Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 1 | pmbaty | 1 | /*++ |
| 2 | |||
| 3 | Copyright (c) Microsoft Corporation. All rights reserved. |
||
| 4 | |||
| 5 | Module Name: |
||
| 6 | |||
| 7 | xnamathvector.inl |
||
| 8 | |||
| 9 | Abstract: |
||
| 10 | |||
| 11 | XNA math library for Windows and Xbox 360: Vector functions |
||
| 12 | --*/ |
||
| 13 | |||
| 14 | #if defined(_MSC_VER) && (_MSC_VER > 1000) |
||
| 15 | #pragma once |
||
| 16 | #endif |
||
| 17 | |||
| 18 | #ifndef __XNAMATHVECTOR_INL__ |
||
| 19 | #define __XNAMATHVECTOR_INL__ |
||
| 20 | |||
| 21 | #if defined(_XM_NO_INTRINSICS_) |
||
| 22 | #define XMISNAN(x) ((*(UINT*)&(x) & 0x7F800000) == 0x7F800000 && (*(UINT*)&(x) & 0x7FFFFF) != 0) |
||
| 23 | #define XMISINF(x) ((*(UINT*)&(x) & 0x7FFFFFFF) == 0x7F800000) |
||
| 24 | #endif |
||
| 25 | |||
| 26 | /**************************************************************************** |
||
| 27 | * |
||
| 28 | * General Vector |
||
| 29 | * |
||
| 30 | ****************************************************************************/ |
||
| 31 | |||
| 32 | //------------------------------------------------------------------------------ |
||
| 33 | // Assignment operations |
||
| 34 | //------------------------------------------------------------------------------ |
||
| 35 | |||
| 36 | //------------------------------------------------------------------------------ |
||
| 37 | // Return a vector with all elements equaling zero |
||
| 38 | XMFINLINE XMVECTOR XMVectorZero() |
||
| 39 | { |
||
| 40 | #if defined(_XM_NO_INTRINSICS_) |
||
| 41 | XMVECTOR vResult = {0.0f,0.0f,0.0f,0.0f}; |
||
| 42 | return vResult; |
||
| 43 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 44 | return _mm_setzero_ps(); |
||
| 45 | #else // _XM_VMX128_INTRINSICS_ |
||
| 46 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 47 | } |
||
| 48 | |||
| 49 | //------------------------------------------------------------------------------ |
||
| 50 | // Initialize a vector with four floating point values |
||
| 51 | XMFINLINE XMVECTOR XMVectorSet |
||
| 52 | ( |
||
| 53 | FLOAT x, |
||
| 54 | FLOAT y, |
||
| 55 | FLOAT z, |
||
| 56 | FLOAT w |
||
| 57 | ) |
||
| 58 | { |
||
| 59 | #if defined(_XM_NO_INTRINSICS_) |
||
| 60 | XMVECTORF32 vResult = {x,y,z,w}; |
||
| 61 | return vResult.v; |
||
| 62 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 63 | return _mm_set_ps( w, z, y, x ); |
||
| 64 | #else // _XM_VMX128_INTRINSICS_ |
||
| 65 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 66 | } |
||
| 67 | |||
| 68 | //------------------------------------------------------------------------------ |
||
| 69 | // Initialize a vector with four integer values |
||
| 70 | XMFINLINE XMVECTOR XMVectorSetInt |
||
| 71 | ( |
||
| 72 | UINT x, |
||
| 73 | UINT y, |
||
| 74 | UINT z, |
||
| 75 | UINT w |
||
| 76 | ) |
||
| 77 | { |
||
| 78 | #if defined(_XM_NO_INTRINSICS_) |
||
| 79 | XMVECTORU32 vResult = {x,y,z,w}; |
||
| 80 | return vResult.v; |
||
| 81 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 82 | __m128i V = _mm_set_epi32( w, z, y, x ); |
||
| 83 | return reinterpret_cast<__m128 *>(&V)[0]; |
||
| 84 | #else // _XM_VMX128_INTRINSICS_ |
||
| 85 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 86 | } |
||
| 87 | |||
| 88 | //------------------------------------------------------------------------------ |
||
| 89 | // Initialize a vector with a replicated floating point value |
||
| 90 | XMFINLINE XMVECTOR XMVectorReplicate |
||
| 91 | ( |
||
| 92 | FLOAT Value |
||
| 93 | ) |
||
| 94 | { |
||
| 95 | #if defined(_XM_NO_INTRINSICS_) || defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 96 | XMVECTORF32 vResult = {Value,Value,Value,Value}; |
||
| 97 | return vResult.v; |
||
| 98 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 99 | return _mm_set_ps1( Value ); |
||
| 100 | #else // _XM_VMX128_INTRINSICS_ |
||
| 101 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 102 | } |
||
| 103 | |||
| 104 | //------------------------------------------------------------------------------ |
||
| 105 | // Initialize a vector with a replicated floating point value passed by pointer |
||
| 106 | XMFINLINE XMVECTOR XMVectorReplicatePtr |
||
| 107 | ( |
||
| 108 | CONST FLOAT *pValue |
||
| 109 | ) |
||
| 110 | { |
||
| 111 | #if defined(_XM_NO_INTRINSICS_) || defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 112 | FLOAT Value = pValue[0]; |
||
| 113 | XMVECTORF32 vResult = {Value,Value,Value,Value}; |
||
| 114 | return vResult.v; |
||
| 115 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 116 | return _mm_load_ps1( pValue ); |
||
| 117 | #else // _XM_VMX128_INTRINSICS_ |
||
| 118 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 119 | } |
||
| 120 | |||
| 121 | //------------------------------------------------------------------------------ |
||
| 122 | // Initialize a vector with a replicated integer value |
||
| 123 | XMFINLINE XMVECTOR XMVectorReplicateInt |
||
| 124 | ( |
||
| 125 | UINT Value |
||
| 126 | ) |
||
| 127 | { |
||
| 128 | #if defined(_XM_NO_INTRINSICS_) || defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 129 | XMVECTORU32 vResult = {Value,Value,Value,Value}; |
||
| 130 | return vResult.v; |
||
| 131 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 132 | __m128i vTemp = _mm_set1_epi32( Value ); |
||
| 133 | return reinterpret_cast<const __m128 *>(&vTemp)[0]; |
||
| 134 | #else // _XM_VMX128_INTRINSICS_ |
||
| 135 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 136 | } |
||
| 137 | |||
| 138 | //------------------------------------------------------------------------------ |
||
| 139 | // Initialize a vector with a replicated integer value passed by pointer |
||
| 140 | XMFINLINE XMVECTOR XMVectorReplicateIntPtr |
||
| 141 | ( |
||
| 142 | CONST UINT *pValue |
||
| 143 | ) |
||
| 144 | { |
||
| 145 | #if defined(_XM_NO_INTRINSICS_) || defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 146 | UINT Value = pValue[0]; |
||
| 147 | XMVECTORU32 vResult = {Value,Value,Value,Value}; |
||
| 148 | return vResult.v; |
||
| 149 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 150 | return _mm_load_ps1(reinterpret_cast<const float *>(pValue)); |
||
| 151 | #else // _XM_VMX128_INTRINSICS_ |
||
| 152 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 153 | } |
||
| 154 | |||
| 155 | //------------------------------------------------------------------------------ |
||
| 156 | // Initialize a vector with all bits set (true mask) |
||
| 157 | XMFINLINE XMVECTOR XMVectorTrueInt() |
||
| 158 | { |
||
| 159 | #if defined(_XM_NO_INTRINSICS_) |
||
| 160 | XMVECTORU32 vResult = {0xFFFFFFFFU,0xFFFFFFFFU,0xFFFFFFFFU,0xFFFFFFFFU}; |
||
| 161 | return vResult.v; |
||
| 162 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 163 | __m128i V = _mm_set1_epi32(-1); |
||
| 164 | return reinterpret_cast<__m128 *>(&V)[0]; |
||
| 165 | #else // _XM_VMX128_INTRINSICS_ |
||
| 166 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 167 | } |
||
| 168 | |||
| 169 | //------------------------------------------------------------------------------ |
||
| 170 | // Initialize a vector with all bits clear (false mask) |
||
| 171 | XMFINLINE XMVECTOR XMVectorFalseInt() |
||
| 172 | { |
||
| 173 | #if defined(_XM_NO_INTRINSICS_) |
||
| 174 | XMVECTOR vResult = {0.0f,0.0f,0.0f,0.0f}; |
||
| 175 | return vResult; |
||
| 176 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 177 | return _mm_setzero_ps(); |
||
| 178 | #else // _XM_VMX128_INTRINSICS_ |
||
| 179 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 180 | } |
||
| 181 | |||
| 182 | //------------------------------------------------------------------------------ |
||
| 183 | // Replicate the x component of the vector |
||
| 184 | XMFINLINE XMVECTOR XMVectorSplatX |
||
| 185 | ( |
||
| 186 | FXMVECTOR V |
||
| 187 | ) |
||
| 188 | { |
||
| 189 | #if defined(_XM_NO_INTRINSICS_) |
||
| 190 | XMVECTOR vResult; |
||
| 191 | vResult.vector4_f32[0] = |
||
| 192 | vResult.vector4_f32[1] = |
||
| 193 | vResult.vector4_f32[2] = |
||
| 194 | vResult.vector4_f32[3] = V.vector4_f32[0]; |
||
| 195 | return vResult; |
||
| 196 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 197 | return _mm_shuffle_ps( V, V, _MM_SHUFFLE(0, 0, 0, 0) ); |
||
| 198 | #else // _XM_VMX128_INTRINSICS_ |
||
| 199 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 200 | } |
||
| 201 | |||
| 202 | //------------------------------------------------------------------------------ |
||
| 203 | // Replicate the y component of the vector |
||
| 204 | XMFINLINE XMVECTOR XMVectorSplatY |
||
| 205 | ( |
||
| 206 | FXMVECTOR V |
||
| 207 | ) |
||
| 208 | { |
||
| 209 | #if defined(_XM_NO_INTRINSICS_) |
||
| 210 | XMVECTOR vResult; |
||
| 211 | vResult.vector4_f32[0] = |
||
| 212 | vResult.vector4_f32[1] = |
||
| 213 | vResult.vector4_f32[2] = |
||
| 214 | vResult.vector4_f32[3] = V.vector4_f32[1]; |
||
| 215 | return vResult; |
||
| 216 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 217 | return _mm_shuffle_ps( V, V, _MM_SHUFFLE(1, 1, 1, 1) ); |
||
| 218 | #else // _XM_VMX128_INTRINSICS_ |
||
| 219 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 220 | } |
||
| 221 | |||
| 222 | //------------------------------------------------------------------------------ |
||
| 223 | // Replicate the z component of the vector |
||
| 224 | XMFINLINE XMVECTOR XMVectorSplatZ |
||
| 225 | ( |
||
| 226 | FXMVECTOR V |
||
| 227 | ) |
||
| 228 | { |
||
| 229 | #if defined(_XM_NO_INTRINSICS_) |
||
| 230 | XMVECTOR vResult; |
||
| 231 | vResult.vector4_f32[0] = |
||
| 232 | vResult.vector4_f32[1] = |
||
| 233 | vResult.vector4_f32[2] = |
||
| 234 | vResult.vector4_f32[3] = V.vector4_f32[2]; |
||
| 235 | return vResult; |
||
| 236 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 237 | return _mm_shuffle_ps( V, V, _MM_SHUFFLE(2, 2, 2, 2) ); |
||
| 238 | #else // _XM_VMX128_INTRINSICS_ |
||
| 239 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 240 | } |
||
| 241 | |||
| 242 | //------------------------------------------------------------------------------ |
||
| 243 | // Replicate the w component of the vector |
||
| 244 | XMFINLINE XMVECTOR XMVectorSplatW |
||
| 245 | ( |
||
| 246 | FXMVECTOR V |
||
| 247 | ) |
||
| 248 | { |
||
| 249 | #if defined(_XM_NO_INTRINSICS_) |
||
| 250 | XMVECTOR vResult; |
||
| 251 | vResult.vector4_f32[0] = |
||
| 252 | vResult.vector4_f32[1] = |
||
| 253 | vResult.vector4_f32[2] = |
||
| 254 | vResult.vector4_f32[3] = V.vector4_f32[3]; |
||
| 255 | return vResult; |
||
| 256 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 257 | return _mm_shuffle_ps( V, V, _MM_SHUFFLE(3, 3, 3, 3) ); |
||
| 258 | #else // _XM_VMX128_INTRINSICS_ |
||
| 259 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 260 | } |
||
| 261 | |||
| 262 | //------------------------------------------------------------------------------ |
||
| 263 | // Return a vector of 1.0f,1.0f,1.0f,1.0f |
||
| 264 | XMFINLINE XMVECTOR XMVectorSplatOne() |
||
| 265 | { |
||
| 266 | #if defined(_XM_NO_INTRINSICS_) |
||
| 267 | XMVECTOR vResult; |
||
| 268 | vResult.vector4_f32[0] = |
||
| 269 | vResult.vector4_f32[1] = |
||
| 270 | vResult.vector4_f32[2] = |
||
| 271 | vResult.vector4_f32[3] = 1.0f; |
||
| 272 | return vResult; |
||
| 273 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 274 | return g_XMOne; |
||
| 275 | #else // _XM_VMX128_INTRINSICS_ |
||
| 276 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 277 | } |
||
| 278 | |||
| 279 | //------------------------------------------------------------------------------ |
||
| 280 | // Return a vector of INF,INF,INF,INF |
||
| 281 | XMFINLINE XMVECTOR XMVectorSplatInfinity() |
||
| 282 | { |
||
| 283 | #if defined(_XM_NO_INTRINSICS_) |
||
| 284 | XMVECTOR vResult; |
||
| 285 | vResult.vector4_u32[0] = |
||
| 286 | vResult.vector4_u32[1] = |
||
| 287 | vResult.vector4_u32[2] = |
||
| 288 | vResult.vector4_u32[3] = 0x7F800000; |
||
| 289 | return vResult; |
||
| 290 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 291 | return g_XMInfinity; |
||
| 292 | #else // _XM_VMX128_INTRINSICS_ |
||
| 293 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 294 | } |
||
| 295 | |||
| 296 | //------------------------------------------------------------------------------ |
||
| 297 | // Return a vector of Q_NAN,Q_NAN,Q_NAN,Q_NAN |
||
| 298 | XMFINLINE XMVECTOR XMVectorSplatQNaN() |
||
| 299 | { |
||
| 300 | #if defined(_XM_NO_INTRINSICS_) |
||
| 301 | XMVECTOR vResult; |
||
| 302 | vResult.vector4_u32[0] = |
||
| 303 | vResult.vector4_u32[1] = |
||
| 304 | vResult.vector4_u32[2] = |
||
| 305 | vResult.vector4_u32[3] = 0x7FC00000; |
||
| 306 | return vResult; |
||
| 307 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 308 | return g_XMQNaN; |
||
| 309 | #else // _XM_VMX128_INTRINSICS_ |
||
| 310 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 311 | } |
||
| 312 | |||
| 313 | //------------------------------------------------------------------------------ |
||
| 314 | // Return a vector of 1.192092896e-7f,1.192092896e-7f,1.192092896e-7f,1.192092896e-7f |
