4_t __rev0_778;  __rev0_778 = __builtin_shufflevector(__s0_778, __s0_778, 3, 2, 1, 0); \
  64403.   float16x8_t __rev1_778;  __rev1_778 = __builtin_shufflevector(__s1_778, __s1_778, 7, 6, 5, 4, 3, 2, 1, 0); \
  64404.   __ret_778 = __noswap_vmulx_f16(__rev0_778, __noswap_splat_laneq_f16(__rev1_778, __p2_778)); \
  64405.   __ret_778 = __builtin_shufflevector(__ret_778, __ret_778, 3, 2, 1, 0); \
  64406.   __ret_778; \
  64407. })
  64408. #endif
  64409.  
  64410. #ifdef __LITTLE_ENDIAN__
  64411. #define vmulxq_n_f16(__p0, __p1) __extension__ ({ \
  64412.   float16x8_t __ret; \
  64413.   float16x8_t __s0 = __p0; \
  64414.   float16_t __s1 = __p1; \
  64415.   __ret = vmulxq_f16(__s0, (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1}); \
  64416.   __ret; \
  64417. })
  64418. #else
  64419. #define vmulxq_n_f16(__p0, __p1) __extension__ ({ \
  64420.   float16x8_t __ret; \
  64421.   float16x8_t __s0 = __p0; \
  64422.   float16_t __s1 = __p1; \
  64423.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
  64424.   __ret = __noswap_vmulxq_f16(__rev0, (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1}); \
  64425.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
  64426.   __ret; \
  64427. })
  64428. #endif
  64429.  
  64430. #ifdef __LITTLE_ENDIAN__
  64431. #define vmulx_n_f16(__p0, __p1) __extension__ ({ \
  64432.   float16x4_t __ret; \
  64433.   float16x4_t __s0 = __p0; \
  64434.   float16_t __s1 = __p1; \
  64435.   __ret = vmulx_f16(__s0, (float16x4_t) {__s1, __s1, __s1, __s1}); \
  64436.   __ret; \
  64437. })
  64438. #else
  64439. #define vmulx_n_f16(__p0, __p1) __extension__ ({ \
  64440.   float16x4_t __ret; \
  64441.   float16x4_t __s0 = __p0; \
  64442.   float16_t __s1 = __p1; \
  64443.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  64444.   __ret = __noswap_vmulx_f16(__rev0, (float16x4_t) {__s1, __s1, __s1, __s1}); \
  64445.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  64446.   __ret; \
  64447. })
  64448. #endif
  64449.  
  64450. #ifdef __LITTLE_ENDIAN__
  64451. __ai __attribute__((target("fullfp16"))) float16x8_t vpaddq_f16(float16x8_t __p0, float16x8_t __p1) {
  64452.   float16x8_t __ret;
  64453.   __ret = (float16x8_t) __builtin_neon_vpaddq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  64454.   return __ret;
  64455. }
  64456. #else
  64457. __ai __attribute__((target("fullfp16"))) float16x8_t vpaddq_f16(float16x8_t __p0, float16x8_t __p1) {
  64458.   float16x8_t __ret;
  64459.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64460.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64461.   __ret = (float16x8_t) __builtin_neon_vpaddq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  64462.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64463.   return __ret;
  64464. }
  64465. #endif
  64466.  
  64467. #ifdef __LITTLE_ENDIAN__
  64468. __ai __attribute__((target("fullfp16"))) float16x8_t vpmaxq_f16(float16x8_t __p0, float16x8_t __p1) {
  64469.   float16x8_t __ret;
  64470.   __ret = (float16x8_t) __builtin_neon_vpmaxq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  64471.   return __ret;
  64472. }
  64473. #else
  64474. __ai __attribute__((target("fullfp16"))) float16x8_t vpmaxq_f16(float16x8_t __p0, float16x8_t __p1) {
  64475.   float16x8_t __ret;
  64476.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64477.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64478.   __ret = (float16x8_t) __builtin_neon_vpmaxq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  64479.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64480.   return __ret;
  64481. }
  64482. #endif
  64483.  
  64484. #ifdef __LITTLE_ENDIAN__
  64485. __ai __attribute__((target("fullfp16"))) float16x8_t vpmaxnmq_f16(float16x8_t __p0, float16x8_t __p1) {
  64486.   float16x8_t __ret;
  64487.   __ret = (float16x8_t) __builtin_neon_vpmaxnmq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  64488.   return __ret;
  64489. }
  64490. #else
  64491. __ai __attribute__((target("fullfp16"))) float16x8_t vpmaxnmq_f16(float16x8_t __p0, float16x8_t __p1) {
  64492.   float16x8_t __ret;
  64493.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64494.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64495.   __ret = (float16x8_t) __builtin_neon_vpmaxnmq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  64496.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64497.   return __ret;
  64498. }
  64499. #endif
  64500.  
  64501. #ifdef __LITTLE_ENDIAN__
  64502. __ai __attribute__((target("fullfp16"))) float16x4_t vpmaxnm_f16(float16x4_t __p0, float16x4_t __p1) {
  64503.   float16x4_t __ret;
  64504.   __ret = (float16x4_t) __builtin_neon_vpmaxnm_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  64505.   return __ret;
  64506. }
  64507. #else
  64508. __ai __attribute__((target("fullfp16"))) float16x4_t vpmaxnm_f16(float16x4_t __p0, float16x4_t __p1) {
  64509.   float16x4_t __ret;
  64510.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64511.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64512.   __ret = (float16x4_t) __builtin_neon_vpmaxnm_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  64513.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64514.   return __ret;
  64515. }
  64516. #endif
  64517.  
  64518. #ifdef __LITTLE_ENDIAN__
  64519. __ai __attribute__((target("fullfp16"))) float16x8_t vpminq_f16(float16x8_t __p0, float16x8_t __p1) {
  64520.   float16x8_t __ret;
  64521.   __ret = (float16x8_t) __builtin_neon_vpminq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  64522.   return __ret;
  64523. }
  64524. #else
  64525. __ai __attribute__((target("fullfp16"))) float16x8_t vpminq_f16(float16x8_t __p0, float16x8_t __p1) {
  64526.   float16x8_t __ret;
  64527.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64528.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64529.   __ret = (float16x8_t) __builtin_neon_vpminq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  64530.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64531.   return __ret;
  64532. }
  64533. #endif
  64534.  
  64535. #ifdef __LITTLE_ENDIAN__
  64536. __ai __attribute__((target("fullfp16"))) float16x8_t vpminnmq_f16(float16x8_t __p0, float16x8_t __p1) {
  64537.   float16x8_t __ret;
  64538.   __ret = (float16x8_t) __builtin_neon_vpminnmq_f16((int8x16_t)__p0, (int8x16_t)__p1, 40);
  64539.   return __ret;
  64540. }
  64541. #else
  64542. __ai __attribute__((target("fullfp16"))) float16x8_t vpminnmq_f16(float16x8_t __p0, float16x8_t __p1) {
  64543.   float16x8_t __ret;
  64544.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64545.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64546.   __ret = (float16x8_t) __builtin_neon_vpminnmq_f16((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
  64547.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64548.   return __ret;
  64549. }
  64550. #endif
  64551.  
  64552. #ifdef __LITTLE_ENDIAN__
  64553. __ai __attribute__((target("fullfp16"))) float16x4_t vpminnm_f16(float16x4_t __p0, float16x4_t __p1) {
  64554.   float16x4_t __ret;
  64555.   __ret = (float16x4_t) __builtin_neon_vpminnm_f16((int8x8_t)__p0, (int8x8_t)__p1, 8);
  64556.   return __ret;
  64557. }
  64558. #else
  64559. __ai __attribute__((target("fullfp16"))) float16x4_t vpminnm_f16(float16x4_t __p0, float16x4_t __p1) {
  64560.   float16x4_t __ret;
  64561.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64562.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64563.   __ret = (float16x4_t) __builtin_neon_vpminnm_f16((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
  64564.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64565.   return __ret;
  64566. }
  64567. #endif
  64568.  
  64569. #ifdef __LITTLE_ENDIAN__
  64570. __ai __attribute__((target("fullfp16"))) float16x8_t vrndiq_f16(float16x8_t __p0) {
  64571.   float16x8_t __ret;
  64572.   __ret = (float16x8_t) __builtin_neon_vrndiq_f16((int8x16_t)__p0, 40);
  64573.   return __ret;
  64574. }
  64575. #else
  64576. __ai __attribute__((target("fullfp16"))) float16x8_t vrndiq_f16(float16x8_t __p0) {
  64577.   float16x8_t __ret;
  64578.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64579.   __ret = (float16x8_t) __builtin_neon_vrndiq_f16((int8x16_t)__rev0, 40);
  64580.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64581.   return __ret;
  64582. }
  64583. #endif
  64584.  
  64585. #ifdef __LITTLE_ENDIAN__
  64586. __ai __attribute__((target("fullfp16"))) float16x4_t vrndi_f16(float16x4_t __p0) {
  64587.   float16x4_t __ret;
  64588.   __ret = (float16x4_t) __builtin_neon_vrndi_f16((int8x8_t)__p0, 8);
  64589.   return __ret;
  64590. }
  64591. #else
  64592. __ai __attribute__((target("fullfp16"))) float16x4_t vrndi_f16(float16x4_t __p0) {
  64593.   float16x4_t __ret;
  64594.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64595.   __ret = (float16x4_t) __builtin_neon_vrndi_f16((int8x8_t)__rev0, 8);
  64596.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64597.   return __ret;
  64598. }
  64599. #endif
  64600.  
  64601. #ifdef __LITTLE_ENDIAN__
  64602. __ai __attribute__((target("fullfp16"))) float16x8_t vsqrtq_f16(float16x8_t __p0) {
  64603.   float16x8_t __ret;
  64604.   __ret = (float16x8_t) __builtin_neon_vsqrtq_f16((int8x16_t)__p0, 40);
  64605.   return __ret;
  64606. }
  64607. #else
  64608. __ai __attribute__((target("fullfp16"))) float16x8_t vsqrtq_f16(float16x8_t __p0) {
  64609.   float16x8_t __ret;
  64610.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64611.   __ret = (float16x8_t) __builtin_neon_vsqrtq_f16((int8x16_t)__rev0, 40);
  64612.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64613.   return __ret;
  64614. }
  64615. #endif
  64616.  
  64617. #ifdef __LITTLE_ENDIAN__
  64618. __ai __attribute__((target("fullfp16"))) float16x4_t vsqrt_f16(float16x4_t __p0) {
  64619.   float16x4_t __ret;
  64620.   __ret = (float16x4_t) __builtin_neon_vsqrt_f16((int8x8_t)__p0, 8);
  64621.   return __ret;
  64622. }
  64623. #else
  64624. __ai __attribute__((target("fullfp16"))) float16x4_t vsqrt_f16(float16x4_t __p0) {
  64625.   float16x4_t __ret;
  64626.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64627.   __ret = (float16x4_t) __builtin_neon_vsqrt_f16((int8x8_t)__rev0, 8);
  64628.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64629.   return __ret;
  64630. }
  64631. #endif
  64632.  
  64633. #ifdef __LITTLE_ENDIAN__
  64634. __ai __attribute__((target("fullfp16"))) float16x8_t vtrn1q_f16(float16x8_t __p0, float16x8_t __p1) {
  64635.   float16x8_t __ret;
  64636.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
  64637.   return __ret;
  64638. }
  64639. #else
  64640. __ai __attribute__((target("fullfp16"))) float16x8_t vtrn1q_f16(float16x8_t __p0, float16x8_t __p1) {
  64641.   float16x8_t __ret;
  64642.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64643.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64644.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
  64645.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64646.   return __ret;
  64647. }
  64648. #endif
  64649.  
  64650. #ifdef __LITTLE_ENDIAN__
  64651. __ai __attribute__((target("fullfp16"))) float16x4_t vtrn1_f16(float16x4_t __p0, float16x4_t __p1) {
  64652.   float16x4_t __ret;
  64653.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
  64654.   return __ret;
  64655. }
  64656. #else
  64657. __ai __attribute__((target("fullfp16"))) float16x4_t vtrn1_f16(float16x4_t __p0, float16x4_t __p1) {
  64658.   float16x4_t __ret;
  64659.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64660.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64661.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
  64662.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64663.   return __ret;
  64664. }
  64665. #endif
  64666.  
  64667. #ifdef __LITTLE_ENDIAN__
  64668. __ai __attribute__((target("fullfp16"))) float16x8_t vtrn2q_f16(float16x8_t __p0, float16x8_t __p1) {
  64669.   float16x8_t __ret;
  64670.   __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
  64671.   return __ret;
  64672. }
  64673. #else
  64674. __ai __attribute__((target("fullfp16"))) float16x8_t vtrn2q_f16(float16x8_t __p0, float16x8_t __p1) {
  64675.   float16x8_t __ret;
  64676.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64677.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64678.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
  64679.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64680.   return __ret;
  64681. }
  64682. #endif
  64683.  
  64684. #ifdef __LITTLE_ENDIAN__
  64685. __ai __attribute__((target("fullfp16"))) float16x4_t vtrn2_f16(float16x4_t __p0, float16x4_t __p1) {
  64686.   float16x4_t __ret;
  64687.   __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
  64688.   return __ret;
  64689. }
  64690. #else
  64691. __ai __attribute__((target("fullfp16"))) float16x4_t vtrn2_f16(float16x4_t __p0, float16x4_t __p1) {
  64692.   float16x4_t __ret;
  64693.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64694.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64695.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
  64696.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64697.   return __ret;
  64698. }
  64699. #endif
  64700.  
  64701. #ifdef __LITTLE_ENDIAN__
  64702. __ai __attribute__((target("fullfp16"))) float16x8_t vuzp1q_f16(float16x8_t __p0, float16x8_t __p1) {
  64703.   float16x8_t __ret;
  64704.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
  64705.   return __ret;
  64706. }
  64707. #else
  64708. __ai __attribute__((target("fullfp16"))) float16x8_t vuzp1q_f16(float16x8_t __p0, float16x8_t __p1) {
  64709.   float16x8_t __ret;
  64710.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64711.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64712.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
  64713.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64714.   return __ret;
  64715. }
  64716. #endif
  64717.  
  64718. #ifdef __LITTLE_ENDIAN__
  64719. __ai __attribute__((target("fullfp16"))) float16x4_t vuzp1_f16(float16x4_t __p0, float16x4_t __p1) {
  64720.   float16x4_t __ret;
  64721.   __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
  64722.   return __ret;
  64723. }
  64724. #else
  64725. __ai __attribute__((target("fullfp16"))) float16x4_t vuzp1_f16(float16x4_t __p0, float16x4_t __p1) {
  64726.   float16x4_t __ret;
  64727.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64728.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64729.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
  64730.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64731.   return __ret;
  64732. }
  64733. #endif
  64734.  
  64735. #ifdef __LITTLE_ENDIAN__
  64736. __ai __attribute__((target("fullfp16"))) float16x8_t vuzp2q_f16(float16x8_t __p0, float16x8_t __p1) {
  64737.   float16x8_t __ret;
  64738.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
  64739.   return __ret;
  64740. }
  64741. #else
  64742. __ai __attribute__((target("fullfp16"))) float16x8_t vuzp2q_f16(float16x8_t __p0, float16x8_t __p1) {
  64743.   float16x8_t __ret;
  64744.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64745.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64746.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
  64747.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64748.   return __ret;
  64749. }
  64750. #endif
  64751.  
  64752. #ifdef __LITTLE_ENDIAN__
  64753. __ai __attribute__((target("fullfp16"))) float16x4_t vuzp2_f16(float16x4_t __p0, float16x4_t __p1) {
  64754.   float16x4_t __ret;
  64755.   __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
  64756.   return __ret;
  64757. }
  64758. #else
  64759. __ai __attribute__((target("fullfp16"))) float16x4_t vuzp2_f16(float16x4_t __p0, float16x4_t __p1) {
  64760.   float16x4_t __ret;
  64761.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64762.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64763.   __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
  64764.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64765.   return __ret;
  64766. }
  64767. #endif
  64768.  
  64769. #ifdef __LITTLE_ENDIAN__
  64770. __ai __attribute__((target("fullfp16"))) float16x8_t vzip1q_f16(float16x8_t __p0, float16x8_t __p1) {
  64771.   float16x8_t __ret;
  64772.   __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
  64773.   return __ret;
  64774. }
  64775. #else
  64776. __ai __attribute__((target("fullfp16"))) float16x8_t vzip1q_f16(float16x8_t __p0, float16x8_t __p1) {
  64777.   float16x8_t __ret;
  64778.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64779.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64780.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
  64781.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64782.   return __ret;
  64783. }
  64784. #endif
  64785.  
  64786. #ifdef __LITTLE_ENDIAN__
  64787. __ai __attribute__((target("fullfp16"))) float16x4_t vzip1_f16(float16x4_t __p0, float16x4_t __p1) {
  64788.   float16x4_t __ret;
  64789.   __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
  64790.   return __ret;
  64791. }
  64792. #else
  64793. __ai __attribute__((target("fullfp16"))) float16x4_t vzip1_f16(float16x4_t __p0, float16x4_t __p1) {
  64794.   float16x4_t __ret;
  64795.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64796.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64797.   __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
  64798.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64799.   return __ret;
  64800. }
  64801. #endif
  64802.  
  64803. #ifdef __LITTLE_ENDIAN__
  64804. __ai __attribute__((target("fullfp16"))) float16x8_t vzip2q_f16(float16x8_t __p0, float16x8_t __p1) {
  64805.   float16x8_t __ret;
  64806.   __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
  64807.   return __ret;
  64808. }
  64809. #else
  64810. __ai __attribute__((target("fullfp16"))) float16x8_t vzip2q_f16(float16x8_t __p0, float16x8_t __p1) {
  64811.   float16x8_t __ret;
  64812.   float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  64813.   float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  64814.   __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
  64815.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  64816.   return __ret;
  64817. }
  64818. #endif
  64819.  
  64820. #ifdef __LITTLE_ENDIAN__
  64821. __ai __attribute__((target("fullfp16"))) float16x4_t vzip2_f16(float16x4_t __p0, float16x4_t __p1) {
  64822.   float16x4_t __ret;
  64823.   __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
  64824.   return __ret;
  64825. }
  64826. #else
  64827. __ai __attribute__((target("fullfp16"))) float16x4_t vzip2_f16(float16x4_t __p0, float16x4_t __p1) {
  64828.   float16x4_t __ret;
  64829.   float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64830.   float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64831.   __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
  64832.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64833.   return __ret;
  64834. }
  64835. #endif
  64836.  
  64837. #ifdef __LITTLE_ENDIAN__
  64838. #define vsudotq_laneq_s32(__p0_779, __p1_779, __p2_779, __p3_779) __extension__ ({ \
  64839.   int32x4_t __ret_779; \
  64840.   int32x4_t __s0_779 = __p0_779; \
  64841.   int8x16_t __s1_779 = __p1_779; \
  64842.   uint8x16_t __s2_779 = __p2_779; \
  64843. uint8x16_t __reint_779 = __s2_779; \
  64844.   __ret_779 = vusdotq_s32(__s0_779, (uint8x16_t)(splatq_laneq_s32(*(int32x4_t *) &__reint_779, __p3_779)), __s1_779); \
  64845.   __ret_779; \
  64846. })
  64847. #else
  64848. #define vsudotq_laneq_s32(__p0_780, __p1_780, __p2_780, __p3_780) __extension__ ({ \
  64849.   int32x4_t __ret_780; \
  64850.   int32x4_t __s0_780 = __p0_780; \
  64851.   int8x16_t __s1_780 = __p1_780; \
  64852.   uint8x16_t __s2_780 = __p2_780; \
  64853.   int32x4_t __rev0_780;  __rev0_780 = __builtin_shufflevector(__s0_780, __s0_780, 3, 2, 1, 0); \
  64854.   int8x16_t __rev1_780;  __rev1_780 = __builtin_shufflevector(__s1_780, __s1_780, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  64855.   uint8x16_t __rev2_780;  __rev2_780 = __builtin_shufflevector(__s2_780, __s2_780, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  64856. uint8x16_t __reint_780 = __rev2_780; \
  64857.   __ret_780 = __noswap_vusdotq_s32(__rev0_780, (uint8x16_t)(__noswap_splatq_laneq_s32(*(int32x4_t *) &__reint_780, __p3_780)), __rev1_780); \
  64858.   __ret_780 = __builtin_shufflevector(__ret_780, __ret_780, 3, 2, 1, 0); \
  64859.   __ret_780; \
  64860. })
  64861. #endif
  64862.  
  64863. #ifdef __LITTLE_ENDIAN__
  64864. #define vsudot_laneq_s32(__p0_781, __p1_781, __p2_781, __p3_781) __extension__ ({ \
  64865.   int32x2_t __ret_781; \
  64866.   int32x2_t __s0_781 = __p0_781; \
  64867.   int8x8_t __s1_781 = __p1_781; \
  64868.   uint8x16_t __s2_781 = __p2_781; \
  64869. uint8x16_t __reint_781 = __s2_781; \
  64870.   __ret_781 = vusdot_s32(__s0_781, (uint8x8_t)(splat_laneq_s32(*(int32x4_t *) &__reint_781, __p3_781)), __s1_781); \
  64871.   __ret_781; \
  64872. })
  64873. #else
  64874. #define vsudot_laneq_s32(__p0_782, __p1_782, __p2_782, __p3_782) __extension__ ({ \
  64875.   int32x2_t __ret_782; \
  64876.   int32x2_t __s0_782 = __p0_782; \
  64877.   int8x8_t __s1_782 = __p1_782; \
  64878.   uint8x16_t __s2_782 = __p2_782; \
  64879.   int32x2_t __rev0_782;  __rev0_782 = __builtin_shufflevector(__s0_782, __s0_782, 1, 0); \
  64880.   int8x8_t __rev1_782;  __rev1_782 = __builtin_shufflevector(__s1_782, __s1_782, 7, 6, 5, 4, 3, 2, 1, 0); \
  64881.   uint8x16_t __rev2_782;  __rev2_782 = __builtin_shufflevector(__s2_782, __s2_782, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  64882. uint8x16_t __reint_782 = __rev2_782; \
  64883.   __ret_782 = __noswap_vusdot_s32(__rev0_782, (uint8x8_t)(__noswap_splat_laneq_s32(*(int32x4_t *) &__reint_782, __p3_782)), __rev1_782); \
  64884.   __ret_782 = __builtin_shufflevector(__ret_782, __ret_782, 1, 0); \
  64885.   __ret_782; \
  64886. })
  64887. #endif
  64888.  
  64889. #ifdef __LITTLE_ENDIAN__
  64890. #define vusdotq_laneq_s32(__p0_783, __p1_783, __p2_783, __p3_783) __extension__ ({ \
  64891.   int32x4_t __ret_783; \
  64892.   int32x4_t __s0_783 = __p0_783; \
  64893.   uint8x16_t __s1_783 = __p1_783; \
  64894.   int8x16_t __s2_783 = __p2_783; \
  64895. int8x16_t __reint_783 = __s2_783; \
  64896.   __ret_783 = vusdotq_s32(__s0_783, __s1_783, (int8x16_t)(splatq_laneq_s32(*(int32x4_t *) &__reint_783, __p3_783))); \
  64897.   __ret_783; \
  64898. })
  64899. #else
  64900. #define vusdotq_laneq_s32(__p0_784, __p1_784, __p2_784, __p3_784) __extension__ ({ \
  64901.   int32x4_t __ret_784; \
  64902.   int32x4_t __s0_784 = __p0_784; \
  64903.   uint8x16_t __s1_784 = __p1_784; \
  64904.   int8x16_t __s2_784 = __p2_784; \
  64905.   int32x4_t __rev0_784;  __rev0_784 = __builtin_shufflevector(__s0_784, __s0_784, 3, 2, 1, 0); \
  64906.   uint8x16_t __rev1_784;  __rev1_784 = __builtin_shufflevector(__s1_784, __s1_784, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  64907.   int8x16_t __rev2_784;  __rev2_784 = __builtin_shufflevector(__s2_784, __s2_784, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  64908. int8x16_t __reint_784 = __rev2_784; \
  64909.   __ret_784 = __noswap_vusdotq_s32(__rev0_784, __rev1_784, (int8x16_t)(__noswap_splatq_laneq_s32(*(int32x4_t *) &__reint_784, __p3_784))); \
  64910.   __ret_784 = __builtin_shufflevector(__ret_784, __ret_784, 3, 2, 1, 0); \
  64911.   __ret_784; \
  64912. })
  64913. #endif
  64914.  
  64915. #ifdef __LITTLE_ENDIAN__
  64916. #define vusdot_laneq_s32(__p0_785, __p1_785, __p2_785, __p3_785) __extension__ ({ \
  64917.   int32x2_t __ret_785; \
  64918.   int32x2_t __s0_785 = __p0_785; \
  64919.   uint8x8_t __s1_785 = __p1_785; \
  64920.   int8x16_t __s2_785 = __p2_785; \
  64921. int8x16_t __reint_785 = __s2_785; \
  64922.   __ret_785 = vusdot_s32(__s0_785, __s1_785, (int8x8_t)(splat_laneq_s32(*(int32x4_t *) &__reint_785, __p3_785))); \
  64923.   __ret_785; \
  64924. })
  64925. #else
  64926. #define vusdot_laneq_s32(__p0_786, __p1_786, __p2_786, __p3_786) __extension__ ({ \
  64927.   int32x2_t __ret_786; \
  64928.   int32x2_t __s0_786 = __p0_786; \
  64929.   uint8x8_t __s1_786 = __p1_786; \
  64930.   int8x16_t __s2_786 = __p2_786; \
  64931.   int32x2_t __rev0_786;  __rev0_786 = __builtin_shufflevector(__s0_786, __s0_786, 1, 0); \
  64932.   uint8x8_t __rev1_786;  __rev1_786 = __builtin_shufflevector(__s1_786, __s1_786, 7, 6, 5, 4, 3, 2, 1, 0); \
  64933.   int8x16_t __rev2_786;  __rev2_786 = __builtin_shufflevector(__s2_786, __s2_786, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  64934. int8x16_t __reint_786 = __rev2_786; \
  64935.   __ret_786 = __noswap_vusdot_s32(__rev0_786, __rev1_786, (int8x8_t)(__noswap_splat_laneq_s32(*(int32x4_t *) &__reint_786, __p3_786))); \
  64936.   __ret_786 = __builtin_shufflevector(__ret_786, __ret_786, 1, 0); \
  64937.   __ret_786; \
  64938. })
  64939. #endif
  64940.  
  64941. #ifdef __LITTLE_ENDIAN__
  64942. __ai __attribute__((target("sha3"))) uint8x16_t vbcaxq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  64943.   uint8x16_t __ret;
  64944.   __ret = (uint8x16_t) __builtin_neon_vbcaxq_u8((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 48);
  64945.   return __ret;
  64946. }
  64947. #else
  64948. __ai __attribute__((target("sha3"))) uint8x16_t vbcaxq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  64949.   uint8x16_t __ret;
  64950.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  64951.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  64952.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  64953.   __ret = (uint8x16_t) __builtin_neon_vbcaxq_u8((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 48);
  64954.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  64955.   return __ret;
  64956. }
  64957. #endif
  64958.  
  64959. #ifdef __LITTLE_ENDIAN__
  64960. __ai __attribute__((target("sha3"))) uint32x4_t vbcaxq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  64961.   uint32x4_t __ret;
  64962.   __ret = (uint32x4_t) __builtin_neon_vbcaxq_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  64963.   return __ret;
  64964. }
  64965. #else
  64966. __ai __attribute__((target("sha3"))) uint32x4_t vbcaxq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  64967.   uint32x4_t __ret;
  64968.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  64969.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  64970.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  64971.   __ret = (uint32x4_t) __builtin_neon_vbcaxq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  64972.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  64973.   return __ret;
  64974. }
  64975. #endif
  64976.  
  64977. #ifdef __LITTLE_ENDIAN__
  64978. __ai __attribute__((target("sha3"))) uint64x2_t vbcaxq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  64979.   uint64x2_t __ret;
  64980.   __ret = (uint64x2_t) __builtin_neon_vbcaxq_u64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 51);
  64981.   return __ret;
  64982. }
  64983. #else
  64984. __ai __attribute__((target("sha3"))) uint64x2_t vbcaxq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  64985.   uint64x2_t __ret;
  64986.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  64987.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  64988.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  64989.   __ret = (uint64x2_t) __builtin_neon_vbcaxq_u64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 51);
  64990.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  64991.   return __ret;
  64992. }
  64993. #endif
  64994.  
  64995. #ifdef __LITTLE_ENDIAN__
  64996. __ai __attribute__((target("sha3"))) uint16x8_t vbcaxq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  64997.   uint16x8_t __ret;
  64998.   __ret = (uint16x8_t) __builtin_neon_vbcaxq_u16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 49);
  64999.   return __ret;
  65000. }
  65001. #else
  65002. __ai __attribute__((target("sha3"))) uint16x8_t vbcaxq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  65003.   uint16x8_t __ret;
  65004.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  65005.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  65006.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  65007.   __ret = (uint16x8_t) __builtin_neon_vbcaxq_u16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 49);
  65008.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  65009.   return __ret;
  65010. }
  65011. #endif
  65012.  
  65013. #ifdef __LITTLE_ENDIAN__
  65014. __ai __attribute__((target("sha3"))) int8x16_t vbcaxq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  65015.   int8x16_t __ret;
  65016.   __ret = (int8x16_t) __builtin_neon_vbcaxq_s8((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 32);
  65017.   return __ret;
  65018. }
  65019. #else
  65020. __ai __attribute__((target("sha3"))) int8x16_t vbcaxq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  65021.   int8x16_t __ret;
  65022.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65023.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65024.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65025.   __ret = (int8x16_t) __builtin_neon_vbcaxq_s8((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 32);
  65026.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65027.   return __ret;
  65028. }
  65029. #endif
  65030.  
  65031. #ifdef __LITTLE_ENDIAN__
  65032. __ai __attribute__((target("sha3"))) int32x4_t vbcaxq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  65033.   int32x4_t __ret;
  65034.   __ret = (int32x4_t) __builtin_neon_vbcaxq_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  65035.   return __ret;
  65036. }
  65037. #else
  65038. __ai __attribute__((target("sha3"))) int32x4_t vbcaxq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  65039.   int32x4_t __ret;
  65040.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  65041.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  65042.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  65043.   __ret = (int32x4_t) __builtin_neon_vbcaxq_s32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  65044.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  65045.   return __ret;
  65046. }
  65047. #endif
  65048.  
  65049. #ifdef __LITTLE_ENDIAN__
  65050. __ai __attribute__((target("sha3"))) int64x2_t vbcaxq_s64(int64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  65051.   int64x2_t __ret;
  65052.   __ret = (int64x2_t) __builtin_neon_vbcaxq_s64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 35);
  65053.   return __ret;
  65054. }
  65055. #else
  65056. __ai __attribute__((target("sha3"))) int64x2_t vbcaxq_s64(int64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  65057.   int64x2_t __ret;
  65058.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65059.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65060.   int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  65061.   __ret = (int64x2_t) __builtin_neon_vbcaxq_s64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 35);
  65062.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65063.   return __ret;
  65064. }
  65065. #endif
  65066.  
  65067. #ifdef __LITTLE_ENDIAN__
  65068. __ai __attribute__((target("sha3"))) int16x8_t vbcaxq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  65069.   int16x8_t __ret;
  65070.   __ret = (int16x8_t) __builtin_neon_vbcaxq_s16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
  65071.   return __ret;
  65072. }
  65073. #else
  65074. __ai __attribute__((target("sha3"))) int16x8_t vbcaxq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  65075.   int16x8_t __ret;
  65076.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  65077.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  65078.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  65079.   __ret = (int16x8_t) __builtin_neon_vbcaxq_s16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 33);
  65080.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  65081.   return __ret;
  65082. }
  65083. #endif
  65084.  
  65085. #ifdef __LITTLE_ENDIAN__
  65086. __ai __attribute__((target("sha3"))) uint8x16_t veor3q_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  65087.   uint8x16_t __ret;
  65088.   __ret = (uint8x16_t) __builtin_neon_veor3q_u8((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 48);
  65089.   return __ret;
  65090. }
  65091. #else
  65092. __ai __attribute__((target("sha3"))) uint8x16_t veor3q_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  65093.   uint8x16_t __ret;
  65094.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65095.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65096.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65097.   __ret = (uint8x16_t) __builtin_neon_veor3q_u8((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 48);
  65098.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65099.   return __ret;
  65100. }
  65101. #endif
  65102.  
  65103. #ifdef __LITTLE_ENDIAN__
  65104. __ai __attribute__((target("sha3"))) uint32x4_t veor3q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  65105.   uint32x4_t __ret;
  65106.   __ret = (uint32x4_t) __builtin_neon_veor3q_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  65107.   return __ret;
  65108. }
  65109. #else
  65110. __ai __attribute__((target("sha3"))) uint32x4_t veor3q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  65111.   uint32x4_t __ret;
  65112.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  65113.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  65114.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  65115.   __ret = (uint32x4_t) __builtin_neon_veor3q_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  65116.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  65117.   return __ret;
  65118. }
  65119. #endif
  65120.  
  65121. #ifdef __LITTLE_ENDIAN__
  65122. __ai __attribute__((target("sha3"))) uint64x2_t veor3q_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  65123.   uint64x2_t __ret;
  65124.   __ret = (uint64x2_t) __builtin_neon_veor3q_u64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 51);
  65125.   return __ret;
  65126. }
  65127. #else
  65128. __ai __attribute__((target("sha3"))) uint64x2_t veor3q_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  65129.   uint64x2_t __ret;
  65130.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65131.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65132.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  65133.   __ret = (uint64x2_t) __builtin_neon_veor3q_u64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 51);
  65134.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65135.   return __ret;
  65136. }
  65137. #endif
  65138.  
  65139. #ifdef __LITTLE_ENDIAN__
  65140. __ai __attribute__((target("sha3"))) uint16x8_t veor3q_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  65141.   uint16x8_t __ret;
  65142.   __ret = (uint16x8_t) __builtin_neon_veor3q_u16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 49);
  65143.   return __ret;
  65144. }
  65145. #else
  65146. __ai __attribute__((target("sha3"))) uint16x8_t veor3q_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  65147.   uint16x8_t __ret;
  65148.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  65149.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  65150.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  65151.   __ret = (uint16x8_t) __builtin_neon_veor3q_u16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 49);
  65152.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  65153.   return __ret;
  65154. }
  65155. #endif
  65156.  
  65157. #ifdef __LITTLE_ENDIAN__
  65158. __ai __attribute__((target("sha3"))) int8x16_t veor3q_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  65159.   int8x16_t __ret;
  65160.   __ret = (int8x16_t) __builtin_neon_veor3q_s8((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 32);
  65161.   return __ret;
  65162. }
  65163. #else
  65164. __ai __attribute__((target("sha3"))) int8x16_t veor3q_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  65165.   int8x16_t __ret;
  65166.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65167.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65168.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65169.   __ret = (int8x16_t) __builtin_neon_veor3q_s8((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 32);
  65170.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  65171.   return __ret;
  65172. }
  65173. #endif
  65174.  
  65175. #ifdef __LITTLE_ENDIAN__
  65176. __ai __attribute__((target("sha3"))) int32x4_t veor3q_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  65177.   int32x4_t __ret;
  65178.   __ret = (int32x4_t) __builtin_neon_veor3q_s32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
  65179.   return __ret;
  65180. }
  65181. #else
  65182. __ai __attribute__((target("sha3"))) int32x4_t veor3q_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  65183.   int32x4_t __ret;
  65184.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  65185.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  65186.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  65187.   __ret = (int32x4_t) __builtin_neon_veor3q_s32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
  65188.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  65189.   return __ret;
  65190. }
  65191. #endif
  65192.  