||
| 315 | XMFINLINE XMVECTOR XMVectorSplatEpsilon() |
||
| 316 | { |
||
| 317 | #if defined(_XM_NO_INTRINSICS_) |
||
| 318 | XMVECTOR vResult; |
||
| 319 | vResult.vector4_u32[0] = |
||
| 320 | vResult.vector4_u32[1] = |
||
| 321 | vResult.vector4_u32[2] = |
||
| 322 | vResult.vector4_u32[3] = 0x34000000; |
||
| 323 | return vResult; |
||
| 324 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 325 | return g_XMEpsilon; |
||
| 326 | #else // _XM_VMX128_INTRINSICS_ |
||
| 327 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 328 | } |
||
| 329 | |||
| 330 | //------------------------------------------------------------------------------ |
||
| 331 | // Return a vector of -0.0f (0x80000000),-0.0f,-0.0f,-0.0f |
||
| 332 | XMFINLINE XMVECTOR XMVectorSplatSignMask() |
||
| 333 | { |
||
| 334 | #if defined(_XM_NO_INTRINSICS_) |
||
| 335 | XMVECTOR vResult; |
||
| 336 | vResult.vector4_u32[0] = |
||
| 337 | vResult.vector4_u32[1] = |
||
| 338 | vResult.vector4_u32[2] = |
||
| 339 | vResult.vector4_u32[3] = 0x80000000U; |
||
| 340 | return vResult; |
||
| 341 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 342 | __m128i V = _mm_set1_epi32( 0x80000000 ); |
||
| 343 | return reinterpret_cast<__m128*>(&V)[0]; |
||
| 344 | #else // _XM_VMX128_INTRINSICS_ |
||
| 345 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 346 | } |
||
| 347 | |||
| 348 | //------------------------------------------------------------------------------ |
||
| 349 | // Return a floating point value via an index. This is not a recommended |
||
| 350 | // function to use due to performance loss. |
||
| 351 | XMFINLINE FLOAT XMVectorGetByIndex(FXMVECTOR V,UINT i) |
||
| 352 | { |
||
| 353 | XMASSERT( i <= 3 ); |
||
| 354 | #if defined(_XM_NO_INTRINSICS_) |
||
| 355 | return V.vector4_f32[i]; |
||
| 356 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 357 | return V.m128_f32[i]; |
||
| 358 | #else // _XM_VMX128_INTRINSICS_ |
||
| 359 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 360 | } |
||
| 361 | |||
| 362 | //------------------------------------------------------------------------------ |
||
| 363 | // Return the X component in an FPU register. |
||
| 364 | // This causes Load/Hit/Store on VMX targets |
||
| 365 | XMFINLINE FLOAT XMVectorGetX(FXMVECTOR V) |
||
| 366 | { |
||
| 367 | #if defined(_XM_NO_INTRINSICS_) |
||
| 368 | return V.vector4_f32[0]; |
||
| 369 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 370 | #if defined(_MSC_VER) && (_MSC_VER>=1500) |
||
| 371 | return _mm_cvtss_f32(V); |
||
| 372 | #else |
||
| 373 | return V.m128_f32[0]; |
||
| 374 | #endif |
||
| 375 | #else // _XM_VMX128_INTRINSICS_ |
||
| 376 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 377 | } |
||
| 378 | |||
| 379 | // Return the Y component in an FPU register. |
||
| 380 | // This causes Load/Hit/Store on VMX targets |
||
| 381 | XMFINLINE FLOAT XMVectorGetY(FXMVECTOR V) |
||
| 382 | { |
||
| 383 | #if defined(_XM_NO_INTRINSICS_) |
||
| 384 | return V.vector4_f32[1]; |
||
| 385 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 386 | #if defined(_MSC_VER) && (_MSC_VER>=1500) |
||
| 387 | XMVECTOR vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 388 | return _mm_cvtss_f32(vTemp); |
||
| 389 | #else |
||
| 390 | return V.m128_f32[1]; |
||
| 391 | #endif |
||
| 392 | #else // _XM_VMX128_INTRINSICS_ |
||
| 393 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 394 | } |
||
| 395 | |||
| 396 | // Return the Z component in an FPU register. |
||
| 397 | // This causes Load/Hit/Store on VMX targets |
||
| 398 | XMFINLINE FLOAT XMVectorGetZ(FXMVECTOR V) |
||
| 399 | { |
||
| 400 | #if defined(_XM_NO_INTRINSICS_) |
||
| 401 | return V.vector4_f32[2]; |
||
| 402 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 403 | #if defined(_MSC_VER) && (_MSC_VER>=1500) |
||
| 404 | XMVECTOR vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(2,2,2,2)); |
||
| 405 | return _mm_cvtss_f32(vTemp); |
||
| 406 | #else |
||
| 407 | return V.m128_f32[2]; |
||
| 408 | #endif |
||
| 409 | #else // _XM_VMX128_INTRINSICS_ |
||
| 410 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 411 | } |
||
| 412 | |||
| 413 | // Return the W component in an FPU register. |
||
| 414 | // This causes Load/Hit/Store on VMX targets |
||
| 415 | XMFINLINE FLOAT XMVectorGetW(FXMVECTOR V) |
||
| 416 | { |
||
| 417 | #if defined(_XM_NO_INTRINSICS_) |
||
| 418 | return V.vector4_f32[3]; |
||
| 419 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 420 | #if defined(_MSC_VER) && (_MSC_VER>=1500) |
||
| 421 | XMVECTOR vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,3,3,3)); |
||
| 422 | return _mm_cvtss_f32(vTemp); |
||
| 423 | #else |
||
| 424 | return V.m128_f32[3]; |
||
| 425 | #endif |
||
| 426 | #else // _XM_VMX128_INTRINSICS_ |
||
| 427 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 428 | } |
||
| 429 | |||
| 430 | //------------------------------------------------------------------------------ |
||
| 431 | |||
| 432 | // Store a component indexed by i into a 32 bit float location in memory. |
||
| 433 | // This causes Load/Hit/Store on VMX targets |
||
| 434 | XMFINLINE VOID XMVectorGetByIndexPtr(FLOAT *f,FXMVECTOR V,UINT i) |
||
| 435 | { |
||
| 436 | XMASSERT( f != 0 ); |
||
| 437 | XMASSERT( i < 4 ); |
||
| 438 | #if defined(_XM_NO_INTRINSICS_) |
||
| 439 | *f = V.vector4_f32[i]; |
||
| 440 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 441 | *f = V.m128_f32[i]; |
||
| 442 | #else // _XM_VMX128_INTRINSICS_ |
||
| 443 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 444 | } |
||
| 445 | |||
| 446 | //------------------------------------------------------------------------------ |
||
| 447 | |||
| 448 | // Store the X component into a 32 bit float location in memory. |
||
| 449 | XMFINLINE VOID XMVectorGetXPtr(FLOAT *x,FXMVECTOR V) |
||
| 450 | { |
||
| 451 | XMASSERT( x != 0 ); |
||
| 452 | #if defined(_XM_NO_INTRINSICS_) |
||
| 453 | *x = V.vector4_f32[0]; |
||
| 454 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 455 | _mm_store_ss(x,V); |
||
| 456 | #else // _XM_VMX128_INTRINSICS_ |
||
| 457 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 458 | } |
||
| 459 | |||
| 460 | // Store the Y component into a 32 bit float location in memory. |
||
| 461 | XMFINLINE VOID XMVectorGetYPtr(FLOAT *y,FXMVECTOR V) |
||
| 462 | { |
||
| 463 | XMASSERT( y != 0 ); |
||
| 464 | #if defined(_XM_NO_INTRINSICS_) |
||
| 465 | *y = V.vector4_f32[1]; |
||
| 466 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 467 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 468 | _mm_store_ss(y,vResult); |
||
| 469 | #else // _XM_VMX128_INTRINSICS_ |
||
| 470 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 471 | } |
||
| 472 | |||
| 473 | // Store the Z component into a 32 bit float location in memory. |
||
| 474 | XMFINLINE VOID XMVectorGetZPtr(FLOAT *z,FXMVECTOR V) |
||
| 475 | { |
||
| 476 | XMASSERT( z != 0 ); |
||
| 477 | #if defined(_XM_NO_INTRINSICS_) |
||
| 478 | *z = V.vector4_f32[2]; |
||
| 479 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 480 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(2,2,2,2)); |
||
| 481 | _mm_store_ss(z,vResult); |
||
| 482 | #else // _XM_VMX128_INTRINSICS_ |
||
| 483 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 484 | } |
||
| 485 | |||
| 486 | // Store the W component into a 32 bit float location in memory. |
||
| 487 | XMFINLINE VOID XMVectorGetWPtr(FLOAT *w,FXMVECTOR V) |
||
| 488 | { |
||
| 489 | XMASSERT( w != 0 ); |
||
| 490 | #if defined(_XM_NO_INTRINSICS_) |
||
| 491 | *w = V.vector4_f32[3]; |
||
| 492 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 493 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,3,3,3)); |
||
| 494 | _mm_store_ss(w,vResult); |
||
| 495 | #else // _XM_VMX128_INTRINSICS_ |
||
| 496 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 497 | } |
||
| 498 | |||
| 499 | //------------------------------------------------------------------------------ |
||
| 500 | |||
| 501 | // Return an integer value via an index. This is not a recommended |
||
| 502 | // function to use due to performance loss. |
||
| 503 | XMFINLINE UINT XMVectorGetIntByIndex(FXMVECTOR V, UINT i) |
||
| 504 | { |
||
| 505 | XMASSERT( i < 4 ); |
||
| 506 | #if defined(_XM_NO_INTRINSICS_) |
||
| 507 | return V.vector4_u32[i]; |
||
| 508 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 509 | #if defined(_MSC_VER) && (_MSC_VER<1400) |
||
| 510 | XMVECTORU32 tmp; |
||
| 511 | tmp.v = V; |
||
| 512 | return tmp.u[i]; |
||
| 513 | #else |
||
| 514 | return V.m128_u32[i]; |
||
| 515 | #endif |
||
| 516 | #else // _XM_VMX128_INTRINSICS_ |
||
| 517 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 518 | } |
||
| 519 | |||
| 520 | //------------------------------------------------------------------------------ |
||
| 521 | |||
| 522 | // Return the X component in an integer register. |
||
| 523 | // This causes Load/Hit/Store on VMX targets |
||
| 524 | XMFINLINE UINT XMVectorGetIntX(FXMVECTOR V) |
||
| 525 | { |
||
| 526 | #if defined(_XM_NO_INTRINSICS_) |
||
| 527 | return V.vector4_u32[0]; |
||
| 528 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 529 | return static_cast<UINT>(_mm_cvtsi128_si32(reinterpret_cast<const __m128i *>(&V)[0])); |
||
| 530 | #else // _XM_VMX128_INTRINSICS_ |
||
| 531 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 532 | } |
||
| 533 | |||
| 534 | // Return the Y component in an integer register. |
||
| 535 | // This causes Load/Hit/Store on VMX targets |
||
| 536 | XMFINLINE UINT XMVectorGetIntY(FXMVECTOR V) |
||
| 537 | { |
||
| 538 | #if defined(_XM_NO_INTRINSICS_) |
||
| 539 | return V.vector4_u32[1]; |
||
| 540 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 541 | __m128i vResulti = _mm_shuffle_epi32(reinterpret_cast<const __m128i *>(&V)[0],_MM_SHUFFLE(1,1,1,1)); |
||
| 542 | return static_cast<UINT>(_mm_cvtsi128_si32(vResulti)); |
||
| 543 | #else // _XM_VMX128_INTRINSICS_ |
||
| 544 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 545 | } |
||
| 546 | |||
| 547 | // Return the Z component in an integer register. |
||
| 548 | // This causes Load/Hit/Store on VMX targets |
||
| 549 | XMFINLINE UINT XMVectorGetIntZ(FXMVECTOR V) |
||
| 550 | { |
||
| 551 | #if defined(_XM_NO_INTRINSICS_) |
||
| 552 | return V.vector4_u32[2]; |
||
| 553 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 554 | __m128i vResulti = _mm_shuffle_epi32(reinterpret_cast<const __m128i *>(&V)[0],_MM_SHUFFLE(2,2,2,2)); |
||
| 555 | return static_cast<UINT>(_mm_cvtsi128_si32(vResulti)); |
||
| 556 | #else // _XM_VMX128_INTRINSICS_ |
||
| 557 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 558 | } |
||
| 559 | |||
| 560 | // Return the W component in an integer register. |
||
| 561 | // This causes Load/Hit/Store on VMX targets |
||
| 562 | XMFINLINE UINT XMVectorGetIntW(FXMVECTOR V) |
||
| 563 | { |
||
| 564 | #if defined(_XM_NO_INTRINSICS_) |
||
| 565 | return V.vector4_u32[3]; |
||
| 566 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 567 | __m128i vResulti = _mm_shuffle_epi32(reinterpret_cast<const __m128i *>(&V)[0],_MM_SHUFFLE(3,3,3,3)); |
||
| 568 | return static_cast<UINT>(_mm_cvtsi128_si32(vResulti)); |
||
| 569 | #else // _XM_VMX128_INTRINSICS_ |
||
| 570 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 571 | } |
||
| 572 | |||
| 573 | //------------------------------------------------------------------------------ |
||
| 574 | |||
| 575 | // Store a component indexed by i into a 32 bit integer location in memory. |
||
| 576 | // This causes Load/Hit/Store on VMX targets |
||
| 577 | XMFINLINE VOID XMVectorGetIntByIndexPtr(UINT *x,FXMVECTOR V,UINT i) |
||
| 578 | { |
||
| 579 | XMASSERT( x != 0 ); |
||
| 580 | XMASSERT( i < 4 ); |
||
| 581 | #if defined(_XM_NO_INTRINSICS_) |
||
| 582 | *x = V.vector4_u32[i]; |
||
| 583 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 584 | #if defined(_MSC_VER) && (_MSC_VER<1400) |
||
| 585 | XMVECTORU32 tmp; |
||
| 586 | tmp.v = V; |
||
| 587 | *x = tmp.u[i]; |
||
| 588 | #else |
||
| 589 | *x = V.m128_u32[i]; |
||
| 590 | #endif |
||
| 591 | #else // _XM_VMX128_INTRINSICS_ |
||
| 592 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 593 | } |
||
| 594 | |||
| 595 | //------------------------------------------------------------------------------ |
||
| 596 | |||
| 597 | // Store the X component into a 32 bit integer location in memory. |
||
| 598 | XMFINLINE VOID XMVectorGetIntXPtr(UINT *x,FXMVECTOR V) |
||
| 599 | { |
||
| 600 | XMASSERT( x != 0 ); |
||
| 601 | #if defined(_XM_NO_INTRINSICS_) |
||
| 602 | *x = V.vector4_u32[0]; |
||
| 603 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 604 | _mm_store_ss(reinterpret_cast<float *>(x),V); |
||
| 605 | #else // _XM_VMX128_INTRINSICS_ |
||
| 606 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 607 | } |
||
| 608 | |||
| 609 | // Store the Y component into a 32 bit integer location in memory. |
||
| 610 | XMFINLINE VOID XMVectorGetIntYPtr(UINT *y,FXMVECTOR V) |
||
| 611 | { |
||
| 612 | XMASSERT( y != 0 ); |
||
| 613 | #if defined(_XM_NO_INTRINSICS_) |
||
| 614 | *y = V.vector4_u32[1]; |
||
| 615 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 616 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 617 | _mm_store_ss(reinterpret_cast<float *>(y),vResult); |
||
| 618 | #else // _XM_VMX128_INTRINSICS_ |
||
| 619 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 620 | } |
||
| 621 | |||
| 622 | // Store the Z component into a 32 bit integer locaCantion in memory. |
||
| 623 | XMFINLINE VOID XMVectorGetIntZPtr(UINT *z,FXMVECTOR V) |
||
| 624 | { |
||
| 625 | XMASSERT( z != 0 ); |
||
| 626 | #if defined(_XM_NO_INTRINSICS_) |
||
| 627 | *z = V.vector4_u32[2]; |
||
| 628 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 629 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(2,2,2,2)); |
||
| 630 | _mm_store_ss(reinterpret_cast<float *>(z),vResult); |
||
| 631 | #else // _XM_VMX128_INTRINSICS_ |
||
| 632 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 633 | } |
||
| 634 | |||
| 635 | // Store the W component into a 32 bit integer location in memory. |
||
| 636 | XMFINLINE VOID XMVectorGetIntWPtr(UINT *w,FXMVECTOR V) |
||
| 637 | { |
||
| 638 | XMASSERT( w != 0 ); |
||
| 639 | #if defined(_XM_NO_INTRINSICS_) |
||
| 640 | *w = V.vector4_u32[3]; |
||
| 641 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 642 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,3,3,3)); |
||
| 643 | _mm_store_ss(reinterpret_cast<float *>(w),vResult); |
||
| 644 | #else // _XM_VMX128_INTRINSICS_ |
||
| 645 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 646 | } |
||
| 647 | |||
| 648 | //------------------------------------------------------------------------------ |
||
| 649 | |||
| 650 | // Set a single indexed floating point component |
||
| 651 | // This causes Load/Hit/Store on VMX targets |
||
| 652 | XMFINLINE XMVECTOR XMVectorSetByIndex(FXMVECTOR V, FLOAT f,UINT i) |
||
| 653 | { |
||
| 654 | #if defined(_XM_NO_INTRINSICS_) |
||
| 655 | XMVECTOR U; |
||
| 656 | XMASSERT( i <= 3 ); |
||
| 657 | U = V; |
||
| 658 | U.vector4_f32[i] = f; |
||
| 659 | return U; |
||
| 660 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 661 | XMASSERT( i <= 3 ); |
||
| 662 | XMVECTOR U = V; |
||
| 663 | U.m128_f32[i] = f; |
||
| 664 | return U; |
||
| 665 | #else // _XM_VMX128_INTRINSICS_ |
||
| 666 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 667 | } |
||
| 668 | |||
| 669 | //------------------------------------------------------------------------------ |
||
| 670 | |||
| 671 | // Sets the X component of a vector to a passed floating point value |
||
| 672 | // This causes Load/Hit/Store on VMX targets |
||
| 673 | XMFINLINE XMVECTOR XMVectorSetX(FXMVECTOR V, FLOAT x) |
||
| 674 | { |
||
| 675 | #if defined(_XM_NO_INTRINSICS_) |
||
| 676 | XMVECTOR U; |
||
| 677 | U.vector4_f32[0] = x; |
||
| 678 | U.vector4_f32[1] = V.vector4_f32[1]; |
||
| 679 | U.vector4_f32[2] = V.vector4_f32[2]; |
||
| 680 | U.vector4_f32[3] = V.vector4_f32[3]; |
||
| 681 | return U; |
||
| 682 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 683 | #if defined(_XM_ISVS2005_) |
||
| 684 | XMVECTOR vResult = V; |
||
| 685 | vResult.m128_f32[0] = x; |
||
| 686 | return vResult; |
||
| 687 | #else |
||
| 688 | XMVECTOR vResult = _mm_set_ss(x); |
||
| 689 | vResult = _mm_move_ss(V,vResult); |
||
| 690 | return vResult; |
||
| 691 | #endif // _XM_ISVS2005_ |
||
| 692 | #else // _XM_VMX128_INTRINSICS_ |
||
| 693 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 694 | } |
||
| 695 | |||
| 696 | // Sets the Y component of a vector to a passed floating point value |
||
| 697 | // This causes Load/Hit/Store on VMX targets |
||
| 698 | XMFINLINE XMVECTOR XMVectorSetY(FXMVECTOR V, FLOAT y) |
||
| 699 | { |
||
| 700 | #if defined(_XM_NO_INTRINSICS_) |
||
| 701 | XMVECTOR U; |
||
| 702 | U.vector4_f32[0] = V.vector4_f32[0]; |
||
| 703 | U.vector4_f32[1] = y; |
||
| 704 | U.vector4_f32[2] = V.vector4_f32[2]; |
||
| 705 | U.vector4_f32[3] = V.vector4_f32[3]; |
||
| 706 | return U; |
||
| 707 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 708 | #if defined(_XM_ISVS2005_) |
||
| 709 | XMVECTOR vResult = V; |
||
| 710 | vResult.m128_f32[1] = y; |
||
| 711 | return vResult; |
||
| 712 | #else |
||
| 713 | // Swap y and x |
||
| 714 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,2,0,1)); |
||
| 715 | // Convert input to vector |
||
| 716 | XMVECTOR vTemp = _mm_set_ss(y); |
||
| 717 | // Replace the x component |
||
| 718 | vResult = _mm_move_ss(vResult,vTemp); |
||
| 719 | // Swap y and x again |
||
| 720 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,2,0,1)); |
||
| 721 | return vResult; |
||
| 722 | #endif // _XM_ISVS2005_ |
||
| 723 | #else // _XM_VMX128_INTRINSICS_ |
||
| 724 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 725 | } |
||
| 726 | // Sets the Z component of a vector to a passed floating point value |
||
| 727 | // This causes Load/Hit/Store on VMX targets |
||
| 728 | XMFINLINE XMVECTOR XMVectorSetZ(FXMVECTOR V, FLOAT z) |
||
| 729 | { |
||
| 730 | #if defined(_XM_NO_INTRINSICS_) |
||
| 731 | XMVECTOR U; |
||
| 732 | U.vector4_f32[0] = V.vector4_f32[0]; |
||
| 733 | U.vector4_f32[1] = V.vector4_f32[1]; |
||
| 734 | U.vector4_f32[2] = z; |
||
| 735 | U.vector4_f32[3] = V.vector4_f32[3]; |
||
| 736 | return U; |
||
| 737 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 738 | #if defined(_XM_ISVS2005_) |
||
| 739 | XMVECTOR vResult = V; |
||
| 740 | vResult.m128_f32[2] = z; |
||
| 741 | return vResult; |
||
| 742 | #else |
||
| 743 | // Swap z and x |
||
| 744 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,0,1,2)); |
||
| 745 | // Convert input to vector |
||
| 746 | XMVECTOR vTemp = _mm_set_ss(z); |
||
| 747 | // Replace the x component |
||
| 748 | vResult = _mm_move_ss(vResult,vTemp); |
||
| 749 | // Swap z and x again |
||
| 750 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,0,1,2)); |
||
| 751 | return vResult; |
||
| 752 | #endif // _XM_ISVS2005_ |
||
| 753 | #else // _XM_VMX128_INTRINSICS_ |
||
| 754 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 755 | } |
||
| 756 | |||
| 757 | // Sets the W component of a vector to a passed floating point value |
||
| 758 | // This causes Load/Hit/Store on VMX targets |
||
| 759 | XMFINLINE XMVECTOR XMVectorSetW(FXMVECTOR V, FLOAT w) |
||
| 760 | { |
||
| 761 | #if defined(_XM_NO_INTRINSICS_) |
||
| 762 | XMVECTOR U; |
||
| 763 | U.vector4_f32[0] = V.vector4_f32[0]; |
||
| 764 | U.vector4_f32[1] = V.vector4_f32[1]; |
||
| 765 | U.vector4_f32[2] = V.vector4_f32[2]; |
||
| 766 | U.vector4_f32[3] = w; |
||
| 767 | return U; |
||
| 768 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 769 | #if defined(_XM_ISVS2005_) |
||
| 770 | XMVECTOR vResult = V; |
||
| 771 | vResult.m128_f32[3] = w; |
||
| 772 | return vResult; |
||
| 773 | #else |
||
| 774 | // Swap w and x |
||
| 775 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,2,1,3)); |
||
| 776 | // Convert input to vector |
||
| 777 | XMVECTOR vTemp = _mm_set_ss(w); |
||
| 778 | // Replace the x component |
||
| 779 | vResult = _mm_move_ss(vResult,vTemp); |
||
| 780 | // Swap w and x again |
||
| 781 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(0,2,1,3)); |
||
| 782 | return vResult; |
||
| 783 | #endif // _XM_ISVS2005_ |
||
| 784 | #else // _XM_VMX128_INTRINSICS_ |
||
| 785 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 786 | } |
||
| 787 | |||
| 788 | //------------------------------------------------------------------------------ |
||
| 789 | |||
| 790 | // Sets a component of a vector to a floating point value passed by pointer |
||
| 791 | // This causes Load/Hit/Store on VMX targets |
||
| 792 | XMFINLINE XMVECTOR XMVectorSetByIndexPtr(FXMVECTOR V,CONST FLOAT *f,UINT i) |
||
| 793 | { |
||
| 794 | #if defined(_XM_NO_INTRINSICS_) |
||
| 795 | XMVECTOR U; |
||
| 796 | XMASSERT( f != 0 ); |
||
| 797 | XMASSERT( i <= 3 ); |
||
| 798 | U = V; |
||
| 799 | U.vector4_f32[i] = *f; |
||
| 800 | return U; |
||
| 801 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 802 | XMASSERT( f != 0 ); |
||
| 803 | XMASSERT( i <= 3 ); |
||
| 804 | XMVECTOR U = V; |
||
| 805 | U.m128_f32[i] = *f; |
||
| 806 | return U; |
||
| 807 | #else // _XM_VMX128_INTRINSICS_ |
||
| 808 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 809 | } |
||
| 810 | |||
| 811 | //------------------------------------------------------------------------------ |
||
| 812 | |||
| 813 | // Sets the X component of a vector to a floating point value passed by pointer |
||
| 814 | XMFINLINE XMVECTOR XMVectorSetXPtr(FXMVECTOR V,CONST FLOAT *x) |
||
| 815 | { |
||
| 816 | #if defined(_XM_NO_INTRINSICS_) |
||
| 817 | XMVECTOR U; |
||
| 818 | XMASSERT( x != 0 ); |
||
| 819 | U.vector4_f32[0] = *x; |
||
| 820 | U.vector4_f32[1] = V.vector4_f32[1]; |
||
| 821 | U.vector4_f32[2] = V.vector4_f32[2]; |
||
| 822 | U.vector4_f32[3] = V.vector4_f32[3]; |
||
| 823 | return U; |
||
| 824 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 825 | XMASSERT( x != 0 ); |
||
| 826 | XMVECTOR vResult = _mm_load_ss(x); |
||
| 827 | vResult = _mm_move_ss(V,vResult); |
||
| 828 | return vResult; |
||
| 829 | #else // _XM_VMX128_INTRINSICS_ |
||
| 830 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 831 | } |
||
| 832 | |||
| 833 | // Sets the Y component of a vector to a floating point value passed by pointer |
||
| 834 | XMFINLINE XMVECTOR XMVectorSetYPtr(FXMVECTOR V,CONST FLOAT *y) |
||
| 835 | { |
||
| 836 | #if defined(_XM_NO_INTRINSICS_) |
||
| 837 | XMVECTOR U; |
||
| 838 | XMASSERT( y != 0 ); |
||
| 839 | U.vector4_f32[0] = V.vector4_f32[0]; |
||
| 840 | U.vector4_f32[1] = *y; |
||
| 841 | U.vector4_f32[2] = V.vector4_f32[2]; |
||
| 842 | U.vector4_f32[3] = V.vector4_f32[3]; |
||
| 843 | return U; |
||
| 844 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 845 | XMASSERT( y != 0 ); |
||
| 846 | // Swap y and x |
||
| 847 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,2,0,1)); |
||
| 848 | // Convert input to vector |
||
| 849 | XMVECTOR vTemp = _mm_load_ss(y); |
||
| 850 | // Replace the x component |
||
| 851 | vResult = _mm_move_ss(vResult,vTemp); |
||
| 852 | // Swap y and x again |
||
| 853 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,2,0,1)); |
||
| 854 | return vResult; |
||
| 855 | #else // _XM_VMX128_INTRINSICS_ |
||
| 856 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 857 | } |
||
| 858 | |||
| 859 | // Sets the Z component of a vector to a floating point value passed by pointer |
||
| 860 | XMFINLINE XMVECTOR XMVectorSetZPtr(FXMVECTOR V,CONST FLOAT *z) |
||
| 861 | { |
||
| 862 | #if defined(_XM_NO_INTRINSICS_) |
||
| 863 | XMVECTOR U; |
||
| 864 | XMASSERT( z != 0 ); |
||
| 865 | U.vector4_f32[0] = V.vector4_f32[0]; |
||
| 866 | U.vector4_f32[1] = V.vector4_f32[1]; |
||
| 867 | U.vector4_f32[2] = *z; |
||
| 868 | U.vector4_f32[3] = V.vector4_f32[3]; |
||
| 869 | return U; |
||
| 870 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 871 | XMASSERT( z != 0 ); |
||
| 872 | // Swap z and x |
||
| 873 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,0,1,2)); |
||
| 874 | // Convert input to vector |
||
| 875 | XMVECTOR vTemp = _mm_load_ss(z); |
||
| 876 | // Replace the x component |
||
| 877 | vResult = _mm_move_ss(vResult,vTemp); |
||
| 878 | // Swap z and x again |
||
| 879 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,0,1,2)); |
||
| 880 | return vResult; |
||
| 881 | #else // _XM_VMX128_INTRINSICS_ |
||
| 882 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 883 | } |
||
| 884 | |||
| 885 | // Sets the W component of a vector to a floating point value passed by pointer |
||
| 886 | XMFINLINE XMVECTOR XMVectorSetWPtr(FXMVECTOR V,CONST FLOAT *w) |
||
| 887 | { |
||
| 888 | #if defined(_XM_NO_INTRINSICS_) |
||
| 889 | XMVECTOR U; |
||
| 890 | XMASSERT( w != 0 ); |
||
| 891 | U.vector4_f32[0] = V.vector4_f32[0]; |
||
| 892 | U.vector4_f32[1] = V.vector4_f32[1]; |
||
| 893 | U.vector4_f32[2] = V.vector4_f32[2]; |
||
| 894 | U.vector4_f32[3] = *w; |
||
| 895 | return U; |
||
| 896 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 897 | XMASSERT( w != 0 ); |
||
| 898 | // Swap w and x |
||
| 899 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,2,1,3)); |
||
| 900 | // Convert input to vector |
||
| 901 | XMVECTOR vTemp = _mm_load_ss(w); |
||
| 902 | // Replace the x component |
||
| 903 | vResult = _mm_move_ss(vResult,vTemp); |
||
| 904 | // Swap w and x again |
||
| 905 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(0,2,1,3)); |
||
| 906 | return vResult; |
||
| 907 | #else // _XM_VMX128_INTRINSICS_ |
||
| 908 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 909 | } |
||
| 910 | |||
| 911 | //------------------------------------------------------------------------------ |
||
| 912 | |||
| 913 | // Sets a component of a vector to an integer passed by value |
||
| 914 | // This causes Load/Hit/Store on VMX targets |
||
| 915 | XMFINLINE XMVECTOR XMVectorSetIntByIndex(FXMVECTOR V, UINT x, UINT i) |
||
| 916 | { |
||
| 917 | #if defined(_XM_NO_INTRINSICS_) |