  65193. #ifdef __LITTLE_ENDIAN__
  65194. __ai __attribute__((target("sha3"))) int64x2_t veor3q_s64(int64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  65195.   int64x2_t __ret;
  65196.   __ret = (int64x2_t) __builtin_neon_veor3q_s64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 35);
  65197.   return __ret;
  65198. }
  65199. #else
  65200. __ai __attribute__((target("sha3"))) int64x2_t veor3q_s64(int64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
  65201.   int64x2_t __ret;
  65202.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65203.   int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65204.   int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  65205.   __ret = (int64x2_t) __builtin_neon_veor3q_s64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 35);
  65206.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65207.   return __ret;
  65208. }
  65209. #endif
  65210.  
  65211. #ifdef __LITTLE_ENDIAN__
  65212. __ai __attribute__((target("sha3"))) int16x8_t veor3q_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  65213.   int16x8_t __ret;
  65214.   __ret = (int16x8_t) __builtin_neon_veor3q_s16((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
  65215.   return __ret;
  65216. }
  65217. #else
  65218. __ai __attribute__((target("sha3"))) int16x8_t veor3q_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  65219.   int16x8_t __ret;
  65220.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  65221.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  65222.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  65223.   __ret = (int16x8_t) __builtin_neon_veor3q_s16((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 33);
  65224.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  65225.   return __ret;
  65226. }
  65227. #endif
  65228.  
  65229. #ifdef __LITTLE_ENDIAN__
  65230. __ai __attribute__((target("sha3"))) uint64x2_t vrax1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  65231.   uint64x2_t __ret;
  65232.   __ret = (uint64x2_t) __builtin_neon_vrax1q_u64((int8x16_t)__p0, (int8x16_t)__p1, 51);
  65233.   return __ret;
  65234. }
  65235. #else
  65236. __ai __attribute__((target("sha3"))) uint64x2_t vrax1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  65237.   uint64x2_t __ret;
  65238.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65239.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65240.   __ret = (uint64x2_t) __builtin_neon_vrax1q_u64((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  65241.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65242.   return __ret;
  65243. }
  65244. #endif
  65245.  
  65246. #ifdef __LITTLE_ENDIAN__
  65247. __ai __attribute__((target("sha3"))) uint64x2_t vsha512hq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  65248.   uint64x2_t __ret;
  65249.   __ret = (uint64x2_t) __builtin_neon_vsha512hq_u64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 51);
  65250.   return __ret;
  65251. }
  65252. #else
  65253. __ai __attribute__((target("sha3"))) uint64x2_t vsha512hq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  65254.   uint64x2_t __ret;
  65255.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65256.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65257.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  65258.   __ret = (uint64x2_t) __builtin_neon_vsha512hq_u64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 51);
  65259.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65260.   return __ret;
  65261. }
  65262. #endif
  65263.  
  65264. #ifdef __LITTLE_ENDIAN__
  65265. __ai __attribute__((target("sha3"))) uint64x2_t vsha512h2q_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  65266.   uint64x2_t __ret;
  65267.   __ret = (uint64x2_t) __builtin_neon_vsha512h2q_u64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 51);
  65268.   return __ret;
  65269. }
  65270. #else
  65271. __ai __attribute__((target("sha3"))) uint64x2_t vsha512h2q_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  65272.   uint64x2_t __ret;
  65273.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65274.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65275.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  65276.   __ret = (uint64x2_t) __builtin_neon_vsha512h2q_u64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 51);
  65277.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65278.   return __ret;
  65279. }
  65280. #endif
  65281.  
  65282. #ifdef __LITTLE_ENDIAN__
  65283. __ai __attribute__((target("sha3"))) uint64x2_t vsha512su0q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  65284.   uint64x2_t __ret;
  65285.   __ret = (uint64x2_t) __builtin_neon_vsha512su0q_u64((int8x16_t)__p0, (int8x16_t)__p1, 51);
  65286.   return __ret;
  65287. }
  65288. #else
  65289. __ai __attribute__((target("sha3"))) uint64x2_t vsha512su0q_u64(uint64x2_t __p0, uint64x2_t __p1) {
  65290.   uint64x2_t __ret;
  65291.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65292.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65293.   __ret = (uint64x2_t) __builtin_neon_vsha512su0q_u64((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
  65294.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65295.   return __ret;
  65296. }
  65297. #endif
  65298.  
  65299. #ifdef __LITTLE_ENDIAN__
  65300. __ai __attribute__((target("sha3"))) uint64x2_t vsha512su1q_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  65301.   uint64x2_t __ret;
  65302.   __ret = (uint64x2_t) __builtin_neon_vsha512su1q_u64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 51);
  65303.   return __ret;
  65304. }
  65305. #else
  65306. __ai __attribute__((target("sha3"))) uint64x2_t vsha512su1q_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
  65307.   uint64x2_t __ret;
  65308.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65309.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65310.   uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  65311.   __ret = (uint64x2_t) __builtin_neon_vsha512su1q_u64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 51);
  65312.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65313.   return __ret;
  65314. }
  65315. #endif
  65316.  
  65317. #ifdef __LITTLE_ENDIAN__
  65318. #define vxarq_u64(__p0, __p1, __p2) __extension__ ({ \
  65319.   uint64x2_t __ret; \
  65320.   uint64x2_t __s0 = __p0; \
  65321.   uint64x2_t __s1 = __p1; \
  65322.   __ret = (uint64x2_t) __builtin_neon_vxarq_u64((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
  65323.   __ret; \
  65324. })
  65325. #else
  65326. #define vxarq_u64(__p0, __p1, __p2) __extension__ ({ \
  65327.   uint64x2_t __ret; \
  65328.   uint64x2_t __s0 = __p0; \
  65329.   uint64x2_t __s1 = __p1; \
  65330.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
  65331.   uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
  65332.   __ret = (uint64x2_t) __builtin_neon_vxarq_u64((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
  65333.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
  65334.   __ret; \
  65335. })
  65336. #endif
  65337.  
  65338. #ifdef __LITTLE_ENDIAN__
  65339. __ai __attribute__((target("sm4"))) uint32x4_t vsm3partw1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  65340.   uint32x4_t __ret;
  65341.   __ret = (uint32x4_t) __builtin_neon_vsm3partw1q_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  65342.   return __ret;
  65343. }
  65344. #else
  65345. __ai __attribute__((target("sm4"))) uint32x4_t vsm3partw1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  65346.   uint32x4_t __ret;
  65347.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  65348.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  65349.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  65350.   __ret = (uint32x4_t) __builtin_neon_vsm3partw1q_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  65351.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  65352.   return __ret;
  65353. }
  65354. #endif
  65355.  
  65356. #ifdef __LITTLE_ENDIAN__
  65357. __ai __attribute__((target("sm4"))) uint32x4_t vsm3partw2q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  65358.   uint32x4_t __ret;
  65359.   __ret = (uint32x4_t) __builtin_neon_vsm3partw2q_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  65360.   return __ret;
  65361. }
  65362. #else
  65363. __ai __attribute__((target("sm4"))) uint32x4_t vsm3partw2q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  65364.   uint32x4_t __ret;
  65365.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  65366.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  65367.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  65368.   __ret = (uint32x4_t) __builtin_neon_vsm3partw2q_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  65369.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  65370.   return __ret;
  65371. }
  65372. #endif
  65373.  
  65374. #ifdef __LITTLE_ENDIAN__
  65375. __ai __attribute__((target("sm4"))) uint32x4_t vsm3ss1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  65376.   uint32x4_t __ret;
  65377.   __ret = (uint32x4_t) __builtin_neon_vsm3ss1q_u32((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
  65378.   return __ret;
  65379. }
  65380. #else
  65381. __ai __attribute__((target("sm4"))) uint32x4_t vsm3ss1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  65382.   uint32x4_t __ret;
  65383.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  65384.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  65385.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  65386.   __ret = (uint32x4_t) __builtin_neon_vsm3ss1q_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
  65387.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  65388.   return __ret;
  65389. }
  65390. #endif
  65391.  
  65392. #ifdef __LITTLE_ENDIAN__
  65393. #define vsm3tt1aq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
  65394.   uint32x4_t __ret; \
  65395.   uint32x4_t __s0 = __p0; \
  65396.   uint32x4_t __s1 = __p1; \
  65397.   uint32x4_t __s2 = __p2; \
  65398.   __ret = (uint32x4_t) __builtin_neon_vsm3tt1aq_u32((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 50); \
  65399.   __ret; \
  65400. })
  65401. #else
  65402. #define vsm3tt1aq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
  65403.   uint32x4_t __ret; \
  65404.   uint32x4_t __s0 = __p0; \
  65405.   uint32x4_t __s1 = __p1; \
  65406.   uint32x4_t __s2 = __p2; \
  65407.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  65408.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  65409.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  65410.   __ret = (uint32x4_t) __builtin_neon_vsm3tt1aq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 50); \
  65411.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  65412.   __ret; \
  65413. })
  65414. #endif
  65415.  
  65416. #ifdef __LITTLE_ENDIAN__
  65417. #define vsm3tt1bq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
  65418.   uint32x4_t __ret; \
  65419.   uint32x4_t __s0 = __p0; \
  65420.   uint32x4_t __s1 = __p1; \
  65421.   uint32x4_t __s2 = __p2; \
  65422.   __ret = (uint32x4_t) __builtin_neon_vsm3tt1bq_u32((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 50); \
  65423.   __ret; \
  65424. })
  65425. #else
  65426. #define vsm3tt1bq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
  65427.   uint32x4_t __ret; \
  65428.   uint32x4_t __s0 = __p0; \
  65429.   uint32x4_t __s1 = __p1; \
  65430.   uint32x4_t __s2 = __p2; \
  65431.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  65432.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  65433.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  65434.   __ret = (uint32x4_t) __builtin_neon_vsm3tt1bq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 50); \
  65435.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  65436.   __ret; \
  65437. })
  65438. #endif
  65439.  
  65440. #ifdef __LITTLE_ENDIAN__
  65441. #define vsm3tt2aq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
  65442.   uint32x4_t __ret; \
  65443.   uint32x4_t __s0 = __p0; \
  65444.   uint32x4_t __s1 = __p1; \
  65445.   uint32x4_t __s2 = __p2; \
  65446.   __ret = (uint32x4_t) __builtin_neon_vsm3tt2aq_u32((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 50); \
  65447.   __ret; \
  65448. })
  65449. #else
  65450. #define vsm3tt2aq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
  65451.   uint32x4_t __ret; \
  65452.   uint32x4_t __s0 = __p0; \
  65453.   uint32x4_t __s1 = __p1; \
  65454.   uint32x4_t __s2 = __p2; \
  65455.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  65456.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  65457.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  65458.   __ret = (uint32x4_t) __builtin_neon_vsm3tt2aq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 50); \
  65459.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  65460.   __ret; \
  65461. })
  65462. #endif
  65463.  
  65464. #ifdef __LITTLE_ENDIAN__
  65465. #define vsm3tt2bq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
  65466.   uint32x4_t __ret; \
  65467.   uint32x4_t __s0 = __p0; \
  65468.   uint32x4_t __s1 = __p1; \
  65469.   uint32x4_t __s2 = __p2; \
  65470.   __ret = (uint32x4_t) __builtin_neon_vsm3tt2bq_u32((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 50); \
  65471.   __ret; \
  65472. })
  65473. #else
  65474. #define vsm3tt2bq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
  65475.   uint32x4_t __ret; \
  65476.   uint32x4_t __s0 = __p0; \
  65477.   uint32x4_t __s1 = __p1; \
  65478.   uint32x4_t __s2 = __p2; \
  65479.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
  65480.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
  65481.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
  65482.   __ret = (uint32x4_t) __builtin_neon_vsm3tt2bq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 50); \
  65483.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
  65484.   __ret; \
  65485. })
  65486. #endif
  65487.  
  65488. #ifdef __LITTLE_ENDIAN__
  65489. __ai __attribute__((target("sm4"))) uint32x4_t vsm4eq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  65490.   uint32x4_t __ret;
  65491.   __ret = (uint32x4_t) __builtin_neon_vsm4eq_u32((int8x16_t)__p0, (int8x16_t)__p1, 50);
  65492.   return __ret;
  65493. }
  65494. #else
  65495. __ai __attribute__((target("sm4"))) uint32x4_t vsm4eq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  65496.   uint32x4_t __ret;
  65497.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  65498.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  65499.   __ret = (uint32x4_t) __builtin_neon_vsm4eq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  65500.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  65501.   return __ret;
  65502. }
  65503. #endif
  65504.  
  65505. #ifdef __LITTLE_ENDIAN__
  65506. __ai __attribute__((target("sm4"))) uint32x4_t vsm4ekeyq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  65507.   uint32x4_t __ret;
  65508.   __ret = (uint32x4_t) __builtin_neon_vsm4ekeyq_u32((int8x16_t)__p0, (int8x16_t)__p1, 50);
  65509.   return __ret;
  65510. }
  65511. #else
  65512. __ai __attribute__((target("sm4"))) uint32x4_t vsm4ekeyq_u32(uint32x4_t __p0, uint32x4_t __p1) {
  65513.   uint32x4_t __ret;
  65514.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  65515.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  65516.   __ret = (uint32x4_t) __builtin_neon_vsm4ekeyq_u32((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
  65517.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  65518.   return __ret;
  65519. }
  65520. #endif
  65521.  
  65522. __ai __attribute__((target("v8.1a"))) int32_t vqrdmlahs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
  65523.   int32_t __ret;
  65524.   __ret = (int32_t) __builtin_neon_vqrdmlahs_s32(__p0, __p1, __p2);
  65525.   return __ret;
  65526. }
  65527. __ai __attribute__((target("v8.1a"))) int16_t vqrdmlahh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
  65528.   int16_t __ret;
  65529.   __ret = (int16_t) __builtin_neon_vqrdmlahh_s16(__p0, __p1, __p2);
  65530.   return __ret;
  65531. }
  65532. #ifdef __LITTLE_ENDIAN__
  65533. #define vqrdmlahs_lane_s32(__p0_787, __p1_787, __p2_787, __p3_787) __extension__ ({ \
  65534.   int32_t __ret_787; \
  65535.   int32_t __s0_787 = __p0_787; \
  65536.   int32_t __s1_787 = __p1_787; \
  65537.   int32x2_t __s2_787 = __p2_787; \
  65538.   __ret_787 = vqrdmlahs_s32(__s0_787, __s1_787, vget_lane_s32(__s2_787, __p3_787)); \
  65539.   __ret_787; \
  65540. })
  65541. #else
  65542. #define vqrdmlahs_lane_s32(__p0_788, __p1_788, __p2_788, __p3_788) __extension__ ({ \
  65543.   int32_t __ret_788; \
  65544.   int32_t __s0_788 = __p0_788; \
  65545.   int32_t __s1_788 = __p1_788; \
  65546.   int32x2_t __s2_788 = __p2_788; \
  65547.   int32x2_t __rev2_788;  __rev2_788 = __builtin_shufflevector(__s2_788, __s2_788, 1, 0); \
  65548.   __ret_788 = vqrdmlahs_s32(__s0_788, __s1_788, __noswap_vget_lane_s32(__rev2_788, __p3_788)); \
  65549.   __ret_788; \
  65550. })
  65551. #endif
  65552.  
  65553. #ifdef __LITTLE_ENDIAN__
  65554. #define vqrdmlahh_lane_s16(__p0_789, __p1_789, __p2_789, __p3_789) __extension__ ({ \
  65555.   int16_t __ret_789; \
  65556.   int16_t __s0_789 = __p0_789; \
  65557.   int16_t __s1_789 = __p1_789; \
  65558.   int16x4_t __s2_789 = __p2_789; \
  65559.   __ret_789 = vqrdmlahh_s16(__s0_789, __s1_789, vget_lane_s16(__s2_789, __p3_789)); \
  65560.   __ret_789; \
  65561. })
  65562. #else
  65563. #define vqrdmlahh_lane_s16(__p0_790, __p1_790, __p2_790, __p3_790) __extension__ ({ \
  65564.   int16_t __ret_790; \
  65565.   int16_t __s0_790 = __p0_790; \
  65566.   int16_t __s1_790 = __p1_790; \
  65567.   int16x4_t __s2_790 = __p2_790; \
  65568.   int16x4_t __rev2_790;  __rev2_790 = __builtin_shufflevector(__s2_790, __s2_790, 3, 2, 1, 0); \
  65569.   __ret_790 = vqrdmlahh_s16(__s0_790, __s1_790, __noswap_vget_lane_s16(__rev2_790, __p3_790)); \
  65570.   __ret_790; \
  65571. })
  65572. #endif
  65573.  
  65574. #ifdef __LITTLE_ENDIAN__
  65575. #define vqrdmlahs_laneq_s32(__p0_791, __p1_791, __p2_791, __p3_791) __extension__ ({ \
  65576.   int32_t __ret_791; \
  65577.   int32_t __s0_791 = __p0_791; \
  65578.   int32_t __s1_791 = __p1_791; \
  65579.   int32x4_t __s2_791 = __p2_791; \
  65580.   __ret_791 = vqrdmlahs_s32(__s0_791, __s1_791, vgetq_lane_s32(__s2_791, __p3_791)); \
  65581.   __ret_791; \
  65582. })
  65583. #else
  65584. #define vqrdmlahs_laneq_s32(__p0_792, __p1_792, __p2_792, __p3_792) __extension__ ({ \
  65585.   int32_t __ret_792; \
  65586.   int32_t __s0_792 = __p0_792; \
  65587.   int32_t __s1_792 = __p1_792; \
  65588.   int32x4_t __s2_792 = __p2_792; \
  65589.   int32x4_t __rev2_792;  __rev2_792 = __builtin_shufflevector(__s2_792, __s2_792, 3, 2, 1, 0); \
  65590.   __ret_792 = vqrdmlahs_s32(__s0_792, __s1_792, __noswap_vgetq_lane_s32(__rev2_792, __p3_792)); \
  65591.   __ret_792; \
  65592. })
  65593. #endif
  65594.  
  65595. #ifdef __LITTLE_ENDIAN__
  65596. #define vqrdmlahh_laneq_s16(__p0_793, __p1_793, __p2_793, __p3_793) __extension__ ({ \
  65597.   int16_t __ret_793; \
  65598.   int16_t __s0_793 = __p0_793; \
  65599.   int16_t __s1_793 = __p1_793; \
  65600.   int16x8_t __s2_793 = __p2_793; \
  65601.   __ret_793 = vqrdmlahh_s16(__s0_793, __s1_793, vgetq_lane_s16(__s2_793, __p3_793)); \
  65602.   __ret_793; \
  65603. })
  65604. #else
  65605. #define vqrdmlahh_laneq_s16(__p0_794, __p1_794, __p2_794, __p3_794) __extension__ ({ \
  65606.   int16_t __ret_794; \
  65607.   int16_t __s0_794 = __p0_794; \
  65608.   int16_t __s1_794 = __p1_794; \
  65609.   int16x8_t __s2_794 = __p2_794; \
  65610.   int16x8_t __rev2_794;  __rev2_794 = __builtin_shufflevector(__s2_794, __s2_794, 7, 6, 5, 4, 3, 2, 1, 0); \
  65611.   __ret_794 = vqrdmlahh_s16(__s0_794, __s1_794, __noswap_vgetq_lane_s16(__rev2_794, __p3_794)); \
  65612.   __ret_794; \
  65613. })
  65614. #endif
  65615.  
  65616. #ifdef __LITTLE_ENDIAN__
  65617. #define vqrdmlahq_laneq_s32(__p0_795, __p1_795, __p2_795, __p3_795) __extension__ ({ \
  65618.   int32x4_t __ret_795; \
  65619.   int32x4_t __s0_795 = __p0_795; \
  65620.   int32x4_t __s1_795 = __p1_795; \
  65621.   int32x4_t __s2_795 = __p2_795; \
  65622.   __ret_795 = vqrdmlahq_s32(__s0_795, __s1_795, splatq_laneq_s32(__s2_795, __p3_795)); \
  65623.   __ret_795; \
  65624. })
  65625. #else
  65626. #define vqrdmlahq_laneq_s32(__p0_796, __p1_796, __p2_796, __p3_796) __extension__ ({ \
  65627.   int32x4_t __ret_796; \
  65628.   int32x4_t __s0_796 = __p0_796; \
  65629.   int32x4_t __s1_796 = __p1_796; \
  65630.   int32x4_t __s2_796 = __p2_796; \
  65631.   int32x4_t __rev0_796;  __rev0_796 = __builtin_shufflevector(__s0_796, __s0_796, 3, 2, 1, 0); \
  65632.   int32x4_t __rev1_796;  __rev1_796 = __builtin_shufflevector(__s1_796, __s1_796, 3, 2, 1, 0); \
  65633.   int32x4_t __rev2_796;  __rev2_796 = __builtin_shufflevector(__s2_796, __s2_796, 3, 2, 1, 0); \
  65634.   __ret_796 = __noswap_vqrdmlahq_s32(__rev0_796, __rev1_796, __noswap_splatq_laneq_s32(__rev2_796, __p3_796)); \
  65635.   __ret_796 = __builtin_shufflevector(__ret_796, __ret_796, 3, 2, 1, 0); \
  65636.   __ret_796; \
  65637. })
  65638. #endif
  65639.  
  65640. #ifdef __LITTLE_ENDIAN__
  65641. #define vqrdmlahq_laneq_s16(__p0_797, __p1_797, __p2_797, __p3_797) __extension__ ({ \
  65642.   int16x8_t __ret_797; \
  65643.   int16x8_t __s0_797 = __p0_797; \
  65644.   int16x8_t __s1_797 = __p1_797; \
  65645.   int16x8_t __s2_797 = __p2_797; \
  65646.   __ret_797 = vqrdmlahq_s16(__s0_797, __s1_797, splatq_laneq_s16(__s2_797, __p3_797)); \
  65647.   __ret_797; \
  65648. })
  65649. #else
  65650. #define vqrdmlahq_laneq_s16(__p0_798, __p1_798, __p2_798, __p3_798) __extension__ ({ \
  65651.   int16x8_t __ret_798; \
  65652.   int16x8_t __s0_798 = __p0_798; \
  65653.   int16x8_t __s1_798 = __p1_798; \
  65654.   int16x8_t __s2_798 = __p2_798; \
  65655.   int16x8_t __rev0_798;  __rev0_798 = __builtin_shufflevector(__s0_798, __s0_798, 7, 6, 5, 4, 3, 2, 1, 0); \
  65656.   int16x8_t __rev1_798;  __rev1_798 = __builtin_shufflevector(__s1_798, __s1_798, 7, 6, 5, 4, 3, 2, 1, 0); \
  65657.   int16x8_t __rev2_798;  __rev2_798 = __builtin_shufflevector(__s2_798, __s2_798, 7, 6, 5, 4, 3, 2, 1, 0); \
  65658.   __ret_798 = __noswap_vqrdmlahq_s16(__rev0_798, __rev1_798, __noswap_splatq_laneq_s16(__rev2_798, __p3_798)); \
  65659.   __ret_798 = __builtin_shufflevector(__ret_798, __ret_798, 7, 6, 5, 4, 3, 2, 1, 0); \
  65660.   __ret_798; \
  65661. })
  65662. #endif
  65663.  
  65664. #ifdef __LITTLE_ENDIAN__
  65665. #define vqrdmlah_laneq_s32(__p0_799, __p1_799, __p2_799, __p3_799) __extension__ ({ \
  65666.   int32x2_t __ret_799; \
  65667.   int32x2_t __s0_799 = __p0_799; \
  65668.   int32x2_t __s1_799 = __p1_799; \
  65669.   int32x4_t __s2_799 = __p2_799; \
  65670.   __ret_799 = vqrdmlah_s32(__s0_799, __s1_799, splat_laneq_s32(__s2_799, __p3_799)); \
  65671.   __ret_799; \
  65672. })
  65673. #else
  65674. #define vqrdmlah_laneq_s32(__p0_800, __p1_800, __p2_800, __p3_800) __extension__ ({ \
  65675.   int32x2_t __ret_800; \
  65676.   int32x2_t __s0_800 = __p0_800; \
  65677.   int32x2_t __s1_800 = __p1_800; \
  65678.   int32x4_t __s2_800 = __p2_800; \
  65679.   int32x2_t __rev0_800;  __rev0_800 = __builtin_shufflevector(__s0_800, __s0_800, 1, 0); \
  65680.   int32x2_t __rev1_800;  __rev1_800 = __builtin_shufflevector(__s1_800, __s1_800, 1, 0); \
  65681.   int32x4_t __rev2_800;  __rev2_800 = __builtin_shufflevector(__s2_800, __s2_800, 3, 2, 1, 0); \
  65682.   __ret_800 = __noswap_vqrdmlah_s32(__rev0_800, __rev1_800, __noswap_splat_laneq_s32(__rev2_800, __p3_800)); \
  65683.   __ret_800 = __builtin_shufflevector(__ret_800, __ret_800, 1, 0); \
  65684.   __ret_800; \
  65685. })
  65686. #endif
  65687.  
  65688. #ifdef __LITTLE_ENDIAN__
  65689. #define vqrdmlah_laneq_s16(__p0_801, __p1_801, __p2_801, __p3_801) __extension__ ({ \
  65690.   int16x4_t __ret_801; \
  65691.   int16x4_t __s0_801 = __p0_801; \
  65692.   int16x4_t __s1_801 = __p1_801; \
  65693.   int16x8_t __s2_801 = __p2_801; \
  65694.   __ret_801 = vqrdmlah_s16(__s0_801, __s1_801, splat_laneq_s16(__s2_801, __p3_801)); \
  65695.   __ret_801; \
  65696. })
  65697. #else
  65698. #define vqrdmlah_laneq_s16(__p0_802, __p1_802, __p2_802, __p3_802) __extension__ ({ \
  65699.   int16x4_t __ret_802; \
  65700.   int16x4_t __s0_802 = __p0_802; \
  65701.   int16x4_t __s1_802 = __p1_802; \
  65702.   int16x8_t __s2_802 = __p2_802; \
  65703.   int16x4_t __rev0_802;  __rev0_802 = __builtin_shufflevector(__s0_802, __s0_802, 3, 2, 1, 0); \
  65704.   int16x4_t __rev1_802;  __rev1_802 = __builtin_shufflevector(__s1_802, __s1_802, 3, 2, 1, 0); \
  65705.   int16x8_t __rev2_802;  __rev2_802 = __builtin_shufflevector(__s2_802, __s2_802, 7, 6, 5, 4, 3, 2, 1, 0); \
  65706.   __ret_802 = __noswap_vqrdmlah_s16(__rev0_802, __rev1_802, __noswap_splat_laneq_s16(__rev2_802, __p3_802)); \
  65707.   __ret_802 = __builtin_shufflevector(__ret_802, __ret_802, 3, 2, 1, 0); \
  65708.   __ret_802; \
  65709. })
  65710. #endif
  65711.  
  65712. __ai __attribute__((target("v8.1a"))) int32_t vqrdmlshs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
  65713.   int32_t __ret;
  65714.   __ret = (int32_t) __builtin_neon_vqrdmlshs_s32(__p0, __p1, __p2);
  65715.   return __ret;
  65716. }
  65717. __ai __attribute__((target("v8.1a"))) int16_t vqrdmlshh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
  65718.   int16_t __ret;
  65719.   __ret = (int16_t) __builtin_neon_vqrdmlshh_s16(__p0, __p1, __p2);
  65720.   return __ret;
  65721. }
  65722. #ifdef __LITTLE_ENDIAN__
  65723. #define vqrdmlshs_lane_s32(__p0_803, __p1_803, __p2_803, __p3_803) __extension__ ({ \
  65724.   int32_t __ret_803; \
  65725.   int32_t __s0_803 = __p0_803; \
  65726.   int32_t __s1_803 = __p1_803; \
  65727.   int32x2_t __s2_803 = __p2_803; \
  65728.   __ret_803 = vqrdmlshs_s32(__s0_803, __s1_803, vget_lane_s32(__s2_803, __p3_803)); \
  65729.   __ret_803; \
  65730. })
  65731. #else
  65732. #define vqrdmlshs_lane_s32(__p0_804, __p1_804, __p2_804, __p3_804) __extension__ ({ \
  65733.   int32_t __ret_804; \
  65734.   int32_t __s0_804 = __p0_804; \
  65735.   int32_t __s1_804 = __p1_804; \
  65736.   int32x2_t __s2_804 = __p2_804; \
  65737.   int32x2_t __rev2_804;  __rev2_804 = __builtin_shufflevector(__s2_804, __s2_804, 1, 0); \
  65738.   __ret_804 = vqrdmlshs_s32(__s0_804, __s1_804, __noswap_vget_lane_s32(__rev2_804, __p3_804)); \
  65739.   __ret_804; \
  65740. })
  65741. #endif
  65742.  
  65743. #ifdef __LITTLE_ENDIAN__
  65744. #define vqrdmlshh_lane_s16(__p0_805, __p1_805, __p2_805, __p3_805) __extension__ ({ \
  65745.   int16_t __ret_805; \
  65746.   int16_t __s0_805 = __p0_805; \
  65747.   int16_t __s1_805 = __p1_805; \
  65748.   int16x4_t __s2_805 = __p2_805; \
  65749.   __ret_805 = vqrdmlshh_s16(__s0_805, __s1_805, vget_lane_s16(__s2_805, __p3_805)); \
  65750.   __ret_805; \
  65751. })
  65752. #else
  65753. #define vqrdmlshh_lane_s16(__p0_806, __p1_806, __p2_806, __p3_806) __extension__ ({ \
  65754.   int16_t __ret_806; \
  65755.   int16_t __s0_806 = __p0_806; \
  65756.   int16_t __s1_806 = __p1_806; \
  65757.   int16x4_t __s2_806 = __p2_806; \
  65758.   int16x4_t __rev2_806;  __rev2_806 = __builtin_shufflevector(__s2_806, __s2_806, 3, 2, 1, 0); \
  65759.   __ret_806 = vqrdmlshh_s16(__s0_806, __s1_806, __noswap_vget_lane_s16(__rev2_806, __p3_806)); \
  65760.   __ret_806; \
  65761. })
  65762. #endif
  65763.  
  65764. #ifdef __LITTLE_ENDIAN__
  65765. #define vqrdmlshs_laneq_s32(__p0_807, __p1_807, __p2_807, __p3_807) __extension__ ({ \
  65766.   int32_t __ret_807; \
  65767.   int32_t __s0_807 = __p0_807; \
  65768.   int32_t __s1_807 = __p1_807; \
  65769.   int32x4_t __s2_807 = __p2_807; \
  65770.   __ret_807 = vqrdmlshs_s32(__s0_807, __s1_807, vgetq_lane_s32(__s2_807, __p3_807)); \
  65771.   __ret_807; \
  65772. })
  65773. #else
  65774. #define vqrdmlshs_laneq_s32(__p0_808, __p1_808, __p2_808, __p3_808) __extension__ ({ \
  65775.   int32_t __ret_808; \
  65776.   int32_t __s0_808 = __p0_808; \
  65777.   int32_t __s1_808 = __p1_808; \
  65778.   int32x4_t __s2_808 = __p2_808; \
  65779.   int32x4_t __rev2_808;  __rev2_808 = __builtin_shufflevector(__s2_808, __s2_808, 3, 2, 1, 0); \
  65780.   __ret_808 = vqrdmlshs_s32(__s0_808, __s1_808, __noswap_vgetq_lane_s32(__rev2_808, __p3_808)); \
  65781.   __ret_808; \
  65782. })
  65783. #endif
  65784.  
  65785. #ifdef __LITTLE_ENDIAN__
  65786. #define vqrdmlshh_laneq_s16(__p0_809, __p1_809, __p2_809, __p3_809) __extension__ ({ \
  65787.   int16_t __ret_809; \
  65788.   int16_t __s0_809 = __p0_809; \
  65789.   int16_t __s1_809 = __p1_809; \
  65790.   int16x8_t __s2_809 = __p2_809; \
  65791.   __ret_809 = vqrdmlshh_s16(__s0_809, __s1_809, vgetq_lane_s16(__s2_809, __p3_809)); \
  65792.   __ret_809; \
  65793. })
  65794. #else
  65795. #define vqrdmlshh_laneq_s16(__p0_810, __p1_810, __p2_810, __p3_810) __extension__ ({ \
  65796.   int16_t __ret_810; \
  65797.   int16_t __s0_810 = __p0_810; \
  65798.   int16_t __s1_810 = __p1_810; \
  65799.   int16x8_t __s2_810 = __p2_810; \
  65800.   int16x8_t __rev2_810;  __rev2_810 = __builtin_shufflevector(__s2_810, __s2_810, 7, 6, 5, 4, 3, 2, 1, 0); \
  65801.   __ret_810 = vqrdmlshh_s16(__s0_810, __s1_810, __noswap_vgetq_lane_s16(__rev2_810, __p3_810)); \
  65802.   __ret_810; \
  65803. })
  65804. #endif
  65805.  
  65806. #ifdef __LITTLE_ENDIAN__
  65807. #define vqrdmlshq_laneq_s32(__p0_811, __p1_811, __p2_811, __p3_811) __extension__ ({ \
  65808.   int32x4_t __ret_811; \
  65809.   int32x4_t __s0_811 = __p0_811; \
  65810.   int32x4_t __s1_811 = __p1_811; \
  65811.   int32x4_t __s2_811 = __p2_811; \
  65812.   __ret_811 = vqrdmlshq_s32(__s0_811, __s1_811, splatq_laneq_s32(__s2_811, __p3_811)); \
  65813.   __ret_811; \
  65814. })
  65815. #else
  65816. #define vqrdmlshq_laneq_s32(__p0_812, __p1_812, __p2_812, __p3_812) __extension__ ({ \
  65817.   int32x4_t __ret_812; \
  65818.   int32x4_t __s0_812 = __p0_812; \
  65819.   int32x4_t __s1_812 = __p1_812; \
  65820.   int32x4_t __s2_812 = __p2_812; \
  65821.   int32x4_t __rev0_812;  __rev0_812 = __builtin_shufflevector(__s0_812, __s0_812, 3, 2, 1, 0); \
  65822.   int32x4_t __rev1_812;  __rev1_812 = __builtin_shufflevector(__s1_812, __s1_812, 3, 2, 1, 0); \
  65823.   int32x4_t __rev2_812;  __rev2_812 = __builtin_shufflevector(__s2_812, __s2_812, 3, 2, 1, 0); \
  65824.   __ret_812 = __noswap_vqrdmlshq_s32(__rev0_812, __rev1_812, __noswap_splatq_laneq_s32(__rev2_812, __p3_812)); \
  65825.   __ret_812 = __builtin_shufflevector(__ret_812, __ret_812, 3, 2, 1, 0); \
  65826.   __ret_812; \
  65827. })
  65828. #endif
  65829.  