||
| 918 | XMVECTOR U; |
||
| 919 | XMASSERT( i <= 3 ); |
||
| 920 | U = V; |
||
| 921 | U.vector4_u32[i] = x; |
||
| 922 | return U; |
||
| 923 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 924 | XMASSERT( i <= 3 ); |
||
| 925 | XMVECTORU32 tmp; |
||
| 926 | tmp.v = V; |
||
| 927 | tmp.u[i] = x; |
||
| 928 | return tmp; |
||
| 929 | #else // _XM_VMX128_INTRINSICS_ |
||
| 930 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 931 | } |
||
| 932 | |||
| 933 | //------------------------------------------------------------------------------ |
||
| 934 | |||
| 935 | // Sets the X component of a vector to an integer passed by value |
||
| 936 | // This causes Load/Hit/Store on VMX targets |
||
| 937 | XMFINLINE XMVECTOR XMVectorSetIntX(FXMVECTOR V, UINT x) |
||
| 938 | { |
||
| 939 | #if defined(_XM_NO_INTRINSICS_) |
||
| 940 | XMVECTOR U; |
||
| 941 | U.vector4_u32[0] = x; |
||
| 942 | U.vector4_u32[1] = V.vector4_u32[1]; |
||
| 943 | U.vector4_u32[2] = V.vector4_u32[2]; |
||
| 944 | U.vector4_u32[3] = V.vector4_u32[3]; |
||
| 945 | return U; |
||
| 946 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 947 | #if defined(_XM_ISVS2005_) |
||
| 948 | XMVECTOR vResult = V; |
||
| 949 | vResult.m128_i32[0] = x; |
||
| 950 | return vResult; |
||
| 951 | #else |
||
| 952 | __m128i vTemp = _mm_cvtsi32_si128(x); |
||
| 953 | XMVECTOR vResult = _mm_move_ss(V,reinterpret_cast<const __m128 *>(&vTemp)[0]); |
||
| 954 | return vResult; |
||
| 955 | #endif // _XM_ISVS2005_ |
||
| 956 | #else // _XM_VMX128_INTRINSICS_ |
||
| 957 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 958 | } |
||
| 959 | |||
| 960 | // Sets the Y component of a vector to an integer passed by value |
||
| 961 | // This causes Load/Hit/Store on VMX targets |
||
| 962 | XMFINLINE XMVECTOR XMVectorSetIntY(FXMVECTOR V, UINT y) |
||
| 963 | { |
||
| 964 | #if defined(_XM_NO_INTRINSICS_) |
||
| 965 | XMVECTOR U; |
||
| 966 | U.vector4_u32[0] = V.vector4_u32[0]; |
||
| 967 | U.vector4_u32[1] = y; |
||
| 968 | U.vector4_u32[2] = V.vector4_u32[2]; |
||
| 969 | U.vector4_u32[3] = V.vector4_u32[3]; |
||
| 970 | return U; |
||
| 971 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 972 | #if defined(_XM_ISVS2005_) |
||
| 973 | XMVECTOR vResult = V; |
||
| 974 | vResult.m128_i32[1] = y; |
||
| 975 | return vResult; |
||
| 976 | #else // Swap y and x |
||
| 977 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,2,0,1)); |
||
| 978 | // Convert input to vector |
||
| 979 | __m128i vTemp = _mm_cvtsi32_si128(y); |
||
| 980 | // Replace the x component |
||
| 981 | vResult = _mm_move_ss(vResult,reinterpret_cast<const __m128 *>(&vTemp)[0]); |
||
| 982 | // Swap y and x again |
||
| 983 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,2,0,1)); |
||
| 984 | return vResult; |
||
| 985 | #endif // _XM_ISVS2005_ |
||
| 986 | #else // _XM_VMX128_INTRINSICS_ |
||
| 987 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 988 | } |
||
| 989 | |||
| 990 | // Sets the Z component of a vector to an integer passed by value |
||
| 991 | // This causes Load/Hit/Store on VMX targets |
||
| 992 | XMFINLINE XMVECTOR XMVectorSetIntZ(FXMVECTOR V, UINT z) |
||
| 993 | { |
||
| 994 | #if defined(_XM_NO_INTRINSICS_) |
||
| 995 | XMVECTOR U; |
||
| 996 | U.vector4_u32[0] = V.vector4_u32[0]; |
||
| 997 | U.vector4_u32[1] = V.vector4_u32[1]; |
||
| 998 | U.vector4_u32[2] = z; |
||
| 999 | U.vector4_u32[3] = V.vector4_u32[3]; |
||
| 1000 | return U; |
||
| 1001 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1002 | #if defined(_XM_ISVS2005_) |
||
| 1003 | XMVECTOR vResult = V; |
||
| 1004 | vResult.m128_i32[2] = z; |
||
| 1005 | return vResult; |
||
| 1006 | #else |
||
| 1007 | // Swap z and x |
||
| 1008 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,0,1,2)); |
||
| 1009 | // Convert input to vector |
||
| 1010 | __m128i vTemp = _mm_cvtsi32_si128(z); |
||
| 1011 | // Replace the x component |
||
| 1012 | vResult = _mm_move_ss(vResult,reinterpret_cast<const __m128 *>(&vTemp)[0]); |
||
| 1013 | // Swap z and x again |
||
| 1014 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,0,1,2)); |
||
| 1015 | return vResult; |
||
| 1016 | #endif // _XM_ISVS2005_ |
||
| 1017 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1018 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1019 | } |
||
| 1020 | |||
| 1021 | // Sets the W component of a vector to an integer passed by value |
||
| 1022 | // This causes Load/Hit/Store on VMX targets |
||
| 1023 | XMFINLINE XMVECTOR XMVectorSetIntW(FXMVECTOR V, UINT w) |
||
| 1024 | { |
||
| 1025 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1026 | XMVECTOR U; |
||
| 1027 | U.vector4_u32[0] = V.vector4_u32[0]; |
||
| 1028 | U.vector4_u32[1] = V.vector4_u32[1]; |
||
| 1029 | U.vector4_u32[2] = V.vector4_u32[2]; |
||
| 1030 | U.vector4_u32[3] = w; |
||
| 1031 | return U; |
||
| 1032 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1033 | #if defined(_XM_ISVS2005_) |
||
| 1034 | XMVECTOR vResult = V; |
||
| 1035 | vResult.m128_i32[3] = w; |
||
| 1036 | return vResult; |
||
| 1037 | #else |
||
| 1038 | // Swap w and x |
||
| 1039 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,2,1,3)); |
||
| 1040 | // Convert input to vector |
||
| 1041 | __m128i vTemp = _mm_cvtsi32_si128(w); |
||
| 1042 | // Replace the x component |
||
| 1043 | vResult = _mm_move_ss(vResult,reinterpret_cast<const __m128 *>(&vTemp)[0]); |
||
| 1044 | // Swap w and x again |
||
| 1045 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(0,2,1,3)); |
||
| 1046 | return vResult; |
||
| 1047 | #endif // _XM_ISVS2005_ |
||
| 1048 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1049 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1050 | } |
||
| 1051 | |||
| 1052 | //------------------------------------------------------------------------------ |
||
| 1053 | |||
| 1054 | // Sets a component of a vector to an integer value passed by pointer |
||
| 1055 | // This causes Load/Hit/Store on VMX targets |
||
| 1056 | XMFINLINE XMVECTOR XMVectorSetIntByIndexPtr(FXMVECTOR V, CONST UINT *x,UINT i) |
||
| 1057 | { |
||
| 1058 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1059 | XMVECTOR U; |
||
| 1060 | XMASSERT( x != 0 ); |
||
| 1061 | XMASSERT( i <= 3 ); |
||
| 1062 | U = V; |
||
| 1063 | U.vector4_u32[i] = *x; |
||
| 1064 | return U; |
||
| 1065 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1066 | XMASSERT( x != 0 ); |
||
| 1067 | XMASSERT( i <= 3 ); |
||
| 1068 | XMVECTORU32 tmp; |
||
| 1069 | tmp.v = V; |
||
| 1070 | tmp.u[i] = *x; |
||
| 1071 | return tmp; |
||
| 1072 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1073 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1074 | } |
||
| 1075 | |||
| 1076 | //------------------------------------------------------------------------------ |
||
| 1077 | |||
| 1078 | // Sets the X component of a vector to an integer value passed by pointer |
||
| 1079 | XMFINLINE XMVECTOR XMVectorSetIntXPtr(FXMVECTOR V,CONST UINT *x) |
||
| 1080 | { |
||
| 1081 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1082 | XMVECTOR U; |
||
| 1083 | XMASSERT( x != 0 ); |
||
| 1084 | U.vector4_u32[0] = *x; |
||
| 1085 | U.vector4_u32[1] = V.vector4_u32[1]; |
||
| 1086 | U.vector4_u32[2] = V.vector4_u32[2]; |
||
| 1087 | U.vector4_u32[3] = V.vector4_u32[3]; |
||
| 1088 | return U; |
||
| 1089 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1090 | XMASSERT( x != 0 ); |
||
| 1091 | XMVECTOR vTemp = _mm_load_ss(reinterpret_cast<const float *>(x)); |
||
| 1092 | XMVECTOR vResult = _mm_move_ss(V,vTemp); |
||
| 1093 | return vResult; |
||
| 1094 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1095 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1096 | } |
||
| 1097 | |||
| 1098 | // Sets the Y component of a vector to an integer value passed by pointer |
||
| 1099 | XMFINLINE XMVECTOR XMVectorSetIntYPtr(FXMVECTOR V,CONST UINT *y) |
||
| 1100 | { |
||
| 1101 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1102 | XMVECTOR U; |
||
| 1103 | XMASSERT( y != 0 ); |
||
| 1104 | U.vector4_u32[0] = V.vector4_u32[0]; |
||
| 1105 | U.vector4_u32[1] = *y; |
||
| 1106 | U.vector4_u32[2] = V.vector4_u32[2]; |
||
| 1107 | U.vector4_u32[3] = V.vector4_u32[3]; |
||
| 1108 | return U; |
||
| 1109 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1110 | XMASSERT( y != 0 ); |
||
| 1111 | // Swap y and x |
||
| 1112 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,2,0,1)); |
||
| 1113 | // Convert input to vector |
||
| 1114 | XMVECTOR vTemp = _mm_load_ss(reinterpret_cast<const float *>(y)); |
||
| 1115 | // Replace the x component |
||
| 1116 | vResult = _mm_move_ss(vResult,vTemp); |
||
| 1117 | // Swap y and x again |
||
| 1118 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,2,0,1)); |
||
| 1119 | return vResult; |
||
| 1120 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1121 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1122 | } |
||
| 1123 | |||
| 1124 | // Sets the Z component of a vector to an integer value passed by pointer |
||
| 1125 | XMFINLINE XMVECTOR XMVectorSetIntZPtr(FXMVECTOR V,CONST UINT *z) |
||
| 1126 | { |
||
| 1127 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1128 | XMVECTOR U; |
||
| 1129 | XMASSERT( z != 0 ); |
||
| 1130 | U.vector4_u32[0] = V.vector4_u32[0]; |
||
| 1131 | U.vector4_u32[1] = V.vector4_u32[1]; |
||
| 1132 | U.vector4_u32[2] = *z; |
||
| 1133 | U.vector4_u32[3] = V.vector4_u32[3]; |
||
| 1134 | return U; |
||
| 1135 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1136 | XMASSERT( z != 0 ); |
||
| 1137 | // Swap z and x |
||
| 1138 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,0,1,2)); |
||
| 1139 | // Convert input to vector |
||
| 1140 | XMVECTOR vTemp = _mm_load_ss(reinterpret_cast<const float *>(z)); |
||
| 1141 | // Replace the x component |
||
| 1142 | vResult = _mm_move_ss(vResult,vTemp); |
||
| 1143 | // Swap z and x again |
||
| 1144 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,0,1,2)); |
||
| 1145 | return vResult; |
||
| 1146 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1147 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1148 | } |
||
| 1149 | |||
| 1150 | // Sets the W component of a vector to an integer value passed by pointer |
||
| 1151 | XMFINLINE XMVECTOR XMVectorSetIntWPtr(FXMVECTOR V,CONST UINT *w) |
||
| 1152 | { |
||
| 1153 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1154 | XMVECTOR U; |
||
| 1155 | XMASSERT( w != 0 ); |
||
| 1156 | U.vector4_u32[0] = V.vector4_u32[0]; |
||
| 1157 | U.vector4_u32[1] = V.vector4_u32[1]; |
||
| 1158 | U.vector4_u32[2] = V.vector4_u32[2]; |
||
| 1159 | U.vector4_u32[3] = *w; |
||
| 1160 | return U; |
||
| 1161 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1162 | XMASSERT( w != 0 ); |
||
| 1163 | // Swap w and x |
||
| 1164 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,2,1,3)); |
||
| 1165 | // Convert input to vector |
||
| 1166 | XMVECTOR vTemp = _mm_load_ss(reinterpret_cast<const float *>(w)); |
||
| 1167 | // Replace the x component |
||
| 1168 | vResult = _mm_move_ss(vResult,vTemp); |
||
| 1169 | // Swap w and x again |
||
| 1170 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(0,2,1,3)); |
||
| 1171 | return vResult; |
||
| 1172 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1173 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1174 | } |
||
| 1175 | |||
| 1176 | //------------------------------------------------------------------------------ |
||
| 1177 | // Define a control vector to be used in XMVectorPermute |
||
| 1178 | // operations. Visualize the two vectors V1 and V2 given |
||
| 1179 | // in a permute as arranged back to back in a linear fashion, |
||
| 1180 | // such that they form an array of 8 floating point values. |
||
| 1181 | // The four integers specified in XMVectorPermuteControl |
||
| 1182 | // will serve as indices into the array to select components |
||
| 1183 | // from the two vectors. ElementIndex0 is used to select |
||
| 1184 | // an element from the vectors to be placed in the first |
||
| 1185 | // component of the resulting vector, ElementIndex1 is used |
||
| 1186 | // to select an element for the second component, etc. |
||
| 1187 | |||
| 1188 | XMFINLINE XMVECTOR XMVectorPermuteControl |
||
| 1189 | ( |
||
| 1190 | UINT ElementIndex0, |
||
| 1191 | UINT ElementIndex1, |
||
| 1192 | UINT ElementIndex2, |
||
| 1193 | UINT ElementIndex3 |
||
| 1194 | ) |
||
| 1195 | { |
||
| 1196 | #if defined(_XM_SSE_INTRINSICS_) || defined(_XM_NO_INTRINSICS_) |
||
| 1197 | XMVECTORU32 vControl; |
||
| 1198 | static CONST UINT ControlElement[] = { |
||
| 1199 | XM_PERMUTE_0X, |
||
| 1200 | XM_PERMUTE_0Y, |
||
| 1201 | XM_PERMUTE_0Z, |
||
| 1202 | XM_PERMUTE_0W, |
||
| 1203 | XM_PERMUTE_1X, |
||
| 1204 | XM_PERMUTE_1Y, |
||
| 1205 | XM_PERMUTE_1Z, |
||
| 1206 | XM_PERMUTE_1W |
||
| 1207 | }; |
||
| 1208 | XMASSERT(ElementIndex0 < 8); |
||
| 1209 | XMASSERT(ElementIndex1 < 8); |
||
| 1210 | XMASSERT(ElementIndex2 < 8); |
||
| 1211 | XMASSERT(ElementIndex3 < 8); |
||
| 1212 | |||
| 1213 | vControl.u[0] = ControlElement[ElementIndex0]; |
||
| 1214 | vControl.u[1] = ControlElement[ElementIndex1]; |
||
| 1215 | vControl.u[2] = ControlElement[ElementIndex2]; |
||
| 1216 | vControl.u[3] = ControlElement[ElementIndex3]; |
||
| 1217 | return vControl.v; |
||
| 1218 | #else |
||
| 1219 | #endif |
||
| 1220 | } |
||
| 1221 | |||
| 1222 | //------------------------------------------------------------------------------ |
||
| 1223 | |||
| 1224 | // Using a control vector made up of 16 bytes from 0-31, remap V1 and V2's byte |
||
| 1225 | // entries into a single 16 byte vector and return it. Index 0-15 = V1, |
||
| 1226 | // 16-31 = V2 |
||
| 1227 | XMFINLINE XMVECTOR XMVectorPermute |
||
| 1228 | ( |
||
| 1229 | FXMVECTOR V1, |
||
| 1230 | FXMVECTOR V2, |
||
| 1231 | FXMVECTOR Control |
||
| 1232 | ) |
||
| 1233 | { |
||
| 1234 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1235 | const BYTE *aByte[2]; |
||
| 1236 | XMVECTOR Result; |
||
| 1237 | UINT i, uIndex, VectorIndex; |
||
| 1238 | const BYTE *pControl; |
||
| 1239 | BYTE *pWork; |
||
| 1240 | |||
| 1241 | // Indices must be in range from 0 to 31 |
||
| 1242 | XMASSERT((Control.vector4_u32[0] & 0xE0E0E0E0) == 0); |
||
| 1243 | XMASSERT((Control.vector4_u32[1] & 0xE0E0E0E0) == 0); |
||
| 1244 | XMASSERT((Control.vector4_u32[2] & 0xE0E0E0E0) == 0); |
||
| 1245 | XMASSERT((Control.vector4_u32[3] & 0xE0E0E0E0) == 0); |
||
| 1246 | |||
| 1247 | // 0-15 = V1, 16-31 = V2 |
||
| 1248 | aByte[0] = (const BYTE*)(&V1); |
||
| 1249 | aByte[1] = (const BYTE*)(&V2); |
||
| 1250 | i = 16; |
||
| 1251 | pControl = (const BYTE *)(&Control); |
||
| 1252 | pWork = (BYTE *)(&Result); |
||
| 1253 | do { |
||
| 1254 | // Get the byte to map from |
||
| 1255 | uIndex = pControl[0]; |
||
| 1256 | ++pControl; |
||
| 1257 | VectorIndex = (uIndex>>4)&1; |
||
| 1258 | uIndex &= 0x0F; |
||
| 1259 | #if defined(_XM_X86_) || defined(_XM_X64_) |
||
| 1260 | uIndex ^= 3; // Swap byte ordering on little endian machines |
||
| 1261 | #endif |
||
| 1262 | pWork[0] = aByte[VectorIndex][uIndex]; |
||
| 1263 | ++pWork; |
||
| 1264 | } while (--i); |
||
| 1265 | return Result; |
||
| 1266 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1267 | #if defined(_PREFAST_) || defined(XMDEBUG) |
||
| 1268 | // Indices must be in range from 0 to 31 |
||
| 1269 | static const XMVECTORI32 PremuteTest = {0xE0E0E0E0,0xE0E0E0E0,0xE0E0E0E0,0xE0E0E0E0}; |
||
| 1270 | XMVECTOR vAssert = _mm_and_ps(Control,PremuteTest); |
||
| 1271 | __m128i vAsserti = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&vAssert)[0],g_XMZero); |
||
| 1272 | XMASSERT(_mm_movemask_ps(*reinterpret_cast<const __m128 *>(&vAsserti)) == 0xf); |
||
| 1273 | #endif |
||
| 1274 | // Store the vectors onto local memory on the stack |
||
| 1275 | XMVECTOR Array[2]; |
||
| 1276 | Array[0] = V1; |
||
| 1277 | Array[1] = V2; |
||
| 1278 | // Output vector, on the stack |
||
| 1279 | XMVECTORU8 vResult; |
||
| 1280 | // Get pointer to the two vectors on the stack |
||
| 1281 | const BYTE *pInput = reinterpret_cast<const BYTE *>(Array); |
||
| 1282 | // Store the Control vector on the stack to access the bytes |
||
| 1283 | // don't use Control, it can cause a register variable to spill on the stack. |
||
| 1284 | XMVECTORU8 vControl; |
||
| 1285 | vControl.v = Control; // Write to memory |
||
| 1286 | UINT i = 0; |
||
| 1287 | do { |
||
| 1288 | UINT ComponentIndex = vControl.u[i] & 0x1FU; |
||
| 1289 | ComponentIndex ^= 3; // Swap byte ordering |
||
| 1290 | vResult.u[i] = pInput[ComponentIndex]; |
||
| 1291 | } while (++i<16); |
||
| 1292 | return vResult; |
||
| 1293 | #else // _XM_SSE_INTRINSICS_ |
||
| 1294 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1295 | } |
||
| 1296 | |||
| 1297 | //------------------------------------------------------------------------------ |
||
| 1298 | // Define a control vector to be used in XMVectorSelect |
||
| 1299 | // operations. The four integers specified in XMVectorSelectControl |
||
| 1300 | // serve as indices to select between components in two vectors. |
||
| 1301 | // The first index controls selection for the first component of |
||
| 1302 | // the vectors involved in a select operation, the second index |
||
| 1303 | // controls selection for the second component etc. A value of |
||
| 1304 | // zero for an index causes the corresponding component from the first |
||
| 1305 | // vector to be selected whereas a one causes the component from the |
||
| 1306 | // second vector to be selected instead. |
||
| 1307 | |||
| 1308 | XMFINLINE XMVECTOR XMVectorSelectControl |
||
| 1309 | ( |
||
| 1310 | UINT VectorIndex0, |
||
| 1311 | UINT VectorIndex1, |
||
| 1312 | UINT VectorIndex2, |
||
| 1313 | UINT VectorIndex3 |
||
| 1314 | ) |
||
| 1315 | { |
||
| 1316 | #if defined(_XM_SSE_INTRINSICS_) && !defined(_XM_NO_INTRINSICS_) |
||
| 1317 | // x=Index0,y=Index1,z=Index2,w=Index3 |
||
| 1318 | __m128i vTemp = _mm_set_epi32(VectorIndex3,VectorIndex2,VectorIndex1,VectorIndex0); |
||
| 1319 | // Any non-zero entries become 0xFFFFFFFF else 0 |
||
| 1320 | vTemp = _mm_cmpgt_epi32(vTemp,g_XMZero); |
||
| 1321 | return reinterpret_cast<__m128 *>(&vTemp)[0]; |
||
| 1322 | #else |
||
| 1323 | XMVECTOR ControlVector; |
||
| 1324 | CONST UINT ControlElement[] = |
||
| 1325 | { |
||
| 1326 | XM_SELECT_0, |
||
| 1327 | XM_SELECT_1 |
||
| 1328 | }; |
||
| 1329 | |||
| 1330 | XMASSERT(VectorIndex0 < 2); |
||
| 1331 | XMASSERT(VectorIndex1 < 2); |
||
| 1332 | XMASSERT(VectorIndex2 < 2); |
||
| 1333 | XMASSERT(VectorIndex3 < 2); |
||
| 1334 | |||
| 1335 | ControlVector.vector4_u32[0] = ControlElement[VectorIndex0]; |
||
| 1336 | ControlVector.vector4_u32[1] = ControlElement[VectorIndex1]; |
||
| 1337 | ControlVector.vector4_u32[2] = ControlElement[VectorIndex2]; |
||
| 1338 | ControlVector.vector4_u32[3] = ControlElement[VectorIndex3]; |
||
| 1339 | |||
| 1340 | return ControlVector; |
||
| 1341 | |||
| 1342 | #endif |
||
| 1343 | } |
||
| 1344 | |||
| 1345 | //------------------------------------------------------------------------------ |
||
| 1346 | |||
| 1347 | XMFINLINE XMVECTOR XMVectorSelect |
||
| 1348 | ( |
||
| 1349 | FXMVECTOR V1, |
||
| 1350 | FXMVECTOR V2, |
||
| 1351 | FXMVECTOR Control |
||
| 1352 | ) |
||
| 1353 | { |
||
| 1354 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1355 | |||
| 1356 | XMVECTOR Result; |
||
| 1357 | |||
| 1358 | Result.vector4_u32[0] = (V1.vector4_u32[0] & ~Control.vector4_u32[0]) | (V2.vector4_u32[0] & Control.vector4_u32[0]); |
||
| 1359 | Result.vector4_u32[1] = (V1.vector4_u32[1] & ~Control.vector4_u32[1]) | (V2.vector4_u32[1] & Control.vector4_u32[1]); |
||
| 1360 | Result.vector4_u32[2] = (V1.vector4_u32[2] & ~Control.vector4_u32[2]) | (V2.vector4_u32[2] & Control.vector4_u32[2]); |
||
| 1361 | Result.vector4_u32[3] = (V1.vector4_u32[3] & ~Control.vector4_u32[3]) | (V2.vector4_u32[3] & Control.vector4_u32[3]); |
||
| 1362 | |||
| 1363 | return Result; |
||
| 1364 | |||
| 1365 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1366 | XMVECTOR vTemp1 = _mm_andnot_ps(Control,V1); |
||
| 1367 | XMVECTOR vTemp2 = _mm_and_ps(V2,Control); |
||
| 1368 | return _mm_or_ps(vTemp1,vTemp2); |
||
| 1369 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1370 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1371 | } |
||
| 1372 | |||
| 1373 | //------------------------------------------------------------------------------ |
||
| 1374 | |||
| 1375 | XMFINLINE XMVECTOR XMVectorMergeXY |
||
| 1376 | ( |
||
| 1377 | FXMVECTOR V1, |
||
| 1378 | FXMVECTOR V2 |
||
| 1379 | ) |
||
| 1380 | { |
||
| 1381 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1382 | |||
| 1383 | XMVECTOR Result; |
||
| 1384 | |||
| 1385 | Result.vector4_u32[0] = V1.vector4_u32[0]; |
||
| 1386 | Result.vector4_u32[1] = V2.vector4_u32[0]; |
||
| 1387 | Result.vector4_u32[2] = V1.vector4_u32[1]; |
||
| 1388 | Result.vector4_u32[3] = V2.vector4_u32[1]; |
||
| 1389 | |||
| 1390 | return Result; |
||
| 1391 | |||
| 1392 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1393 | return _mm_unpacklo_ps( V1, V2 ); |
||
| 1394 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1395 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1396 | } |
||
| 1397 | |||
| 1398 | //------------------------------------------------------------------------------ |
||
| 1399 | |||
| 1400 | XMFINLINE XMVECTOR XMVectorMergeZW |
||
| 1401 | ( |
||
| 1402 | FXMVECTOR V1, |
||
| 1403 | FXMVECTOR V2 |
||
| 1404 | ) |
||
| 1405 | { |
||
| 1406 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1407 | |||
| 1408 | XMVECTOR Result; |
||
| 1409 | |||
| 1410 | Result.vector4_u32[0] = V1.vector4_u32[2]; |
||
| 1411 | Result.vector4_u32[1] = V2.vector4_u32[2]; |
||
| 1412 | Result.vector4_u32[2] = V1.vector4_u32[3]; |
||
| 1413 | Result.vector4_u32[3] = V2.vector4_u32[3]; |
||
| 1414 | |||
| 1415 | return Result; |
||
| 1416 | |||
| 1417 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1418 | return _mm_unpackhi_ps( V1, V2 ); |
||
| 1419 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1420 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1421 | } |
||
| 1422 | |||
| 1423 | //------------------------------------------------------------------------------ |
||
| 1424 | // Comparison operations |
||
| 1425 | //------------------------------------------------------------------------------ |
||
| 1426 | |||
| 1427 | //------------------------------------------------------------------------------ |
||
| 1428 | |||
| 1429 | XMFINLINE XMVECTOR XMVectorEqual |
||
| 1430 | ( |
||
| 1431 | FXMVECTOR V1, |
||
| 1432 | FXMVECTOR V2 |
||
| 1433 | ) |
||
| 1434 | { |
||
| 1435 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1436 | |||
| 1437 | XMVECTOR Control; |
||
| 1438 | |||
| 1439 | Control.vector4_u32[0] = (V1.vector4_f32[0] == V2.vector4_f32[0]) ? 0xFFFFFFFF : 0; |
||
| 1440 | Control.vector4_u32[1] = (V1.vector4_f32[1] == V2.vector4_f32[1]) ? 0xFFFFFFFF : 0; |
||
| 1441 | Control.vector4_u32[2] = (V1.vector4_f32[2] == V2.vector4_f32[2]) ? 0xFFFFFFFF : 0; |
||
| 1442 | Control.vector4_u32[3] = (V1.vector4_f32[3] == V2.vector4_f32[3]) ? 0xFFFFFFFF : 0; |
||
| 1443 | |||
| 1444 | return Control; |
||
| 1445 | |||
| 1446 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1447 | return _mm_cmpeq_ps( V1, V2 ); |
||
| 1448 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1449 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1450 | } |
||
| 1451 | |||
| 1452 | //------------------------------------------------------------------------------ |
||
| 1453 | |||
| 1454 | XMFINLINE XMVECTOR XMVectorEqualR |
||
| 1455 | ( |
||
| 1456 | UINT* pCR, |
||
| 1457 | FXMVECTOR V1, |
||
| 1458 | FXMVECTOR V2 |
||
| 1459 | ) |
||
| 1460 | { |
||
| 1461 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1462 | UINT ux, uy, uz, uw, CR; |
||
| 1463 | XMVECTOR Control; |
||
| 1464 | |||
| 1465 | XMASSERT( pCR ); |
||
| 1466 | |||
| 1467 | ux = (V1.vector4_f32[0] == V2.vector4_f32[0]) ? 0xFFFFFFFFU : 0; |
||
| 1468 | uy = (V1.vector4_f32[1] == V2.vector4_f32[1]) ? 0xFFFFFFFFU : 0; |
||
| 1469 | uz = (V1.vector4_f32[2] == V2.vector4_f32[2]) ? 0xFFFFFFFFU : 0; |
||
| 1470 | uw = (V1.vector4_f32[3] == V2.vector4_f32[3]) ? 0xFFFFFFFFU : 0; |
||
| 1471 | CR = 0; |
||
| 1472 | if (ux&uy&uz&uw) |
||
| 1473 | { |
||
| 1474 | // All elements are greater |
||
| 1475 | CR = XM_CRMASK_CR6TRUE; |
||
| 1476 | } |
||
| 1477 | else if (!(ux|uy|uz|uw)) |
||
| 1478 | { |
||
| 1479 | // All elements are not greater |
||
| 1480 | CR = XM_CRMASK_CR6FALSE; |
||
| 1481 | } |
||
| 1482 | *pCR = CR; |
||
| 1483 | Control.vector4_u32[0] = ux; |
||
| 1484 | Control.vector4_u32[1] = uy; |
||
| 1485 | Control.vector4_u32[2] = uz; |
||
| 1486 | Control.vector4_u32[3] = uw; |
||
| 1487 | return Control; |
||
| 1488 | |||
| 1489 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1490 | XMASSERT( pCR ); |
||
| 1491 | XMVECTOR vTemp = _mm_cmpeq_ps(V1,V2); |
||
| 1492 | UINT CR = 0; |
||
| 1493 | int iTest = _mm_movemask_ps(vTemp); |
||
| 1494 | if (iTest==0xf) |
||
| 1495 | { |
||
| 1496 | CR = XM_CRMASK_CR6TRUE; |
||
| 1497 | } |
||
| 1498 | else if (!iTest) |
||
| 1499 | { |
||
| 1500 | // All elements are not greater |
||
| 1501 | CR = XM_CRMASK_CR6FALSE; |
||
| 1502 | } |
||
| 1503 | *pCR = CR; |
||
| 1504 | return vTemp; |
||
| 1505 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1506 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1507 | } |
||
| 1508 | |||
| 1509 | //------------------------------------------------------------------------------ |
||
| 1510 | // Treat the components of the vectors as unsigned integers and |
||
| 1511 | // compare individual bits between the two. This is useful for |
||
| 1512 | // comparing control vectors and result vectors returned from |
||
| 1513 | // other comparison operations. |
||
| 1514 | |||
| 1515 | XMFINLINE XMVECTOR XMVectorEqualInt |
||
| 1516 | ( |
||
| 1517 | FXMVECTOR V1, |
||
| 1518 | FXMVECTOR V2 |
||
| 1519 | ) |
||
| 1520 | { |
||
| 1521 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1522 | |||