  65830. #ifdef __LITTLE_ENDIAN__
  65831. #define vqrdmlshq_laneq_s16(__p0_813, __p1_813, __p2_813, __p3_813) __extension__ ({ \
  65832.   int16x8_t __ret_813; \
  65833.   int16x8_t __s0_813 = __p0_813; \
  65834.   int16x8_t __s1_813 = __p1_813; \
  65835.   int16x8_t __s2_813 = __p2_813; \
  65836.   __ret_813 = vqrdmlshq_s16(__s0_813, __s1_813, splatq_laneq_s16(__s2_813, __p3_813)); \
  65837.   __ret_813; \
  65838. })
  65839. #else
  65840. #define vqrdmlshq_laneq_s16(__p0_814, __p1_814, __p2_814, __p3_814) __extension__ ({ \
  65841.   int16x8_t __ret_814; \
  65842.   int16x8_t __s0_814 = __p0_814; \
  65843.   int16x8_t __s1_814 = __p1_814; \
  65844.   int16x8_t __s2_814 = __p2_814; \
  65845.   int16x8_t __rev0_814;  __rev0_814 = __builtin_shufflevector(__s0_814, __s0_814, 7, 6, 5, 4, 3, 2, 1, 0); \
  65846.   int16x8_t __rev1_814;  __rev1_814 = __builtin_shufflevector(__s1_814, __s1_814, 7, 6, 5, 4, 3, 2, 1, 0); \
  65847.   int16x8_t __rev2_814;  __rev2_814 = __builtin_shufflevector(__s2_814, __s2_814, 7, 6, 5, 4, 3, 2, 1, 0); \
  65848.   __ret_814 = __noswap_vqrdmlshq_s16(__rev0_814, __rev1_814, __noswap_splatq_laneq_s16(__rev2_814, __p3_814)); \
  65849.   __ret_814 = __builtin_shufflevector(__ret_814, __ret_814, 7, 6, 5, 4, 3, 2, 1, 0); \
  65850.   __ret_814; \
  65851. })
  65852. #endif
  65853.  
  65854. #ifdef __LITTLE_ENDIAN__
  65855. #define vqrdmlsh_laneq_s32(__p0_815, __p1_815, __p2_815, __p3_815) __extension__ ({ \
  65856.   int32x2_t __ret_815; \
  65857.   int32x2_t __s0_815 = __p0_815; \
  65858.   int32x2_t __s1_815 = __p1_815; \
  65859.   int32x4_t __s2_815 = __p2_815; \
  65860.   __ret_815 = vqrdmlsh_s32(__s0_815, __s1_815, splat_laneq_s32(__s2_815, __p3_815)); \
  65861.   __ret_815; \
  65862. })
  65863. #else
  65864. #define vqrdmlsh_laneq_s32(__p0_816, __p1_816, __p2_816, __p3_816) __extension__ ({ \
  65865.   int32x2_t __ret_816; \
  65866.   int32x2_t __s0_816 = __p0_816; \
  65867.   int32x2_t __s1_816 = __p1_816; \
  65868.   int32x4_t __s2_816 = __p2_816; \
  65869.   int32x2_t __rev0_816;  __rev0_816 = __builtin_shufflevector(__s0_816, __s0_816, 1, 0); \
  65870.   int32x2_t __rev1_816;  __rev1_816 = __builtin_shufflevector(__s1_816, __s1_816, 1, 0); \
  65871.   int32x4_t __rev2_816;  __rev2_816 = __builtin_shufflevector(__s2_816, __s2_816, 3, 2, 1, 0); \
  65872.   __ret_816 = __noswap_vqrdmlsh_s32(__rev0_816, __rev1_816, __noswap_splat_laneq_s32(__rev2_816, __p3_816)); \
  65873.   __ret_816 = __builtin_shufflevector(__ret_816, __ret_816, 1, 0); \
  65874.   __ret_816; \
  65875. })
  65876. #endif
  65877.  
  65878. #ifdef __LITTLE_ENDIAN__
  65879. #define vqrdmlsh_laneq_s16(__p0_817, __p1_817, __p2_817, __p3_817) __extension__ ({ \
  65880.   int16x4_t __ret_817; \
  65881.   int16x4_t __s0_817 = __p0_817; \
  65882.   int16x4_t __s1_817 = __p1_817; \
  65883.   int16x8_t __s2_817 = __p2_817; \
  65884.   __ret_817 = vqrdmlsh_s16(__s0_817, __s1_817, splat_laneq_s16(__s2_817, __p3_817)); \
  65885.   __ret_817; \
  65886. })
  65887. #else
  65888. #define vqrdmlsh_laneq_s16(__p0_818, __p1_818, __p2_818, __p3_818) __extension__ ({ \
  65889.   int16x4_t __ret_818; \
  65890.   int16x4_t __s0_818 = __p0_818; \
  65891.   int16x4_t __s1_818 = __p1_818; \
  65892.   int16x8_t __s2_818 = __p2_818; \
  65893.   int16x4_t __rev0_818;  __rev0_818 = __builtin_shufflevector(__s0_818, __s0_818, 3, 2, 1, 0); \
  65894.   int16x4_t __rev1_818;  __rev1_818 = __builtin_shufflevector(__s1_818, __s1_818, 3, 2, 1, 0); \
  65895.   int16x8_t __rev2_818;  __rev2_818 = __builtin_shufflevector(__s2_818, __s2_818, 7, 6, 5, 4, 3, 2, 1, 0); \
  65896.   __ret_818 = __noswap_vqrdmlsh_s16(__rev0_818, __rev1_818, __noswap_splat_laneq_s16(__rev2_818, __p3_818)); \
  65897.   __ret_818 = __builtin_shufflevector(__ret_818, __ret_818, 3, 2, 1, 0); \
  65898.   __ret_818; \
  65899. })
  65900. #endif
  65901.  
  65902. #ifdef __LITTLE_ENDIAN__
  65903. __ai __attribute__((target("v8.3a"))) float64x2_t vcaddq_rot270_f64(float64x2_t __p0, float64x2_t __p1) {
  65904.   float64x2_t __ret;
  65905.   __ret = (float64x2_t) __builtin_neon_vcaddq_rot270_f64((int8x16_t)__p0, (int8x16_t)__p1, 42);
  65906.   return __ret;
  65907. }
  65908. #else
  65909. __ai __attribute__((target("v8.3a"))) float64x2_t vcaddq_rot270_f64(float64x2_t __p0, float64x2_t __p1) {
  65910.   float64x2_t __ret;
  65911.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65912.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65913.   __ret = (float64x2_t) __builtin_neon_vcaddq_rot270_f64((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  65914.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65915.   return __ret;
  65916. }
  65917. #endif
  65918.  
  65919. #ifdef __LITTLE_ENDIAN__
  65920. __ai __attribute__((target("v8.3a"))) float64x2_t vcaddq_rot90_f64(float64x2_t __p0, float64x2_t __p1) {
  65921.   float64x2_t __ret;
  65922.   __ret = (float64x2_t) __builtin_neon_vcaddq_rot90_f64((int8x16_t)__p0, (int8x16_t)__p1, 42);
  65923.   return __ret;
  65924. }
  65925. #else
  65926. __ai __attribute__((target("v8.3a"))) float64x2_t vcaddq_rot90_f64(float64x2_t __p0, float64x2_t __p1) {
  65927.   float64x2_t __ret;
  65928.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65929.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65930.   __ret = (float64x2_t) __builtin_neon_vcaddq_rot90_f64((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  65931.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65932.   return __ret;
  65933. }
  65934. #endif
  65935.  
  65936. #ifdef __LITTLE_ENDIAN__
  65937. __ai __attribute__((target("v8.3a"))) float64x2_t vcmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  65938.   float64x2_t __ret;
  65939.   __ret = (float64x2_t) __builtin_neon_vcmlaq_f64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  65940.   return __ret;
  65941. }
  65942. #else
  65943. __ai __attribute__((target("v8.3a"))) float64x2_t vcmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  65944.   float64x2_t __ret;
  65945.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  65946.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  65947.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  65948.   __ret = (float64x2_t) __builtin_neon_vcmlaq_f64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
  65949.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  65950.   return __ret;
  65951. }
  65952. __ai __attribute__((target("v8.3a"))) float64x2_t __noswap_vcmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  65953.   float64x2_t __ret;
  65954.   __ret = (float64x2_t) __builtin_neon_vcmlaq_f64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  65955.   return __ret;
  65956. }
  65957. #endif
  65958.  
  65959. __ai __attribute__((target("v8.3a"))) float64x1_t vcmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
  65960.   float64x1_t __ret;
  65961.   __ret = (float64x1_t) __builtin_neon_vcmla_f64((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
  65962.   return __ret;
  65963. }
  65964. #define vcmla_lane_f64(__p0_819, __p1_819, __p2_819, __p3_819) __extension__ ({ \
  65965.   float64x1_t __ret_819; \
  65966.   float64x1_t __s0_819 = __p0_819; \
  65967.   float64x1_t __s1_819 = __p1_819; \
  65968.   float64x1_t __s2_819 = __p2_819; \
  65969. float64x1_t __reint_819 = __s2_819; \
  65970. uint64x2_t __reint1_819 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_819, __p3_819), vgetq_lane_u64(*(uint64x2_t *) &__reint_819, __p3_819)}; \
  65971.   __ret_819 = vcmla_f64(__s0_819, __s1_819, *(float64x1_t *) &__reint1_819); \
  65972.   __ret_819; \
  65973. })
  65974. #ifdef __LITTLE_ENDIAN__
  65975. #define vcmlaq_lane_f64(__p0_820, __p1_820, __p2_820, __p3_820) __extension__ ({ \
  65976.   float64x2_t __ret_820; \
  65977.   float64x2_t __s0_820 = __p0_820; \
  65978.   float64x2_t __s1_820 = __p1_820; \
  65979.   float64x1_t __s2_820 = __p2_820; \
  65980. float64x1_t __reint_820 = __s2_820; \
  65981. uint64x2_t __reint1_820 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_820, __p3_820), vgetq_lane_u64(*(uint64x2_t *) &__reint_820, __p3_820)}; \
  65982.   __ret_820 = vcmlaq_f64(__s0_820, __s1_820, *(float64x2_t *) &__reint1_820); \
  65983.   __ret_820; \
  65984. })
  65985. #else
  65986. #define vcmlaq_lane_f64(__p0_821, __p1_821, __p2_821, __p3_821) __extension__ ({ \
  65987.   float64x2_t __ret_821; \
  65988.   float64x2_t __s0_821 = __p0_821; \
  65989.   float64x2_t __s1_821 = __p1_821; \
  65990.   float64x1_t __s2_821 = __p2_821; \
  65991.   float64x2_t __rev0_821;  __rev0_821 = __builtin_shufflevector(__s0_821, __s0_821, 1, 0); \
  65992.   float64x2_t __rev1_821;  __rev1_821 = __builtin_shufflevector(__s1_821, __s1_821, 1, 0); \
  65993. float64x1_t __reint_821 = __s2_821; \
  65994. uint64x2_t __reint1_821 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_821, __p3_821), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_821, __p3_821)}; \
  65995.   __ret_821 = __noswap_vcmlaq_f64(__rev0_821, __rev1_821, *(float64x2_t *) &__reint1_821); \
  65996.   __ret_821 = __builtin_shufflevector(__ret_821, __ret_821, 1, 0); \
  65997.   __ret_821; \
  65998. })
  65999. #endif
  66000.  
  66001. #ifdef __LITTLE_ENDIAN__
  66002. #define vcmla_laneq_f64(__p0_822, __p1_822, __p2_822, __p3_822) __extension__ ({ \
  66003.   float64x1_t __ret_822; \
  66004.   float64x1_t __s0_822 = __p0_822; \
  66005.   float64x1_t __s1_822 = __p1_822; \
  66006.   float64x2_t __s2_822 = __p2_822; \
  66007. float64x2_t __reint_822 = __s2_822; \
  66008. uint64x2_t __reint1_822 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_822, __p3_822), vgetq_lane_u64(*(uint64x2_t *) &__reint_822, __p3_822)}; \
  66009.   __ret_822 = vcmla_f64(__s0_822, __s1_822, *(float64x1_t *) &__reint1_822); \
  66010.   __ret_822; \
  66011. })
  66012. #else
  66013. #define vcmla_laneq_f64(__p0_823, __p1_823, __p2_823, __p3_823) __extension__ ({ \
  66014.   float64x1_t __ret_823; \
  66015.   float64x1_t __s0_823 = __p0_823; \
  66016.   float64x1_t __s1_823 = __p1_823; \
  66017.   float64x2_t __s2_823 = __p2_823; \
  66018.   float64x2_t __rev2_823;  __rev2_823 = __builtin_shufflevector(__s2_823, __s2_823, 1, 0); \
  66019. float64x2_t __reint_823 = __rev2_823; \
  66020. uint64x2_t __reint1_823 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_823, __p3_823), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_823, __p3_823)}; \
  66021.   __ret_823 = vcmla_f64(__s0_823, __s1_823, *(float64x1_t *) &__reint1_823); \
  66022.   __ret_823; \
  66023. })
  66024. #endif
  66025.  
  66026. #ifdef __LITTLE_ENDIAN__
  66027. #define vcmlaq_laneq_f64(__p0_824, __p1_824, __p2_824, __p3_824) __extension__ ({ \
  66028.   float64x2_t __ret_824; \
  66029.   float64x2_t __s0_824 = __p0_824; \
  66030.   float64x2_t __s1_824 = __p1_824; \
  66031.   float64x2_t __s2_824 = __p2_824; \
  66032. float64x2_t __reint_824 = __s2_824; \
  66033. uint64x2_t __reint1_824 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_824, __p3_824), vgetq_lane_u64(*(uint64x2_t *) &__reint_824, __p3_824)}; \
  66034.   __ret_824 = vcmlaq_f64(__s0_824, __s1_824, *(float64x2_t *) &__reint1_824); \
  66035.   __ret_824; \
  66036. })
  66037. #else
  66038. #define vcmlaq_laneq_f64(__p0_825, __p1_825, __p2_825, __p3_825) __extension__ ({ \
  66039.   float64x2_t __ret_825; \
  66040.   float64x2_t __s0_825 = __p0_825; \
  66041.   float64x2_t __s1_825 = __p1_825; \
  66042.   float64x2_t __s2_825 = __p2_825; \
  66043.   float64x2_t __rev0_825;  __rev0_825 = __builtin_shufflevector(__s0_825, __s0_825, 1, 0); \
  66044.   float64x2_t __rev1_825;  __rev1_825 = __builtin_shufflevector(__s1_825, __s1_825, 1, 0); \
  66045.   float64x2_t __rev2_825;  __rev2_825 = __builtin_shufflevector(__s2_825, __s2_825, 1, 0); \
  66046. float64x2_t __reint_825 = __rev2_825; \
  66047. uint64x2_t __reint1_825 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_825, __p3_825), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_825, __p3_825)}; \
  66048.   __ret_825 = __noswap_vcmlaq_f64(__rev0_825, __rev1_825, *(float64x2_t *) &__reint1_825); \
  66049.   __ret_825 = __builtin_shufflevector(__ret_825, __ret_825, 1, 0); \
  66050.   __ret_825; \
  66051. })
  66052. #endif
  66053.  
  66054. #ifdef __LITTLE_ENDIAN__
  66055. __ai __attribute__((target("v8.3a"))) float64x2_t vcmlaq_rot180_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  66056.   float64x2_t __ret;
  66057.   __ret = (float64x2_t) __builtin_neon_vcmlaq_rot180_f64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  66058.   return __ret;
  66059. }
  66060. #else
  66061. __ai __attribute__((target("v8.3a"))) float64x2_t vcmlaq_rot180_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  66062.   float64x2_t __ret;
  66063.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66064.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  66065.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  66066.   __ret = (float64x2_t) __builtin_neon_vcmlaq_rot180_f64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
  66067.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66068.   return __ret;
  66069. }
  66070. __ai __attribute__((target("v8.3a"))) float64x2_t __noswap_vcmlaq_rot180_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  66071.   float64x2_t __ret;
  66072.   __ret = (float64x2_t) __builtin_neon_vcmlaq_rot180_f64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  66073.   return __ret;
  66074. }
  66075. #endif
  66076.  
  66077. __ai __attribute__((target("v8.3a"))) float64x1_t vcmla_rot180_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
  66078.   float64x1_t __ret;
  66079.   __ret = (float64x1_t) __builtin_neon_vcmla_rot180_f64((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
  66080.   return __ret;
  66081. }
  66082. #define vcmla_rot180_lane_f64(__p0_826, __p1_826, __p2_826, __p3_826) __extension__ ({ \
  66083.   float64x1_t __ret_826; \
  66084.   float64x1_t __s0_826 = __p0_826; \
  66085.   float64x1_t __s1_826 = __p1_826; \
  66086.   float64x1_t __s2_826 = __p2_826; \
  66087. float64x1_t __reint_826 = __s2_826; \
  66088. uint64x2_t __reint1_826 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_826, __p3_826), vgetq_lane_u64(*(uint64x2_t *) &__reint_826, __p3_826)}; \
  66089.   __ret_826 = vcmla_rot180_f64(__s0_826, __s1_826, *(float64x1_t *) &__reint1_826); \
  66090.   __ret_826; \
  66091. })
  66092. #ifdef __LITTLE_ENDIAN__
  66093. #define vcmlaq_rot180_lane_f64(__p0_827, __p1_827, __p2_827, __p3_827) __extension__ ({ \
  66094.   float64x2_t __ret_827; \
  66095.   float64x2_t __s0_827 = __p0_827; \
  66096.   float64x2_t __s1_827 = __p1_827; \
  66097.   float64x1_t __s2_827 = __p2_827; \
  66098. float64x1_t __reint_827 = __s2_827; \
  66099. uint64x2_t __reint1_827 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_827, __p3_827), vgetq_lane_u64(*(uint64x2_t *) &__reint_827, __p3_827)}; \
  66100.   __ret_827 = vcmlaq_rot180_f64(__s0_827, __s1_827, *(float64x2_t *) &__reint1_827); \
  66101.   __ret_827; \
  66102. })
  66103. #else
  66104. #define vcmlaq_rot180_lane_f64(__p0_828, __p1_828, __p2_828, __p3_828) __extension__ ({ \
  66105.   float64x2_t __ret_828; \
  66106.   float64x2_t __s0_828 = __p0_828; \
  66107.   float64x2_t __s1_828 = __p1_828; \
  66108.   float64x1_t __s2_828 = __p2_828; \
  66109.   float64x2_t __rev0_828;  __rev0_828 = __builtin_shufflevector(__s0_828, __s0_828, 1, 0); \
  66110.   float64x2_t __rev1_828;  __rev1_828 = __builtin_shufflevector(__s1_828, __s1_828, 1, 0); \
  66111. float64x1_t __reint_828 = __s2_828; \
  66112. uint64x2_t __reint1_828 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_828, __p3_828), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_828, __p3_828)}; \
  66113.   __ret_828 = __noswap_vcmlaq_rot180_f64(__rev0_828, __rev1_828, *(float64x2_t *) &__reint1_828); \
  66114.   __ret_828 = __builtin_shufflevector(__ret_828, __ret_828, 1, 0); \
  66115.   __ret_828; \
  66116. })
  66117. #endif
  66118.  
  66119. #ifdef __LITTLE_ENDIAN__
  66120. #define vcmla_rot180_laneq_f64(__p0_829, __p1_829, __p2_829, __p3_829) __extension__ ({ \
  66121.   float64x1_t __ret_829; \
  66122.   float64x1_t __s0_829 = __p0_829; \
  66123.   float64x1_t __s1_829 = __p1_829; \
  66124.   float64x2_t __s2_829 = __p2_829; \
  66125. float64x2_t __reint_829 = __s2_829; \
  66126. uint64x2_t __reint1_829 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_829, __p3_829), vgetq_lane_u64(*(uint64x2_t *) &__reint_829, __p3_829)}; \
  66127.   __ret_829 = vcmla_rot180_f64(__s0_829, __s1_829, *(float64x1_t *) &__reint1_829); \
  66128.   __ret_829; \
  66129. })
  66130. #else
  66131. #define vcmla_rot180_laneq_f64(__p0_830, __p1_830, __p2_830, __p3_830) __extension__ ({ \
  66132.   float64x1_t __ret_830; \
  66133.   float64x1_t __s0_830 = __p0_830; \
  66134.   float64x1_t __s1_830 = __p1_830; \
  66135.   float64x2_t __s2_830 = __p2_830; \
  66136.   float64x2_t __rev2_830;  __rev2_830 = __builtin_shufflevector(__s2_830, __s2_830, 1, 0); \
  66137. float64x2_t __reint_830 = __rev2_830; \
  66138. uint64x2_t __reint1_830 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_830, __p3_830), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_830, __p3_830)}; \
  66139.   __ret_830 = vcmla_rot180_f64(__s0_830, __s1_830, *(float64x1_t *) &__reint1_830); \
  66140.   __ret_830; \
  66141. })
  66142. #endif
  66143.  
  66144. #ifdef __LITTLE_ENDIAN__
  66145. #define vcmlaq_rot180_laneq_f64(__p0_831, __p1_831, __p2_831, __p3_831) __extension__ ({ \
  66146.   float64x2_t __ret_831; \
  66147.   float64x2_t __s0_831 = __p0_831; \
  66148.   float64x2_t __s1_831 = __p1_831; \
  66149.   float64x2_t __s2_831 = __p2_831; \
  66150. float64x2_t __reint_831 = __s2_831; \
  66151. uint64x2_t __reint1_831 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_831, __p3_831), vgetq_lane_u64(*(uint64x2_t *) &__reint_831, __p3_831)}; \
  66152.   __ret_831 = vcmlaq_rot180_f64(__s0_831, __s1_831, *(float64x2_t *) &__reint1_831); \
  66153.   __ret_831; \
  66154. })
  66155. #else
  66156. #define vcmlaq_rot180_laneq_f64(__p0_832, __p1_832, __p2_832, __p3_832) __extension__ ({ \
  66157.   float64x2_t __ret_832; \
  66158.   float64x2_t __s0_832 = __p0_832; \
  66159.   float64x2_t __s1_832 = __p1_832; \
  66160.   float64x2_t __s2_832 = __p2_832; \
  66161.   float64x2_t __rev0_832;  __rev0_832 = __builtin_shufflevector(__s0_832, __s0_832, 1, 0); \
  66162.   float64x2_t __rev1_832;  __rev1_832 = __builtin_shufflevector(__s1_832, __s1_832, 1, 0); \
  66163.   float64x2_t __rev2_832;  __rev2_832 = __builtin_shufflevector(__s2_832, __s2_832, 1, 0); \
  66164. float64x2_t __reint_832 = __rev2_832; \
  66165. uint64x2_t __reint1_832 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_832, __p3_832), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_832, __p3_832)}; \
  66166.   __ret_832 = __noswap_vcmlaq_rot180_f64(__rev0_832, __rev1_832, *(float64x2_t *) &__reint1_832); \
  66167.   __ret_832 = __builtin_shufflevector(__ret_832, __ret_832, 1, 0); \
  66168.   __ret_832; \
  66169. })
  66170. #endif
  66171.  
  66172. #ifdef __LITTLE_ENDIAN__
  66173. __ai __attribute__((target("v8.3a"))) float64x2_t vcmlaq_rot270_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  66174.   float64x2_t __ret;
  66175.   __ret = (float64x2_t) __builtin_neon_vcmlaq_rot270_f64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  66176.   return __ret;
  66177. }
  66178. #else
  66179. __ai __attribute__((target("v8.3a"))) float64x2_t vcmlaq_rot270_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  66180.   float64x2_t __ret;
  66181.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66182.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  66183.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  66184.   __ret = (float64x2_t) __builtin_neon_vcmlaq_rot270_f64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
  66185.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66186.   return __ret;
  66187. }
  66188. __ai __attribute__((target("v8.3a"))) float64x2_t __noswap_vcmlaq_rot270_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  66189.   float64x2_t __ret;
  66190.   __ret = (float64x2_t) __builtin_neon_vcmlaq_rot270_f64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  66191.   return __ret;
  66192. }
  66193. #endif
  66194.  
  66195. __ai __attribute__((target("v8.3a"))) float64x1_t vcmla_rot270_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
  66196.   float64x1_t __ret;
  66197.   __ret = (float64x1_t) __builtin_neon_vcmla_rot270_f64((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
  66198.   return __ret;
  66199. }
  66200. #define vcmla_rot270_lane_f64(__p0_833, __p1_833, __p2_833, __p3_833) __extension__ ({ \
  66201.   float64x1_t __ret_833; \
  66202.   float64x1_t __s0_833 = __p0_833; \
  66203.   float64x1_t __s1_833 = __p1_833; \
  66204.   float64x1_t __s2_833 = __p2_833; \
  66205. float64x1_t __reint_833 = __s2_833; \
  66206. uint64x2_t __reint1_833 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_833, __p3_833), vgetq_lane_u64(*(uint64x2_t *) &__reint_833, __p3_833)}; \
  66207.   __ret_833 = vcmla_rot270_f64(__s0_833, __s1_833, *(float64x1_t *) &__reint1_833); \
  66208.   __ret_833; \
  66209. })
  66210. #ifdef __LITTLE_ENDIAN__
  66211. #define vcmlaq_rot270_lane_f64(__p0_834, __p1_834, __p2_834, __p3_834) __extension__ ({ \
  66212.   float64x2_t __ret_834; \
  66213.   float64x2_t __s0_834 = __p0_834; \
  66214.   float64x2_t __s1_834 = __p1_834; \
  66215.   float64x1_t __s2_834 = __p2_834; \
  66216. float64x1_t __reint_834 = __s2_834; \
  66217. uint64x2_t __reint1_834 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_834, __p3_834), vgetq_lane_u64(*(uint64x2_t *) &__reint_834, __p3_834)}; \
  66218.   __ret_834 = vcmlaq_rot270_f64(__s0_834, __s1_834, *(float64x2_t *) &__reint1_834); \
  66219.   __ret_834; \
  66220. })
  66221. #else
  66222. #define vcmlaq_rot270_lane_f64(__p0_835, __p1_835, __p2_835, __p3_835) __extension__ ({ \
  66223.   float64x2_t __ret_835; \
  66224.   float64x2_t __s0_835 = __p0_835; \
  66225.   float64x2_t __s1_835 = __p1_835; \
  66226.   float64x1_t __s2_835 = __p2_835; \
  66227.   float64x2_t __rev0_835;  __rev0_835 = __builtin_shufflevector(__s0_835, __s0_835, 1, 0); \
  66228.   float64x2_t __rev1_835;  __rev1_835 = __builtin_shufflevector(__s1_835, __s1_835, 1, 0); \
  66229. float64x1_t __reint_835 = __s2_835; \
  66230. uint64x2_t __reint1_835 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_835, __p3_835), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_835, __p3_835)}; \
  66231.   __ret_835 = __noswap_vcmlaq_rot270_f64(__rev0_835, __rev1_835, *(float64x2_t *) &__reint1_835); \
  66232.   __ret_835 = __builtin_shufflevector(__ret_835, __ret_835, 1, 0); \
  66233.   __ret_835; \
  66234. })
  66235. #endif
  66236.  
  66237. #ifdef __LITTLE_ENDIAN__
  66238. #define vcmla_rot270_laneq_f64(__p0_836, __p1_836, __p2_836, __p3_836) __extension__ ({ \
  66239.   float64x1_t __ret_836; \
  66240.   float64x1_t __s0_836 = __p0_836; \
  66241.   float64x1_t __s1_836 = __p1_836; \
  66242.   float64x2_t __s2_836 = __p2_836; \
  66243. float64x2_t __reint_836 = __s2_836; \
  66244. uint64x2_t __reint1_836 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_836, __p3_836), vgetq_lane_u64(*(uint64x2_t *) &__reint_836, __p3_836)}; \
  66245.   __ret_836 = vcmla_rot270_f64(__s0_836, __s1_836, *(float64x1_t *) &__reint1_836); \
  66246.   __ret_836; \
  66247. })
  66248. #else
  66249. #define vcmla_rot270_laneq_f64(__p0_837, __p1_837, __p2_837, __p3_837) __extension__ ({ \
  66250.   float64x1_t __ret_837; \
  66251.   float64x1_t __s0_837 = __p0_837; \
  66252.   float64x1_t __s1_837 = __p1_837; \
  66253.   float64x2_t __s2_837 = __p2_837; \
  66254.   float64x2_t __rev2_837;  __rev2_837 = __builtin_shufflevector(__s2_837, __s2_837, 1, 0); \
  66255. float64x2_t __reint_837 = __rev2_837; \
  66256. uint64x2_t __reint1_837 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_837, __p3_837), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_837, __p3_837)}; \
  66257.   __ret_837 = vcmla_rot270_f64(__s0_837, __s1_837, *(float64x1_t *) &__reint1_837); \
  66258.   __ret_837; \
  66259. })
  66260. #endif
  66261.  
  66262. #ifdef __LITTLE_ENDIAN__
  66263. #define vcmlaq_rot270_laneq_f64(__p0_838, __p1_838, __p2_838, __p3_838) __extension__ ({ \
  66264.   float64x2_t __ret_838; \
  66265.   float64x2_t __s0_838 = __p0_838; \
  66266.   float64x2_t __s1_838 = __p1_838; \
  66267.   float64x2_t __s2_838 = __p2_838; \
  66268. float64x2_t __reint_838 = __s2_838; \
  66269. uint64x2_t __reint1_838 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_838, __p3_838), vgetq_lane_u64(*(uint64x2_t *) &__reint_838, __p3_838)}; \
  66270.   __ret_838 = vcmlaq_rot270_f64(__s0_838, __s1_838, *(float64x2_t *) &__reint1_838); \
  66271.   __ret_838; \
  66272. })
  66273. #else
  66274. #define vcmlaq_rot270_laneq_f64(__p0_839, __p1_839, __p2_839, __p3_839) __extension__ ({ \
  66275.   float64x2_t __ret_839; \
  66276.   float64x2_t __s0_839 = __p0_839; \
  66277.   float64x2_t __s1_839 = __p1_839; \
  66278.   float64x2_t __s2_839 = __p2_839; \
  66279.   float64x2_t __rev0_839;  __rev0_839 = __builtin_shufflevector(__s0_839, __s0_839, 1, 0); \
  66280.   float64x2_t __rev1_839;  __rev1_839 = __builtin_shufflevector(__s1_839, __s1_839, 1, 0); \
  66281.   float64x2_t __rev2_839;  __rev2_839 = __builtin_shufflevector(__s2_839, __s2_839, 1, 0); \
  66282. float64x2_t __reint_839 = __rev2_839; \
  66283. uint64x2_t __reint1_839 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_839, __p3_839), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_839, __p3_839)}; \
  66284.   __ret_839 = __noswap_vcmlaq_rot270_f64(__rev0_839, __rev1_839, *(float64x2_t *) &__reint1_839); \
  66285.   __ret_839 = __builtin_shufflevector(__ret_839, __ret_839, 1, 0); \
  66286.   __ret_839; \
  66287. })
  66288. #endif
  66289.  
  66290. #ifdef __LITTLE_ENDIAN__
  66291. __ai __attribute__((target("v8.3a"))) float64x2_t vcmlaq_rot90_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  66292.   float64x2_t __ret;
  66293.   __ret = (float64x2_t) __builtin_neon_vcmlaq_rot90_f64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  66294.   return __ret;
  66295. }
  66296. #else
  66297. __ai __attribute__((target("v8.3a"))) float64x2_t vcmlaq_rot90_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  66298.   float64x2_t __ret;
  66299.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66300.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  66301.   float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  66302.   __ret = (float64x2_t) __builtin_neon_vcmlaq_rot90_f64((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
  66303.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66304.   return __ret;
  66305. }
  66306. __ai __attribute__((target("v8.3a"))) float64x2_t __noswap_vcmlaq_rot90_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
  66307.   float64x2_t __ret;
  66308.   __ret = (float64x2_t) __builtin_neon_vcmlaq_rot90_f64((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
  66309.   return __ret;
  66310. }
  66311. #endif
  66312.  
  66313. __ai __attribute__((target("v8.3a"))) float64x1_t vcmla_rot90_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
  66314.   float64x1_t __ret;
  66315.   __ret = (float64x1_t) __builtin_neon_vcmla_rot90_f64((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
  66316.   return __ret;
  66317. }
  66318. #define vcmla_rot90_lane_f64(__p0_840, __p1_840, __p2_840, __p3_840) __extension__ ({ \
  66319.   float64x1_t __ret_840; \
  66320.   float64x1_t __s0_840 = __p0_840; \
  66321.   float64x1_t __s1_840 = __p1_840; \
  66322.   float64x1_t __s2_840 = __p2_840; \
  66323. float64x1_t __reint_840 = __s2_840; \
  66324. uint64x2_t __reint1_840 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_840, __p3_840), vgetq_lane_u64(*(uint64x2_t *) &__reint_840, __p3_840)}; \
  66325.   __ret_840 = vcmla_rot90_f64(__s0_840, __s1_840, *(float64x1_t *) &__reint1_840); \
  66326.   __ret_840; \
  66327. })
  66328. #ifdef __LITTLE_ENDIAN__
  66329. #define vcmlaq_rot90_lane_f64(__p0_841, __p1_841, __p2_841, __p3_841) __extension__ ({ \
  66330.   float64x2_t __ret_841; \
  66331.   float64x2_t __s0_841 = __p0_841; \
  66332.   float64x2_t __s1_841 = __p1_841; \
  66333.   float64x1_t __s2_841 = __p2_841; \
  66334. float64x1_t __reint_841 = __s2_841; \
  66335. uint64x2_t __reint1_841 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_841, __p3_841), vgetq_lane_u64(*(uint64x2_t *) &__reint_841, __p3_841)}; \
  66336.   __ret_841 = vcmlaq_rot90_f64(__s0_841, __s1_841, *(float64x2_t *) &__reint1_841); \
  66337.   __ret_841; \
  66338. })
  66339. #else
  66340. #define vcmlaq_rot90_lane_f64(__p0_842, __p1_842, __p2_842, __p3_842) __extension__ ({ \
  66341.   float64x2_t __ret_842; \
  66342.   float64x2_t __s0_842 = __p0_842; \
  66343.   float64x2_t __s1_842 = __p1_842; \
  66344.   float64x1_t __s2_842 = __p2_842; \
  66345.   float64x2_t __rev0_842;  __rev0_842 = __builtin_shufflevector(__s0_842, __s0_842, 1, 0); \
  66346.   float64x2_t __rev1_842;  __rev1_842 = __builtin_shufflevector(__s1_842, __s1_842, 1, 0); \
  66347. float64x1_t __reint_842 = __s2_842; \
  66348. uint64x2_t __reint1_842 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_842, __p3_842), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_842, __p3_842)}; \
  66349.   __ret_842 = __noswap_vcmlaq_rot90_f64(__rev0_842, __rev1_842, *(float64x2_t *) &__reint1_842); \
  66350.   __ret_842 = __builtin_shufflevector(__ret_842, __ret_842, 1, 0); \
  66351.   __ret_842; \
  66352. })
  66353. #endif
  66354.  
  66355. #ifdef __LITTLE_ENDIAN__
  66356. #define vcmla_rot90_laneq_f64(__p0_843, __p1_843, __p2_843, __p3_843) __extension__ ({ \
  66357.   float64x1_t __ret_843; \
  66358.   float64x1_t __s0_843 = __p0_843; \
  66359.   float64x1_t __s1_843 = __p1_843; \
  66360.   float64x2_t __s2_843 = __p2_843; \
  66361. float64x2_t __reint_843 = __s2_843; \
  66362. uint64x2_t __reint1_843 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_843, __p3_843), vgetq_lane_u64(*(uint64x2_t *) &__reint_843, __p3_843)}; \
  66363.   __ret_843 = vcmla_rot90_f64(__s0_843, __s1_843, *(float64x1_t *) &__reint1_843); \
  66364.   __ret_843; \
  66365. })
  66366. #else
  66367. #define vcmla_rot90_laneq_f64(__p0_844, __p1_844, __p2_844, __p3_844) __extension__ ({ \
  66368.   float64x1_t __ret_844; \
  66369.   float64x1_t __s0_844 = __p0_844; \
  66370.   float64x1_t __s1_844 = __p1_844; \
  66371.   float64x2_t __s2_844 = __p2_844; \
  66372.   float64x2_t __rev2_844;  __rev2_844 = __builtin_shufflevector(__s2_844, __s2_844, 1, 0); \
  66373. float64x2_t __reint_844 = __rev2_844; \
  66374. uint64x2_t __reint1_844 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_844, __p3_844), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_844, __p3_844)}; \
  66375.   __ret_844 = vcmla_rot90_f64(__s0_844, __s1_844, *(float64x1_t *) &__reint1_844); \
  66376.   __ret_844; \
  66377. })
  66378. #endif
  66379.  