| 1523 | XMVECTOR Control; |
||
| 1524 | |||
| 1525 | Control.vector4_u32[0] = (V1.vector4_u32[0] == V2.vector4_u32[0]) ? 0xFFFFFFFF : 0; |
||
| 1526 | Control.vector4_u32[1] = (V1.vector4_u32[1] == V2.vector4_u32[1]) ? 0xFFFFFFFF : 0; |
||
| 1527 | Control.vector4_u32[2] = (V1.vector4_u32[2] == V2.vector4_u32[2]) ? 0xFFFFFFFF : 0; |
||
| 1528 | Control.vector4_u32[3] = (V1.vector4_u32[3] == V2.vector4_u32[3]) ? 0xFFFFFFFF : 0; |
||
| 1529 | |||
| 1530 | return Control; |
||
| 1531 | |||
| 1532 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1533 | __m128i V = _mm_cmpeq_epi32( reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0] ); |
||
| 1534 | return reinterpret_cast<__m128 *>(&V)[0]; |
||
| 1535 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1536 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1537 | } |
||
| 1538 | |||
| 1539 | //------------------------------------------------------------------------------ |
||
| 1540 | |||
| 1541 | XMFINLINE XMVECTOR XMVectorEqualIntR |
||
| 1542 | ( |
||
| 1543 | UINT* pCR, |
||
| 1544 | FXMVECTOR V1, |
||
| 1545 | FXMVECTOR V2 |
||
| 1546 | ) |
||
| 1547 | { |
||
| 1548 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1549 | |||
| 1550 | XMVECTOR Control; |
||
| 1551 | |||
| 1552 | XMASSERT(pCR); |
||
| 1553 | |||
| 1554 | Control = XMVectorEqualInt(V1, V2); |
||
| 1555 | |||
| 1556 | *pCR = 0; |
||
| 1557 | |||
| 1558 | if (XMVector4EqualInt(Control, XMVectorTrueInt())) |
||
| 1559 | { |
||
| 1560 | // All elements are equal |
||
| 1561 | *pCR |= XM_CRMASK_CR6TRUE; |
||
| 1562 | } |
||
| 1563 | else if (XMVector4EqualInt(Control, XMVectorFalseInt())) |
||
| 1564 | { |
||
| 1565 | // All elements are not equal |
||
| 1566 | *pCR |= XM_CRMASK_CR6FALSE; |
||
| 1567 | } |
||
| 1568 | |||
| 1569 | return Control; |
||
| 1570 | |||
| 1571 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1572 | XMASSERT(pCR); |
||
| 1573 | __m128i V = _mm_cmpeq_epi32( reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0] ); |
||
| 1574 | int iTemp = _mm_movemask_ps(reinterpret_cast<const __m128*>(&V)[0]); |
||
| 1575 | UINT CR = 0; |
||
| 1576 | if (iTemp==0x0F) |
||
| 1577 | { |
||
| 1578 | CR = XM_CRMASK_CR6TRUE; |
||
| 1579 | } |
||
| 1580 | else if (!iTemp) |
||
| 1581 | { |
||
| 1582 | CR = XM_CRMASK_CR6FALSE; |
||
| 1583 | } |
||
| 1584 | *pCR = CR; |
||
| 1585 | return reinterpret_cast<__m128 *>(&V)[0]; |
||
| 1586 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1587 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1588 | } |
||
| 1589 | |||
| 1590 | //------------------------------------------------------------------------------ |
||
| 1591 | |||
| 1592 | XMFINLINE XMVECTOR XMVectorNearEqual |
||
| 1593 | ( |
||
| 1594 | FXMVECTOR V1, |
||
| 1595 | FXMVECTOR V2, |
||
| 1596 | FXMVECTOR Epsilon |
||
| 1597 | ) |
||
| 1598 | { |
||
| 1599 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1600 | |||
| 1601 | FLOAT fDeltax, fDeltay, fDeltaz, fDeltaw; |
||
| 1602 | XMVECTOR Control; |
||
| 1603 | |||
| 1604 | fDeltax = V1.vector4_f32[0]-V2.vector4_f32[0]; |
||
| 1605 | fDeltay = V1.vector4_f32[1]-V2.vector4_f32[1]; |
||
| 1606 | fDeltaz = V1.vector4_f32[2]-V2.vector4_f32[2]; |
||
| 1607 | fDeltaw = V1.vector4_f32[3]-V2.vector4_f32[3]; |
||
| 1608 | |||
| 1609 | fDeltax = fabsf(fDeltax); |
||
| 1610 | fDeltay = fabsf(fDeltay); |
||
| 1611 | fDeltaz = fabsf(fDeltaz); |
||
| 1612 | fDeltaw = fabsf(fDeltaw); |
||
| 1613 | |||
| 1614 | Control.vector4_u32[0] = (fDeltax <= Epsilon.vector4_f32[0]) ? 0xFFFFFFFFU : 0; |
||
| 1615 | Control.vector4_u32[1] = (fDeltay <= Epsilon.vector4_f32[1]) ? 0xFFFFFFFFU : 0; |
||
| 1616 | Control.vector4_u32[2] = (fDeltaz <= Epsilon.vector4_f32[2]) ? 0xFFFFFFFFU : 0; |
||
| 1617 | Control.vector4_u32[3] = (fDeltaw <= Epsilon.vector4_f32[3]) ? 0xFFFFFFFFU : 0; |
||
| 1618 | |||
| 1619 | return Control; |
||
| 1620 | |||
| 1621 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1622 | // Get the difference |
||
| 1623 | XMVECTOR vDelta = _mm_sub_ps(V1,V2); |
||
| 1624 | // Get the absolute value of the difference |
||
| 1625 | XMVECTOR vTemp = _mm_setzero_ps(); |
||
| 1626 | vTemp = _mm_sub_ps(vTemp,vDelta); |
||
| 1627 | vTemp = _mm_max_ps(vTemp,vDelta); |
||
| 1628 | vTemp = _mm_cmple_ps(vTemp,Epsilon); |
||
| 1629 | return vTemp; |
||
| 1630 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1631 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1632 | } |
||
| 1633 | |||
| 1634 | //------------------------------------------------------------------------------ |
||
| 1635 | |||
| 1636 | XMFINLINE XMVECTOR XMVectorNotEqual |
||
| 1637 | ( |
||
| 1638 | FXMVECTOR V1, |
||
| 1639 | FXMVECTOR V2 |
||
| 1640 | ) |
||
| 1641 | { |
||
| 1642 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1643 | |||
| 1644 | XMVECTOR Control; |
||
| 1645 | Control.vector4_u32[0] = (V1.vector4_f32[0] != V2.vector4_f32[0]) ? 0xFFFFFFFF : 0; |
||
| 1646 | Control.vector4_u32[1] = (V1.vector4_f32[1] != V2.vector4_f32[1]) ? 0xFFFFFFFF : 0; |
||
| 1647 | Control.vector4_u32[2] = (V1.vector4_f32[2] != V2.vector4_f32[2]) ? 0xFFFFFFFF : 0; |
||
| 1648 | Control.vector4_u32[3] = (V1.vector4_f32[3] != V2.vector4_f32[3]) ? 0xFFFFFFFF : 0; |
||
| 1649 | return Control; |
||
| 1650 | |||
| 1651 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1652 | return _mm_cmpneq_ps( V1, V2 ); |
||
| 1653 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1654 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1655 | } |
||
| 1656 | |||
| 1657 | //------------------------------------------------------------------------------ |
||
| 1658 | |||
| 1659 | XMFINLINE XMVECTOR XMVectorNotEqualInt |
||
| 1660 | ( |
||
| 1661 | FXMVECTOR V1, |
||
| 1662 | FXMVECTOR V2 |
||
| 1663 | ) |
||
| 1664 | { |
||
| 1665 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1666 | |||
| 1667 | XMVECTOR Control; |
||
| 1668 | Control.vector4_u32[0] = (V1.vector4_u32[0] != V2.vector4_u32[0]) ? 0xFFFFFFFFU : 0; |
||
| 1669 | Control.vector4_u32[1] = (V1.vector4_u32[1] != V2.vector4_u32[1]) ? 0xFFFFFFFFU : 0; |
||
| 1670 | Control.vector4_u32[2] = (V1.vector4_u32[2] != V2.vector4_u32[2]) ? 0xFFFFFFFFU : 0; |
||
| 1671 | Control.vector4_u32[3] = (V1.vector4_u32[3] != V2.vector4_u32[3]) ? 0xFFFFFFFFU : 0; |
||
| 1672 | return Control; |
||
| 1673 | |||
| 1674 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1675 | __m128i V = _mm_cmpeq_epi32( reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0] ); |
||
| 1676 | return _mm_xor_ps(reinterpret_cast<__m128 *>(&V)[0],g_XMNegOneMask); |
||
| 1677 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1678 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1679 | } |
||
| 1680 | |||
| 1681 | //------------------------------------------------------------------------------ |
||
| 1682 | |||
| 1683 | XMFINLINE XMVECTOR XMVectorGreater |
||
| 1684 | ( |
||
| 1685 | FXMVECTOR V1, |
||
| 1686 | FXMVECTOR V2 |
||
| 1687 | ) |
||
| 1688 | { |
||
| 1689 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1690 | |||
| 1691 | XMVECTOR Control; |
||
| 1692 | Control.vector4_u32[0] = (V1.vector4_f32[0] > V2.vector4_f32[0]) ? 0xFFFFFFFF : 0; |
||
| 1693 | Control.vector4_u32[1] = (V1.vector4_f32[1] > V2.vector4_f32[1]) ? 0xFFFFFFFF : 0; |
||
| 1694 | Control.vector4_u32[2] = (V1.vector4_f32[2] > V2.vector4_f32[2]) ? 0xFFFFFFFF : 0; |
||
| 1695 | Control.vector4_u32[3] = (V1.vector4_f32[3] > V2.vector4_f32[3]) ? 0xFFFFFFFF : 0; |
||
| 1696 | return Control; |
||
| 1697 | |||
| 1698 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1699 | return _mm_cmpgt_ps( V1, V2 ); |
||
| 1700 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1701 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1702 | } |
||
| 1703 | |||
| 1704 | //------------------------------------------------------------------------------ |
||
| 1705 | |||
| 1706 | XMFINLINE XMVECTOR XMVectorGreaterR |
||
| 1707 | ( |
||
| 1708 | UINT* pCR, |
||
| 1709 | FXMVECTOR V1, |
||
| 1710 | FXMVECTOR V2 |
||
| 1711 | ) |
||
| 1712 | { |
||
| 1713 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1714 | UINT ux, uy, uz, uw, CR; |
||
| 1715 | XMVECTOR Control; |
||
| 1716 | |||
| 1717 | XMASSERT( pCR ); |
||
| 1718 | |||
| 1719 | ux = (V1.vector4_f32[0] > V2.vector4_f32[0]) ? 0xFFFFFFFFU : 0; |
||
| 1720 | uy = (V1.vector4_f32[1] > V2.vector4_f32[1]) ? 0xFFFFFFFFU : 0; |
||
| 1721 | uz = (V1.vector4_f32[2] > V2.vector4_f32[2]) ? 0xFFFFFFFFU : 0; |
||
| 1722 | uw = (V1.vector4_f32[3] > V2.vector4_f32[3]) ? 0xFFFFFFFFU : 0; |
||
| 1723 | CR = 0; |
||
| 1724 | if (ux&uy&uz&uw) |
||
| 1725 | { |
||
| 1726 | // All elements are greater |
||
| 1727 | CR = XM_CRMASK_CR6TRUE; |
||
| 1728 | } |
||
| 1729 | else if (!(ux|uy|uz|uw)) |
||
| 1730 | { |
||
| 1731 | // All elements are not greater |
||
| 1732 | CR = XM_CRMASK_CR6FALSE; |
||
| 1733 | } |
||
| 1734 | *pCR = CR; |
||
| 1735 | Control.vector4_u32[0] = ux; |
||
| 1736 | Control.vector4_u32[1] = uy; |
||
| 1737 | Control.vector4_u32[2] = uz; |
||
| 1738 | Control.vector4_u32[3] = uw; |
||
| 1739 | return Control; |
||
| 1740 | |||
| 1741 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1742 | XMASSERT( pCR ); |
||
| 1743 | XMVECTOR vTemp = _mm_cmpgt_ps(V1,V2); |
||
| 1744 | UINT CR = 0; |
||
| 1745 | int iTest = _mm_movemask_ps(vTemp); |
||
| 1746 | if (iTest==0xf) |
||
| 1747 | { |
||
| 1748 | CR = XM_CRMASK_CR6TRUE; |
||
| 1749 | } |
||
| 1750 | else if (!iTest) |
||
| 1751 | { |
||
| 1752 | // All elements are not greater |
||
| 1753 | CR = XM_CRMASK_CR6FALSE; |
||
| 1754 | } |
||
| 1755 | *pCR = CR; |
||
| 1756 | return vTemp; |
||
| 1757 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1758 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1759 | } |
||
| 1760 | |||
| 1761 | //------------------------------------------------------------------------------ |
||
| 1762 | |||
| 1763 | XMFINLINE XMVECTOR XMVectorGreaterOrEqual |
||
| 1764 | ( |
||
| 1765 | FXMVECTOR V1, |
||
| 1766 | FXMVECTOR V2 |
||
| 1767 | ) |
||
| 1768 | { |
||
| 1769 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1770 | |||
| 1771 | XMVECTOR Control; |
||
| 1772 | Control.vector4_u32[0] = (V1.vector4_f32[0] >= V2.vector4_f32[0]) ? 0xFFFFFFFF : 0; |
||
| 1773 | Control.vector4_u32[1] = (V1.vector4_f32[1] >= V2.vector4_f32[1]) ? 0xFFFFFFFF : 0; |
||
| 1774 | Control.vector4_u32[2] = (V1.vector4_f32[2] >= V2.vector4_f32[2]) ? 0xFFFFFFFF : 0; |
||
| 1775 | Control.vector4_u32[3] = (V1.vector4_f32[3] >= V2.vector4_f32[3]) ? 0xFFFFFFFF : 0; |
||
| 1776 | return Control; |
||
| 1777 | |||
| 1778 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1779 | return _mm_cmpge_ps( V1, V2 ); |
||
| 1780 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1781 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1782 | } |
||
| 1783 | |||
| 1784 | //------------------------------------------------------------------------------ |
||
| 1785 | |||
| 1786 | XMFINLINE XMVECTOR XMVectorGreaterOrEqualR |
||
| 1787 | ( |
||
| 1788 | UINT* pCR, |
||
| 1789 | FXMVECTOR V1, |
||
| 1790 | FXMVECTOR V2 |
||
| 1791 | ) |
||
| 1792 | { |
||
| 1793 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1794 | UINT ux, uy, uz, uw, CR; |
||
| 1795 | XMVECTOR Control; |
||
| 1796 | |||
| 1797 | XMASSERT( pCR ); |
||
| 1798 | |||
| 1799 | ux = (V1.vector4_f32[0] >= V2.vector4_f32[0]) ? 0xFFFFFFFFU : 0; |
||
| 1800 | uy = (V1.vector4_f32[1] >= V2.vector4_f32[1]) ? 0xFFFFFFFFU : 0; |
||
| 1801 | uz = (V1.vector4_f32[2] >= V2.vector4_f32[2]) ? 0xFFFFFFFFU : 0; |
||
| 1802 | uw = (V1.vector4_f32[3] >= V2.vector4_f32[3]) ? 0xFFFFFFFFU : 0; |
||
| 1803 | CR = 0; |
||
| 1804 | if (ux&uy&uz&uw) |
||
| 1805 | { |
||
| 1806 | // All elements are greater |
||
| 1807 | CR = XM_CRMASK_CR6TRUE; |
||
| 1808 | } |
||
| 1809 | else if (!(ux|uy|uz|uw)) |
||
| 1810 | { |
||
| 1811 | // All elements are not greater |
||
| 1812 | CR = XM_CRMASK_CR6FALSE; |
||
| 1813 | } |
||
| 1814 | *pCR = CR; |
||
| 1815 | Control.vector4_u32[0] = ux; |
||
| 1816 | Control.vector4_u32[1] = uy; |
||
| 1817 | Control.vector4_u32[2] = uz; |
||
| 1818 | Control.vector4_u32[3] = uw; |
||
| 1819 | return Control; |
||
| 1820 | |||
| 1821 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1822 | XMASSERT( pCR ); |
||
| 1823 | XMVECTOR vTemp = _mm_cmpge_ps(V1,V2); |
||
| 1824 | UINT CR = 0; |
||
| 1825 | int iTest = _mm_movemask_ps(vTemp); |
||
| 1826 | if (iTest==0xf) |
||
| 1827 | { |
||
| 1828 | CR = XM_CRMASK_CR6TRUE; |
||
| 1829 | } |
||
| 1830 | else if (!iTest) |
||
| 1831 | { |
||
| 1832 | // All elements are not greater |
||
| 1833 | CR = XM_CRMASK_CR6FALSE; |
||
| 1834 | } |
||
| 1835 | *pCR = CR; |
||
| 1836 | return vTemp; |
||
| 1837 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1838 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1839 | } |
||
| 1840 | |||
| 1841 | //------------------------------------------------------------------------------ |
||
| 1842 | |||
| 1843 | XMFINLINE XMVECTOR XMVectorLess |
||
| 1844 | ( |
||
| 1845 | FXMVECTOR V1, |
||
| 1846 | FXMVECTOR V2 |
||
| 1847 | ) |
||
| 1848 | { |
||
| 1849 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1850 | |||
| 1851 | XMVECTOR Control; |
||
| 1852 | Control.vector4_u32[0] = (V1.vector4_f32[0] < V2.vector4_f32[0]) ? 0xFFFFFFFF : 0; |
||
| 1853 | Control.vector4_u32[1] = (V1.vector4_f32[1] < V2.vector4_f32[1]) ? 0xFFFFFFFF : 0; |
||
| 1854 | Control.vector4_u32[2] = (V1.vector4_f32[2] < V2.vector4_f32[2]) ? 0xFFFFFFFF : 0; |
||
| 1855 | Control.vector4_u32[3] = (V1.vector4_f32[3] < V2.vector4_f32[3]) ? 0xFFFFFFFF : 0; |
||
| 1856 | return Control; |
||
| 1857 | |||
| 1858 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1859 | return _mm_cmplt_ps( V1, V2 ); |
||
| 1860 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1861 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1862 | } |
||
| 1863 | |||
| 1864 | //------------------------------------------------------------------------------ |
||
| 1865 | |||
| 1866 | XMFINLINE XMVECTOR XMVectorLessOrEqual |
||
| 1867 | ( |
||
| 1868 | FXMVECTOR V1, |
||
| 1869 | FXMVECTOR V2 |
||
| 1870 | ) |
||
| 1871 | { |
||
| 1872 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1873 | |||
| 1874 | XMVECTOR Control; |
||
| 1875 | Control.vector4_u32[0] = (V1.vector4_f32[0] <= V2.vector4_f32[0]) ? 0xFFFFFFFF : 0; |
||
| 1876 | Control.vector4_u32[1] = (V1.vector4_f32[1] <= V2.vector4_f32[1]) ? 0xFFFFFFFF : 0; |
||
| 1877 | Control.vector4_u32[2] = (V1.vector4_f32[2] <= V2.vector4_f32[2]) ? 0xFFFFFFFF : 0; |
||
| 1878 | Control.vector4_u32[3] = (V1.vector4_f32[3] <= V2.vector4_f32[3]) ? 0xFFFFFFFF : 0; |
||
| 1879 | return Control; |
||
| 1880 | |||
| 1881 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1882 | return _mm_cmple_ps( V1, V2 ); |
||
| 1883 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1884 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1885 | } |
||
| 1886 | |||
| 1887 | //------------------------------------------------------------------------------ |
||
| 1888 | |||
| 1889 | XMFINLINE XMVECTOR XMVectorInBounds |
||
| 1890 | ( |
||
| 1891 | FXMVECTOR V, |
||
| 1892 | FXMVECTOR Bounds |
||
| 1893 | ) |
||
| 1894 | { |
||
| 1895 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1896 | |||
| 1897 | XMVECTOR Control; |
||
| 1898 | Control.vector4_u32[0] = (V.vector4_f32[0] <= Bounds.vector4_f32[0] && V.vector4_f32[0] >= -Bounds.vector4_f32[0]) ? 0xFFFFFFFF : 0; |
||
| 1899 | Control.vector4_u32[1] = (V.vector4_f32[1] <= Bounds.vector4_f32[1] && V.vector4_f32[1] >= -Bounds.vector4_f32[1]) ? 0xFFFFFFFF : 0; |
||
| 1900 | Control.vector4_u32[2] = (V.vector4_f32[2] <= Bounds.vector4_f32[2] && V.vector4_f32[2] >= -Bounds.vector4_f32[2]) ? 0xFFFFFFFF : 0; |
||
| 1901 | Control.vector4_u32[3] = (V.vector4_f32[3] <= Bounds.vector4_f32[3] && V.vector4_f32[3] >= -Bounds.vector4_f32[3]) ? 0xFFFFFFFF : 0; |
||
| 1902 | return Control; |
||
| 1903 | |||
| 1904 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1905 | // Test if less than or equal |
||
| 1906 | XMVECTOR vTemp1 = _mm_cmple_ps(V,Bounds); |
||
| 1907 | // Negate the bounds |
||
| 1908 | XMVECTOR vTemp2 = _mm_mul_ps(Bounds,g_XMNegativeOne); |
||
| 1909 | // Test if greater or equal (Reversed) |
||
| 1910 | vTemp2 = _mm_cmple_ps(vTemp2,V); |
||
| 1911 | // Blend answers |
||
| 1912 | vTemp1 = _mm_and_ps(vTemp1,vTemp2); |
||
| 1913 | return vTemp1; |
||
| 1914 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1915 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1916 | } |
||
| 1917 | |||
| 1918 | //------------------------------------------------------------------------------ |
||
| 1919 | |||
| 1920 | XMFINLINE XMVECTOR XMVectorInBoundsR |
||
| 1921 | ( |
||
| 1922 | UINT* pCR, |
||
| 1923 | FXMVECTOR V, |
||
| 1924 | FXMVECTOR Bounds |
||
| 1925 | ) |
||
| 1926 | { |
||
| 1927 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1928 | UINT ux, uy, uz, uw, CR; |
||
| 1929 | XMVECTOR Control; |
||
| 1930 | |||
| 1931 | XMASSERT( pCR != 0 ); |
||
| 1932 | |||
| 1933 | ux = (V.vector4_f32[0] <= Bounds.vector4_f32[0] && V.vector4_f32[0] >= -Bounds.vector4_f32[0]) ? 0xFFFFFFFFU : 0; |
||
| 1934 | uy = (V.vector4_f32[1] <= Bounds.vector4_f32[1] && V.vector4_f32[1] >= -Bounds.vector4_f32[1]) ? 0xFFFFFFFFU : 0; |
||
| 1935 | uz = (V.vector4_f32[2] <= Bounds.vector4_f32[2] && V.vector4_f32[2] >= -Bounds.vector4_f32[2]) ? 0xFFFFFFFFU : 0; |
||
| 1936 | uw = (V.vector4_f32[3] <= Bounds.vector4_f32[3] && V.vector4_f32[3] >= -Bounds.vector4_f32[3]) ? 0xFFFFFFFFU : 0; |
||
| 1937 | |||
| 1938 | CR = 0; |
||
| 1939 | |||
| 1940 | if (ux&uy&uz&uw) |
||
| 1941 | { |
||
| 1942 | // All elements are in bounds |
||
| 1943 | CR = XM_CRMASK_CR6BOUNDS; |
||
| 1944 | } |
||
| 1945 | *pCR = CR; |
||
| 1946 | Control.vector4_u32[0] = ux; |
||
| 1947 | Control.vector4_u32[1] = uy; |
||
| 1948 | Control.vector4_u32[2] = uz; |
||
| 1949 | Control.vector4_u32[3] = uw; |
||
| 1950 | return Control; |
||
| 1951 | |||
| 1952 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1953 | XMASSERT( pCR != 0 ); |
||
| 1954 | // Test if less than or equal |
||
| 1955 | XMVECTOR vTemp1 = _mm_cmple_ps(V,Bounds); |
||
| 1956 | // Negate the bounds |
||
| 1957 | XMVECTOR vTemp2 = _mm_mul_ps(Bounds,g_XMNegativeOne); |
||
| 1958 | // Test if greater or equal (Reversed) |
||
| 1959 | vTemp2 = _mm_cmple_ps(vTemp2,V); |
||
| 1960 | // Blend answers |
||
| 1961 | vTemp1 = _mm_and_ps(vTemp1,vTemp2); |
||
| 1962 | |||
| 1963 | UINT CR = 0; |
||
| 1964 | if (_mm_movemask_ps(vTemp1)==0xf) { |
||
| 1965 | // All elements are in bounds |
||
| 1966 | CR = XM_CRMASK_CR6BOUNDS; |
||
| 1967 | } |
||
| 1968 | *pCR = CR; |
||
| 1969 | return vTemp1; |
||
| 1970 | #else // _XM_VMX128_INTRINSICS_ |
||
| 1971 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 1972 | } |
||
| 1973 | |||
| 1974 | //------------------------------------------------------------------------------ |
||
| 1975 | |||
| 1976 | XMFINLINE XMVECTOR XMVectorIsNaN |
||
| 1977 | ( |
||
| 1978 | FXMVECTOR V |
||
| 1979 | ) |
||
| 1980 | { |
||
| 1981 | #if defined(_XM_NO_INTRINSICS_) |
||
| 1982 | |||
| 1983 | XMVECTOR Control; |
||
| 1984 | Control.vector4_u32[0] = XMISNAN(V.vector4_f32[0]) ? 0xFFFFFFFFU : 0; |
||
| 1985 | Control.vector4_u32[1] = XMISNAN(V.vector4_f32[1]) ? 0xFFFFFFFFU : 0; |
||
| 1986 | Control.vector4_u32[2] = XMISNAN(V.vector4_f32[2]) ? 0xFFFFFFFFU : 0; |
||
| 1987 | Control.vector4_u32[3] = XMISNAN(V.vector4_f32[3]) ? 0xFFFFFFFFU : 0; |
||
| 1988 | return Control; |
||
| 1989 | |||
| 1990 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 1991 | // Mask off the exponent |
||
| 1992 | __m128i vTempInf = _mm_and_si128(reinterpret_cast<const __m128i *>(&V)[0],g_XMInfinity); |
||
| 1993 | // Mask off the mantissa |
||
| 1994 | __m128i vTempNan = _mm_and_si128(reinterpret_cast<const __m128i *>(&V)[0],g_XMQNaNTest); |
||
| 1995 | // Are any of the exponents == 0x7F800000? |
||
| 1996 | vTempInf = _mm_cmpeq_epi32(vTempInf,g_XMInfinity); |
||
| 1997 | // Are any of the mantissa's zero? (SSE2 doesn't have a neq test) |
||
| 1998 | vTempNan = _mm_cmpeq_epi32(vTempNan,g_XMZero); |
||
| 1999 | // Perform a not on the NaN test to be true on NON-zero mantissas |
||
| 2000 | vTempNan = _mm_andnot_si128(vTempNan,vTempInf); |
||
| 2001 | // If any are NaN, the signs are true after the merge above |
||
| 2002 | return reinterpret_cast<const XMVECTOR *>(&vTempNan)[0]; |
||
| 2003 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2004 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2005 | } |
||
| 2006 | |||
| 2007 | //------------------------------------------------------------------------------ |
||
| 2008 | |||
| 2009 | XMFINLINE XMVECTOR XMVectorIsInfinite |
||
| 2010 | ( |
||
| 2011 | FXMVECTOR V |
||
| 2012 | ) |
||
| 2013 | { |
||
| 2014 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2015 | |||
| 2016 | XMVECTOR Control; |
||
| 2017 | Control.vector4_u32[0] = XMISINF(V.vector4_f32[0]) ? 0xFFFFFFFFU : 0; |
||
| 2018 | Control.vector4_u32[1] = XMISINF(V.vector4_f32[1]) ? 0xFFFFFFFFU : 0; |
||
| 2019 | Control.vector4_u32[2] = XMISINF(V.vector4_f32[2]) ? 0xFFFFFFFFU : 0; |
||
| 2020 | Control.vector4_u32[3] = XMISINF(V.vector4_f32[3]) ? 0xFFFFFFFFU : 0; |
||
| 2021 | return Control; |
||
| 2022 | |||
| 2023 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2024 | // Mask off the sign bit |
||
| 2025 | __m128 vTemp = _mm_and_ps(V,g_XMAbsMask); |
||
| 2026 | // Compare to infinity |
||
| 2027 | vTemp = _mm_cmpeq_ps(vTemp,g_XMInfinity); |
||
| 2028 | // If any are infinity, the signs are true. |
||
| 2029 | return vTemp; |
||
| 2030 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2031 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2032 | } |
||
| 2033 | |||
| 2034 | //------------------------------------------------------------------------------ |
||
| 2035 | // Rounding and clamping operations |
||
| 2036 | //------------------------------------------------------------------------------ |
||
| 2037 | |||
| 2038 | //------------------------------------------------------------------------------ |
||
| 2039 | |||
| 2040 | XMFINLINE XMVECTOR XMVectorMin |
||
| 2041 | ( |
||
| 2042 | FXMVECTOR V1, |
||
| 2043 | FXMVECTOR V2 |
||
| 2044 | ) |
||
| 2045 | { |
||
| 2046 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2047 | |||
| 2048 | XMVECTOR Result; |
||
| 2049 | Result.vector4_f32[0] = (V1.vector4_f32[0] < V2.vector4_f32[0]) ? V1.vector4_f32[0] : V2.vector4_f32[0]; |
||
| 2050 | Result.vector4_f32[1] = (V1.vector4_f32[1] < V2.vector4_f32[1]) ? V1.vector4_f32[1] : V2.vector4_f32[1]; |
||
| 2051 | Result.vector4_f32[2] = (V1.vector4_f32[2] < V2.vector4_f32[2]) ? V1.vector4_f32[2] : V2.vector4_f32[2]; |
||
| 2052 | Result.vector4_f32[3] = (V1.vector4_f32[3] < V2.vector4_f32[3]) ? V1.vector4_f32[3] : V2.vector4_f32[3]; |
||
| 2053 | return Result; |
||
| 2054 | |||
| 2055 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2056 | return _mm_min_ps( V1, V2 ); |
||
| 2057 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2058 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2059 | } |
||
| 2060 | |||
| 2061 | //------------------------------------------------------------------------------ |
||
| 2062 | |||
| 2063 | XMFINLINE XMVECTOR XMVectorMax |
||
| 2064 | ( |
||
| 2065 | FXMVECTOR V1, |
||
| 2066 | FXMVECTOR V2 |
||
| 2067 | ) |
||
| 2068 | { |
||
| 2069 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2070 | |||
| 2071 | XMVECTOR Result; |
||
| 2072 | Result.vector4_f32[0] = (V1.vector4_f32[0] > V2.vector4_f32[0]) ? V1.vector4_f32[0] : V2.vector4_f32[0]; |
||
| 2073 | Result.vector4_f32[1] = (V1.vector4_f32[1] > V2.vector4_f32[1]) ? V1.vector4_f32[1] : V2.vector4_f32[1]; |
||
| 2074 | Result.vector4_f32[2] = (V1.vector4_f32[2] > V2.vector4_f32[2]) ? V1.vector4_f32[2] : V2.vector4_f32[2]; |
||
| 2075 | Result.vector4_f32[3] = (V1.vector4_f32[3] > V2.vector4_f32[3]) ? V1.vector4_f32[3] : V2.vector4_f32[3]; |
||
| 2076 | return Result; |
||
| 2077 | |||
| 2078 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2079 | return _mm_max_ps( V1, V2 ); |
||
| 2080 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2081 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2082 | } |
||
| 2083 | |||
| 2084 | //------------------------------------------------------------------------------ |
||
| 2085 | |||
| 2086 | XMFINLINE XMVECTOR XMVectorRound |
||
| 2087 | ( |
||
| 2088 | FXMVECTOR V |
||
| 2089 | ) |
||
| 2090 | { |
||
| 2091 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2092 | |||
| 2093 | XMVECTOR Result; |
||
| 2094 | XMVECTOR Bias; |
||
| 2095 | CONST XMVECTOR Zero = XMVectorZero(); |
||
| 2096 | CONST XMVECTOR BiasPos = XMVectorReplicate(0.5f); |
||
| 2097 | CONST XMVECTOR BiasNeg = XMVectorReplicate(-0.5f); |
||
| 2098 | |||
| 2099 | Bias = XMVectorLess(V, Zero); |
||
| 2100 | Bias = XMVectorSelect(BiasPos, BiasNeg, Bias); |
||
| 2101 | Result = XMVectorAdd(V, Bias); |
||
| 2102 | Result = XMVectorTruncate(Result); |
||
| 2103 | |||
| 2104 | return Result; |
||
| 2105 | |||
| 2106 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2107 | // To handle NAN, INF and numbers greater than 8388608, use masking |
||
| 2108 | // Get the abs value |
||
| 2109 | __m128i vTest = _mm_and_si128(reinterpret_cast<const __m128i *>(&V)[0],g_XMAbsMask); |
||
| 2110 | // Test for greater than 8388608 (All floats with NO fractionals, NAN and INF |
||
| 2111 | vTest = _mm_cmplt_epi32(vTest,g_XMNoFraction); |
||
| 2112 | // Convert to int and back to float for rounding |
||
| 2113 | __m128i vInt = _mm_cvtps_epi32(V); |
||
| 2114 | // Convert back to floats |
||
| 2115 | XMVECTOR vResult = _mm_cvtepi32_ps(vInt); |
||
| 2116 | // All numbers less than 8388608 will use the round to int |
||
| 2117 | vResult = _mm_and_ps(vResult,reinterpret_cast<const XMVECTOR *>(&vTest)[0]); |
||
| 2118 | // All others, use the ORIGINAL value |
||
| 2119 | vTest = _mm_andnot_si128(vTest,reinterpret_cast<const __m128i *>(&V)[0]); |
||
| 2120 | vResult = _mm_or_ps(vResult,reinterpret_cast<const XMVECTOR *>(&vTest)[0]); |
||
| 2121 | return vResult; |
||