  66380. #ifdef __LITTLE_ENDIAN__
  66381. #define vcmlaq_rot90_laneq_f64(__p0_845, __p1_845, __p2_845, __p3_845) __extension__ ({ \
  66382.   float64x2_t __ret_845; \
  66383.   float64x2_t __s0_845 = __p0_845; \
  66384.   float64x2_t __s1_845 = __p1_845; \
  66385.   float64x2_t __s2_845 = __p2_845; \
  66386. float64x2_t __reint_845 = __s2_845; \
  66387. uint64x2_t __reint1_845 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_845, __p3_845), vgetq_lane_u64(*(uint64x2_t *) &__reint_845, __p3_845)}; \
  66388.   __ret_845 = vcmlaq_rot90_f64(__s0_845, __s1_845, *(float64x2_t *) &__reint1_845); \
  66389.   __ret_845; \
  66390. })
  66391. #else
  66392. #define vcmlaq_rot90_laneq_f64(__p0_846, __p1_846, __p2_846, __p3_846) __extension__ ({ \
  66393.   float64x2_t __ret_846; \
  66394.   float64x2_t __s0_846 = __p0_846; \
  66395.   float64x2_t __s1_846 = __p1_846; \
  66396.   float64x2_t __s2_846 = __p2_846; \
  66397.   float64x2_t __rev0_846;  __rev0_846 = __builtin_shufflevector(__s0_846, __s0_846, 1, 0); \
  66398.   float64x2_t __rev1_846;  __rev1_846 = __builtin_shufflevector(__s1_846, __s1_846, 1, 0); \
  66399.   float64x2_t __rev2_846;  __rev2_846 = __builtin_shufflevector(__s2_846, __s2_846, 1, 0); \
  66400. float64x2_t __reint_846 = __rev2_846; \
  66401. uint64x2_t __reint1_846 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_846, __p3_846), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_846, __p3_846)}; \
  66402.   __ret_846 = __noswap_vcmlaq_rot90_f64(__rev0_846, __rev1_846, *(float64x2_t *) &__reint1_846); \
  66403.   __ret_846 = __builtin_shufflevector(__ret_846, __ret_846, 1, 0); \
  66404.   __ret_846; \
  66405. })
  66406. #endif
  66407.  
  66408. #ifdef __LITTLE_ENDIAN__
  66409. __ai __attribute__((target("v8.5a"))) float32x4_t vrnd32xq_f32(float32x4_t __p0) {
  66410.   float32x4_t __ret;
  66411.   __ret = (float32x4_t) __builtin_neon_vrnd32xq_f32((int8x16_t)__p0, 41);
  66412.   return __ret;
  66413. }
  66414. #else
  66415. __ai __attribute__((target("v8.5a"))) float32x4_t vrnd32xq_f32(float32x4_t __p0) {
  66416.   float32x4_t __ret;
  66417.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  66418.   __ret = (float32x4_t) __builtin_neon_vrnd32xq_f32((int8x16_t)__rev0, 41);
  66419.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  66420.   return __ret;
  66421. }
  66422. #endif
  66423.  
  66424. #ifdef __LITTLE_ENDIAN__
  66425. __ai __attribute__((target("v8.5a"))) float32x2_t vrnd32x_f32(float32x2_t __p0) {
  66426.   float32x2_t __ret;
  66427.   __ret = (float32x2_t) __builtin_neon_vrnd32x_f32((int8x8_t)__p0, 9);
  66428.   return __ret;
  66429. }
  66430. #else
  66431. __ai __attribute__((target("v8.5a"))) float32x2_t vrnd32x_f32(float32x2_t __p0) {
  66432.   float32x2_t __ret;
  66433.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66434.   __ret = (float32x2_t) __builtin_neon_vrnd32x_f32((int8x8_t)__rev0, 9);
  66435.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66436.   return __ret;
  66437. }
  66438. #endif
  66439.  
  66440. #ifdef __LITTLE_ENDIAN__
  66441. __ai __attribute__((target("v8.5a"))) float32x4_t vrnd32zq_f32(float32x4_t __p0) {
  66442.   float32x4_t __ret;
  66443.   __ret = (float32x4_t) __builtin_neon_vrnd32zq_f32((int8x16_t)__p0, 41);
  66444.   return __ret;
  66445. }
  66446. #else
  66447. __ai __attribute__((target("v8.5a"))) float32x4_t vrnd32zq_f32(float32x4_t __p0) {
  66448.   float32x4_t __ret;
  66449.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  66450.   __ret = (float32x4_t) __builtin_neon_vrnd32zq_f32((int8x16_t)__rev0, 41);
  66451.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  66452.   return __ret;
  66453. }
  66454. #endif
  66455.  
  66456. #ifdef __LITTLE_ENDIAN__
  66457. __ai __attribute__((target("v8.5a"))) float32x2_t vrnd32z_f32(float32x2_t __p0) {
  66458.   float32x2_t __ret;
  66459.   __ret = (float32x2_t) __builtin_neon_vrnd32z_f32((int8x8_t)__p0, 9);
  66460.   return __ret;
  66461. }
  66462. #else
  66463. __ai __attribute__((target("v8.5a"))) float32x2_t vrnd32z_f32(float32x2_t __p0) {
  66464.   float32x2_t __ret;
  66465.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66466.   __ret = (float32x2_t) __builtin_neon_vrnd32z_f32((int8x8_t)__rev0, 9);
  66467.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66468.   return __ret;
  66469. }
  66470. #endif
  66471.  
  66472. #ifdef __LITTLE_ENDIAN__
  66473. __ai __attribute__((target("v8.5a"))) float32x4_t vrnd64xq_f32(float32x4_t __p0) {
  66474.   float32x4_t __ret;
  66475.   __ret = (float32x4_t) __builtin_neon_vrnd64xq_f32((int8x16_t)__p0, 41);
  66476.   return __ret;
  66477. }
  66478. #else
  66479. __ai __attribute__((target("v8.5a"))) float32x4_t vrnd64xq_f32(float32x4_t __p0) {
  66480.   float32x4_t __ret;
  66481.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  66482.   __ret = (float32x4_t) __builtin_neon_vrnd64xq_f32((int8x16_t)__rev0, 41);
  66483.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  66484.   return __ret;
  66485. }
  66486. #endif
  66487.  
  66488. #ifdef __LITTLE_ENDIAN__
  66489. __ai __attribute__((target("v8.5a"))) float32x2_t vrnd64x_f32(float32x2_t __p0) {
  66490.   float32x2_t __ret;
  66491.   __ret = (float32x2_t) __builtin_neon_vrnd64x_f32((int8x8_t)__p0, 9);
  66492.   return __ret;
  66493. }
  66494. #else
  66495. __ai __attribute__((target("v8.5a"))) float32x2_t vrnd64x_f32(float32x2_t __p0) {
  66496.   float32x2_t __ret;
  66497.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66498.   __ret = (float32x2_t) __builtin_neon_vrnd64x_f32((int8x8_t)__rev0, 9);
  66499.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66500.   return __ret;
  66501. }
  66502. #endif
  66503.  
  66504. #ifdef __LITTLE_ENDIAN__
  66505. __ai __attribute__((target("v8.5a"))) float32x4_t vrnd64zq_f32(float32x4_t __p0) {
  66506.   float32x4_t __ret;
  66507.   __ret = (float32x4_t) __builtin_neon_vrnd64zq_f32((int8x16_t)__p0, 41);
  66508.   return __ret;
  66509. }
  66510. #else
  66511. __ai __attribute__((target("v8.5a"))) float32x4_t vrnd64zq_f32(float32x4_t __p0) {
  66512.   float32x4_t __ret;
  66513.   float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  66514.   __ret = (float32x4_t) __builtin_neon_vrnd64zq_f32((int8x16_t)__rev0, 41);
  66515.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  66516.   return __ret;
  66517. }
  66518. #endif
  66519.  
  66520. #ifdef __LITTLE_ENDIAN__
  66521. __ai __attribute__((target("v8.5a"))) float32x2_t vrnd64z_f32(float32x2_t __p0) {
  66522.   float32x2_t __ret;
  66523.   __ret = (float32x2_t) __builtin_neon_vrnd64z_f32((int8x8_t)__p0, 9);
  66524.   return __ret;
  66525. }
  66526. #else
  66527. __ai __attribute__((target("v8.5a"))) float32x2_t vrnd64z_f32(float32x2_t __p0) {
  66528.   float32x2_t __ret;
  66529.   float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66530.   __ret = (float32x2_t) __builtin_neon_vrnd64z_f32((int8x8_t)__rev0, 9);
  66531.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66532.   return __ret;
  66533. }
  66534. #endif
  66535.  
  66536. #endif
  66537. #if defined(__aarch64__) && defined(__ARM_FEATURE_DIRECTED_ROUNDING)
  66538. #ifdef __LITTLE_ENDIAN__
  66539. __ai float64x2_t vrndq_f64(float64x2_t __p0) {
  66540.   float64x2_t __ret;
  66541.   __ret = (float64x2_t) __builtin_neon_vrndq_v((int8x16_t)__p0, 42);
  66542.   return __ret;
  66543. }
  66544. #else
  66545. __ai float64x2_t vrndq_f64(float64x2_t __p0) {
  66546.   float64x2_t __ret;
  66547.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66548.   __ret = (float64x2_t) __builtin_neon_vrndq_v((int8x16_t)__rev0, 42);
  66549.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66550.   return __ret;
  66551. }
  66552. #endif
  66553.  
  66554. __ai float64x1_t vrnd_f64(float64x1_t __p0) {
  66555.   float64x1_t __ret;
  66556.   __ret = (float64x1_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 10);
  66557.   return __ret;
  66558. }
  66559. #ifdef __LITTLE_ENDIAN__
  66560. __ai float64x2_t vrndaq_f64(float64x2_t __p0) {
  66561.   float64x2_t __ret;
  66562.   __ret = (float64x2_t) __builtin_neon_vrndaq_v((int8x16_t)__p0, 42);
  66563.   return __ret;
  66564. }
  66565. #else
  66566. __ai float64x2_t vrndaq_f64(float64x2_t __p0) {
  66567.   float64x2_t __ret;
  66568.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66569.   __ret = (float64x2_t) __builtin_neon_vrndaq_v((int8x16_t)__rev0, 42);
  66570.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66571.   return __ret;
  66572. }
  66573. #endif
  66574.  
  66575. __ai float64x1_t vrnda_f64(float64x1_t __p0) {
  66576.   float64x1_t __ret;
  66577.   __ret = (float64x1_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 10);
  66578.   return __ret;
  66579. }
  66580. #ifdef __LITTLE_ENDIAN__
  66581. __ai float64x2_t vrndiq_f64(float64x2_t __p0) {
  66582.   float64x2_t __ret;
  66583.   __ret = (float64x2_t) __builtin_neon_vrndiq_v((int8x16_t)__p0, 42);
  66584.   return __ret;
  66585. }
  66586. #else
  66587. __ai float64x2_t vrndiq_f64(float64x2_t __p0) {
  66588.   float64x2_t __ret;
  66589.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66590.   __ret = (float64x2_t) __builtin_neon_vrndiq_v((int8x16_t)__rev0, 42);
  66591.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66592.   return __ret;
  66593. }
  66594. #endif
  66595.  
  66596. __ai float64x1_t vrndi_f64(float64x1_t __p0) {
  66597.   float64x1_t __ret;
  66598.   __ret = (float64x1_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 10);
  66599.   return __ret;
  66600. }
  66601. #ifdef __LITTLE_ENDIAN__
  66602. __ai float64x2_t vrndmq_f64(float64x2_t __p0) {
  66603.   float64x2_t __ret;
  66604.   __ret = (float64x2_t) __builtin_neon_vrndmq_v((int8x16_t)__p0, 42);
  66605.   return __ret;
  66606. }
  66607. #else
  66608. __ai float64x2_t vrndmq_f64(float64x2_t __p0) {
  66609.   float64x2_t __ret;
  66610.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66611.   __ret = (float64x2_t) __builtin_neon_vrndmq_v((int8x16_t)__rev0, 42);
  66612.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66613.   return __ret;
  66614. }
  66615. #endif
  66616.  
  66617. __ai float64x1_t vrndm_f64(float64x1_t __p0) {
  66618.   float64x1_t __ret;
  66619.   __ret = (float64x1_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 10);
  66620.   return __ret;
  66621. }
  66622. #ifdef __LITTLE_ENDIAN__
  66623. __ai float64x2_t vrndnq_f64(float64x2_t __p0) {
  66624.   float64x2_t __ret;
  66625.   __ret = (float64x2_t) __builtin_neon_vrndnq_v((int8x16_t)__p0, 42);
  66626.   return __ret;
  66627. }
  66628. #else
  66629. __ai float64x2_t vrndnq_f64(float64x2_t __p0) {
  66630.   float64x2_t __ret;
  66631.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66632.   __ret = (float64x2_t) __builtin_neon_vrndnq_v((int8x16_t)__rev0, 42);
  66633.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66634.   return __ret;
  66635. }
  66636. #endif
  66637.  
  66638. __ai float64x1_t vrndn_f64(float64x1_t __p0) {
  66639.   float64x1_t __ret;
  66640.   __ret = (float64x1_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 10);
  66641.   return __ret;
  66642. }
  66643. #ifdef __LITTLE_ENDIAN__
  66644. __ai float64x2_t vrndpq_f64(float64x2_t __p0) {
  66645.   float64x2_t __ret;
  66646.   __ret = (float64x2_t) __builtin_neon_vrndpq_v((int8x16_t)__p0, 42);
  66647.   return __ret;
  66648. }
  66649. #else
  66650. __ai float64x2_t vrndpq_f64(float64x2_t __p0) {
  66651.   float64x2_t __ret;
  66652.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66653.   __ret = (float64x2_t) __builtin_neon_vrndpq_v((int8x16_t)__rev0, 42);
  66654.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66655.   return __ret;
  66656. }
  66657. #endif
  66658.  
  66659. __ai float64x1_t vrndp_f64(float64x1_t __p0) {
  66660.   float64x1_t __ret;
  66661.   __ret = (float64x1_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 10);
  66662.   return __ret;
  66663. }
  66664. #ifdef __LITTLE_ENDIAN__
  66665. __ai float64x2_t vrndxq_f64(float64x2_t __p0) {
  66666.   float64x2_t __ret;
  66667.   __ret = (float64x2_t) __builtin_neon_vrndxq_v((int8x16_t)__p0, 42);
  66668.   return __ret;
  66669. }
  66670. #else
  66671. __ai float64x2_t vrndxq_f64(float64x2_t __p0) {
  66672.   float64x2_t __ret;
  66673.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66674.   __ret = (float64x2_t) __builtin_neon_vrndxq_v((int8x16_t)__rev0, 42);
  66675.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66676.   return __ret;
  66677. }
  66678. #endif
  66679.  
  66680. __ai float64x1_t vrndx_f64(float64x1_t __p0) {
  66681.   float64x1_t __ret;
  66682.   __ret = (float64x1_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 10);
  66683.   return __ret;
  66684. }
  66685. #endif
  66686. #if defined(__aarch64__) && defined(__ARM_FEATURE_NUMERIC_MAXMIN)
  66687. #ifdef __LITTLE_ENDIAN__
  66688. __ai float64x2_t vmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
  66689.   float64x2_t __ret;
  66690.   __ret = (float64x2_t) __builtin_neon_vmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  66691.   return __ret;
  66692. }
  66693. #else
  66694. __ai float64x2_t vmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
  66695.   float64x2_t __ret;
  66696.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66697.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  66698.   __ret = (float64x2_t) __builtin_neon_vmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  66699.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66700.   return __ret;
  66701. }
  66702. #endif
  66703.  
  66704. __ai float64x1_t vmaxnm_f64(float64x1_t __p0, float64x1_t __p1) {
  66705.   float64x1_t __ret;
  66706.   __ret = (float64x1_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
  66707.   return __ret;
  66708. }
  66709. #ifdef __LITTLE_ENDIAN__
  66710. __ai float64x2_t vminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
  66711.   float64x2_t __ret;
  66712.   __ret = (float64x2_t) __builtin_neon_vminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
  66713.   return __ret;
  66714. }
  66715. #else
  66716. __ai float64x2_t vminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
  66717.   float64x2_t __ret;
  66718.   float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66719.   float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  66720.   __ret = (float64x2_t) __builtin_neon_vminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
  66721.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66722.   return __ret;
  66723. }
  66724. #endif
  66725.  
  66726. __ai float64x1_t vminnm_f64(float64x1_t __p0, float64x1_t __p1) {
  66727.   float64x1_t __ret;
  66728.   __ret = (float64x1_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
  66729.   return __ret;
  66730. }
  66731. #endif
  66732. #ifdef __LITTLE_ENDIAN__
  66733. __ai uint8x16_t vabaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  66734.   uint8x16_t __ret;
  66735.   __ret = __p0 + vabdq_u8(__p1, __p2);
  66736.   return __ret;
  66737. }
  66738. #else
  66739. __ai uint8x16_t vabaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  66740.   uint8x16_t __ret;
  66741.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  66742.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  66743.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  66744.   __ret = __rev0 + __noswap_vabdq_u8(__rev1, __rev2);
  66745.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  66746.   return __ret;
  66747. }
  66748. #endif
  66749.  
  66750. #ifdef __LITTLE_ENDIAN__
  66751. __ai uint32x4_t vabaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  66752.   uint32x4_t __ret;
  66753.   __ret = __p0 + vabdq_u32(__p1, __p2);
  66754.   return __ret;
  66755. }
  66756. #else
  66757. __ai uint32x4_t vabaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  66758.   uint32x4_t __ret;
  66759.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  66760.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  66761.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  66762.   __ret = __rev0 + __noswap_vabdq_u32(__rev1, __rev2);
  66763.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  66764.   return __ret;
  66765. }
  66766. #endif
  66767.  
  66768. #ifdef __LITTLE_ENDIAN__
  66769. __ai uint16x8_t vabaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  66770.   uint16x8_t __ret;
  66771.   __ret = __p0 + vabdq_u16(__p1, __p2);
  66772.   return __ret;
  66773. }
  66774. #else
  66775. __ai uint16x8_t vabaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  66776.   uint16x8_t __ret;
  66777.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  66778.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  66779.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  66780.   __ret = __rev0 + __noswap_vabdq_u16(__rev1, __rev2);
  66781.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  66782.   return __ret;
  66783. }
  66784. #endif
  66785.  
  66786. #ifdef __LITTLE_ENDIAN__
  66787. __ai int8x16_t vabaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  66788.   int8x16_t __ret;
  66789.   __ret = __p0 + vabdq_s8(__p1, __p2);
  66790.   return __ret;
  66791. }
  66792. #else
  66793. __ai int8x16_t vabaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
  66794.   int8x16_t __ret;
  66795.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  66796.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  66797.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  66798.   __ret = __rev0 + __noswap_vabdq_s8(__rev1, __rev2);
  66799.   __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  66800.   return __ret;
  66801. }
  66802. #endif
  66803.  
  66804. #ifdef __LITTLE_ENDIAN__
  66805. __ai int32x4_t vabaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  66806.   int32x4_t __ret;
  66807.   __ret = __p0 + vabdq_s32(__p1, __p2);
  66808.   return __ret;
  66809. }
  66810. #else
  66811. __ai int32x4_t vabaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
  66812.   int32x4_t __ret;
  66813.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  66814.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  66815.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  66816.   __ret = __rev0 + __noswap_vabdq_s32(__rev1, __rev2);
  66817.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  66818.   return __ret;
  66819. }
  66820. #endif
  66821.  
  66822. #ifdef __LITTLE_ENDIAN__
  66823. __ai int16x8_t vabaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  66824.   int16x8_t __ret;
  66825.   __ret = __p0 + vabdq_s16(__p1, __p2);
  66826.   return __ret;
  66827. }
  66828. #else
  66829. __ai int16x8_t vabaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
  66830.   int16x8_t __ret;
  66831.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  66832.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  66833.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  66834.   __ret = __rev0 + __noswap_vabdq_s16(__rev1, __rev2);
  66835.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  66836.   return __ret;
  66837. }
  66838. #endif
  66839.  
  66840. #ifdef __LITTLE_ENDIAN__
  66841. __ai uint8x8_t vaba_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  66842.   uint8x8_t __ret;
  66843.   __ret = __p0 + vabd_u8(__p1, __p2);
  66844.   return __ret;
  66845. }
  66846. #else
  66847. __ai uint8x8_t vaba_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  66848.   uint8x8_t __ret;
  66849.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  66850.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  66851.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  66852.   __ret = __rev0 + __noswap_vabd_u8(__rev1, __rev2);
  66853.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  66854.   return __ret;
  66855. }
  66856. #endif
  66857.  
  66858. #ifdef __LITTLE_ENDIAN__
  66859. __ai uint32x2_t vaba_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  66860.   uint32x2_t __ret;
  66861.   __ret = __p0 + vabd_u32(__p1, __p2);
  66862.   return __ret;
  66863. }
  66864. #else
  66865. __ai uint32x2_t vaba_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  66866.   uint32x2_t __ret;
  66867.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66868.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  66869.   uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  66870.   __ret = __rev0 + __noswap_vabd_u32(__rev1, __rev2);
  66871.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66872.   return __ret;
  66873. }
  66874. #endif
  66875.  
  66876. #ifdef __LITTLE_ENDIAN__
  66877. __ai uint16x4_t vaba_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  66878.   uint16x4_t __ret;
  66879.   __ret = __p0 + vabd_u16(__p1, __p2);
  66880.   return __ret;
  66881. }
  66882. #else
  66883. __ai uint16x4_t vaba_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  66884.   uint16x4_t __ret;
  66885.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  66886.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  66887.   uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  66888.   __ret = __rev0 + __noswap_vabd_u16(__rev1, __rev2);
  66889.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  66890.   return __ret;
  66891. }
  66892. #endif
  66893.  
  66894. #ifdef __LITTLE_ENDIAN__
  66895. __ai int8x8_t vaba_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  66896.   int8x8_t __ret;
  66897.   __ret = __p0 + vabd_s8(__p1, __p2);
  66898.   return __ret;
  66899. }
  66900. #else
  66901. __ai int8x8_t vaba_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  66902.   int8x8_t __ret;
  66903.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  66904.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  66905.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  66906.   __ret = __rev0 + __noswap_vabd_s8(__rev1, __rev2);
  66907.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  66908.   return __ret;
  66909. }
  66910. #endif
  66911.  
  66912. #ifdef __LITTLE_ENDIAN__
  66913. __ai int32x2_t vaba_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  66914.   int32x2_t __ret;
  66915.   __ret = __p0 + vabd_s32(__p1, __p2);
  66916.   return __ret;
  66917. }
  66918. #else
  66919. __ai int32x2_t vaba_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  66920.   int32x2_t __ret;
  66921.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66922.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  66923.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  66924.   __ret = __rev0 + __noswap_vabd_s32(__rev1, __rev2);
  66925.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66926.   return __ret;
  66927. }
  66928. #endif
  66929.  
  66930. #ifdef __LITTLE_ENDIAN__
  66931. __ai int16x4_t vaba_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  66932.   int16x4_t __ret;
  66933.   __ret = __p0 + vabd_s16(__p1, __p2);
  66934.   return __ret;
  66935. }
  66936. #else
  66937. __ai int16x4_t vaba_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  66938.   int16x4_t __ret;
  66939.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  66940.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  66941.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  66942.   __ret = __rev0 + __noswap_vabd_s16(__rev1, __rev2);
  66943.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  66944.   return __ret;
  66945. }
  66946. #endif
  66947.  
  66948. #ifdef __LITTLE_ENDIAN__
  66949. __ai uint16x8_t vabdl_u8(uint8x8_t __p0, uint8x8_t __p1) {
  66950.   uint16x8_t __ret;
  66951.   __ret = (uint16x8_t)(vmovl_u8((uint8x8_t)(vabd_u8(__p0, __p1))));
  66952.   return __ret;
  66953. }
  66954. #else
  66955. __ai uint16x8_t vabdl_u8(uint8x8_t __p0, uint8x8_t __p1) {
  66956.   uint16x8_t __ret;
  66957.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  66958.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  66959.   __ret = (uint16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_u8(__rev0, __rev1))));
  66960.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  66961.   return __ret;
  66962. }
  66963. __ai uint16x8_t __noswap_vabdl_u8(uint8x8_t __p0, uint8x8_t __p1) {
  66964.   uint16x8_t __ret;
  66965.   __ret = (uint16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_u8(__p0, __p1))));
  66966.   return __ret;
  66967. }
  66968. #endif
  66969.  
  66970. #ifdef __LITTLE_ENDIAN__
  66971. __ai uint64x2_t vabdl_u32(uint32x2_t __p0, uint32x2_t __p1) {
  66972.   uint64x2_t __ret;
  66973.   __ret = (uint64x2_t)(vmovl_u32((uint32x2_t)(vabd_u32(__p0, __p1))));
  66974.   return __ret;
  66975. }
  66976. #else
  66977. __ai uint64x2_t vabdl_u32(uint32x2_t __p0, uint32x2_t __p1) {
  66978.   uint64x2_t __ret;
  66979.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  66980.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  66981.   __ret = (uint64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_u32(__rev0, __rev1))));
  66982.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  66983.   return __ret;
  66984. }
  66985. __ai uint64x2_t __noswap_vabdl_u32(uint32x2_t __p0, uint32x2_t __p1) {
  66986.   uint64x2_t __ret;
  66987.   __ret = (uint64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_u32(__p0, __p1))));
  66988.   return __ret;
  66989. }
  66990. #endif
  66991.  
  66992. #ifdef __LITTLE_ENDIAN__
  66993. __ai uint32x4_t vabdl_u16(uint16x4_t __p0, uint16x4_t __p1) {
  66994.   uint32x4_t __ret;
  66995.   __ret = (uint32x4_t)(vmovl_u16((uint16x4_t)(vabd_u16(__p0, __p1))));
  66996.   return __ret;
  66997. }
  66998. #else
  66999. __ai uint32x4_t vabdl_u16(uint16x4_t __p0, uint16x4_t __p1) {
  67000.   uint32x4_t __ret;
  67001.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67002.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67003.   __ret = (uint32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_u16(__rev0, __rev1))));
  67004.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67005.   return __ret;
  67006. }
  67007. __ai uint32x4_t __noswap_vabdl_u16(uint16x4_t __p0, uint16x4_t __p1) {
  67008.   uint32x4_t __ret;
  67009.   __ret = (uint32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_u16(__p0, __p1))));
  67010.   return __ret;
  67011. }
  67012. #endif
  67013.  
  67014. #ifdef __LITTLE_ENDIAN__
  67015. __ai int16x8_t vabdl_s8(int8x8_t __p0, int8x8_t __p1) {
  67016.   int16x8_t __ret;
  67017.   __ret = (int16x8_t)(vmovl_u8((uint8x8_t)(vabd_s8(__p0, __p1))));
  67018.   return __ret;
  67019. }
  67020. #else
  67021. __ai int16x8_t vabdl_s8(int8x8_t __p0, int8x8_t __p1) {
  67022.   int16x8_t __ret;
  67023.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  67024.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  67025.   __ret = (int16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_s8(__rev0, __rev1))));
  67026.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  67027.   return __ret;
  67028. }
  67029. __ai int16x8_t __noswap_vabdl_s8(int8x8_t __p0, int8x8_t __p1) {
  67030.   int16x8_t __ret;
  67031.   __ret = (int16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_s8(__p0, __p1))));
  67032.   return __ret;
  67033. }
  67034. #endif
  67035.  
  67036. #ifdef __LITTLE_ENDIAN__
  67037. __ai int64x2_t vabdl_s32(int32x2_t __p0, int32x2_t __p1) {
  67038.   int64x2_t __ret;
  67039.   __ret = (int64x2_t)(vmovl_u32((uint32x2_t)(vabd_s32(__p0, __p1))));
  67040.   return __ret;
  67041. }
  67042. #else
  67043. __ai int64x2_t vabdl_s32(int32x2_t __p0, int32x2_t __p1) {
  67044.   int64x2_t __ret;
  67045.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67046.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67047.   __ret = (int64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_s32(__rev0, __rev1))));
  67048.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67049.   return __ret;
  67050. }
  67051. __ai int64x2_t __noswap_vabdl_s32(int32x2_t __p0, int32x2_t __p1) {
  67052.   int64x2_t __ret;
  67053.   __ret = (int64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_s32(__p0, __p1))));
  67054.   return __ret;
  67055. }
  67056. #endif
  67057.  
  67058. #ifdef __LITTLE_ENDIAN__
  67059. __ai int32x4_t vabdl_s16(int16x4_t __p0, int16x4_t __p1) {
  67060.   int32x4_t __ret;
  67061.   __ret = (int32x4_t)(vmovl_u16((uint16x4_t)(vabd_s16(__p0, __p1))));
  67062.   return __ret;
  67063. }
  67064. #else
  67065. __ai int32x4_t vabdl_s16(int16x4_t __p0, int16x4_t __p1) {
  67066.   int32x4_t __ret;
  67067.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67068.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67069.   __ret = (int32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_s16(__rev0, __rev1))));
  67070.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67071.   return __ret;
  67072. }
  67073. __ai int32x4_t __noswap_vabdl_s16(int16x4_t __p0, int16x4_t __p1) {
  67074.   int32x4_t __ret;
  67075.   __ret = (int32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_s16(__p0, __p1))));
  67076.   return __ret;
  67077. }
  67078. #endif
  67079.  
  67080. #ifdef __LITTLE_ENDIAN__
  67081. __ai uint16x8_t vaddl_u8(uint8x8_t __p0, uint8x8_t __p1) {
  67082.   uint16x8_t __ret;
  67083.   __ret = vmovl_u8(__p0) + vmovl_u8(__p1);
  67084.   return __ret;
  67085. }
  67086. #else
  67087. __ai uint16x8_t vaddl_u8(uint8x8_t __p0, uint8x8_t __p1) {
  67088.   uint16x8_t __ret;
  67089.   uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  67090.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  67091.   __ret = __noswap_vmovl_u8(__rev0) + __noswap_vmovl_u8(__rev1);
  67092.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  67093.   return __ret;
  67094. }
  67095. #endif
  67096.  
  67097. #ifdef __LITTLE_ENDIAN__
  67098. __ai uint64x2_t vaddl_u32(uint32x2_t __p0, uint32x2_t __p1) {
  67099.   uint64x2_t __ret;
  67100.   __ret = vmovl_u32(__p0) + vmovl_u32(__p1);
  67101.   return __ret;
  67102. }
  67103. #else
  67104. __ai uint64x2_t vaddl_u32(uint32x2_t __p0, uint32x2_t __p1) {
  67105.   uint64x2_t __ret;
  67106.   uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67107.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67108.   __ret = __noswap_vmovl_u32(__rev0) + __noswap_vmovl_u32(__rev1);
  67109.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67110.   return __ret;
  67111. }
  67112. #endif
  67113.  
  67114. #ifdef __LITTLE_ENDIAN__
  67115. __ai uint32x4_t vaddl_u16(uint16x4_t __p0, uint16x4_t __p1) {
  67116.   uint32x4_t __ret;
  67117.   __ret = vmovl_u16(__p0) + vmovl_u16(__p1);
  67118.   return __ret;
  67119. }
  67120. #else
  67121. __ai uint32x4_t vaddl_u16(uint16x4_t __p0, uint16x4_t __p1) {
  67122.   uint32x4_t __ret;
  67123.   uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67124.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67125.   __ret = __noswap_vmovl_u16(__rev0) + __noswap_vmovl_u16(__rev1);
  67126.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67127.   return __ret;
  67128. }
  67129. #endif
  67130.  
  67131. #ifdef __LITTLE_ENDIAN__
  67132. __ai int16x8_t vaddl_s8(int8x8_t __p0, int8x8_t __p1) {
  67133.   int16x8_t __ret;
  67134.   __ret = vmovl_s8(__p0) + vmovl_s8(__p1);
  67135.   return __ret;
  67136. }
  67137. #else
  67138. __ai int16x8_t vaddl_s8(int8x8_t __p0, int8x8_t __p1) {
  67139.   int16x8_t __ret;
  67140.   int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  67141.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  67142.   __ret = __noswap_vmovl_s8(__rev0) + __noswap_vmovl_s8(__rev1);
  67143.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  67144.   return __ret;
  67145. }
  67146. #endif
  67147.  
  67148. #ifdef __LITTLE_ENDIAN__
  67149. __ai int64x2_t vaddl_s32(int32x2_t __p0, int32x2_t __p1) {
  67150.   int64x2_t __ret;
  67151.   __ret = vmovl_s32(__p0) + vmovl_s32(__p1);
  67152.   return __ret;
  67153. }
  67154. #else
  67155. __ai int64x2_t vaddl_s32(int32x2_t __p0, int32x2_t __p1) {
  67156.   int64x2_t __ret;
  67157.   int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67158.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67159.   __ret = __noswap_vmovl_s32(__rev0) + __noswap_vmovl_s32(__rev1);
  67160.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67161.   return __ret;
  67162. }
  67163. #endif
  67164.  
  67165. #ifdef __LITTLE_ENDIAN__
  67166. __ai int32x4_t vaddl_s16(int16x4_t __p0, int16x4_t __p1) {
  67167.   int32x4_t __ret;
  67168.   __ret = vmovl_s16(__p0) + vmovl_s16(__p1);
  67169.   return __ret;
  67170. }
  67171. #else
  67172. __ai int32x4_t vaddl_s16(int16x4_t __p0, int16x4_t __p1) {
  67173.   int32x4_t __ret;
  67174.   int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67175.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67176.   __ret = __noswap_vmovl_s16(__rev0) + __noswap_vmovl_s16(__rev1);
  67177.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67178.   return __ret;
  67179. }
  67180. #endif
  67181.  
  67182. #ifdef __LITTLE_ENDIAN__
  67183. __ai uint16x8_t vaddw_u8(uint16x8_t __p0, uint8x8_t __p1) {
  67184.   uint16x8_t __ret;
  67185.   __ret = __p0 + vmovl_u8(__p1);
  67186.   return __ret;
  67187. }
  67188. #else
  67189. __ai uint16x8_t vaddw_u8(uint16x8_t __p0, uint8x8_t __p1) {
  67190.   uint16x8_t __ret;
  67191.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  67192.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  67193.   __ret = __rev0 + __noswap_vmovl_u8(__rev1);
  67194.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  67195.   return __ret;
  67196. }
  67197. #endif
  67198.  
  67199. #ifdef __LITTLE_ENDIAN__
  67200. __ai uint64x2_t vaddw_u32(uint64x2_t __p0, uint32x2_t __p1) {
  67201.   uint64x2_t __ret;
  67202.   __ret = __p0 + vmovl_u32(__p1);
  67203.   return __ret;
  67204. }
  67205. #else
  67206. __ai uint64x2_t vaddw_u32(uint64x2_t __p0, uint32x2_t __p1) {
  67207.   uint64x2_t __ret;
  67208.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67209.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67210.   __ret = __rev0 + __noswap_vmovl_u32(__rev1);
  67211.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67212.   return __ret;
  67213. }
  67214. #endif
  67215.  
  67216. #ifdef __LITTLE_ENDIAN__
  67217. __ai uint32x4_t vaddw_u16(uint32x4_t __p0, uint16x4_t __p1) {
  67218.   uint32x4_t __ret;
  67219.   __ret = __p0 + vmovl_u16(__p1);
  67220.   return __ret;
  67221. }
  67222. #else
  67223. __ai uint32x4_t vaddw_u16(uint32x4_t __p0, uint16x4_t __p1) {
  67224.   uint32x4_t __ret;
  67225.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67226.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67227.   __ret = __rev0 + __noswap_vmovl_u16(__rev1);
  67228.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67229.   return __ret;
  67230. }
  67231. #endif
  67232.  
  67233. #ifdef __LITTLE_ENDIAN__
  67234. __ai int16x8_t vaddw_s8(int16x8_t __p0, int8x8_t __p1) {
  67235.   int16x8_t __ret;
  67236.   __ret = __p0 + vmovl_s8(__p1);
  67237.   return __ret;
  67238. }
  67239. #else
  67240. __ai int16x8_t vaddw_s8(int16x8_t __p0, int8x8_t __p1) {
  67241.   int16x8_t __ret;
  67242.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  67243.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  67244.   __ret = __rev0 + __noswap_vmovl_s8(__rev1);
  67245.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  67246.   return __ret;
  67247. }
  67248. #endif
  67249.  
  67250. #ifdef __LITTLE_ENDIAN__
  67251. __ai int64x2_t vaddw_s32(int64x2_t __p0, int32x2_t __p1) {
  67252.   int64x2_t __ret;
  67253.   __ret = __p0 + vmovl_s32(__p1);
  67254.   return __ret;
  67255. }
  67256. #else
  67257. __ai int64x2_t vaddw_s32(int64x2_t __p0, int32x2_t __p1) {
  67258.   int64x2_t __ret;
  67259.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67260.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67261.   __ret = __rev0 + __noswap_vmovl_s32(__rev1);
  67262.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67263.   return __ret;
  67264. }
  67265. #endif
  67266.  
  67267. #ifdef __LITTLE_ENDIAN__
  67268. __ai int32x4_t vaddw_s16(int32x4_t __p0, int16x4_t __p1) {
  67269.   int32x4_t __ret;
  67270.   __ret = __p0 + vmovl_s16(__p1);
  67271.   return __ret;
  67272. }
  67273. #else
  67274. __ai int32x4_t vaddw_s16(int32x4_t __p0, int16x4_t __p1) {
  67275.   int32x4_t __ret;
  67276.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67277.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67278.   __ret = __rev0 + __noswap_vmovl_s16(__rev1);
  67279.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67280.   return __ret;
  67281. }
  67282. #endif
  67283.  
  67284. #ifdef __LITTLE_ENDIAN__
  67285. #define vget_lane_f16(__p0_847, __p1_847) __extension__ ({ \
  67286.   float16_t __ret_847; \
  67287.   float16x4_t __s0_847 = __p0_847; \
  67288. float16x4_t __reint_847 = __s0_847; \
  67289. int16_t __reint1_847 = vget_lane_s16(*(int16x4_t *) &__reint_847, __p1_847); \
  67290.   __ret_847 = *(float16_t *) &__reint1_847; \
  67291.   __ret_847; \
  67292. })
  67293. #else
  67294. #define vget_lane_f16(__p0_848, __p1_848) __extension__ ({ \
  67295.   float16_t __ret_848; \
  67296.   float16x4_t __s0_848 = __p0_848; \
  67297.   float16x4_t __rev0_848;  __rev0_848 = __builtin_shufflevector(__s0_848, __s0_848, 3, 2, 1, 0); \
  67298. float16x4_t __reint_848 = __rev0_848; \
  67299. int16_t __reint1_848 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_848, __p1_848); \
  67300.   __ret_848 = *(float16_t *) &__reint1_848; \
  67301.   __ret_848; \
  67302. })
  67303. #define __noswap_vget_lane_f16(__p0_849, __p1_849) __extension__ ({ \
  67304.   float16_t __ret_849; \
  67305.   float16x4_t __s0_849 = __p0_849; \
  67306. float16x4_t __reint_849 = __s0_849; \
  67307. int16_t __reint1_849 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_849, __p1_849); \
  67308.   __ret_849 = *(float16_t *) &__reint1_849; \
  67309.   __ret_849; \
  67310. })
  67311. #endif
  67312.  
  67313. #ifdef __LITTLE_ENDIAN__
  67314. #define vgetq_lane_f16(__p0_850, __p1_850) __extension__ ({ \
  67315.   float16_t __ret_850; \
  67316.   float16x8_t __s0_850 = __p0_850; \
  67317. float16x8_t __reint_850 = __s0_850; \
  67318. int16_t __reint1_850 = vgetq_lane_s16(*(int16x8_t *) &__reint_850, __p1_850); \
  67319.   __ret_850 = *(float16_t *) &__reint1_850; \
  67320.   __ret_850; \
  67321. })
  67322. #else
  67323. #define vgetq_lane_f16(__p0_851, __p1_851) __extension__ ({ \
  67324.   float16_t __ret_851; \
  67325.   float16x8_t __s0_851 = __p0_851; \
  67326.   float16x8_t __rev0_851;  __rev0_851 = __builtin_shufflevector(__s0_851, __s0_851, 7, 6, 5, 4, 3, 2, 1, 0); \
  67327. float16x8_t __reint_851 = __rev0_851; \
  67328. int16_t __reint1_851 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_851, __p1_851); \
  67329.   __ret_851 = *(float16_t *) &__reint1_851; \
  67330.   __ret_851; \
  67331. })
  67332. #define __noswap_vgetq_lane_f16(__p0_852, __p1_852) __extension__ ({ \
  67333.   float16_t __ret_852; \
  67334.   float16x8_t __s0_852 = __p0_852; \
  67335. float16x8_t __reint_852 = __s0_852; \
  67336. int16_t __reint1_852 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_852, __p1_852); \
  67337.   __ret_852 = *(float16_t *) &__reint1_852; \
  67338.   __ret_852; \
  67339. })
  67340. #endif
  67341.  
  67342. #ifdef __LITTLE_ENDIAN__
  67343. __ai uint16x8_t vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  67344.   uint16x8_t __ret;
  67345.   __ret = __p0 + vmull_u8(__p1, __p2);
  67346.   return __ret;
  67347. }
  67348. #else
  67349. __ai uint16x8_t vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  67350.   uint16x8_t __ret;
  67351.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  67352.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  67353.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  67354.   __ret = __rev0 + __noswap_vmull_u8(__rev1, __rev2);
  67355.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  67356.   return __ret;
  67357. }
  67358. __ai uint16x8_t __noswap_vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  67359.   uint16x8_t __ret;
  67360.   __ret = __p0 + __noswap_vmull_u8(__p1, __p2);
  67361.   return __ret;
  67362. }
  67363. #endif
  67364.  
  67365. #ifdef __LITTLE_ENDIAN__
  67366. __ai uint64x2_t vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  67367.   uint64x2_t __ret;
  67368.   __ret = __p0 + vmull_u32(__p1, __p2);
  67369.   return __ret;
  67370. }
  67371. #else
  67372. __ai uint64x2_t vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  67373.   uint64x2_t __ret;
  67374.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67375.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67376.   uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  67377.   __ret = __rev0 + __noswap_vmull_u32(__rev1, __rev2);
  67378.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67379.   return __ret;
  67380. }
  67381. __ai uint64x2_t __noswap_vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  67382.   uint64x2_t __ret;
  67383.   __ret = __p0 + __noswap_vmull_u32(__p1, __p2);
  67384.   return __ret;
  67385. }
  67386. #endif
  67387.  
  67388. #ifdef __LITTLE_ENDIAN__
  67389. __ai uint32x4_t vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  67390.   uint32x4_t __ret;
  67391.   __ret = __p0 + vmull_u16(__p1, __p2);
  67392.   return __ret;
  67393. }
  67394. #else
  67395. __ai uint32x4_t vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  67396.   uint32x4_t __ret;
  67397.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67398.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67399.   uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  67400.   __ret = __rev0 + __noswap_vmull_u16(__rev1, __rev2);
  67401.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67402.   return __ret;
  67403. }
  67404. __ai uint32x4_t __noswap_vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  67405.   uint32x4_t __ret;
  67406.   __ret = __p0 + __noswap_vmull_u16(__p1, __p2);
  67407.   return __ret;
  67408. }
  67409. #endif
  67410.  
  67411. #ifdef __LITTLE_ENDIAN__
  67412. __ai int16x8_t vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  67413.   int16x8_t __ret;
  67414.   __ret = __p0 + vmull_s8(__p1, __p2);
  67415.   return __ret;
  67416. }
  67417. #else
  67418. __ai int16x8_t vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  67419.   int16x8_t __ret;
  67420.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  67421.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  67422.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  67423.   __ret = __rev0 + __noswap_vmull_s8(__rev1, __rev2);
  67424.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  67425.   return __ret;
  67426. }
  67427. __ai int16x8_t __noswap_vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  67428.   int16x8_t __ret;
  67429.   __ret = __p0 + __noswap_vmull_s8(__p1, __p2);
  67430.   return __ret;
  67431. }
  67432. #endif
  67433.  
  67434. #ifdef __LITTLE_ENDIAN__
  67435. __ai int64x2_t vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  67436.   int64x2_t __ret;
  67437.   __ret = __p0 + vmull_s32(__p1, __p2);
  67438.   return __ret;
  67439. }
  67440. #else
  67441. __ai int64x2_t vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  67442.   int64x2_t __ret;
  67443.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67444.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67445.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  67446.   __ret = __rev0 + __noswap_vmull_s32(__rev1, __rev2);
  67447.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67448.   return __ret;
  67449. }
  67450. __ai int64x2_t __noswap_vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  67451.   int64x2_t __ret;
  67452.   __ret = __p0 + __noswap_vmull_s32(__p1, __p2);
  67453.   return __ret;
  67454. }
  67455. #endif
  67456.  
  67457. #ifdef __LITTLE_ENDIAN__
  67458. __ai int32x4_t vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  67459.   int32x4_t __ret;
  67460.   __ret = __p0 + vmull_s16(__p1, __p2);
  67461.   return __ret;
  67462. }
  67463. #else
  67464. __ai int32x4_t vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  67465.   int32x4_t __ret;
  67466.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67467.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67468.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  67469.   __ret = __rev0 + __noswap_vmull_s16(__rev1, __rev2);
  67470.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67471.   return __ret;
  67472. }
  67473. __ai int32x4_t __noswap_vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  67474.   int32x4_t __ret;
  67475.   __ret = __p0 + __noswap_vmull_s16(__p1, __p2);
  67476.   return __ret;
  67477. }
  67478. #endif
  67479.  
  67480. #ifdef __LITTLE_ENDIAN__
  67481. #define vmlal_lane_u32(__p0_853, __p1_853, __p2_853, __p3_853) __extension__ ({ \
  67482.   uint64x2_t __ret_853; \
  67483.   uint64x2_t __s0_853 = __p0_853; \
  67484.   uint32x2_t __s1_853 = __p1_853; \
  67485.   uint32x2_t __s2_853 = __p2_853; \
  67486.   __ret_853 = __s0_853 + vmull_u32(__s1_853, splat_lane_u32(__s2_853, __p3_853)); \
  67487.   __ret_853; \
  67488. })
  67489. #else
  67490. #define vmlal_lane_u32(__p0_854, __p1_854, __p2_854, __p3_854) __extension__ ({ \
  67491.   uint64x2_t __ret_854; \
  67492.   uint64x2_t __s0_854 = __p0_854; \
  67493.   uint32x2_t __s1_854 = __p1_854; \
  67494.   uint32x2_t __s2_854 = __p2_854; \
  67495.   uint64x2_t __rev0_854;  __rev0_854 = __builtin_shufflevector(__s0_854, __s0_854, 1, 0); \
  67496.   uint32x2_t __rev1_854;  __rev1_854 = __builtin_shufflevector(__s1_854, __s1_854, 1, 0); \
  67497.   uint32x2_t __rev2_854;  __rev2_854 = __builtin_shufflevector(__s2_854, __s2_854, 1, 0); \
  67498.   __ret_854 = __rev0_854 + __noswap_vmull_u32(__rev1_854, __noswap_splat_lane_u32(__rev2_854, __p3_854)); \
  67499.   __ret_854 = __builtin_shufflevector(__ret_854, __ret_854, 1, 0); \
  67500.   __ret_854; \
  67501. })
  67502. #endif
  67503.  
  67504. #ifdef __LITTLE_ENDIAN__
  67505. #define vmlal_lane_u16(__p0_855, __p1_855, __p2_855, __p3_855) __extension__ ({ \
  67506.   uint32x4_t __ret_855; \
  67507.   uint32x4_t __s0_855 = __p0_855; \
  67508.   uint16x4_t __s1_855 = __p1_855; \
  67509.   uint16x4_t __s2_855 = __p2_855; \
  67510.   __ret_855 = __s0_855 + vmull_u16(__s1_855, splat_lane_u16(__s2_855, __p3_855)); \
  67511.   __ret_855; \
  67512. })
  67513. #else
  67514. #define vmlal_lane_u16(__p0_856, __p1_856, __p2_856, __p3_856) __extension__ ({ \
  67515.   uint32x4_t __ret_856; \
  67516.   uint32x4_t __s0_856 = __p0_856; \
  67517.   uint16x4_t __s1_856 = __p1_856; \
  67518.   uint16x4_t __s2_856 = __p2_856; \
  67519.   uint32x4_t __rev0_856;  __rev0_856 = __builtin_shufflevector(__s0_856, __s0_856, 3, 2, 1, 0); \
  67520.   uint16x4_t __rev1_856;  __rev1_856 = __builtin_shufflevector(__s1_856, __s1_856, 3, 2, 1, 0); \
  67521.   uint16x4_t __rev2_856;  __rev2_856 = __builtin_shufflevector(__s2_856, __s2_856, 3, 2, 1, 0); \
  67522.   __ret_856 = __rev0_856 + __noswap_vmull_u16(__rev1_856, __noswap_splat_lane_u16(__rev2_856, __p3_856)); \
  67523.   __ret_856 = __builtin_shufflevector(__ret_856, __ret_856, 3, 2, 1, 0); \
  67524.   __ret_856; \
  67525. })
  67526. #endif
  67527.  
  67528. #ifdef __LITTLE_ENDIAN__
  67529. #define vmlal_lane_s32(__p0_857, __p1_857, __p2_857, __p3_857) __extension__ ({ \
  67530.   int64x2_t __ret_857; \
  67531.   int64x2_t __s0_857 = __p0_857; \
  67532.   int32x2_t __s1_857 = __p1_857; \
  67533.   int32x2_t __s2_857 = __p2_857; \
  67534.   __ret_857 = __s0_857 + vmull_s32(__s1_857, splat_lane_s32(__s2_857, __p3_857)); \
  67535.   __ret_857; \
  67536. })
  67537. #else
  67538. #define vmlal_lane_s32(__p0_858, __p1_858, __p2_858, __p3_858) __extension__ ({ \
  67539.   int64x2_t __ret_858; \
  67540.   int64x2_t __s0_858 = __p0_858; \
  67541.   int32x2_t __s1_858 = __p1_858; \
  67542.   int32x2_t __s2_858 = __p2_858; \
  67543.   int64x2_t __rev0_858;  __rev0_858 = __builtin_shufflevector(__s0_858, __s0_858, 1, 0); \
  67544.   int32x2_t __rev1_858;  __rev1_858 = __builtin_shufflevector(__s1_858, __s1_858, 1, 0); \
  67545.   int32x2_t __rev2_858;  __rev2_858 = __builtin_shufflevector(__s2_858, __s2_858, 1, 0); \
  67546.   __ret_858 = __rev0_858 + __noswap_vmull_s32(__rev1_858, __noswap_splat_lane_s32(__rev2_858, __p3_858)); \
  67547.   __ret_858 = __builtin_shufflevector(__ret_858, __ret_858, 1, 0); \
  67548.   __ret_858; \
  67549. })
  67550. #endif
  67551.  
  67552. #ifdef __LITTLE_ENDIAN__
  67553. #define vmlal_lane_s16(__p0_859, __p1_859, __p2_859, __p3_859) __extension__ ({ \
  67554.   int32x4_t __ret_859; \
  67555.   int32x4_t __s0_859 = __p0_859; \
  67556.   int16x4_t __s1_859 = __p1_859; \
  67557.   int16x4_t __s2_859 = __p2_859; \
  67558.   __ret_859 = __s0_859 + vmull_s16(__s1_859, splat_lane_s16(__s2_859, __p3_859)); \
  67559.   __ret_859; \
  67560. })
  67561. #else
  67562. #define vmlal_lane_s16(__p0_860, __p1_860, __p2_860, __p3_860) __extension__ ({ \
  67563.   int32x4_t __ret_860; \
  67564.   int32x4_t __s0_860 = __p0_860; \
  67565.   int16x4_t __s1_860 = __p1_860; \
  67566.   int16x4_t __s2_860 = __p2_860; \
  67567.   int32x4_t __rev0_860;  __rev0_860 = __builtin_shufflevector(__s0_860, __s0_860, 3, 2, 1, 0); \
  67568.   int16x4_t __rev1_860;  __rev1_860 = __builtin_shufflevector(__s1_860, __s1_860, 3, 2, 1, 0); \
  67569.   int16x4_t __rev2_860;  __rev2_860 = __builtin_shufflevector(__s2_860, __s2_860, 3, 2, 1, 0); \
  67570.   __ret_860 = __rev0_860 + __noswap_vmull_s16(__rev1_860, __noswap_splat_lane_s16(__rev2_860, __p3_860)); \
  67571.   __ret_860 = __builtin_shufflevector(__ret_860, __ret_860, 3, 2, 1, 0); \
  67572.   __ret_860; \
  67573. })
  67574. #endif
  67575.  
  67576. #ifdef __LITTLE_ENDIAN__
  67577. __ai uint64x2_t vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  67578.   uint64x2_t __ret;
  67579.   __ret = __p0 + vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
  67580.   return __ret;
  67581. }
  67582. #else
  67583. __ai uint64x2_t vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  67584.   uint64x2_t __ret;
  67585.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67586.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67587.   __ret = __rev0 + __noswap_vmull_u32(__rev1, (uint32x2_t) {__p2, __p2});
  67588.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67589.   return __ret;
  67590. }
  67591. __ai uint64x2_t __noswap_vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  67592.   uint64x2_t __ret;
  67593.   __ret = __p0 + __noswap_vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
  67594.   return __ret;
  67595. }
  67596. #endif
  67597.  
  67598. #ifdef __LITTLE_ENDIAN__
  67599. __ai uint32x4_t vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  67600.   uint32x4_t __ret;
  67601.   __ret = __p0 + vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
  67602.   return __ret;
  67603. }
  67604. #else
  67605. __ai uint32x4_t vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  67606.   uint32x4_t __ret;
  67607.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67608.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67609.   __ret = __rev0 + __noswap_vmull_u16(__rev1, (uint16x4_t) {__p2, __p2, __p2, __p2});
  67610.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67611.   return __ret;
  67612. }
  67613. __ai uint32x4_t __noswap_vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  67614.   uint32x4_t __ret;
  67615.   __ret = __p0 + __noswap_vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
  67616.   return __ret;
  67617. }
  67618. #endif
  67619.  
  67620. #ifdef __LITTLE_ENDIAN__
  67621. __ai int64x2_t vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  67622.   int64x2_t __ret;
  67623.   __ret = __p0 + vmull_s32(__p1, (int32x2_t) {__p2, __p2});
  67624.   return __ret;
  67625. }
  67626. #else
  67627. __ai int64x2_t vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  67628.   int64x2_t __ret;
  67629.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67630.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67631.   __ret = __rev0 + __noswap_vmull_s32(__rev1, (int32x2_t) {__p2, __p2});
  67632.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67633.   return __ret;
  67634. }
  67635. __ai int64x2_t __noswap_vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  67636.   int64x2_t __ret;
  67637.   __ret = __p0 + __noswap_vmull_s32(__p1, (int32x2_t) {__p2, __p2});
  67638.   return __ret;
  67639. }
  67640. #endif
  67641.  
  67642. #ifdef __LITTLE_ENDIAN__
  67643. __ai int32x4_t vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  67644.   int32x4_t __ret;
  67645.   __ret = __p0 + vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
  67646.   return __ret;
  67647. }
  67648. #else
  67649. __ai int32x4_t vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  67650.   int32x4_t __ret;
  67651.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67652.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67653.   __ret = __rev0 + __noswap_vmull_s16(__rev1, (int16x4_t) {__p2, __p2, __p2, __p2});
  67654.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67655.   return __ret;
  67656. }
  67657. __ai int32x4_t __noswap_vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  67658.   int32x4_t __ret;
  67659.   __ret = __p0 + __noswap_vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
  67660.   return __ret;
  67661. }
  67662. #endif
  67663.  
  67664. #ifdef __LITTLE_ENDIAN__
  67665. __ai uint16x8_t vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  67666.   uint16x8_t __ret;
  67667.   __ret = __p0 - vmull_u8(__p1, __p2);
  67668.   return __ret;
  67669. }
  67670. #else
  67671. __ai uint16x8_t vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  67672.   uint16x8_t __ret;
  67673.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  67674.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  67675.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  67676.   __ret = __rev0 - __noswap_vmull_u8(__rev1, __rev2);
  67677.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  67678.   return __ret;
  67679. }
  67680. __ai uint16x8_t __noswap_vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  67681.   uint16x8_t __ret;
  67682.   __ret = __p0 - __noswap_vmull_u8(__p1, __p2);
  67683.   return __ret;
  67684. }
  67685. #endif
  67686.  
  67687. #ifdef __LITTLE_ENDIAN__
  67688. __ai uint64x2_t vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  67689.   uint64x2_t __ret;
  67690.   __ret = __p0 - vmull_u32(__p1, __p2);
  67691.   return __ret;
  67692. }
  67693. #else
  67694. __ai uint64x2_t vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  67695.   uint64x2_t __ret;
  67696.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67697.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67698.   uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  67699.   __ret = __rev0 - __noswap_vmull_u32(__rev1, __rev2);
  67700.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67701.   return __ret;
  67702. }
  67703. __ai uint64x2_t __noswap_vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  67704.   uint64x2_t __ret;
  67705.   __ret = __p0 - __noswap_vmull_u32(__p1, __p2);
  67706.   return __ret;
  67707. }
  67708. #endif
  67709.  
  67710. #ifdef __LITTLE_ENDIAN__
  67711. __ai uint32x4_t vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  67712.   uint32x4_t __ret;
  67713.   __ret = __p0 - vmull_u16(__p1, __p2);
  67714.   return __ret;
  67715. }
  67716. #else
  67717. __ai uint32x4_t vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  67718.   uint32x4_t __ret;
  67719.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67720.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67721.   uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  67722.   __ret = __rev0 - __noswap_vmull_u16(__rev1, __rev2);
  67723.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67724.   return __ret;
  67725. }
  67726. __ai uint32x4_t __noswap_vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  67727.   uint32x4_t __ret;
  67728.   __ret = __p0 - __noswap_vmull_u16(__p1, __p2);
  67729.   return __ret;
  67730. }
  67731. #endif
  67732.  
  67733. #ifdef __LITTLE_ENDIAN__
  67734. __ai int16x8_t vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  67735.   int16x8_t __ret;
  67736.   __ret = __p0 - vmull_s8(__p1, __p2);
  67737.   return __ret;
  67738. }
  67739. #else
  67740. __ai int16x8_t vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  67741.   int16x8_t __ret;
  67742.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  67743.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  67744.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  67745.   __ret = __rev0 - __noswap_vmull_s8(__rev1, __rev2);
  67746.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  67747.   return __ret;
  67748. }
  67749. __ai int16x8_t __noswap_vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  67750.   int16x8_t __ret;
  67751.   __ret = __p0 - __noswap_vmull_s8(__p1, __p2);
  67752.   return __ret;
  67753. }
  67754. #endif
  67755.  
  67756. #ifdef __LITTLE_ENDIAN__
  67757. __ai int64x2_t vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  67758.   int64x2_t __ret;
  67759.   __ret = __p0 - vmull_s32(__p1, __p2);
  67760.   return __ret;
  67761. }
  67762. #else
  67763. __ai int64x2_t vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  67764.   int64x2_t __ret;
  67765.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67766.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67767.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  67768.   __ret = __rev0 - __noswap_vmull_s32(__rev1, __rev2);
  67769.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67770.   return __ret;
  67771. }
  67772. __ai int64x2_t __noswap_vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  67773.   int64x2_t __ret;
  67774.   __ret = __p0 - __noswap_vmull_s32(__p1, __p2);
  67775.   return __ret;
  67776. }
  67777. #endif
  67778.  
  67779. #ifdef __LITTLE_ENDIAN__
  67780. __ai int32x4_t vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  67781.   int32x4_t __ret;
  67782.   __ret = __p0 - vmull_s16(__p1, __p2);
  67783.   return __ret;
  67784. }
  67785. #else
  67786. __ai int32x4_t vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  67787.   int32x4_t __ret;
  67788.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67789.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67790.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  67791.   __ret = __rev0 - __noswap_vmull_s16(__rev1, __rev2);
  67792.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67793.   return __ret;
  67794. }
  67795. __ai int32x4_t __noswap_vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  67796.   int32x4_t __ret;
  67797.   __ret = __p0 - __noswap_vmull_s16(__p1, __p2);
  67798.   return __ret;
  67799. }
  67800. #endif
  67801.  
  67802. #ifdef __LITTLE_ENDIAN__
  67803. #define vmlsl_lane_u32(__p0_861, __p1_861, __p2_861, __p3_861) __extension__ ({ \
  67804.   uint64x2_t __ret_861; \
  67805.   uint64x2_t __s0_861 = __p0_861; \
  67806.   uint32x2_t __s1_861 = __p1_861; \
  67807.   uint32x2_t __s2_861 = __p2_861; \
  67808.   __ret_861 = __s0_861 - vmull_u32(__s1_861, splat_lane_u32(__s2_861, __p3_861)); \
  67809.   __ret_861; \
  67810. })
  67811. #else
  67812. #define vmlsl_lane_u32(__p0_862, __p1_862, __p2_862, __p3_862) __extension__ ({ \
  67813.   uint64x2_t __ret_862; \
  67814.   uint64x2_t __s0_862 = __p0_862; \
  67815.   uint32x2_t __s1_862 = __p1_862; \
  67816.   uint32x2_t __s2_862 = __p2_862; \
  67817.   uint64x2_t __rev0_862;  __rev0_862 = __builtin_shufflevector(__s0_862, __s0_862, 1, 0); \
  67818.   uint32x2_t __rev1_862;  __rev1_862 = __builtin_shufflevector(__s1_862, __s1_862, 1, 0); \
  67819.   uint32x2_t __rev2_862;  __rev2_862 = __builtin_shufflevector(__s2_862, __s2_862, 1, 0); \
  67820.   __ret_862 = __rev0_862 - __noswap_vmull_u32(__rev1_862, __noswap_splat_lane_u32(__rev2_862, __p3_862)); \
  67821.   __ret_862 = __builtin_shufflevector(__ret_862, __ret_862, 1, 0); \
  67822.   __ret_862; \
  67823. })
  67824. #endif
  67825.  
  67826. #ifdef __LITTLE_ENDIAN__
  67827. #define vmlsl_lane_u16(__p0_863, __p1_863, __p2_863, __p3_863) __extension__ ({ \
  67828.   uint32x4_t __ret_863; \
  67829.   uint32x4_t __s0_863 = __p0_863; \
  67830.   uint16x4_t __s1_863 = __p1_863; \
  67831.   uint16x4_t __s2_863 = __p2_863; \
  67832.   __ret_863 = __s0_863 - vmull_u16(__s1_863, splat_lane_u16(__s2_863, __p3_863)); \
  67833.   __ret_863; \
  67834. })
  67835. #else
  67836. #define vmlsl_lane_u16(__p0_864, __p1_864, __p2_864, __p3_864) __extension__ ({ \
  67837.   uint32x4_t __ret_864; \
  67838.   uint32x4_t __s0_864 = __p0_864; \
  67839.   uint16x4_t __s1_864 = __p1_864; \
  67840.   uint16x4_t __s2_864 = __p2_864; \
  67841.   uint32x4_t __rev0_864;  __rev0_864 = __builtin_shufflevector(__s0_864, __s0_864, 3, 2, 1, 0); \
  67842.   uint16x4_t __rev1_864;  __rev1_864 = __builtin_shufflevector(__s1_864, __s1_864, 3, 2, 1, 0); \
  67843.   uint16x4_t __rev2_864;  __rev2_864 = __builtin_shufflevector(__s2_864, __s2_864, 3, 2, 1, 0); \
  67844.   __ret_864 = __rev0_864 - __noswap_vmull_u16(__rev1_864, __noswap_splat_lane_u16(__rev2_864, __p3_864)); \
  67845.   __ret_864 = __builtin_shufflevector(__ret_864, __ret_864, 3, 2, 1, 0); \
  67846.   __ret_864; \
  67847. })
  67848. #endif
  67849.  
  67850. #ifdef __LITTLE_ENDIAN__
  67851. #define vmlsl_lane_s32(__p0_865, __p1_865, __p2_865, __p3_865) __extension__ ({ \
  67852.   int64x2_t __ret_865; \
  67853.   int64x2_t __s0_865 = __p0_865; \
  67854.   int32x2_t __s1_865 = __p1_865; \
  67855.   int32x2_t __s2_865 = __p2_865; \
  67856.   __ret_865 = __s0_865 - vmull_s32(__s1_865, splat_lane_s32(__s2_865, __p3_865)); \
  67857.   __ret_865; \
  67858. })
  67859. #else
  67860. #define vmlsl_lane_s32(__p0_866, __p1_866, __p2_866, __p3_866) __extension__ ({ \
  67861.   int64x2_t __ret_866; \
  67862.   int64x2_t __s0_866 = __p0_866; \
  67863.   int32x2_t __s1_866 = __p1_866; \
  67864.   int32x2_t __s2_866 = __p2_866; \
  67865.   int64x2_t __rev0_866;  __rev0_866 = __builtin_shufflevector(__s0_866, __s0_866, 1, 0); \
  67866.   int32x2_t __rev1_866;  __rev1_866 = __builtin_shufflevector(__s1_866, __s1_866, 1, 0); \
  67867.   int32x2_t __rev2_866;  __rev2_866 = __builtin_shufflevector(__s2_866, __s2_866, 1, 0); \
  67868.   __ret_866 = __rev0_866 - __noswap_vmull_s32(__rev1_866, __noswap_splat_lane_s32(__rev2_866, __p3_866)); \
  67869.   __ret_866 = __builtin_shufflevector(__ret_866, __ret_866, 1, 0); \
  67870.   __ret_866; \
  67871. })
  67872. #endif
  67873.  
  67874. #ifdef __LITTLE_ENDIAN__
  67875. #define vmlsl_lane_s16(__p0_867, __p1_867, __p2_867, __p3_867) __extension__ ({ \
  67876.   int32x4_t __ret_867; \
  67877.   int32x4_t __s0_867 = __p0_867; \
  67878.   int16x4_t __s1_867 = __p1_867; \
  67879.   int16x4_t __s2_867 = __p2_867; \
  67880.   __ret_867 = __s0_867 - vmull_s16(__s1_867, splat_lane_s16(__s2_867, __p3_867)); \
  67881.   __ret_867; \
  67882. })
  67883. #else
  67884. #define vmlsl_lane_s16(__p0_868, __p1_868, __p2_868, __p3_868) __extension__ ({ \
  67885.   int32x4_t __ret_868; \
  67886.   int32x4_t __s0_868 = __p0_868; \
  67887.   int16x4_t __s1_868 = __p1_868; \
  67888.   int16x4_t __s2_868 = __p2_868; \
  67889.   int32x4_t __rev0_868;  __rev0_868 = __builtin_shufflevector(__s0_868, __s0_868, 3, 2, 1, 0); \
  67890.   int16x4_t __rev1_868;  __rev1_868 = __builtin_shufflevector(__s1_868, __s1_868, 3, 2, 1, 0); \
  67891.   int16x4_t __rev2_868;  __rev2_868 = __builtin_shufflevector(__s2_868, __s2_868, 3, 2, 1, 0); \
  67892.   __ret_868 = __rev0_868 - __noswap_vmull_s16(__rev1_868, __noswap_splat_lane_s16(__rev2_868, __p3_868)); \
  67893.   __ret_868 = __builtin_shufflevector(__ret_868, __ret_868, 3, 2, 1, 0); \
  67894.   __ret_868; \
  67895. })
  67896. #endif
  67897.  
  67898. #ifdef __LITTLE_ENDIAN__
  67899. __ai uint64x2_t vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  67900.   uint64x2_t __ret;
  67901.   __ret = __p0 - vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
  67902.   return __ret;
  67903. }
  67904. #else
  67905. __ai uint64x2_t vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  67906.   uint64x2_t __ret;
  67907.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67908.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67909.   __ret = __rev0 - __noswap_vmull_u32(__rev1, (uint32x2_t) {__p2, __p2});
  67910.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67911.   return __ret;
  67912. }
  67913. __ai uint64x2_t __noswap_vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
  67914.   uint64x2_t __ret;
  67915.   __ret = __p0 - __noswap_vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
  67916.   return __ret;
  67917. }
  67918. #endif
  67919.  
  67920. #ifdef __LITTLE_ENDIAN__
  67921. __ai uint32x4_t vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  67922.   uint32x4_t __ret;
  67923.   __ret = __p0 - vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
  67924.   return __ret;
  67925. }
  67926. #else
  67927. __ai uint32x4_t vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  67928.   uint32x4_t __ret;
  67929.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67930.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67931.   __ret = __rev0 - __noswap_vmull_u16(__rev1, (uint16x4_t) {__p2, __p2, __p2, __p2});
  67932.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67933.   return __ret;
  67934. }
  67935. __ai uint32x4_t __noswap_vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
  67936.   uint32x4_t __ret;
  67937.   __ret = __p0 - __noswap_vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
  67938.   return __ret;
  67939. }
  67940. #endif
  67941.  
  67942. #ifdef __LITTLE_ENDIAN__
  67943. __ai int64x2_t vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  67944.   int64x2_t __ret;
  67945.   __ret = __p0 - vmull_s32(__p1, (int32x2_t) {__p2, __p2});
  67946.   return __ret;
  67947. }
  67948. #else
  67949. __ai int64x2_t vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  67950.   int64x2_t __ret;
  67951.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  67952.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  67953.   __ret = __rev0 - __noswap_vmull_s32(__rev1, (int32x2_t) {__p2, __p2});
  67954.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  67955.   return __ret;
  67956. }
  67957. __ai int64x2_t __noswap_vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
  67958.   int64x2_t __ret;
  67959.   __ret = __p0 - __noswap_vmull_s32(__p1, (int32x2_t) {__p2, __p2});
  67960.   return __ret;
  67961. }
  67962. #endif
  67963.  
  67964. #ifdef __LITTLE_ENDIAN__
  67965. __ai int32x4_t vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  67966.   int32x4_t __ret;
  67967.   __ret = __p0 - vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
  67968.   return __ret;
  67969. }
  67970. #else
  67971. __ai int32x4_t vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  67972.   int32x4_t __ret;
  67973.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  67974.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  67975.   __ret = __rev0 - __noswap_vmull_s16(__rev1, (int16x4_t) {__p2, __p2, __p2, __p2});
  67976.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  67977.   return __ret;
  67978. }
  67979. __ai int32x4_t __noswap_vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
  67980.   int32x4_t __ret;
  67981.   __ret = __p0 - __noswap_vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
  67982.   return __ret;
  67983. }
  67984. #endif
  67985.  
  67986. #ifdef __LITTLE_ENDIAN__
  67987. #define vset_lane_f16(__p0_869, __p1_869, __p2_869) __extension__ ({ \
  67988.   float16x4_t __ret_869; \
  67989.   float16_t __s0_869 = __p0_869; \
  67990.   float16x4_t __s1_869 = __p1_869; \
  67991. float16_t __reint_869 = __s0_869; \
  67992. float16x4_t __reint1_869 = __s1_869; \
  67993. int16x4_t __reint2_869 = vset_lane_s16(*(int16_t *) &__reint_869, *(int16x4_t *) &__reint1_869, __p2_869); \
  67994.   __ret_869 = *(float16x4_t *) &__reint2_869; \
  67995.   __ret_869; \
  67996. })
  67997. #else
  67998. #define vset_lane_f16(__p0_870, __p1_870, __p2_870) __extension__ ({ \
  67999.   float16x4_t __ret_870; \
  68000.   float16_t __s0_870 = __p0_870; \
  68001.   float16x4_t __s1_870 = __p1_870; \
  68002.   float16x4_t __rev1_870;  __rev1_870 = __builtin_shufflevector(__s1_870, __s1_870, 3, 2, 1, 0); \
  68003. float16_t __reint_870 = __s0_870; \
  68004. float16x4_t __reint1_870 = __rev1_870; \
  68005. int16x4_t __reint2_870 = __noswap_vset_lane_s16(*(int16_t *) &__reint_870, *(int16x4_t *) &__reint1_870, __p2_870); \
  68006.   __ret_870 = *(float16x4_t *) &__reint2_870; \
  68007.   __ret_870 = __builtin_shufflevector(__ret_870, __ret_870, 3, 2, 1, 0); \
  68008.   __ret_870; \
  68009. })
  68010. #endif
  68011.  
  68012. #ifdef __LITTLE_ENDIAN__
  68013. #define vsetq_lane_f16(__p0_871, __p1_871, __p2_871) __extension__ ({ \
  68014.   float16x8_t __ret_871; \
  68015.   float16_t __s0_871 = __p0_871; \
  68016.   float16x8_t __s1_871 = __p1_871; \
  68017. float16_t __reint_871 = __s0_871; \
  68018. float16x8_t __reint1_871 = __s1_871; \
  68019. int16x8_t __reint2_871 = vsetq_lane_s16(*(int16_t *) &__reint_871, *(int16x8_t *) &__reint1_871, __p2_871); \
  68020.   __ret_871 = *(float16x8_t *) &__reint2_871; \
  68021.   __ret_871; \
  68022. })
  68023. #else
  68024. #define vsetq_lane_f16(__p0_872, __p1_872, __p2_872) __extension__ ({ \
  68025.   float16x8_t __ret_872; \
  68026.   float16_t __s0_872 = __p0_872; \
  68027.   float16x8_t __s1_872 = __p1_872; \
  68028.   float16x8_t __rev1_872;  __rev1_872 = __builtin_shufflevector(__s1_872, __s1_872, 7, 6, 5, 4, 3, 2, 1, 0); \
  68029. float16_t __reint_872 = __s0_872; \
  68030. float16x8_t __reint1_872 = __rev1_872; \
  68031. int16x8_t __reint2_872 = __noswap_vsetq_lane_s16(*(int16_t *) &__reint_872, *(int16x8_t *) &__reint1_872, __p2_872); \
  68032.   __ret_872 = *(float16x8_t *) &__reint2_872; \
  68033.   __ret_872 = __builtin_shufflevector(__ret_872, __ret_872, 7, 6, 5, 4, 3, 2, 1, 0); \
  68034.   __ret_872; \
  68035. })
  68036. #endif
  68037.  
  68038. #ifdef __LITTLE_ENDIAN__
  68039. #define vbfmlalbq_lane_f32(__p0_873, __p1_873, __p2_873, __p3_873) __extension__ ({ \
  68040.   float32x4_t __ret_873; \
  68041.   float32x4_t __s0_873 = __p0_873; \
  68042.   bfloat16x8_t __s1_873 = __p1_873; \
  68043.   bfloat16x4_t __s2_873 = __p2_873; \
  68044.   __ret_873 = vbfmlalbq_f32(__s0_873, __s1_873, (bfloat16x8_t) {vget_lane_bf16(__s2_873, __p3_873), vget_lane_bf16(__s2_873, __p3_873), vget_lane_bf16(__s2_873, __p3_873), vget_lane_bf16(__s2_873, __p3_873), vget_lane_bf16(__s2_873, __p3_873), vget_lane_bf16(__s2_873, __p3_873), vget_lane_bf16(__s2_873, __p3_873), vget_lane_bf16(__s2_873, __p3_873)}); \
  68045.   __ret_873; \
  68046. })
  68047. #else
  68048. #define vbfmlalbq_lane_f32(__p0_874, __p1_874, __p2_874, __p3_874) __extension__ ({ \
  68049.   float32x4_t __ret_874; \
  68050.   float32x4_t __s0_874 = __p0_874; \
  68051.   bfloat16x8_t __s1_874 = __p1_874; \
  68052.   bfloat16x4_t __s2_874 = __p2_874; \
  68053.   float32x4_t __rev0_874;  __rev0_874 = __builtin_shufflevector(__s0_874, __s0_874, 3, 2, 1, 0); \
  68054.   bfloat16x8_t __rev1_874;  __rev1_874 = __builtin_shufflevector(__s1_874, __s1_874, 7, 6, 5, 4, 3, 2, 1, 0); \
  68055.   bfloat16x4_t __rev2_874;  __rev2_874 = __builtin_shufflevector(__s2_874, __s2_874, 3, 2, 1, 0); \
  68056.   __ret_874 = __noswap_vbfmlalbq_f32(__rev0_874, __rev1_874, (bfloat16x8_t) {__noswap_vget_lane_bf16(__rev2_874, __p3_874), __noswap_vget_lane_bf16(__rev2_874, __p3_874), __noswap_vget_lane_bf16(__rev2_874, __p3_874), __noswap_vget_lane_bf16(__rev2_874, __p3_874), __noswap_vget_lane_bf16(__rev2_874, __p3_874), __noswap_vget_lane_bf16(__rev2_874, __p3_874), __noswap_vget_lane_bf16(__rev2_874, __p3_874), __noswap_vget_lane_bf16(__rev2_874, __p3_874)}); \
  68057.   __ret_874 = __builtin_shufflevector(__ret_874, __ret_874, 3, 2, 1, 0); \
  68058.   __ret_874; \
  68059. })
  68060. #endif
  68061.  
  68062. #ifdef __LITTLE_ENDIAN__
  68063. #define vbfmlalbq_laneq_f32(__p0_875, __p1_875, __p2_875, __p3_875) __extension__ ({ \
  68064.   float32x4_t __ret_875; \
  68065.   float32x4_t __s0_875 = __p0_875; \
  68066.   bfloat16x8_t __s1_875 = __p1_875; \
  68067.   bfloat16x8_t __s2_875 = __p2_875; \
  68068.   __ret_875 = vbfmlalbq_f32(__s0_875, __s1_875, (bfloat16x8_t) {vgetq_lane_bf16(__s2_875, __p3_875), vgetq_lane_bf16(__s2_875, __p3_875), vgetq_lane_bf16(__s2_875, __p3_875), vgetq_lane_bf16(__s2_875, __p3_875), vgetq_lane_bf16(__s2_875, __p3_875), vgetq_lane_bf16(__s2_875, __p3_875), vgetq_lane_bf16(__s2_875, __p3_875), vgetq_lane_bf16(__s2_875, __p3_875)}); \
  68069.   __ret_875; \
  68070. })
  68071. #else
  68072. #define vbfmlalbq_laneq_f32(__p0_876, __p1_876, __p2_876, __p3_876) __extension__ ({ \
  68073.   float32x4_t __ret_876; \
  68074.   float32x4_t __s0_876 = __p0_876; \
  68075.   bfloat16x8_t __s1_876 = __p1_876; \
  68076.   bfloat16x8_t __s2_876 = __p2_876; \
  68077.   float32x4_t __rev0_876;  __rev0_876 = __builtin_shufflevector(__s0_876, __s0_876, 3, 2, 1, 0); \
  68078.   bfloat16x8_t __rev1_876;  __rev1_876 = __builtin_shufflevector(__s1_876, __s1_876, 7, 6, 5, 4, 3, 2, 1, 0); \
  68079.   bfloat16x8_t __rev2_876;  __rev2_876 = __builtin_shufflevector(__s2_876, __s2_876, 7, 6, 5, 4, 3, 2, 1, 0); \
  68080.   __ret_876 = __noswap_vbfmlalbq_f32(__rev0_876, __rev1_876, (bfloat16x8_t) {__noswap_vgetq_lane_bf16(__rev2_876, __p3_876), __noswap_vgetq_lane_bf16(__rev2_876, __p3_876), __noswap_vgetq_lane_bf16(__rev2_876, __p3_876), __noswap_vgetq_lane_bf16(__rev2_876, __p3_876), __noswap_vgetq_lane_bf16(__rev2_876, __p3_876), __noswap_vgetq_lane_bf16(__rev2_876, __p3_876), __noswap_vgetq_lane_bf16(__rev2_876, __p3_876), __noswap_vgetq_lane_bf16(__rev2_876, __p3_876)}); \
  68081.   __ret_876 = __builtin_shufflevector(__ret_876, __ret_876, 3, 2, 1, 0); \
  68082.   __ret_876; \
  68083. })
  68084. #endif
  68085.  
  68086. #ifdef __LITTLE_ENDIAN__
  68087. #define vbfmlaltq_lane_f32(__p0_877, __p1_877, __p2_877, __p3_877) __extension__ ({ \
  68088.   float32x4_t __ret_877; \
  68089.   float32x4_t __s0_877 = __p0_877; \
  68090.   bfloat16x8_t __s1_877 = __p1_877; \
  68091.   bfloat16x4_t __s2_877 = __p2_877; \
  68092.   __ret_877 = vbfmlaltq_f32(__s0_877, __s1_877, (bfloat16x8_t) {vget_lane_bf16(__s2_877, __p3_877), vget_lane_bf16(__s2_877, __p3_877), vget_lane_bf16(__s2_877, __p3_877), vget_lane_bf16(__s2_877, __p3_877), vget_lane_bf16(__s2_877, __p3_877), vget_lane_bf16(__s2_877, __p3_877), vget_lane_bf16(__s2_877, __p3_877), vget_lane_bf16(__s2_877, __p3_877)}); \
  68093.   __ret_877; \
  68094. })
  68095. #else
  68096. #define vbfmlaltq_lane_f32(__p0_878, __p1_878, __p2_878, __p3_878) __extension__ ({ \
  68097.   float32x4_t __ret_878; \
  68098.   float32x4_t __s0_878 = __p0_878; \
  68099.   bfloat16x8_t __s1_878 = __p1_878; \
  68100.   bfloat16x4_t __s2_878 = __p2_878; \
  68101.   float32x4_t __rev0_878;  __rev0_878 = __builtin_shufflevector(__s0_878, __s0_878, 3, 2, 1, 0); \
  68102.   bfloat16x8_t __rev1_878;  __rev1_878 = __builtin_shufflevector(__s1_878, __s1_878, 7, 6, 5, 4, 3, 2, 1, 0); \
  68103.   bfloat16x4_t __rev2_878;  __rev2_878 = __builtin_shufflevector(__s2_878, __s2_878, 3, 2, 1, 0); \
  68104.   __ret_878 = __noswap_vbfmlaltq_f32(__rev0_878, __rev1_878, (bfloat16x8_t) {__noswap_vget_lane_bf16(__rev2_878, __p3_878), __noswap_vget_lane_bf16(__rev2_878, __p3_878), __noswap_vget_lane_bf16(__rev2_878, __p3_878), __noswap_vget_lane_bf16(__rev2_878, __p3_878), __noswap_vget_lane_bf16(__rev2_878, __p3_878), __noswap_vget_lane_bf16(__rev2_878, __p3_878), __noswap_vget_lane_bf16(__rev2_878, __p3_878), __noswap_vget_lane_bf16(__rev2_878, __p3_878)}); \
  68105.   __ret_878 = __builtin_shufflevector(__ret_878, __ret_878, 3, 2, 1, 0); \
  68106.   __ret_878; \
  68107. })
  68108. #endif
  68109.  
  68110. #ifdef __LITTLE_ENDIAN__
  68111. #define vbfmlaltq_laneq_f32(__p0_879, __p1_879, __p2_879, __p3_879) __extension__ ({ \
  68112.   float32x4_t __ret_879; \
  68113.   float32x4_t __s0_879 = __p0_879; \
  68114.   bfloat16x8_t __s1_879 = __p1_879; \
  68115.   bfloat16x8_t __s2_879 = __p2_879; \
  68116.   __ret_879 = vbfmlaltq_f32(__s0_879, __s1_879, (bfloat16x8_t) {vgetq_lane_bf16(__s2_879, __p3_879), vgetq_lane_bf16(__s2_879, __p3_879), vgetq_lane_bf16(__s2_879, __p3_879), vgetq_lane_bf16(__s2_879, __p3_879), vgetq_lane_bf16(__s2_879, __p3_879), vgetq_lane_bf16(__s2_879, __p3_879), vgetq_lane_bf16(__s2_879, __p3_879), vgetq_lane_bf16(__s2_879, __p3_879)}); \
  68117.   __ret_879; \
  68118. })
  68119. #else
  68120. #define vbfmlaltq_laneq_f32(__p0_880, __p1_880, __p2_880, __p3_880) __extension__ ({ \
  68121.   float32x4_t __ret_880; \
  68122.   float32x4_t __s0_880 = __p0_880; \
  68123.   bfloat16x8_t __s1_880 = __p1_880; \
  68124.   bfloat16x8_t __s2_880 = __p2_880; \
  68125.   float32x4_t __rev0_880;  __rev0_880 = __builtin_shufflevector(__s0_880, __s0_880, 3, 2, 1, 0); \
  68126.   bfloat16x8_t __rev1_880;  __rev1_880 = __builtin_shufflevector(__s1_880, __s1_880, 7, 6, 5, 4, 3, 2, 1, 0); \
  68127.   bfloat16x8_t __rev2_880;  __rev2_880 = __builtin_shufflevector(__s2_880, __s2_880, 7, 6, 5, 4, 3, 2, 1, 0); \
  68128.   __ret_880 = __noswap_vbfmlaltq_f32(__rev0_880, __rev1_880, (bfloat16x8_t) {__noswap_vgetq_lane_bf16(__rev2_880, __p3_880), __noswap_vgetq_lane_bf16(__rev2_880, __p3_880), __noswap_vgetq_lane_bf16(__rev2_880, __p3_880), __noswap_vgetq_lane_bf16(__rev2_880, __p3_880), __noswap_vgetq_lane_bf16(__rev2_880, __p3_880), __noswap_vgetq_lane_bf16(__rev2_880, __p3_880), __noswap_vgetq_lane_bf16(__rev2_880, __p3_880), __noswap_vgetq_lane_bf16(__rev2_880, __p3_880)}); \
  68129.   __ret_880 = __builtin_shufflevector(__ret_880, __ret_880, 3, 2, 1, 0); \
  68130.   __ret_880; \
  68131. })
  68132. #endif
  68133.  
  68134. #ifdef __LITTLE_ENDIAN__
  68135. __ai __attribute__((target("bf16"))) float32x4_t vcvtq_high_f32_bf16(bfloat16x8_t __p0) {
  68136.   float32x4_t __ret;
  68137.   __ret = vcvt_f32_bf16(vget_high_bf16(__p0));
  68138.   return __ret;
  68139. }
  68140. #else
  68141. __ai __attribute__((target("bf16"))) float32x4_t vcvtq_high_f32_bf16(bfloat16x8_t __p0) {
  68142.   float32x4_t __ret;
  68143.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68144.   __ret = __noswap_vcvt_f32_bf16(__noswap_vget_high_bf16(__rev0));
  68145.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68146.   return __ret;
  68147. }
  68148. #endif
  68149.  
  68150. #ifdef __LITTLE_ENDIAN__
  68151. __ai __attribute__((target("bf16"))) float32x4_t vcvtq_low_f32_bf16(bfloat16x8_t __p0) {
  68152.   float32x4_t __ret;
  68153.   __ret = vcvt_f32_bf16(vget_low_bf16(__p0));
  68154.   return __ret;
  68155. }
  68156. #else
  68157. __ai __attribute__((target("bf16"))) float32x4_t vcvtq_low_f32_bf16(bfloat16x8_t __p0) {
  68158.   float32x4_t __ret;
  68159.   bfloat16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68160.   __ret = __noswap_vcvt_f32_bf16(__noswap_vget_low_bf16(__rev0));
  68161.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68162.   return __ret;
  68163. }
  68164. #endif
  68165.  
  68166. #ifdef __LITTLE_ENDIAN__
  68167. #define vsudotq_lane_s32(__p0_881, __p1_881, __p2_881, __p3_881) __extension__ ({ \
  68168.   int32x4_t __ret_881; \
  68169.   int32x4_t __s0_881 = __p0_881; \
  68170.   int8x16_t __s1_881 = __p1_881; \
  68171.   uint8x8_t __s2_881 = __p2_881; \
  68172. uint8x8_t __reint_881 = __s2_881; \
  68173.   __ret_881 = vusdotq_s32(__s0_881, (uint8x16_t)(splatq_lane_s32(*(int32x2_t *) &__reint_881, __p3_881)), __s1_881); \
  68174.   __ret_881; \
  68175. })
  68176. #else
  68177. #define vsudotq_lane_s32(__p0_882, __p1_882, __p2_882, __p3_882) __extension__ ({ \
  68178.   int32x4_t __ret_882; \
  68179.   int32x4_t __s0_882 = __p0_882; \
  68180.   int8x16_t __s1_882 = __p1_882; \
  68181.   uint8x8_t __s2_882 = __p2_882; \
  68182.   int32x4_t __rev0_882;  __rev0_882 = __builtin_shufflevector(__s0_882, __s0_882, 3, 2, 1, 0); \
  68183.   int8x16_t __rev1_882;  __rev1_882 = __builtin_shufflevector(__s1_882, __s1_882, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
  68184.   uint8x8_t __rev2_882;  __rev2_882 = __builtin_shufflevector(__s2_882, __s2_882, 7, 6, 5, 4, 3, 2, 1, 0); \
  68185. uint8x8_t __reint_882 = __rev2_882; \
  68186.   __ret_882 = __noswap_vusdotq_s32(__rev0_882, (uint8x16_t)(__noswap_splatq_lane_s32(*(int32x2_t *) &__reint_882, __p3_882)), __rev1_882); \
  68187.   __ret_882 = __builtin_shufflevector(__ret_882, __ret_882, 3, 2, 1, 0); \
  68188.   __ret_882; \
  68189. })
  68190. #endif
  68191.  
  68192. #ifdef __LITTLE_ENDIAN__
  68193. #define vsudot_lane_s32(__p0_883, __p1_883, __p2_883, __p3_883) __extension__ ({ \
  68194.   int32x2_t __ret_883; \
  68195.   int32x2_t __s0_883 = __p0_883; \
  68196.   int8x8_t __s1_883 = __p1_883; \
  68197.   uint8x8_t __s2_883 = __p2_883; \
  68198. uint8x8_t __reint_883 = __s2_883; \
  68199.   __ret_883 = vusdot_s32(__s0_883, (uint8x8_t)(splat_lane_s32(*(int32x2_t *) &__reint_883, __p3_883)), __s1_883); \
  68200.   __ret_883; \
  68201. })
  68202. #else
  68203. #define vsudot_lane_s32(__p0_884, __p1_884, __p2_884, __p3_884) __extension__ ({ \
  68204.   int32x2_t __ret_884; \
  68205.   int32x2_t __s0_884 = __p0_884; \
  68206.   int8x8_t __s1_884 = __p1_884; \
  68207.   uint8x8_t __s2_884 = __p2_884; \
  68208.   int32x2_t __rev0_884;  __rev0_884 = __builtin_shufflevector(__s0_884, __s0_884, 1, 0); \
  68209.   int8x8_t __rev1_884;  __rev1_884 = __builtin_shufflevector(__s1_884, __s1_884, 7, 6, 5, 4, 3, 2, 1, 0); \
  68210.   uint8x8_t __rev2_884;  __rev2_884 = __builtin_shufflevector(__s2_884, __s2_884, 7, 6, 5, 4, 3, 2, 1, 0); \
  68211. uint8x8_t __reint_884 = __rev2_884; \
  68212.   __ret_884 = __noswap_vusdot_s32(__rev0_884, (uint8x8_t)(__noswap_splat_lane_s32(*(int32x2_t *) &__reint_884, __p3_884)), __rev1_884); \
  68213.   __ret_884 = __builtin_shufflevector(__ret_884, __ret_884, 1, 0); \
  68214.   __ret_884; \
  68215. })
  68216. #endif
  68217.  
  68218. #if defined(__aarch64__)
  68219. #ifdef __LITTLE_ENDIAN__
  68220. __ai uint16x8_t vabdl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
  68221.   uint16x8_t __ret;
  68222.   __ret = vabdl_u8(vget_high_u8(__p0), vget_high_u8(__p1));
  68223.   return __ret;
  68224. }
  68225. #else
  68226. __ai uint16x8_t vabdl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
  68227.   uint16x8_t __ret;
  68228.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68229.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68230.   __ret = __noswap_vabdl_u8(__noswap_vget_high_u8(__rev0), __noswap_vget_high_u8(__rev1));
  68231.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68232.   return __ret;
  68233. }
  68234. #endif
  68235.  
  68236. #ifdef __LITTLE_ENDIAN__
  68237. __ai uint64x2_t vabdl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
  68238.   uint64x2_t __ret;
  68239.   __ret = vabdl_u32(vget_high_u32(__p0), vget_high_u32(__p1));
  68240.   return __ret;
  68241. }
  68242. #else
  68243. __ai uint64x2_t vabdl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
  68244.   uint64x2_t __ret;
  68245.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68246.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68247.   __ret = __noswap_vabdl_u32(__noswap_vget_high_u32(__rev0), __noswap_vget_high_u32(__rev1));
  68248.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68249.   return __ret;
  68250. }
  68251. #endif
  68252.  
  68253. #ifdef __LITTLE_ENDIAN__
  68254. __ai uint32x4_t vabdl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
  68255.   uint32x4_t __ret;
  68256.   __ret = vabdl_u16(vget_high_u16(__p0), vget_high_u16(__p1));
  68257.   return __ret;
  68258. }
  68259. #else
  68260. __ai uint32x4_t vabdl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
  68261.   uint32x4_t __ret;
  68262.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68263.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68264.   __ret = __noswap_vabdl_u16(__noswap_vget_high_u16(__rev0), __noswap_vget_high_u16(__rev1));
  68265.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68266.   return __ret;
  68267. }
  68268. #endif
  68269.  
  68270. #ifdef __LITTLE_ENDIAN__
  68271. __ai int16x8_t vabdl_high_s8(int8x16_t __p0, int8x16_t __p1) {
  68272.   int16x8_t __ret;
  68273.   __ret = vabdl_s8(vget_high_s8(__p0), vget_high_s8(__p1));
  68274.   return __ret;
  68275. }
  68276. #else
  68277. __ai int16x8_t vabdl_high_s8(int8x16_t __p0, int8x16_t __p1) {
  68278.   int16x8_t __ret;
  68279.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68280.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68281.   __ret = __noswap_vabdl_s8(__noswap_vget_high_s8(__rev0), __noswap_vget_high_s8(__rev1));
  68282.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68283.   return __ret;
  68284. }
  68285. #endif
  68286.  
  68287. #ifdef __LITTLE_ENDIAN__
  68288. __ai int64x2_t vabdl_high_s32(int32x4_t __p0, int32x4_t __p1) {
  68289.   int64x2_t __ret;
  68290.   __ret = vabdl_s32(vget_high_s32(__p0), vget_high_s32(__p1));
  68291.   return __ret;
  68292. }
  68293. #else
  68294. __ai int64x2_t vabdl_high_s32(int32x4_t __p0, int32x4_t __p1) {
  68295.   int64x2_t __ret;
  68296.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68297.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68298.   __ret = __noswap_vabdl_s32(__noswap_vget_high_s32(__rev0), __noswap_vget_high_s32(__rev1));
  68299.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68300.   return __ret;
  68301. }
  68302. #endif
  68303.  
  68304. #ifdef __LITTLE_ENDIAN__
  68305. __ai int32x4_t vabdl_high_s16(int16x8_t __p0, int16x8_t __p1) {
  68306.   int32x4_t __ret;
  68307.   __ret = vabdl_s16(vget_high_s16(__p0), vget_high_s16(__p1));
  68308.   return __ret;
  68309. }
  68310. #else
  68311. __ai int32x4_t vabdl_high_s16(int16x8_t __p0, int16x8_t __p1) {
  68312.   int32x4_t __ret;
  68313.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68314.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68315.   __ret = __noswap_vabdl_s16(__noswap_vget_high_s16(__rev0), __noswap_vget_high_s16(__rev1));
  68316.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68317.   return __ret;
  68318. }
  68319. #endif
  68320.  
  68321. #ifdef __LITTLE_ENDIAN__
  68322. __ai uint16x8_t vaddl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
  68323.   uint16x8_t __ret;
  68324.   __ret = vmovl_high_u8(__p0) + vmovl_high_u8(__p1);
  68325.   return __ret;
  68326. }
  68327. #else
  68328. __ai uint16x8_t vaddl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
  68329.   uint16x8_t __ret;
  68330.   uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68331.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68332.   __ret = __noswap_vmovl_high_u8(__rev0) + __noswap_vmovl_high_u8(__rev1);
  68333.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68334.   return __ret;
  68335. }
  68336. #endif
  68337.  
  68338. #ifdef __LITTLE_ENDIAN__
  68339. __ai uint64x2_t vaddl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
  68340.   uint64x2_t __ret;
  68341.   __ret = vmovl_high_u32(__p0) + vmovl_high_u32(__p1);
  68342.   return __ret;
  68343. }
  68344. #else
  68345. __ai uint64x2_t vaddl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
  68346.   uint64x2_t __ret;
  68347.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68348.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68349.   __ret = __noswap_vmovl_high_u32(__rev0) + __noswap_vmovl_high_u32(__rev1);
  68350.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68351.   return __ret;
  68352. }
  68353. #endif
  68354.  
  68355. #ifdef __LITTLE_ENDIAN__
  68356. __ai uint32x4_t vaddl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
  68357.   uint32x4_t __ret;
  68358.   __ret = vmovl_high_u16(__p0) + vmovl_high_u16(__p1);
  68359.   return __ret;
  68360. }
  68361. #else
  68362. __ai uint32x4_t vaddl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
  68363.   uint32x4_t __ret;
  68364.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68365.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68366.   __ret = __noswap_vmovl_high_u16(__rev0) + __noswap_vmovl_high_u16(__rev1);
  68367.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68368.   return __ret;
  68369. }
  68370. #endif
  68371.  
  68372. #ifdef __LITTLE_ENDIAN__
  68373. __ai int16x8_t vaddl_high_s8(int8x16_t __p0, int8x16_t __p1) {
  68374.   int16x8_t __ret;
  68375.   __ret = vmovl_high_s8(__p0) + vmovl_high_s8(__p1);
  68376.   return __ret;
  68377. }
  68378. #else
  68379. __ai int16x8_t vaddl_high_s8(int8x16_t __p0, int8x16_t __p1) {
  68380.   int16x8_t __ret;
  68381.   int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68382.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68383.   __ret = __noswap_vmovl_high_s8(__rev0) + __noswap_vmovl_high_s8(__rev1);
  68384.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68385.   return __ret;
  68386. }
  68387. #endif
  68388.  
  68389. #ifdef __LITTLE_ENDIAN__
  68390. __ai int64x2_t vaddl_high_s32(int32x4_t __p0, int32x4_t __p1) {
  68391.   int64x2_t __ret;
  68392.   __ret = vmovl_high_s32(__p0) + vmovl_high_s32(__p1);
  68393.   return __ret;
  68394. }
  68395. #else
  68396. __ai int64x2_t vaddl_high_s32(int32x4_t __p0, int32x4_t __p1) {
  68397.   int64x2_t __ret;
  68398.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68399.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68400.   __ret = __noswap_vmovl_high_s32(__rev0) + __noswap_vmovl_high_s32(__rev1);
  68401.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68402.   return __ret;
  68403. }
  68404. #endif
  68405.  
  68406. #ifdef __LITTLE_ENDIAN__
  68407. __ai int32x4_t vaddl_high_s16(int16x8_t __p0, int16x8_t __p1) {
  68408.   int32x4_t __ret;
  68409.   __ret = vmovl_high_s16(__p0) + vmovl_high_s16(__p1);
  68410.   return __ret;
  68411. }
  68412. #else
  68413. __ai int32x4_t vaddl_high_s16(int16x8_t __p0, int16x8_t __p1) {
  68414.   int32x4_t __ret;
  68415.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68416.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68417.   __ret = __noswap_vmovl_high_s16(__rev0) + __noswap_vmovl_high_s16(__rev1);
  68418.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68419.   return __ret;
  68420. }
  68421. #endif
  68422.  
  68423. #ifdef __LITTLE_ENDIAN__
  68424. __ai uint16x8_t vaddw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
  68425.   uint16x8_t __ret;
  68426.   __ret = __p0 + vmovl_high_u8(__p1);
  68427.   return __ret;
  68428. }
  68429. #else
  68430. __ai uint16x8_t vaddw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
  68431.   uint16x8_t __ret;
  68432.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68433.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68434.   __ret = __rev0 + __noswap_vmovl_high_u8(__rev1);
  68435.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68436.   return __ret;
  68437. }
  68438. #endif
  68439.  
  68440. #ifdef __LITTLE_ENDIAN__
  68441. __ai uint64x2_t vaddw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
  68442.   uint64x2_t __ret;
  68443.   __ret = __p0 + vmovl_high_u32(__p1);
  68444.   return __ret;
  68445. }
  68446. #else
  68447. __ai uint64x2_t vaddw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
  68448.   uint64x2_t __ret;
  68449.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68450.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68451.   __ret = __rev0 + __noswap_vmovl_high_u32(__rev1);
  68452.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68453.   return __ret;
  68454. }
  68455. #endif
  68456.  
  68457. #ifdef __LITTLE_ENDIAN__
  68458. __ai uint32x4_t vaddw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
  68459.   uint32x4_t __ret;
  68460.   __ret = __p0 + vmovl_high_u16(__p1);
  68461.   return __ret;
  68462. }
  68463. #else
  68464. __ai uint32x4_t vaddw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
  68465.   uint32x4_t __ret;
  68466.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68467.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68468.   __ret = __rev0 + __noswap_vmovl_high_u16(__rev1);
  68469.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68470.   return __ret;
  68471. }
  68472. #endif
  68473.  
  68474. #ifdef __LITTLE_ENDIAN__
  68475. __ai int16x8_t vaddw_high_s8(int16x8_t __p0, int8x16_t __p1) {
  68476.   int16x8_t __ret;
  68477.   __ret = __p0 + vmovl_high_s8(__p1);
  68478.   return __ret;
  68479. }
  68480. #else
  68481. __ai int16x8_t vaddw_high_s8(int16x8_t __p0, int8x16_t __p1) {
  68482.   int16x8_t __ret;
  68483.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68484.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68485.   __ret = __rev0 + __noswap_vmovl_high_s8(__rev1);
  68486.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68487.   return __ret;
  68488. }
  68489. #endif
  68490.  
  68491. #ifdef __LITTLE_ENDIAN__
  68492. __ai int64x2_t vaddw_high_s32(int64x2_t __p0, int32x4_t __p1) {
  68493.   int64x2_t __ret;
  68494.   __ret = __p0 + vmovl_high_s32(__p1);
  68495.   return __ret;
  68496. }
  68497. #else
  68498. __ai int64x2_t vaddw_high_s32(int64x2_t __p0, int32x4_t __p1) {
  68499.   int64x2_t __ret;
  68500.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68501.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68502.   __ret = __rev0 + __noswap_vmovl_high_s32(__rev1);
  68503.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68504.   return __ret;
  68505. }
  68506. #endif
  68507.  
  68508. #ifdef __LITTLE_ENDIAN__
  68509. __ai int32x4_t vaddw_high_s16(int32x4_t __p0, int16x8_t __p1) {
  68510.   int32x4_t __ret;
  68511.   __ret = __p0 + vmovl_high_s16(__p1);
  68512.   return __ret;
  68513. }
  68514. #else
  68515. __ai int32x4_t vaddw_high_s16(int32x4_t __p0, int16x8_t __p1) {
  68516.   int32x4_t __ret;
  68517.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68518.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68519.   __ret = __rev0 + __noswap_vmovl_high_s16(__rev1);
  68520.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68521.   return __ret;
  68522. }
  68523. #endif
  68524.  
  68525. #ifdef __LITTLE_ENDIAN__
  68526. #define vcopyq_lane_p64(__p0_885, __p1_885, __p2_885, __p3_885) __extension__ ({ \
  68527.   poly64x2_t __ret_885; \
  68528.   poly64x2_t __s0_885 = __p0_885; \
  68529.   poly64x1_t __s2_885 = __p2_885; \
  68530.   __ret_885 = vsetq_lane_p64(vget_lane_p64(__s2_885, __p3_885), __s0_885, __p1_885); \
  68531.   __ret_885; \
  68532. })
  68533. #else
  68534. #define vcopyq_lane_p64(__p0_886, __p1_886, __p2_886, __p3_886) __extension__ ({ \
  68535.   poly64x2_t __ret_886; \
  68536.   poly64x2_t __s0_886 = __p0_886; \
  68537.   poly64x1_t __s2_886 = __p2_886; \
  68538.   poly64x2_t __rev0_886;  __rev0_886 = __builtin_shufflevector(__s0_886, __s0_886, 1, 0); \
  68539.   __ret_886 = __noswap_vsetq_lane_p64(vget_lane_p64(__s2_886, __p3_886), __rev0_886, __p1_886); \
  68540.   __ret_886 = __builtin_shufflevector(__ret_886, __ret_886, 1, 0); \
  68541.   __ret_886; \
  68542. })
  68543. #endif
  68544.  
  68545. #ifdef __LITTLE_ENDIAN__
  68546. #define vcopyq_lane_f64(__p0_887, __p1_887, __p2_887, __p3_887) __extension__ ({ \
  68547.   float64x2_t __ret_887; \
  68548.   float64x2_t __s0_887 = __p0_887; \
  68549.   float64x1_t __s2_887 = __p2_887; \
  68550.   __ret_887 = vsetq_lane_f64(vget_lane_f64(__s2_887, __p3_887), __s0_887, __p1_887); \
  68551.   __ret_887; \
  68552. })
  68553. #else
  68554. #define vcopyq_lane_f64(__p0_888, __p1_888, __p2_888, __p3_888) __extension__ ({ \
  68555.   float64x2_t __ret_888; \
  68556.   float64x2_t __s0_888 = __p0_888; \
  68557.   float64x1_t __s2_888 = __p2_888; \
  68558.   float64x2_t __rev0_888;  __rev0_888 = __builtin_shufflevector(__s0_888, __s0_888, 1, 0); \
  68559.   __ret_888 = __noswap_vsetq_lane_f64(vget_lane_f64(__s2_888, __p3_888), __rev0_888, __p1_888); \
  68560.   __ret_888 = __builtin_shufflevector(__ret_888, __ret_888, 1, 0); \
  68561.   __ret_888; \
  68562. })
  68563. #endif
  68564.  
  68565. #define vcopy_lane_p64(__p0_889, __p1_889, __p2_889, __p3_889) __extension__ ({ \
  68566.   poly64x1_t __ret_889; \
  68567.   poly64x1_t __s0_889 = __p0_889; \
  68568.   poly64x1_t __s2_889 = __p2_889; \
  68569.   __ret_889 = vset_lane_p64(vget_lane_p64(__s2_889, __p3_889), __s0_889, __p1_889); \
  68570.   __ret_889; \
  68571. })
  68572. #define vcopy_lane_f64(__p0_890, __p1_890, __p2_890, __p3_890) __extension__ ({ \
  68573.   float64x1_t __ret_890; \
  68574.   float64x1_t __s0_890 = __p0_890; \
  68575.   float64x1_t __s2_890 = __p2_890; \
  68576.   __ret_890 = vset_lane_f64(vget_lane_f64(__s2_890, __p3_890), __s0_890, __p1_890); \
  68577.   __ret_890; \
  68578. })
  68579. #ifdef __LITTLE_ENDIAN__
  68580. #define vcopyq_laneq_p64(__p0_891, __p1_891, __p2_891, __p3_891) __extension__ ({ \
  68581.   poly64x2_t __ret_891; \
  68582.   poly64x2_t __s0_891 = __p0_891; \
  68583.   poly64x2_t __s2_891 = __p2_891; \
  68584.   __ret_891 = vsetq_lane_p64(vgetq_lane_p64(__s2_891, __p3_891), __s0_891, __p1_891); \
  68585.   __ret_891; \
  68586. })
  68587. #else
  68588. #define vcopyq_laneq_p64(__p0_892, __p1_892, __p2_892, __p3_892) __extension__ ({ \
  68589.   poly64x2_t __ret_892; \
  68590.   poly64x2_t __s0_892 = __p0_892; \
  68591.   poly64x2_t __s2_892 = __p2_892; \
  68592.   poly64x2_t __rev0_892;  __rev0_892 = __builtin_shufflevector(__s0_892, __s0_892, 1, 0); \
  68593.   poly64x2_t __rev2_892;  __rev2_892 = __builtin_shufflevector(__s2_892, __s2_892, 1, 0); \
  68594.   __ret_892 = __noswap_vsetq_lane_p64(__noswap_vgetq_lane_p64(__rev2_892, __p3_892), __rev0_892, __p1_892); \
  68595.   __ret_892 = __builtin_shufflevector(__ret_892, __ret_892, 1, 0); \
  68596.   __ret_892; \
  68597. })
  68598. #endif
  68599.  
  68600. #ifdef __LITTLE_ENDIAN__
  68601. #define vcopyq_laneq_f64(__p0_893, __p1_893, __p2_893, __p3_893) __extension__ ({ \
  68602.   float64x2_t __ret_893; \
  68603.   float64x2_t __s0_893 = __p0_893; \
  68604.   float64x2_t __s2_893 = __p2_893; \
  68605.   __ret_893 = vsetq_lane_f64(vgetq_lane_f64(__s2_893, __p3_893), __s0_893, __p1_893); \
  68606.   __ret_893; \
  68607. })
  68608. #else
  68609. #define vcopyq_laneq_f64(__p0_894, __p1_894, __p2_894, __p3_894) __extension__ ({ \
  68610.   float64x2_t __ret_894; \
  68611.   float64x2_t __s0_894 = __p0_894; \
  68612.   float64x2_t __s2_894 = __p2_894; \
  68613.   float64x2_t __rev0_894;  __rev0_894 = __builtin_shufflevector(__s0_894, __s0_894, 1, 0); \
  68614.   float64x2_t __rev2_894;  __rev2_894 = __builtin_shufflevector(__s2_894, __s2_894, 1, 0); \
  68615.   __ret_894 = __noswap_vsetq_lane_f64(__noswap_vgetq_lane_f64(__rev2_894, __p3_894), __rev0_894, __p1_894); \
  68616.   __ret_894 = __builtin_shufflevector(__ret_894, __ret_894, 1, 0); \
  68617.   __ret_894; \
  68618. })
  68619. #endif
  68620.  
  68621. #ifdef __LITTLE_ENDIAN__
  68622. #define vcopy_laneq_p64(__p0_895, __p1_895, __p2_895, __p3_895) __extension__ ({ \
  68623.   poly64x1_t __ret_895; \
  68624.   poly64x1_t __s0_895 = __p0_895; \
  68625.   poly64x2_t __s2_895 = __p2_895; \
  68626.   __ret_895 = vset_lane_p64(vgetq_lane_p64(__s2_895, __p3_895), __s0_895, __p1_895); \
  68627.   __ret_895; \
  68628. })
  68629. #else
  68630. #define vcopy_laneq_p64(__p0_896, __p1_896, __p2_896, __p3_896) __extension__ ({ \
  68631.   poly64x1_t __ret_896; \
  68632.   poly64x1_t __s0_896 = __p0_896; \
  68633.   poly64x2_t __s2_896 = __p2_896; \
  68634.   poly64x2_t __rev2_896;  __rev2_896 = __builtin_shufflevector(__s2_896, __s2_896, 1, 0); \
  68635.   __ret_896 = vset_lane_p64(__noswap_vgetq_lane_p64(__rev2_896, __p3_896), __s0_896, __p1_896); \
  68636.   __ret_896; \
  68637. })
  68638. #endif
  68639.  
  68640. #ifdef __LITTLE_ENDIAN__
  68641. #define vcopy_laneq_f64(__p0_897, __p1_897, __p2_897, __p3_897) __extension__ ({ \
  68642.   float64x1_t __ret_897; \
  68643.   float64x1_t __s0_897 = __p0_897; \
  68644.   float64x2_t __s2_897 = __p2_897; \
  68645.   __ret_897 = vset_lane_f64(vgetq_lane_f64(__s2_897, __p3_897), __s0_897, __p1_897); \
  68646.   __ret_897; \
  68647. })
  68648. #else
  68649. #define vcopy_laneq_f64(__p0_898, __p1_898, __p2_898, __p3_898) __extension__ ({ \
  68650.   float64x1_t __ret_898; \
  68651.   float64x1_t __s0_898 = __p0_898; \
  68652.   float64x2_t __s2_898 = __p2_898; \
  68653.   float64x2_t __rev2_898;  __rev2_898 = __builtin_shufflevector(__s2_898, __s2_898, 1, 0); \
  68654.   __ret_898 = vset_lane_f64(__noswap_vgetq_lane_f64(__rev2_898, __p3_898), __s0_898, __p1_898); \
  68655.   __ret_898; \
  68656. })
  68657. #endif
  68658.  
  68659. #ifdef __LITTLE_ENDIAN__
  68660. __ai uint16x8_t vmlal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  68661.   uint16x8_t __ret;
  68662.   __ret = vmlal_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
  68663.   return __ret;
  68664. }
  68665. #else
  68666. __ai uint16x8_t vmlal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  68667.   uint16x8_t __ret;
  68668.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68669.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68670.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68671.   __ret = __noswap_vmlal_u8(__rev0, __noswap_vget_high_u8(__rev1), __noswap_vget_high_u8(__rev2));
  68672.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68673.   return __ret;
  68674. }
  68675. #endif
  68676.  
  68677. #ifdef __LITTLE_ENDIAN__
  68678. __ai uint64x2_t vmlal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  68679.   uint64x2_t __ret;
  68680.   __ret = vmlal_u32(__p0, vget_high_u32(__p1), vget_high_u32(__p2));
  68681.   return __ret;
  68682. }
  68683. #else
  68684. __ai uint64x2_t vmlal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  68685.   uint64x2_t __ret;
  68686.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68687.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68688.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  68689.   __ret = __noswap_vmlal_u32(__rev0, __noswap_vget_high_u32(__rev1), __noswap_vget_high_u32(__rev2));
  68690.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68691.   return __ret;
  68692. }
  68693. #endif
  68694.  
  68695. #ifdef __LITTLE_ENDIAN__
  68696. __ai uint32x4_t vmlal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  68697.   uint32x4_t __ret;
  68698.   __ret = vmlal_u16(__p0, vget_high_u16(__p1), vget_high_u16(__p2));
  68699.   return __ret;
  68700. }
  68701. #else
  68702. __ai uint32x4_t vmlal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  68703.   uint32x4_t __ret;
  68704.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68705.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68706.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  68707.   __ret = __noswap_vmlal_u16(__rev0, __noswap_vget_high_u16(__rev1), __noswap_vget_high_u16(__rev2));
  68708.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68709.   return __ret;
  68710. }
  68711. #endif
  68712.  
  68713. #ifdef __LITTLE_ENDIAN__
  68714. __ai int16x8_t vmlal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
  68715.   int16x8_t __ret;
  68716.   __ret = vmlal_s8(__p0, vget_high_s8(__p1), vget_high_s8(__p2));
  68717.   return __ret;
  68718. }
  68719. #else
  68720. __ai int16x8_t vmlal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
  68721.   int16x8_t __ret;
  68722.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68723.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68724.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68725.   __ret = __noswap_vmlal_s8(__rev0, __noswap_vget_high_s8(__rev1), __noswap_vget_high_s8(__rev2));
  68726.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68727.   return __ret;
  68728. }
  68729. #endif
  68730.  
  68731. #ifdef __LITTLE_ENDIAN__
  68732. __ai int64x2_t vmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  68733.   int64x2_t __ret;
  68734.   __ret = vmlal_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
  68735.   return __ret;
  68736. }
  68737. #else
  68738. __ai int64x2_t vmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  68739.   int64x2_t __ret;
  68740.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68741.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68742.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  68743.   __ret = __noswap_vmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
  68744.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68745.   return __ret;
  68746. }
  68747. #endif
  68748.  
  68749. #ifdef __LITTLE_ENDIAN__
  68750. __ai int32x4_t vmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  68751.   int32x4_t __ret;
  68752.   __ret = vmlal_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
  68753.   return __ret;
  68754. }
  68755. #else
  68756. __ai int32x4_t vmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  68757.   int32x4_t __ret;
  68758.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68759.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68760.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  68761.   __ret = __noswap_vmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
  68762.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68763.   return __ret;
  68764. }
  68765. #endif
  68766.  
  68767. #ifdef __LITTLE_ENDIAN__
  68768. __ai uint64x2_t vmlal_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
  68769.   uint64x2_t __ret;
  68770.   __ret = vmlal_n_u32(__p0, vget_high_u32(__p1), __p2);
  68771.   return __ret;
  68772. }
  68773. #else
  68774. __ai uint64x2_t vmlal_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
  68775.   uint64x2_t __ret;
  68776.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68777.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68778.   __ret = __noswap_vmlal_n_u32(__rev0, __noswap_vget_high_u32(__rev1), __p2);
  68779.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68780.   return __ret;
  68781. }
  68782. #endif
  68783.  
  68784. #ifdef __LITTLE_ENDIAN__
  68785. __ai uint32x4_t vmlal_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
  68786.   uint32x4_t __ret;
  68787.   __ret = vmlal_n_u16(__p0, vget_high_u16(__p1), __p2);
  68788.   return __ret;
  68789. }
  68790. #else
  68791. __ai uint32x4_t vmlal_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
  68792.   uint32x4_t __ret;
  68793.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68794.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68795.   __ret = __noswap_vmlal_n_u16(__rev0, __noswap_vget_high_u16(__rev1), __p2);
  68796.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68797.   return __ret;
  68798. }
  68799. #endif
  68800.  
  68801. #ifdef __LITTLE_ENDIAN__
  68802. __ai int64x2_t vmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
  68803.   int64x2_t __ret;
  68804.   __ret = vmlal_n_s32(__p0, vget_high_s32(__p1), __p2);
  68805.   return __ret;
  68806. }
  68807. #else
  68808. __ai int64x2_t vmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
  68809.   int64x2_t __ret;
  68810.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68811.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68812.   __ret = __noswap_vmlal_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
  68813.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68814.   return __ret;
  68815. }
  68816. #endif
  68817.  
  68818. #ifdef __LITTLE_ENDIAN__
  68819. __ai int32x4_t vmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
  68820.   int32x4_t __ret;
  68821.   __ret = vmlal_n_s16(__p0, vget_high_s16(__p1), __p2);
  68822.   return __ret;
  68823. }
  68824. #else
  68825. __ai int32x4_t vmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
  68826.   int32x4_t __ret;
  68827.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68828.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68829.   __ret = __noswap_vmlal_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
  68830.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68831.   return __ret;
  68832. }
  68833. #endif
  68834.  
  68835. #ifdef __LITTLE_ENDIAN__
  68836. __ai uint16x8_t vmlsl_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  68837.   uint16x8_t __ret;
  68838.   __ret = vmlsl_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
  68839.   return __ret;
  68840. }
  68841. #else
  68842. __ai uint16x8_t vmlsl_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  68843.   uint16x8_t __ret;
  68844.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68845.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68846.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68847.   __ret = __noswap_vmlsl_u8(__rev0, __noswap_vget_high_u8(__rev1), __noswap_vget_high_u8(__rev2));
  68848.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68849.   return __ret;
  68850. }
  68851. #endif
  68852.  
  68853. #ifdef __LITTLE_ENDIAN__
  68854. __ai uint64x2_t vmlsl_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  68855.   uint64x2_t __ret;
  68856.   __ret = vmlsl_u32(__p0, vget_high_u32(__p1), vget_high_u32(__p2));
  68857.   return __ret;
  68858. }
  68859. #else
  68860. __ai uint64x2_t vmlsl_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  68861.   uint64x2_t __ret;
  68862.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68863.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68864.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  68865.   __ret = __noswap_vmlsl_u32(__rev0, __noswap_vget_high_u32(__rev1), __noswap_vget_high_u32(__rev2));
  68866.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68867.   return __ret;
  68868. }
  68869. #endif
  68870.  
  68871. #ifdef __LITTLE_ENDIAN__
  68872. __ai uint32x4_t vmlsl_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  68873.   uint32x4_t __ret;
  68874.   __ret = vmlsl_u16(__p0, vget_high_u16(__p1), vget_high_u16(__p2));
  68875.   return __ret;
  68876. }
  68877. #else
  68878. __ai uint32x4_t vmlsl_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  68879.   uint32x4_t __ret;
  68880.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68881.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68882.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  68883.   __ret = __noswap_vmlsl_u16(__rev0, __noswap_vget_high_u16(__rev1), __noswap_vget_high_u16(__rev2));
  68884.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68885.   return __ret;
  68886. }
  68887. #endif
  68888.  
  68889. #ifdef __LITTLE_ENDIAN__
  68890. __ai int16x8_t vmlsl_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
  68891.   int16x8_t __ret;
  68892.   __ret = vmlsl_s8(__p0, vget_high_s8(__p1), vget_high_s8(__p2));
  68893.   return __ret;
  68894. }
  68895. #else
  68896. __ai int16x8_t vmlsl_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
  68897.   int16x8_t __ret;
  68898.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  68899.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68900.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  68901.   __ret = __noswap_vmlsl_s8(__rev0, __noswap_vget_high_s8(__rev1), __noswap_vget_high_s8(__rev2));
  68902.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  68903.   return __ret;
  68904. }
  68905. #endif
  68906.  
  68907. #ifdef __LITTLE_ENDIAN__
  68908. __ai int64x2_t vmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  68909.   int64x2_t __ret;
  68910.   __ret = vmlsl_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
  68911.   return __ret;
  68912. }
  68913. #else
  68914. __ai int64x2_t vmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  68915.   int64x2_t __ret;
  68916.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68917.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68918.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  68919.   __ret = __noswap_vmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
  68920.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68921.   return __ret;
  68922. }
  68923. #endif
  68924.  
  68925. #ifdef __LITTLE_ENDIAN__
  68926. __ai int32x4_t vmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  68927.   int32x4_t __ret;
  68928.   __ret = vmlsl_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
  68929.   return __ret;
  68930. }
  68931. #else
  68932. __ai int32x4_t vmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  68933.   int32x4_t __ret;
  68934.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68935.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68936.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  68937.   __ret = __noswap_vmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
  68938.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68939.   return __ret;
  68940. }
  68941. #endif
  68942.  
  68943. #ifdef __LITTLE_ENDIAN__
  68944. __ai uint64x2_t vmlsl_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
  68945.   uint64x2_t __ret;
  68946.   __ret = vmlsl_n_u32(__p0, vget_high_u32(__p1), __p2);
  68947.   return __ret;
  68948. }
  68949. #else
  68950. __ai uint64x2_t vmlsl_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
  68951.   uint64x2_t __ret;
  68952.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68953.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68954.   __ret = __noswap_vmlsl_n_u32(__rev0, __noswap_vget_high_u32(__rev1), __p2);
  68955.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68956.   return __ret;
  68957. }
  68958. #endif
  68959.  
  68960. #ifdef __LITTLE_ENDIAN__
  68961. __ai uint32x4_t vmlsl_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
  68962.   uint32x4_t __ret;
  68963.   __ret = vmlsl_n_u16(__p0, vget_high_u16(__p1), __p2);
  68964.   return __ret;
  68965. }
  68966. #else
  68967. __ai uint32x4_t vmlsl_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
  68968.   uint32x4_t __ret;
  68969.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  68970.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  68971.   __ret = __noswap_vmlsl_n_u16(__rev0, __noswap_vget_high_u16(__rev1), __p2);
  68972.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  68973.   return __ret;
  68974. }
  68975. #endif
  68976.  
  68977. #ifdef __LITTLE_ENDIAN__
  68978. __ai int64x2_t vmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
  68979.   int64x2_t __ret;
  68980.   __ret = vmlsl_n_s32(__p0, vget_high_s32(__p1), __p2);
  68981.   return __ret;
  68982. }
  68983. #else
  68984. __ai int64x2_t vmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
  68985.   int64x2_t __ret;
  68986.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  68987.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  68988.   __ret = __noswap_vmlsl_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
  68989.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  68990.   return __ret;
  68991. }
  68992. #endif
  68993.  
  68994. #ifdef __LITTLE_ENDIAN__
  68995. __ai int32x4_t vmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
  68996.   int32x4_t __ret;
  68997.   __ret = vmlsl_n_s16(__p0, vget_high_s16(__p1), __p2);
  68998.   return __ret;
  68999. }
  69000. #else
  69001. __ai int32x4_t vmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
  69002.   int32x4_t __ret;
  69003.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  69004.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  69005.   __ret = __noswap_vmlsl_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
  69006.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  69007.   return __ret;
  69008. }
  69009. #endif
  69010.  
  69011. #define vmulx_lane_f64(__p0_899, __p1_899, __p2_899) __extension__ ({ \
  69012.   float64x1_t __ret_899; \
  69013.   float64x1_t __s0_899 = __p0_899; \
  69014.   float64x1_t __s1_899 = __p1_899; \
  69015.   float64_t __x_899 = vget_lane_f64(__s0_899, 0); \
  69016.   float64_t __y_899 = vget_lane_f64(__s1_899, __p2_899); \
  69017.   float64_t __z_899 = vmulxd_f64(__x_899, __y_899); \
  69018.   __ret_899 = vset_lane_f64(__z_899, __s0_899, __p2_899); \
  69019.   __ret_899; \
  69020. })
  69021. #ifdef __LITTLE_ENDIAN__
  69022. #define vmulx_laneq_f64(__p0_900, __p1_900, __p2_900) __extension__ ({ \
  69023.   float64x1_t __ret_900; \
  69024.   float64x1_t __s0_900 = __p0_900; \
  69025.   float64x2_t __s1_900 = __p1_900; \
  69026.   float64_t __x_900 = vget_lane_f64(__s0_900, 0); \
  69027.   float64_t __y_900 = vgetq_lane_f64(__s1_900, __p2_900); \
  69028.   float64_t __z_900 = vmulxd_f64(__x_900, __y_900); \
  69029.   __ret_900 = vset_lane_f64(__z_900, __s0_900, 0); \
  69030.   __ret_900; \
  69031. })
  69032. #else
  69033. #define vmulx_laneq_f64(__p0_901, __p1_901, __p2_901) __extension__ ({ \
  69034.   float64x1_t __ret_901; \
  69035.   float64x1_t __s0_901 = __p0_901; \
  69036.   float64x2_t __s1_901 = __p1_901; \
  69037.   float64x2_t __rev1_901;  __rev1_901 = __builtin_shufflevector(__s1_901, __s1_901, 1, 0); \
  69038.   float64_t __x_901 = vget_lane_f64(__s0_901, 0); \
  69039.   float64_t __y_901 = __noswap_vgetq_lane_f64(__rev1_901, __p2_901); \
  69040.   float64_t __z_901 = vmulxd_f64(__x_901, __y_901); \
  69041.   __ret_901 = vset_lane_f64(__z_901, __s0_901, 0); \
  69042.   __ret_901; \
  69043. })
  69044. #endif
  69045.  
  69046. #ifdef __LITTLE_ENDIAN__
  69047. #define vfmlalq_lane_high_f16(__p0_902, __p1_902, __p2_902, __p3_902) __extension__ ({ \
  69048.   float32x4_t __ret_902; \
  69049.   float32x4_t __s0_902 = __p0_902; \
  69050.   float16x8_t __s1_902 = __p1_902; \
  69051.   float16x4_t __s2_902 = __p2_902; \
  69052.   __ret_902 = vfmlalq_high_f16(__s0_902, __s1_902, (float16x8_t) {vget_lane_f16(__s2_902, __p3_902), vget_lane_f16(__s2_902, __p3_902), vget_lane_f16(__s2_902, __p3_902), vget_lane_f16(__s2_902, __p3_902), vget_lane_f16(__s2_902, __p3_902), vget_lane_f16(__s2_902, __p3_902), vget_lane_f16(__s2_902, __p3_902), vget_lane_f16(__s2_902, __p3_902)}); \
  69053.   __ret_902; \
  69054. })
  69055. #else
  69056. #define vfmlalq_lane_high_f16(__p0_903, __p1_903, __p2_903, __p3_903) __extension__ ({ \
  69057.   float32x4_t __ret_903; \
  69058.   float32x4_t __s0_903 = __p0_903; \
  69059.   float16x8_t __s1_903 = __p1_903; \
  69060.   float16x4_t __s2_903 = __p2_903; \
  69061.   float32x4_t __rev0_903;  __rev0_903 = __builtin_shufflevector(__s0_903, __s0_903, 3, 2, 1, 0); \
  69062.   float16x8_t __rev1_903;  __rev1_903 = __builtin_shufflevector(__s1_903, __s1_903, 7, 6, 5, 4, 3, 2, 1, 0); \
  69063.   float16x4_t __rev2_903;  __rev2_903 = __builtin_shufflevector(__s2_903, __s2_903, 3, 2, 1, 0); \
  69064.   __ret_903 = __noswap_vfmlalq_high_f16(__rev0_903, __rev1_903, (float16x8_t) {__noswap_vget_lane_f16(__rev2_903, __p3_903), __noswap_vget_lane_f16(__rev2_903, __p3_903), __noswap_vget_lane_f16(__rev2_903, __p3_903), __noswap_vget_lane_f16(__rev2_903, __p3_903), __noswap_vget_lane_f16(__rev2_903, __p3_903), __noswap_vget_lane_f16(__rev2_903, __p3_903), __noswap_vget_lane_f16(__rev2_903, __p3_903), __noswap_vget_lane_f16(__rev2_903, __p3_903)}); \
  69065.   __ret_903 = __builtin_shufflevector(__ret_903, __ret_903, 3, 2, 1, 0); \
  69066.   __ret_903; \
  69067. })
  69068. #endif
  69069.  
  69070. #ifdef __LITTLE_ENDIAN__
  69071. #define vfmlal_lane_high_f16(__p0_904, __p1_904, __p2_904, __p3_904) __extension__ ({ \
  69072.   float32x2_t __ret_904; \
  69073.   float32x2_t __s0_904 = __p0_904; \
  69074.   float16x4_t __s1_904 = __p1_904; \
  69075.   float16x4_t __s2_904 = __p2_904; \
  69076.   __ret_904 = vfmlal_high_f16(__s0_904, __s1_904, (float16x4_t) {vget_lane_f16(__s2_904, __p3_904), vget_lane_f16(__s2_904, __p3_904), vget_lane_f16(__s2_904, __p3_904), vget_lane_f16(__s2_904, __p3_904)}); \
  69077.   __ret_904; \
  69078. })
  69079. #else
  69080. #define vfmlal_lane_high_f16(__p0_905, __p1_905, __p2_905, __p3_905) __extension__ ({ \
  69081.   float32x2_t __ret_905; \
  69082.   float32x2_t __s0_905 = __p0_905; \
  69083.   float16x4_t __s1_905 = __p1_905; \
  69084.   float16x4_t __s2_905 = __p2_905; \
  69085.   float32x2_t __rev0_905;  __rev0_905 = __builtin_shufflevector(__s0_905, __s0_905, 1, 0); \
  69086.   float16x4_t __rev1_905;  __rev1_905 = __builtin_shufflevector(__s1_905, __s1_905, 3, 2, 1, 0); \
  69087.   float16x4_t __rev2_905;  __rev2_905 = __builtin_shufflevector(__s2_905, __s2_905, 3, 2, 1, 0); \
  69088.   __ret_905 = __noswap_vfmlal_high_f16(__rev0_905, __rev1_905, (float16x4_t) {__noswap_vget_lane_f16(__rev2_905, __p3_905), __noswap_vget_lane_f16(__rev2_905, __p3_905), __noswap_vget_lane_f16(__rev2_905, __p3_905), __noswap_vget_lane_f16(__rev2_905, __p3_905)}); \
  69089.   __ret_905 = __builtin_shufflevector(__ret_905, __ret_905, 1, 0); \
  69090.   __ret_905; \
  69091. })
  69092. #endif
  69093.  
  69094. #ifdef __LITTLE_ENDIAN__
  69095. #define vfmlalq_lane_low_f16(__p0_906, __p1_906, __p2_906, __p3_906) __extension__ ({ \
  69096.   float32x4_t __ret_906; \
  69097.   float32x4_t __s0_906 = __p0_906; \
  69098.   float16x8_t __s1_906 = __p1_906; \
  69099.   float16x4_t __s2_906 = __p2_906; \
  69100.   __ret_906 = vfmlalq_low_f16(__s0_906, __s1_906, (float16x8_t) {vget_lane_f16(__s2_906, __p3_906), vget_lane_f16(__s2_906, __p3_906), vget_lane_f16(__s2_906, __p3_906), vget_lane_f16(__s2_906, __p3_906), vget_lane_f16(__s2_906, __p3_906), vget_lane_f16(__s2_906, __p3_906), vget_lane_f16(__s2_906, __p3_906), vget_lane_f16(__s2_906, __p3_906)}); \
  69101.   __ret_906; \
  69102. })
  69103. #else
  69104. #define vfmlalq_lane_low_f16(__p0_907, __p1_907, __p2_907, __p3_907) __extension__ ({ \
  69105.   float32x4_t __ret_907; \
  69106.   float32x4_t __s0_907 = __p0_907; \
  69107.   float16x8_t __s1_907 = __p1_907; \
  69108.   float16x4_t __s2_907 = __p2_907; \
  69109.   float32x4_t __rev0_907;  __rev0_907 = __builtin_shufflevector(__s0_907, __s0_907, 3, 2, 1, 0); \
  69110.   float16x8_t __rev1_907;  __rev1_907 = __builtin_shufflevector(__s1_907, __s1_907, 7, 6, 5, 4, 3, 2, 1, 0); \
  69111.   float16x4_t __rev2_907;  __rev2_907 = __builtin_shufflevector(__s2_907, __s2_907, 3, 2, 1, 0); \
  69112.   __ret_907 = __noswap_vfmlalq_low_f16(__rev0_907, __rev1_907, (float16x8_t) {__noswap_vget_lane_f16(__rev2_907, __p3_907), __noswap_vget_lane_f16(__rev2_907, __p3_907), __noswap_vget_lane_f16(__rev2_907, __p3_907), __noswap_vget_lane_f16(__rev2_907, __p3_907), __noswap_vget_lane_f16(__rev2_907, __p3_907), __noswap_vget_lane_f16(__rev2_907, __p3_907), __noswap_vget_lane_f16(__rev2_907, __p3_907), __noswap_vget_lane_f16(__rev2_907, __p3_907)}); \
  69113.   __ret_907 = __builtin_shufflevector(__ret_907, __ret_907, 3, 2, 1, 0); \
  69114.   __ret_907; \
  69115. })
  69116. #endif
  69117.  
  69118. #ifdef __LITTLE_ENDIAN__
  69119. #define vfmlal_lane_low_f16(__p0_908, __p1_908, __p2_908, __p3_908) __extension__ ({ \
  69120.   float32x2_t __ret_908; \
  69121.   float32x2_t __s0_908 = __p0_908; \
  69122.   float16x4_t __s1_908 = __p1_908; \
  69123.   float16x4_t __s2_908 = __p2_908; \
  69124.   __ret_908 = vfmlal_low_f16(__s0_908, __s1_908, (float16x4_t) {vget_lane_f16(__s2_908, __p3_908), vget_lane_f16(__s2_908, __p3_908), vget_lane_f16(__s2_908, __p3_908), vget_lane_f16(__s2_908, __p3_908)}); \
  69125.   __ret_908; \
  69126. })
  69127. #else
  69128. #define vfmlal_lane_low_f16(__p0_909, __p1_909, __p2_909, __p3_909) __extension__ ({ \
  69129.   float32x2_t __ret_909; \
  69130.   float32x2_t __s0_909 = __p0_909; \
  69131.   float16x4_t __s1_909 = __p1_909; \
  69132.   float16x4_t __s2_909 = __p2_909; \
  69133.   float32x2_t __rev0_909;  __rev0_909 = __builtin_shufflevector(__s0_909, __s0_909, 1, 0); \
  69134.   float16x4_t __rev1_909;  __rev1_909 = __builtin_shufflevector(__s1_909, __s1_909, 3, 2, 1, 0); \
  69135.   float16x4_t __rev2_909;  __rev2_909 = __builtin_shufflevector(__s2_909, __s2_909, 3, 2, 1, 0); \
  69136.   __ret_909 = __noswap_vfmlal_low_f16(__rev0_909, __rev1_909, (float16x4_t) {__noswap_vget_lane_f16(__rev2_909, __p3_909), __noswap_vget_lane_f16(__rev2_909, __p3_909), __noswap_vget_lane_f16(__rev2_909, __p3_909), __noswap_vget_lane_f16(__rev2_909, __p3_909)}); \
  69137.   __ret_909 = __builtin_shufflevector(__ret_909, __ret_909, 1, 0); \
  69138.   __ret_909; \
  69139. })
  69140. #endif
  69141.  
  69142. #ifdef __LITTLE_ENDIAN__
  69143. #define vfmlalq_laneq_high_f16(__p0_910, __p1_910, __p2_910, __p3_910) __extension__ ({ \
  69144.   float32x4_t __ret_910; \
  69145.   float32x4_t __s0_910 = __p0_910; \
  69146.   float16x8_t __s1_910 = __p1_910; \
  69147.   float16x8_t __s2_910 = __p2_910; \
  69148.   __ret_910 = vfmlalq_high_f16(__s0_910, __s1_910, (float16x8_t) {vgetq_lane_f16(__s2_910, __p3_910), vgetq_lane_f16(__s2_910, __p3_910), vgetq_lane_f16(__s2_910, __p3_910), vgetq_lane_f16(__s2_910, __p3_910), vgetq_lane_f16(__s2_910, __p3_910), vgetq_lane_f16(__s2_910, __p3_910), vgetq_lane_f16(__s2_910, __p3_910), vgetq_lane_f16(__s2_910, __p3_910)}); \
  69149.   __ret_910; \
  69150. })
  69151. #else
  69152. #define vfmlalq_laneq_high_f16(__p0_911, __p1_911, __p2_911, __p3_911) __extension__ ({ \
  69153.   float32x4_t __ret_911; \
  69154.   float32x4_t __s0_911 = __p0_911; \
  69155.   float16x8_t __s1_911 = __p1_911; \
  69156.   float16x8_t __s2_911 = __p2_911; \
  69157.   float32x4_t __rev0_911;  __rev0_911 = __builtin_shufflevector(__s0_911, __s0_911, 3, 2, 1, 0); \
  69158.   float16x8_t __rev1_911;  __rev1_911 = __builtin_shufflevector(__s1_911, __s1_911, 7, 6, 5, 4, 3, 2, 1, 0); \
  69159.   float16x8_t __rev2_911;  __rev2_911 = __builtin_shufflevector(__s2_911, __s2_911, 7, 6, 5, 4, 3, 2, 1, 0); \
  69160.   __ret_911 = __noswap_vfmlalq_high_f16(__rev0_911, __rev1_911, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_911, __p3_911), __noswap_vgetq_lane_f16(__rev2_911, __p3_911), __noswap_vgetq_lane_f16(__rev2_911, __p3_911), __noswap_vgetq_lane_f16(__rev2_911, __p3_911), __noswap_vgetq_lane_f16(__rev2_911, __p3_911), __noswap_vgetq_lane_f16(__rev2_911, __p3_911), __noswap_vgetq_lane_f16(__rev2_911, __p3_911), __noswap_vgetq_lane_f16(__rev2_911, __p3_911)}); \
  69161.   __ret_911 = __builtin_shufflevector(__ret_911, __ret_911, 3, 2, 1, 0); \
  69162.   __ret_911; \
  69163. })
  69164. #endif
  69165.  
  69166. #ifdef __LITTLE_ENDIAN__
  69167. #define vfmlal_laneq_high_f16(__p0_912, __p1_912, __p2_912, __p3_912) __extension__ ({ \
  69168.   float32x2_t __ret_912; \
  69169.   float32x2_t __s0_912 = __p0_912; \
  69170.   float16x4_t __s1_912 = __p1_912; \
  69171.   float16x8_t __s2_912 = __p2_912; \
  69172.   __ret_912 = vfmlal_high_f16(__s0_912, __s1_912, (float16x4_t) {vgetq_lane_f16(__s2_912, __p3_912), vgetq_lane_f16(__s2_912, __p3_912), vgetq_lane_f16(__s2_912, __p3_912), vgetq_lane_f16(__s2_912, __p3_912)}); \
  69173.   __ret_912; \
  69174. })
  69175. #else
  69176. #define vfmlal_laneq_high_f16(__p0_913, __p1_913, __p2_913, __p3_913) __extension__ ({ \
  69177.   float32x2_t __ret_913; \
  69178.   float32x2_t __s0_913 = __p0_913; \
  69179.   float16x4_t __s1_913 = __p1_913; \
  69180.   float16x8_t __s2_913 = __p2_913; \
  69181.   float32x2_t __rev0_913;  __rev0_913 = __builtin_shufflevector(__s0_913, __s0_913, 1, 0); \
  69182.   float16x4_t __rev1_913;  __rev1_913 = __builtin_shufflevector(__s1_913, __s1_913, 3, 2, 1, 0); \
  69183.   float16x8_t __rev2_913;  __rev2_913 = __builtin_shufflevector(__s2_913, __s2_913, 7, 6, 5, 4, 3, 2, 1, 0); \
  69184.   __ret_913 = __noswap_vfmlal_high_f16(__rev0_913, __rev1_913, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_913, __p3_913), __noswap_vgetq_lane_f16(__rev2_913, __p3_913), __noswap_vgetq_lane_f16(__rev2_913, __p3_913), __noswap_vgetq_lane_f16(__rev2_913, __p3_913)}); \
  69185.   __ret_913 = __builtin_shufflevector(__ret_913, __ret_913, 1, 0); \
  69186.   __ret_913; \
  69187. })
  69188. #endif
  69189.  
  69190. #ifdef __LITTLE_ENDIAN__
  69191. #define vfmlalq_laneq_low_f16(__p0_914, __p1_914, __p2_914, __p3_914) __extension__ ({ \
  69192.   float32x4_t __ret_914; \
  69193.   float32x4_t __s0_914 = __p0_914; \
  69194.   float16x8_t __s1_914 = __p1_914; \
  69195.   float16x8_t __s2_914 = __p2_914; \
  69196.   __ret_914 = vfmlalq_low_f16(__s0_914, __s1_914, (float16x8_t) {vgetq_lane_f16(__s2_914, __p3_914), vgetq_lane_f16(__s2_914, __p3_914), vgetq_lane_f16(__s2_914, __p3_914), vgetq_lane_f16(__s2_914, __p3_914), vgetq_lane_f16(__s2_914, __p3_914), vgetq_lane_f16(__s2_914, __p3_914), vgetq_lane_f16(__s2_914, __p3_914), vgetq_lane_f16(__s2_914, __p3_914)}); \
  69197.   __ret_914; \
  69198. })
  69199. #else
  69200. #define vfmlalq_laneq_low_f16(__p0_915, __p1_915, __p2_915, __p3_915) __extension__ ({ \
  69201.   float32x4_t __ret_915; \
  69202.   float32x4_t __s0_915 = __p0_915; \
  69203.   float16x8_t __s1_915 = __p1_915; \
  69204.   float16x8_t __s2_915 = __p2_915; \
  69205.   float32x4_t __rev0_915;  __rev0_915 = __builtin_shufflevector(__s0_915, __s0_915, 3, 2, 1, 0); \
  69206.   float16x8_t __rev1_915;  __rev1_915 = __builtin_shufflevector(__s1_915, __s1_915, 7, 6, 5, 4, 3, 2, 1, 0); \
  69207.   float16x8_t __rev2_915;  __rev2_915 = __builtin_shufflevector(__s2_915, __s2_915, 7, 6, 5, 4, 3, 2, 1, 0); \
  69208.   __ret_915 = __noswap_vfmlalq_low_f16(__rev0_915, __rev1_915, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_915, __p3_915), __noswap_vgetq_lane_f16(__rev2_915, __p3_915), __noswap_vgetq_lane_f16(__rev2_915, __p3_915), __noswap_vgetq_lane_f16(__rev2_915, __p3_915), __noswap_vgetq_lane_f16(__rev2_915, __p3_915), __noswap_vgetq_lane_f16(__rev2_915, __p3_915), __noswap_vgetq_lane_f16(__rev2_915, __p3_915), __noswap_vgetq_lane_f16(__rev2_915, __p3_915)}); \
  69209.   __ret_915 = __builtin_shufflevector(__ret_915, __ret_915, 3, 2, 1, 0); \
  69210.   __ret_915; \
  69211. })
  69212. #endif
  69213.  
  69214. #ifdef __LITTLE_ENDIAN__
  69215. #define vfmlal_laneq_low_f16(__p0_916, __p1_916, __p2_916, __p3_916) __extension__ ({ \
  69216.   float32x2_t __ret_916; \
  69217.   float32x2_t __s0_916 = __p0_916; \
  69218.   float16x4_t __s1_916 = __p1_916; \
  69219.   float16x8_t __s2_916 = __p2_916; \
  69220.   __ret_916 = vfmlal_low_f16(__s0_916, __s1_916, (float16x4_t) {vgetq_lane_f16(__s2_916, __p3_916), vgetq_lane_f16(__s2_916, __p3_916), vgetq_lane_f16(__s2_916, __p3_916), vgetq_lane_f16(__s2_916, __p3_916)}); \
  69221.   __ret_916; \
  69222. })
  69223. #else
  69224. #define vfmlal_laneq_low_f16(__p0_917, __p1_917, __p2_917, __p3_917) __extension__ ({ \
  69225.   float32x2_t __ret_917; \
  69226.   float32x2_t __s0_917 = __p0_917; \
  69227.   float16x4_t __s1_917 = __p1_917; \
  69228.   float16x8_t __s2_917 = __p2_917; \
  69229.   float32x2_t __rev0_917;  __rev0_917 = __builtin_shufflevector(__s0_917, __s0_917, 1, 0); \
  69230.   float16x4_t __rev1_917;  __rev1_917 = __builtin_shufflevector(__s1_917, __s1_917, 3, 2, 1, 0); \
  69231.   float16x8_t __rev2_917;  __rev2_917 = __builtin_shufflevector(__s2_917, __s2_917, 7, 6, 5, 4, 3, 2, 1, 0); \
  69232.   __ret_917 = __noswap_vfmlal_low_f16(__rev0_917, __rev1_917, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_917, __p3_917), __noswap_vgetq_lane_f16(__rev2_917, __p3_917), __noswap_vgetq_lane_f16(__rev2_917, __p3_917), __noswap_vgetq_lane_f16(__rev2_917, __p3_917)}); \
  69233.   __ret_917 = __builtin_shufflevector(__ret_917, __ret_917, 1, 0); \
  69234.   __ret_917; \
  69235. })
  69236. #endif
  69237.  
  69238. #ifdef __LITTLE_ENDIAN__
  69239. #define vfmlslq_lane_high_f16(__p0_918, __p1_918, __p2_918, __p3_918) __extension__ ({ \
  69240.   float32x4_t __ret_918; \
  69241.   float32x4_t __s0_918 = __p0_918; \
  69242.   float16x8_t __s1_918 = __p1_918; \
  69243.   float16x4_t __s2_918 = __p2_918; \
  69244.   __ret_918 = vfmlslq_high_f16(__s0_918, __s1_918, (float16x8_t) {vget_lane_f16(__s2_918, __p3_918), vget_lane_f16(__s2_918, __p3_918), vget_lane_f16(__s2_918, __p3_918), vget_lane_f16(__s2_918, __p3_918), vget_lane_f16(__s2_918, __p3_918), vget_lane_f16(__s2_918, __p3_918), vget_lane_f16(__s2_918, __p3_918), vget_lane_f16(__s2_918, __p3_918)}); \
  69245.   __ret_918; \
  69246. })
  69247. #else
  69248. #define vfmlslq_lane_high_f16(__p0_919, __p1_919, __p2_919, __p3_919) __extension__ ({ \
  69249.   float32x4_t __ret_919; \
  69250.   float32x4_t __s0_919 = __p0_919; \
  69251.   float16x8_t __s1_919 = __p1_919; \
  69252.   float16x4_t __s2_919 = __p2_919; \
  69253.   float32x4_t __rev0_919;  __rev0_919 = __builtin_shufflevector(__s0_919, __s0_919, 3, 2, 1, 0); \
  69254.   float16x8_t __rev1_919;  __rev1_919 = __builtin_shufflevector(__s1_919, __s1_919, 7, 6, 5, 4, 3, 2, 1, 0); \
  69255.   float16x4_t __rev2_919;  __rev2_919 = __builtin_shufflevector(__s2_919, __s2_919, 3, 2, 1, 0); \
  69256.   __ret_919 = __noswap_vfmlslq_high_f16(__rev0_919, __rev1_919, (float16x8_t) {__noswap_vget_lane_f16(__rev2_919, __p3_919), __noswap_vget_lane_f16(__rev2_919, __p3_919), __noswap_vget_lane_f16(__rev2_919, __p3_919), __noswap_vget_lane_f16(__rev2_919, __p3_919), __noswap_vget_lane_f16(__rev2_919, __p3_919), __noswap_vget_lane_f16(__rev2_919, __p3_919), __noswap_vget_lane_f16(__rev2_919, __p3_919), __noswap_vget_lane_f16(__rev2_919, __p3_919)}); \
  69257.   __ret_919 = __builtin_shufflevector(__ret_919, __ret_919, 3, 2, 1, 0); \
  69258.   __ret_919; \
  69259. })
  69260. #endif
  69261.  
  69262. #ifdef __LITTLE_ENDIAN__
  69263. #define vfmlsl_lane_high_f16(__p0_920, __p1_920, __p2_920, __p3_920) __extension__ ({ \
  69264.   float32x2_t __ret_920; \
  69265.   float32x2_t __s0_920 = __p0_920; \
  69266.   float16x4_t __s1_920 = __p1_920; \
  69267.   float16x4_t __s2_920 = __p2_920; \
  69268.   __ret_920 = vfmlsl_high_f16(__s0_920, __s1_920, (float16x4_t) {vget_lane_f16(__s2_920, __p3_920), vget_lane_f16(__s2_920, __p3_920), vget_lane_f16(__s2_920, __p3_920), vget_lane_f16(__s2_920, __p3_920)}); \
  69269.   __ret_920; \
  69270. })
  69271. #else
  69272. #define vfmlsl_lane_high_f16(__p0_921, __p1_921, __p2_921, __p3_921) __extension__ ({ \
  69273.   float32x2_t __ret_921; \
  69274.   float32x2_t __s0_921 = __p0_921; \
  69275.   float16x4_t __s1_921 = __p1_921; \
  69276.   float16x4_t __s2_921 = __p2_921; \
  69277.   float32x2_t __rev0_921;  __rev0_921 = __builtin_shufflevector(__s0_921, __s0_921, 1, 0); \
  69278.   float16x4_t __rev1_921;  __rev1_921 = __builtin_shufflevector(__s1_921, __s1_921, 3, 2, 1, 0); \
  69279.   float16x4_t __rev2_921;  __rev2_921 = __builtin_shufflevector(__s2_921, __s2_921, 3, 2, 1, 0); \
  69280.   __ret_921 = __noswap_vfmlsl_high_f16(__rev0_921, __rev1_921, (float16x4_t) {__noswap_vget_lane_f16(__rev2_921, __p3_921), __noswap_vget_lane_f16(__rev2_921, __p3_921), __noswap_vget_lane_f16(__rev2_921, __p3_921), __noswap_vget_lane_f16(__rev2_921, __p3_921)}); \
  69281.   __ret_921 = __builtin_shufflevector(__ret_921, __ret_921, 1, 0); \
  69282.   __ret_921; \
  69283. })
  69284. #endif
  69285.  
  69286. #ifdef __LITTLE_ENDIAN__
  69287. #define vfmlslq_lane_low_f16(__p0_922, __p1_922, __p2_922, __p3_922) __extension__ ({ \
  69288.   float32x4_t __ret_922; \
  69289.   float32x4_t __s0_922 = __p0_922; \
  69290.   float16x8_t __s1_922 = __p1_922; \
  69291.   float16x4_t __s2_922 = __p2_922; \
  69292.   __ret_922 = vfmlslq_low_f16(__s0_922, __s1_922, (float16x8_t) {vget_lane_f16(__s2_922, __p3_922), vget_lane_f16(__s2_922, __p3_922), vget_lane_f16(__s2_922, __p3_922), vget_lane_f16(__s2_922, __p3_922), vget_lane_f16(__s2_922, __p3_922), vget_lane_f16(__s2_922, __p3_922), vget_lane_f16(__s2_922, __p3_922), vget_lane_f16(__s2_922, __p3_922)}); \
  69293.   __ret_922; \
  69294. })
  69295. #else
  69296. #define vfmlslq_lane_low_f16(__p0_923, __p1_923, __p2_923, __p3_923) __extension__ ({ \
  69297.   float32x4_t __ret_923; \
  69298.   float32x4_t __s0_923 = __p0_923; \
  69299.   float16x8_t __s1_923 = __p1_923; \
  69300.   float16x4_t __s2_923 = __p2_923; \
  69301.   float32x4_t __rev0_923;  __rev0_923 = __builtin_shufflevector(__s0_923, __s0_923, 3, 2, 1, 0); \
  69302.   float16x8_t __rev1_923;  __rev1_923 = __builtin_shufflevector(__s1_923, __s1_923, 7, 6, 5, 4, 3, 2, 1, 0); \
  69303.   float16x4_t __rev2_923;  __rev2_923 = __builtin_shufflevector(__s2_923, __s2_923, 3, 2, 1, 0); \
  69304.   __ret_923 = __noswap_vfmlslq_low_f16(__rev0_923, __rev1_923, (float16x8_t) {__noswap_vget_lane_f16(__rev2_923, __p3_923), __noswap_vget_lane_f16(__rev2_923, __p3_923), __noswap_vget_lane_f16(__rev2_923, __p3_923), __noswap_vget_lane_f16(__rev2_923, __p3_923), __noswap_vget_lane_f16(__rev2_923, __p3_923), __noswap_vget_lane_f16(__rev2_923, __p3_923), __noswap_vget_lane_f16(__rev2_923, __p3_923), __noswap_vget_lane_f16(__rev2_923, __p3_923)}); \
  69305.   __ret_923 = __builtin_shufflevector(__ret_923, __ret_923, 3, 2, 1, 0); \
  69306.   __ret_923; \
  69307. })
  69308. #endif
  69309.  
  69310. #ifdef __LITTLE_ENDIAN__
  69311. #define vfmlsl_lane_low_f16(__p0_924, __p1_924, __p2_924, __p3_924) __extension__ ({ \
  69312.   float32x2_t __ret_924; \
  69313.   float32x2_t __s0_924 = __p0_924; \
  69314.   float16x4_t __s1_924 = __p1_924; \
  69315.   float16x4_t __s2_924 = __p2_924; \
  69316.   __ret_924 = vfmlsl_low_f16(__s0_924, __s1_924, (float16x4_t) {vget_lane_f16(__s2_924, __p3_924), vget_lane_f16(__s2_924, __p3_924), vget_lane_f16(__s2_924, __p3_924), vget_lane_f16(__s2_924, __p3_924)}); \
  69317.   __ret_924; \
  69318. })
  69319. #else
  69320. #define vfmlsl_lane_low_f16(__p0_925, __p1_925, __p2_925, __p3_925) __extension__ ({ \
  69321.   float32x2_t __ret_925; \
  69322.   float32x2_t __s0_925 = __p0_925; \
  69323.   float16x4_t __s1_925 = __p1_925; \
  69324.   float16x4_t __s2_925 = __p2_925; \
  69325.   float32x2_t __rev0_925;  __rev0_925 = __builtin_shufflevector(__s0_925, __s0_925, 1, 0); \
  69326.   float16x4_t __rev1_925;  __rev1_925 = __builtin_shufflevector(__s1_925, __s1_925, 3, 2, 1, 0); \
  69327.   float16x4_t __rev2_925;  __rev2_925 = __builtin_shufflevector(__s2_925, __s2_925, 3, 2, 1, 0); \
  69328.   __ret_925 = __noswap_vfmlsl_low_f16(__rev0_925, __rev1_925, (float16x4_t) {__noswap_vget_lane_f16(__rev2_925, __p3_925), __noswap_vget_lane_f16(__rev2_925, __p3_925), __noswap_vget_lane_f16(__rev2_925, __p3_925), __noswap_vget_lane_f16(__rev2_925, __p3_925)}); \
  69329.   __ret_925 = __builtin_shufflevector(__ret_925, __ret_925, 1, 0); \
  69330.   __ret_925; \
  69331. })
  69332. #endif
  69333.  
  69334. #ifdef __LITTLE_ENDIAN__
  69335. #define vfmlslq_laneq_high_f16(__p0_926, __p1_926, __p2_926, __p3_926) __extension__ ({ \
  69336.   float32x4_t __ret_926; \
  69337.   float32x4_t __s0_926 = __p0_926; \
  69338.   float16x8_t __s1_926 = __p1_926; \
  69339.   float16x8_t __s2_926 = __p2_926; \
  69340.   __ret_926 = vfmlslq_high_f16(__s0_926, __s1_926, (float16x8_t) {vgetq_lane_f16(__s2_926, __p3_926), vgetq_lane_f16(__s2_926, __p3_926), vgetq_lane_f16(__s2_926, __p3_926), vgetq_lane_f16(__s2_926, __p3_926), vgetq_lane_f16(__s2_926, __p3_926), vgetq_lane_f16(__s2_926, __p3_926), vgetq_lane_f16(__s2_926, __p3_926), vgetq_lane_f16(__s2_926, __p3_926)}); \
  69341.   __ret_926; \
  69342. })
  69343. #else
  69344. #define vfmlslq_laneq_high_f16(__p0_927, __p1_927, __p2_927, __p3_927) __extension__ ({ \
  69345.   float32x4_t __ret_927; \
  69346.   float32x4_t __s0_927 = __p0_927; \
  69347.   float16x8_t __s1_927 = __p1_927; \
  69348.   float16x8_t __s2_927 = __p2_927; \
  69349.   float32x4_t __rev0_927;  __rev0_927 = __builtin_shufflevector(__s0_927, __s0_927, 3, 2, 1, 0); \
  69350.   float16x8_t __rev1_927;  __rev1_927 = __builtin_shufflevector(__s1_927, __s1_927, 7, 6, 5, 4, 3, 2, 1, 0); \
  69351.   float16x8_t __rev2_927;  __rev2_927 = __builtin_shufflevector(__s2_927, __s2_927, 7, 6, 5, 4, 3, 2, 1, 0); \
  69352.   __ret_927 = __noswap_vfmlslq_high_f16(__rev0_927, __rev1_927, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_927, __p3_927), __noswap_vgetq_lane_f16(__rev2_927, __p3_927), __noswap_vgetq_lane_f16(__rev2_927, __p3_927), __noswap_vgetq_lane_f16(__rev2_927, __p3_927), __noswap_vgetq_lane_f16(__rev2_927, __p3_927), __noswap_vgetq_lane_f16(__rev2_927, __p3_927), __noswap_vgetq_lane_f16(__rev2_927, __p3_927), __noswap_vgetq_lane_f16(__rev2_927, __p3_927)}); \
  69353.   __ret_927 = __builtin_shufflevector(__ret_927, __ret_927, 3, 2, 1, 0); \
  69354.   __ret_927; \
  69355. })
  69356. #endif
  69357.  
  69358. #ifdef __LITTLE_ENDIAN__
  69359. #define vfmlsl_laneq_high_f16(__p0_928, __p1_928, __p2_928, __p3_928) __extension__ ({ \
  69360.   float32x2_t __ret_928; \
  69361.   float32x2_t __s0_928 = __p0_928; \
  69362.   float16x4_t __s1_928 = __p1_928; \
  69363.   float16x8_t __s2_928 = __p2_928; \
  69364.   __ret_928 = vfmlsl_high_f16(__s0_928, __s1_928, (float16x4_t) {vgetq_lane_f16(__s2_928, __p3_928), vgetq_lane_f16(__s2_928, __p3_928), vgetq_lane_f16(__s2_928, __p3_928), vgetq_lane_f16(__s2_928, __p3_928)}); \
  69365.   __ret_928; \
  69366. })
  69367. #else
  69368. #define vfmlsl_laneq_high_f16(__p0_929, __p1_929, __p2_929, __p3_929) __extension__ ({ \
  69369.   float32x2_t __ret_929; \
  69370.   float32x2_t __s0_929 = __p0_929; \
  69371.   float16x4_t __s1_929 = __p1_929; \
  69372.   float16x8_t __s2_929 = __p2_929; \
  69373.   float32x2_t __rev0_929;  __rev0_929 = __builtin_shufflevector(__s0_929, __s0_929, 1, 0); \
  69374.   float16x4_t __rev1_929;  __rev1_929 = __builtin_shufflevector(__s1_929, __s1_929, 3, 2, 1, 0); \
  69375.   float16x8_t __rev2_929;  __rev2_929 = __builtin_shufflevector(__s2_929, __s2_929, 7, 6, 5, 4, 3, 2, 1, 0); \
  69376.   __ret_929 = __noswap_vfmlsl_high_f16(__rev0_929, __rev1_929, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_929, __p3_929), __noswap_vgetq_lane_f16(__rev2_929, __p3_929), __noswap_vgetq_lane_f16(__rev2_929, __p3_929), __noswap_vgetq_lane_f16(__rev2_929, __p3_929)}); \
  69377.   __ret_929 = __builtin_shufflevector(__ret_929, __ret_929, 1, 0); \
  69378.   __ret_929; \
  69379. })
  69380. #endif
  69381.  
  69382. #ifdef __LITTLE_ENDIAN__
  69383. #define vfmlslq_laneq_low_f16(__p0_930, __p1_930, __p2_930, __p3_930) __extension__ ({ \
  69384.   float32x4_t __ret_930; \
  69385.   float32x4_t __s0_930 = __p0_930; \
  69386.   float16x8_t __s1_930 = __p1_930; \
  69387.   float16x8_t __s2_930 = __p2_930; \
  69388.   __ret_930 = vfmlslq_low_f16(__s0_930, __s1_930, (float16x8_t) {vgetq_lane_f16(__s2_930, __p3_930), vgetq_lane_f16(__s2_930, __p3_930), vgetq_lane_f16(__s2_930, __p3_930), vgetq_lane_f16(__s2_930, __p3_930), vgetq_lane_f16(__s2_930, __p3_930), vgetq_lane_f16(__s2_930, __p3_930), vgetq_lane_f16(__s2_930, __p3_930), vgetq_lane_f16(__s2_930, __p3_930)}); \
  69389.   __ret_930; \
  69390. })
  69391. #else
  69392. #define vfmlslq_laneq_low_f16(__p0_931, __p1_931, __p2_931, __p3_931) __extension__ ({ \
  69393.   float32x4_t __ret_931; \
  69394.   float32x4_t __s0_931 = __p0_931; \
  69395.   float16x8_t __s1_931 = __p1_931; \
  69396.   float16x8_t __s2_931 = __p2_931; \
  69397.   float32x4_t __rev0_931;  __rev0_931 = __builtin_shufflevector(__s0_931, __s0_931, 3, 2, 1, 0); \
  69398.   float16x8_t __rev1_931;  __rev1_931 = __builtin_shufflevector(__s1_931, __s1_931, 7, 6, 5, 4, 3, 2, 1, 0); \
  69399.   float16x8_t __rev2_931;  __rev2_931 = __builtin_shufflevector(__s2_931, __s2_931, 7, 6, 5, 4, 3, 2, 1, 0); \
  69400.   __ret_931 = __noswap_vfmlslq_low_f16(__rev0_931, __rev1_931, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_931, __p3_931), __noswap_vgetq_lane_f16(__rev2_931, __p3_931), __noswap_vgetq_lane_f16(__rev2_931, __p3_931), __noswap_vgetq_lane_f16(__rev2_931, __p3_931), __noswap_vgetq_lane_f16(__rev2_931, __p3_931), __noswap_vgetq_lane_f16(__rev2_931, __p3_931), __noswap_vgetq_lane_f16(__rev2_931, __p3_931), __noswap_vgetq_lane_f16(__rev2_931, __p3_931)}); \
  69401.   __ret_931 = __builtin_shufflevector(__ret_931, __ret_931, 3, 2, 1, 0); \
  69402.   __ret_931; \
  69403. })
  69404. #endif
  69405.  
  69406. #ifdef __LITTLE_ENDIAN__
  69407. #define vfmlsl_laneq_low_f16(__p0_932, __p1_932, __p2_932, __p3_932) __extension__ ({ \
  69408.   float32x2_t __ret_932; \
  69409.   float32x2_t __s0_932 = __p0_932; \
  69410.   float16x4_t __s1_932 = __p1_932; \
  69411.   float16x8_t __s2_932 = __p2_932; \
  69412.   __ret_932 = vfmlsl_low_f16(__s0_932, __s1_932, (float16x4_t) {vgetq_lane_f16(__s2_932, __p3_932), vgetq_lane_f16(__s2_932, __p3_932), vgetq_lane_f16(__s2_932, __p3_932), vgetq_lane_f16(__s2_932, __p3_932)}); \
  69413.   __ret_932; \
  69414. })
  69415. #else
  69416. #define vfmlsl_laneq_low_f16(__p0_933, __p1_933, __p2_933, __p3_933) __extension__ ({ \
  69417.   float32x2_t __ret_933; \
  69418.   float32x2_t __s0_933 = __p0_933; \
  69419.   float16x4_t __s1_933 = __p1_933; \
  69420.   float16x8_t __s2_933 = __p2_933; \
  69421.   float32x2_t __rev0_933;  __rev0_933 = __builtin_shufflevector(__s0_933, __s0_933, 1, 0); \
  69422.   float16x4_t __rev1_933;  __rev1_933 = __builtin_shufflevector(__s1_933, __s1_933, 3, 2, 1, 0); \
  69423.   float16x8_t __rev2_933;  __rev2_933 = __builtin_shufflevector(__s2_933, __s2_933, 7, 6, 5, 4, 3, 2, 1, 0); \
  69424.   __ret_933 = __noswap_vfmlsl_low_f16(__rev0_933, __rev1_933, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_933, __p3_933), __noswap_vgetq_lane_f16(__rev2_933, __p3_933), __noswap_vgetq_lane_f16(__rev2_933, __p3_933), __noswap_vgetq_lane_f16(__rev2_933, __p3_933)}); \
  69425.   __ret_933 = __builtin_shufflevector(__ret_933, __ret_933, 1, 0); \
  69426.   __ret_933; \
  69427. })
  69428. #endif
  69429.  
  69430. #ifdef __LITTLE_ENDIAN__
  69431. #define vmulh_lane_f16(__p0_934, __p1_934, __p2_934) __extension__ ({ \
  69432.   float16_t __ret_934; \
  69433.   float16_t __s0_934 = __p0_934; \
  69434.   float16x4_t __s1_934 = __p1_934; \
  69435.   __ret_934 = __s0_934 * vget_lane_f16(__s1_934, __p2_934); \
  69436.   __ret_934; \
  69437. })
  69438. #else
  69439. #define vmulh_lane_f16(__p0_935, __p1_935, __p2_935) __extension__ ({ \
  69440.   float16_t __ret_935; \
  69441.   float16_t __s0_935 = __p0_935; \
  69442.   float16x4_t __s1_935 = __p1_935; \
  69443.   float16x4_t __rev1_935;  __rev1_935 = __builtin_shufflevector(__s1_935, __s1_935, 3, 2, 1, 0); \
  69444.   __ret_935 = __s0_935 * __noswap_vget_lane_f16(__rev1_935, __p2_935); \
  69445.   __ret_935; \
  69446. })
  69447. #endif
  69448.  
  69449. #ifdef __LITTLE_ENDIAN__
  69450. #define vmulh_laneq_f16(__p0_936, __p1_936, __p2_936) __extension__ ({ \
  69451.   float16_t __ret_936; \
  69452.   float16_t __s0_936 = __p0_936; \
  69453.   float16x8_t __s1_936 = __p1_936; \
  69454.   __ret_936 = __s0_936 * vgetq_lane_f16(__s1_936, __p2_936); \
  69455.   __ret_936; \
  69456. })
  69457. #else
  69458. #define vmulh_laneq_f16(__p0_937, __p1_937, __p2_937) __extension__ ({ \
  69459.   float16_t __ret_937; \
  69460.   float16_t __s0_937 = __p0_937; \
  69461.   float16x8_t __s1_937 = __p1_937; \
  69462.   float16x8_t __rev1_937;  __rev1_937 = __builtin_shufflevector(__s1_937, __s1_937, 7, 6, 5, 4, 3, 2, 1, 0); \
  69463.   __ret_937 = __s0_937 * __noswap_vgetq_lane_f16(__rev1_937, __p2_937); \
  69464.   __ret_937; \
  69465. })
  69466. #endif
  69467.  
  69468. #endif
  69469. #ifdef __LITTLE_ENDIAN__
  69470. __ai uint16x8_t vabal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  69471.   uint16x8_t __ret;
  69472.   __ret = __p0 + vabdl_u8(__p1, __p2);
  69473.   return __ret;
  69474. }
  69475. #else
  69476. __ai uint16x8_t vabal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  69477.   uint16x8_t __ret;
  69478.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  69479.   uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  69480.   uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  69481.   __ret = __rev0 + __noswap_vabdl_u8(__rev1, __rev2);
  69482.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  69483.   return __ret;
  69484. }
  69485. __ai uint16x8_t __noswap_vabal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
  69486.   uint16x8_t __ret;
  69487.   __ret = __p0 + __noswap_vabdl_u8(__p1, __p2);
  69488.   return __ret;
  69489. }
  69490. #endif
  69491.  
  69492. #ifdef __LITTLE_ENDIAN__
  69493. __ai uint64x2_t vabal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  69494.   uint64x2_t __ret;
  69495.   __ret = __p0 + vabdl_u32(__p1, __p2);
  69496.   return __ret;
  69497. }
  69498. #else
  69499. __ai uint64x2_t vabal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  69500.   uint64x2_t __ret;
  69501.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  69502.   uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  69503.   uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  69504.   __ret = __rev0 + __noswap_vabdl_u32(__rev1, __rev2);
  69505.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  69506.   return __ret;
  69507. }
  69508. __ai uint64x2_t __noswap_vabal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
  69509.   uint64x2_t __ret;
  69510.   __ret = __p0 + __noswap_vabdl_u32(__p1, __p2);
  69511.   return __ret;
  69512. }
  69513. #endif
  69514.  
  69515. #ifdef __LITTLE_ENDIAN__
  69516. __ai uint32x4_t vabal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  69517.   uint32x4_t __ret;
  69518.   __ret = __p0 + vabdl_u16(__p1, __p2);
  69519.   return __ret;
  69520. }
  69521. #else
  69522. __ai uint32x4_t vabal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  69523.   uint32x4_t __ret;
  69524.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  69525.   uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  69526.   uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  69527.   __ret = __rev0 + __noswap_vabdl_u16(__rev1, __rev2);
  69528.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  69529.   return __ret;
  69530. }
  69531. __ai uint32x4_t __noswap_vabal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
  69532.   uint32x4_t __ret;
  69533.   __ret = __p0 + __noswap_vabdl_u16(__p1, __p2);
  69534.   return __ret;
  69535. }
  69536. #endif
  69537.  
  69538. #ifdef __LITTLE_ENDIAN__
  69539. __ai int16x8_t vabal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  69540.   int16x8_t __ret;
  69541.   __ret = __p0 + vabdl_s8(__p1, __p2);
  69542.   return __ret;
  69543. }
  69544. #else
  69545. __ai int16x8_t vabal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  69546.   int16x8_t __ret;
  69547.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  69548.   int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  69549.   int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  69550.   __ret = __rev0 + __noswap_vabdl_s8(__rev1, __rev2);
  69551.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  69552.   return __ret;
  69553. }
  69554. __ai int16x8_t __noswap_vabal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
  69555.   int16x8_t __ret;
  69556.   __ret = __p0 + __noswap_vabdl_s8(__p1, __p2);
  69557.   return __ret;
  69558. }
  69559. #endif
  69560.  
  69561. #ifdef __LITTLE_ENDIAN__
  69562. __ai int64x2_t vabal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  69563.   int64x2_t __ret;
  69564.   __ret = __p0 + vabdl_s32(__p1, __p2);
  69565.   return __ret;
  69566. }
  69567. #else
  69568. __ai int64x2_t vabal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  69569.   int64x2_t __ret;
  69570.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  69571.   int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
  69572.   int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
  69573.   __ret = __rev0 + __noswap_vabdl_s32(__rev1, __rev2);
  69574.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  69575.   return __ret;
  69576. }
  69577. __ai int64x2_t __noswap_vabal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
  69578.   int64x2_t __ret;
  69579.   __ret = __p0 + __noswap_vabdl_s32(__p1, __p2);
  69580.   return __ret;
  69581. }
  69582. #endif
  69583.  
  69584. #ifdef __LITTLE_ENDIAN__
  69585. __ai int32x4_t vabal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  69586.   int32x4_t __ret;
  69587.   __ret = __p0 + vabdl_s16(__p1, __p2);
  69588.   return __ret;
  69589. }
  69590. #else
  69591. __ai int32x4_t vabal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  69592.   int32x4_t __ret;
  69593.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  69594.   int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  69595.   int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  69596.   __ret = __rev0 + __noswap_vabdl_s16(__rev1, __rev2);
  69597.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  69598.   return __ret;
  69599. }
  69600. __ai int32x4_t __noswap_vabal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
  69601.   int32x4_t __ret;
  69602.   __ret = __p0 + __noswap_vabdl_s16(__p1, __p2);
  69603.   return __ret;
  69604. }
  69605. #endif
  69606.  
  69607. #if defined(__aarch64__)
  69608. #ifdef __LITTLE_ENDIAN__
  69609. __ai uint16x8_t vabal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  69610.   uint16x8_t __ret;
  69611.   __ret = vabal_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
  69612.   return __ret;
  69613. }
  69614. #else
  69615. __ai uint16x8_t vabal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
  69616.   uint16x8_t __ret;
  69617.   uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  69618.   uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  69619.   uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  69620.   __ret = __noswap_vabal_u8(__rev0, __noswap_vget_high_u8(__rev1), __noswap_vget_high_u8(__rev2));
  69621.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  69622.   return __ret;
  69623. }
  69624. #endif
  69625.  
  69626. #ifdef __LITTLE_ENDIAN__
  69627. __ai uint64x2_t vabal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  69628.   uint64x2_t __ret;
  69629.   __ret = vabal_u32(__p0, vget_high_u32(__p1), vget_high_u32(__p2));
  69630.   return __ret;
  69631. }
  69632. #else
  69633. __ai uint64x2_t vabal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
  69634.   uint64x2_t __ret;
  69635.   uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  69636.   uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  69637.   uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  69638.   __ret = __noswap_vabal_u32(__rev0, __noswap_vget_high_u32(__rev1), __noswap_vget_high_u32(__rev2));
  69639.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  69640.   return __ret;
  69641. }
  69642. #endif
  69643.  
  69644. #ifdef __LITTLE_ENDIAN__
  69645. __ai uint32x4_t vabal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  69646.   uint32x4_t __ret;
  69647.   __ret = vabal_u16(__p0, vget_high_u16(__p1), vget_high_u16(__p2));
  69648.   return __ret;
  69649. }
  69650. #else
  69651. __ai uint32x4_t vabal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
  69652.   uint32x4_t __ret;
  69653.   uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  69654.   uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  69655.   uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  69656.   __ret = __noswap_vabal_u16(__rev0, __noswap_vget_high_u16(__rev1), __noswap_vget_high_u16(__rev2));
  69657.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  69658.   return __ret;
  69659. }
  69660. #endif
  69661.  
  69662. #ifdef __LITTLE_ENDIAN__
  69663. __ai int16x8_t vabal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
  69664.   int16x8_t __ret;
  69665.   __ret = vabal_s8(__p0, vget_high_s8(__p1), vget_high_s8(__p2));
  69666.   return __ret;
  69667. }
  69668. #else
  69669. __ai int16x8_t vabal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
  69670.   int16x8_t __ret;
  69671.   int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
  69672.   int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  69673.   int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
  69674.   __ret = __noswap_vabal_s8(__rev0, __noswap_vget_high_s8(__rev1), __noswap_vget_high_s8(__rev2));
  69675.   __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
  69676.   return __ret;
  69677. }
  69678. #endif
  69679.  
  69680. #ifdef __LITTLE_ENDIAN__
  69681. __ai int64x2_t vabal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  69682.   int64x2_t __ret;
  69683.   __ret = vabal_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
  69684.   return __ret;
  69685. }
  69686. #else
  69687. __ai int64x2_t vabal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
  69688.   int64x2_t __ret;
  69689.   int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
  69690.   int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
  69691.   int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
  69692.   __ret = __noswap_vabal_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
  69693.   __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
  69694.   return __ret;
  69695. }
  69696. #endif
  69697.  
  69698. #ifdef __LITTLE_ENDIAN__
  69699. __ai int32x4_t vabal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  69700.   int32x4_t __ret;
  69701.   __ret = vabal_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
  69702.   return __ret;
  69703. }
  69704. #else
  69705. __ai int32x4_t vabal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
  69706.   int32x4_t __ret;
  69707.   int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
  69708.   int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
  69709.   int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
  69710.   __ret = __noswap_vabal_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
  69711.   __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
  69712.   return __ret;
  69713. }
  69714. #endif
  69715.  
  69716. #endif
  69717.  
  69718. #undef __ai
  69719.  
  69720. #endif /* if !defined(__ARM_NEON) */
  69721. #endif /* ifndef __ARM_FP */
  69722. #endif /* __ARM_NEON_H */
  69723.