| 2122 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2123 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2124 | } |
||
| 2125 | |||
| 2126 | //------------------------------------------------------------------------------ |
||
| 2127 | |||
| 2128 | XMFINLINE XMVECTOR XMVectorTruncate |
||
| 2129 | ( |
||
| 2130 | FXMVECTOR V |
||
| 2131 | ) |
||
| 2132 | { |
||
| 2133 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2134 | XMVECTOR Result; |
||
| 2135 | Result.vector4_f32[0] = (FLOAT)((INT)V.vector4_f32[0]); |
||
| 2136 | Result.vector4_f32[1] = (FLOAT)((INT)V.vector4_f32[1]); |
||
| 2137 | Result.vector4_f32[2] = (FLOAT)((INT)V.vector4_f32[2]); |
||
| 2138 | Result.vector4_f32[3] = (FLOAT)((INT)V.vector4_f32[3]); |
||
| 2139 | |||
| 2140 | return Result; |
||
| 2141 | |||
| 2142 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2143 | // To handle NAN, INF and numbers greater than 8388608, use masking |
||
| 2144 | // Get the abs value |
||
| 2145 | __m128i vTest = _mm_and_si128(reinterpret_cast<const __m128i *>(&V)[0],g_XMAbsMask); |
||
| 2146 | // Test for greater than 8388608 (All floats with NO fractionals, NAN and INF |
||
| 2147 | vTest = _mm_cmplt_epi32(vTest,g_XMNoFraction); |
||
| 2148 | // Convert to int and back to float for rounding with truncation |
||
| 2149 | __m128i vInt = _mm_cvttps_epi32(V); |
||
| 2150 | // Convert back to floats |
||
| 2151 | XMVECTOR vResult = _mm_cvtepi32_ps(vInt); |
||
| 2152 | // All numbers less than 8388608 will use the round to int |
||
| 2153 | vResult = _mm_and_ps(vResult,reinterpret_cast<const XMVECTOR *>(&vTest)[0]); |
||
| 2154 | // All others, use the ORIGINAL value |
||
| 2155 | vTest = _mm_andnot_si128(vTest,reinterpret_cast<const __m128i *>(&V)[0]); |
||
| 2156 | vResult = _mm_or_ps(vResult,reinterpret_cast<const XMVECTOR *>(&vTest)[0]); |
||
| 2157 | return vResult; |
||
| 2158 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2159 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2160 | } |
||
| 2161 | |||
| 2162 | //------------------------------------------------------------------------------ |
||
| 2163 | |||
| 2164 | XMFINLINE XMVECTOR XMVectorFloor |
||
| 2165 | ( |
||
| 2166 | FXMVECTOR V |
||
| 2167 | ) |
||
| 2168 | { |
||
| 2169 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2170 | |||
| 2171 | XMVECTOR vResult = { |
||
| 2172 | floorf(V.vector4_f32[0]), |
||
| 2173 | floorf(V.vector4_f32[1]), |
||
| 2174 | floorf(V.vector4_f32[2]), |
||
| 2175 | floorf(V.vector4_f32[3]) |
||
| 2176 | }; |
||
| 2177 | return vResult; |
||
| 2178 | |||
| 2179 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2180 | XMVECTOR vResult = _mm_sub_ps(V,g_XMOneHalfMinusEpsilon); |
||
| 2181 | __m128i vInt = _mm_cvtps_epi32(vResult); |
||
| 2182 | vResult = _mm_cvtepi32_ps(vInt); |
||
| 2183 | return vResult; |
||
| 2184 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2185 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2186 | } |
||
| 2187 | |||
| 2188 | //------------------------------------------------------------------------------ |
||
| 2189 | |||
| 2190 | XMFINLINE XMVECTOR XMVectorCeiling |
||
| 2191 | ( |
||
| 2192 | FXMVECTOR V |
||
| 2193 | ) |
||
| 2194 | { |
||
| 2195 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2196 | XMVECTOR vResult = { |
||
| 2197 | ceilf(V.vector4_f32[0]), |
||
| 2198 | ceilf(V.vector4_f32[1]), |
||
| 2199 | ceilf(V.vector4_f32[2]), |
||
| 2200 | ceilf(V.vector4_f32[3]) |
||
| 2201 | }; |
||
| 2202 | return vResult; |
||
| 2203 | |||
| 2204 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2205 | XMVECTOR vResult = _mm_add_ps(V,g_XMOneHalfMinusEpsilon); |
||
| 2206 | __m128i vInt = _mm_cvtps_epi32(vResult); |
||
| 2207 | vResult = _mm_cvtepi32_ps(vInt); |
||
| 2208 | return vResult; |
||
| 2209 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2210 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2211 | } |
||
| 2212 | |||
| 2213 | //------------------------------------------------------------------------------ |
||
| 2214 | |||
| 2215 | XMFINLINE XMVECTOR XMVectorClamp |
||
| 2216 | ( |
||
| 2217 | FXMVECTOR V, |
||
| 2218 | FXMVECTOR Min, |
||
| 2219 | FXMVECTOR Max |
||
| 2220 | ) |
||
| 2221 | { |
||
| 2222 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2223 | |||
| 2224 | XMVECTOR Result; |
||
| 2225 | |||
| 2226 | XMASSERT(XMVector4LessOrEqual(Min, Max)); |
||
| 2227 | |||
| 2228 | Result = XMVectorMax(Min, V); |
||
| 2229 | Result = XMVectorMin(Max, Result); |
||
| 2230 | |||
| 2231 | return Result; |
||
| 2232 | |||
| 2233 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2234 | XMVECTOR vResult; |
||
| 2235 | XMASSERT(XMVector4LessOrEqual(Min, Max)); |
||
| 2236 | vResult = _mm_max_ps(Min,V); |
||
| 2237 | vResult = _mm_min_ps(vResult,Max); |
||
| 2238 | return vResult; |
||
| 2239 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2240 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2241 | } |
||
| 2242 | |||
| 2243 | //------------------------------------------------------------------------------ |
||
| 2244 | |||
| 2245 | XMFINLINE XMVECTOR XMVectorSaturate |
||
| 2246 | ( |
||
| 2247 | FXMVECTOR V |
||
| 2248 | ) |
||
| 2249 | { |
||
| 2250 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2251 | |||
| 2252 | CONST XMVECTOR Zero = XMVectorZero(); |
||
| 2253 | |||
| 2254 | return XMVectorClamp(V, Zero, g_XMOne.v); |
||
| 2255 | |||
| 2256 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2257 | // Set <0 to 0 |
||
| 2258 | XMVECTOR vResult = _mm_max_ps(V,g_XMZero); |
||
| 2259 | // Set>1 to 1 |
||
| 2260 | return _mm_min_ps(vResult,g_XMOne); |
||
| 2261 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2262 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2263 | } |
||
| 2264 | |||
| 2265 | //------------------------------------------------------------------------------ |
||
| 2266 | // Bitwise logical operations |
||
| 2267 | //------------------------------------------------------------------------------ |
||
| 2268 | |||
| 2269 | XMFINLINE XMVECTOR XMVectorAndInt |
||
| 2270 | ( |
||
| 2271 | FXMVECTOR V1, |
||
| 2272 | FXMVECTOR V2 |
||
| 2273 | ) |
||
| 2274 | { |
||
| 2275 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2276 | |||
| 2277 | XMVECTOR Result; |
||
| 2278 | |||
| 2279 | Result.vector4_u32[0] = V1.vector4_u32[0] & V2.vector4_u32[0]; |
||
| 2280 | Result.vector4_u32[1] = V1.vector4_u32[1] & V2.vector4_u32[1]; |
||
| 2281 | Result.vector4_u32[2] = V1.vector4_u32[2] & V2.vector4_u32[2]; |
||
| 2282 | Result.vector4_u32[3] = V1.vector4_u32[3] & V2.vector4_u32[3]; |
||
| 2283 | return Result; |
||
| 2284 | |||
| 2285 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2286 | return _mm_and_ps(V1,V2); |
||
| 2287 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2288 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2289 | } |
||
| 2290 | |||
| 2291 | //------------------------------------------------------------------------------ |
||
| 2292 | |||
| 2293 | XMFINLINE XMVECTOR XMVectorAndCInt |
||
| 2294 | ( |
||
| 2295 | FXMVECTOR V1, |
||
| 2296 | FXMVECTOR V2 |
||
| 2297 | ) |
||
| 2298 | { |
||
| 2299 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2300 | |||
| 2301 | XMVECTOR Result; |
||
| 2302 | |||
| 2303 | Result.vector4_u32[0] = V1.vector4_u32[0] & ~V2.vector4_u32[0]; |
||
| 2304 | Result.vector4_u32[1] = V1.vector4_u32[1] & ~V2.vector4_u32[1]; |
||
| 2305 | Result.vector4_u32[2] = V1.vector4_u32[2] & ~V2.vector4_u32[2]; |
||
| 2306 | Result.vector4_u32[3] = V1.vector4_u32[3] & ~V2.vector4_u32[3]; |
||
| 2307 | |||
| 2308 | return Result; |
||
| 2309 | |||
| 2310 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2311 | __m128i V = _mm_andnot_si128( reinterpret_cast<const __m128i *>(&V2)[0], reinterpret_cast<const __m128i *>(&V1)[0] ); |
||
| 2312 | return reinterpret_cast<__m128 *>(&V)[0]; |
||
| 2313 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2314 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2315 | } |
||
| 2316 | |||
| 2317 | //------------------------------------------------------------------------------ |
||
| 2318 | |||
| 2319 | XMFINLINE XMVECTOR XMVectorOrInt |
||
| 2320 | ( |
||
| 2321 | FXMVECTOR V1, |
||
| 2322 | FXMVECTOR V2 |
||
| 2323 | ) |
||
| 2324 | { |
||
| 2325 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2326 | |||
| 2327 | XMVECTOR Result; |
||
| 2328 | |||
| 2329 | Result.vector4_u32[0] = V1.vector4_u32[0] | V2.vector4_u32[0]; |
||
| 2330 | Result.vector4_u32[1] = V1.vector4_u32[1] | V2.vector4_u32[1]; |
||
| 2331 | Result.vector4_u32[2] = V1.vector4_u32[2] | V2.vector4_u32[2]; |
||
| 2332 | Result.vector4_u32[3] = V1.vector4_u32[3] | V2.vector4_u32[3]; |
||
| 2333 | |||
| 2334 | return Result; |
||
| 2335 | |||
| 2336 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2337 | __m128i V = _mm_or_si128( reinterpret_cast<const __m128i *>(&V1)[0], reinterpret_cast<const __m128i *>(&V2)[0] ); |
||
| 2338 | return reinterpret_cast<__m128 *>(&V)[0]; |
||
| 2339 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2340 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2341 | } |
||
| 2342 | |||
| 2343 | //------------------------------------------------------------------------------ |
||
| 2344 | |||
| 2345 | XMFINLINE XMVECTOR XMVectorNorInt |
||
| 2346 | ( |
||
| 2347 | FXMVECTOR V1, |
||
| 2348 | FXMVECTOR V2 |
||
| 2349 | ) |
||
| 2350 | { |
||
| 2351 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2352 | |||
| 2353 | XMVECTOR Result; |
||
| 2354 | |||
| 2355 | Result.vector4_u32[0] = ~(V1.vector4_u32[0] | V2.vector4_u32[0]); |
||
| 2356 | Result.vector4_u32[1] = ~(V1.vector4_u32[1] | V2.vector4_u32[1]); |
||
| 2357 | Result.vector4_u32[2] = ~(V1.vector4_u32[2] | V2.vector4_u32[2]); |
||
| 2358 | Result.vector4_u32[3] = ~(V1.vector4_u32[3] | V2.vector4_u32[3]); |
||
| 2359 | |||
| 2360 | return Result; |
||
| 2361 | |||
| 2362 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2363 | __m128i Result; |
||
| 2364 | Result = _mm_or_si128( reinterpret_cast<const __m128i *>(&V1)[0], reinterpret_cast<const __m128i *>(&V2)[0] ); |
||
| 2365 | Result = _mm_andnot_si128( Result,g_XMNegOneMask); |
||
| 2366 | return reinterpret_cast<__m128 *>(&Result)[0]; |
||
| 2367 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2368 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2369 | } |
||
| 2370 | |||
| 2371 | //------------------------------------------------------------------------------ |
||
| 2372 | |||
| 2373 | XMFINLINE XMVECTOR XMVectorXorInt |
||
| 2374 | ( |
||
| 2375 | FXMVECTOR V1, |
||
| 2376 | FXMVECTOR V2 |
||
| 2377 | ) |
||
| 2378 | { |
||
| 2379 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2380 | |||
| 2381 | XMVECTOR Result; |
||
| 2382 | |||
| 2383 | Result.vector4_u32[0] = V1.vector4_u32[0] ^ V2.vector4_u32[0]; |
||
| 2384 | Result.vector4_u32[1] = V1.vector4_u32[1] ^ V2.vector4_u32[1]; |
||
| 2385 | Result.vector4_u32[2] = V1.vector4_u32[2] ^ V2.vector4_u32[2]; |
||
| 2386 | Result.vector4_u32[3] = V1.vector4_u32[3] ^ V2.vector4_u32[3]; |
||
| 2387 | |||
| 2388 | return Result; |
||
| 2389 | |||
| 2390 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2391 | __m128i V = _mm_xor_si128( reinterpret_cast<const __m128i *>(&V1)[0], reinterpret_cast<const __m128i *>(&V2)[0] ); |
||
| 2392 | return reinterpret_cast<__m128 *>(&V)[0]; |
||
| 2393 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2394 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2395 | } |
||
| 2396 | |||
| 2397 | //------------------------------------------------------------------------------ |
||
| 2398 | // Computation operations |
||
| 2399 | //------------------------------------------------------------------------------ |
||
| 2400 | |||
| 2401 | //------------------------------------------------------------------------------ |
||
| 2402 | |||
| 2403 | XMFINLINE XMVECTOR XMVectorNegate |
||
| 2404 | ( |
||
| 2405 | FXMVECTOR V |
||
| 2406 | ) |
||
| 2407 | { |
||
| 2408 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2409 | |||
| 2410 | XMVECTOR Result; |
||
| 2411 | |||
| 2412 | Result.vector4_f32[0] = -V.vector4_f32[0]; |
||
| 2413 | Result.vector4_f32[1] = -V.vector4_f32[1]; |
||
| 2414 | Result.vector4_f32[2] = -V.vector4_f32[2]; |
||
| 2415 | Result.vector4_f32[3] = -V.vector4_f32[3]; |
||
| 2416 | |||
| 2417 | return Result; |
||
| 2418 | |||
| 2419 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2420 | XMVECTOR Z; |
||
| 2421 | |||
| 2422 | Z = _mm_setzero_ps(); |
||
| 2423 | |||
| 2424 | return _mm_sub_ps( Z, V ); |
||
| 2425 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2426 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2427 | } |
||
| 2428 | |||
| 2429 | //------------------------------------------------------------------------------ |
||
| 2430 | |||
| 2431 | XMFINLINE XMVECTOR XMVectorAdd |
||
| 2432 | ( |
||
| 2433 | FXMVECTOR V1, |
||
| 2434 | FXMVECTOR V2 |
||
| 2435 | ) |
||
| 2436 | { |
||
| 2437 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2438 | |||
| 2439 | XMVECTOR Result; |
||
| 2440 | |||
| 2441 | Result.vector4_f32[0] = V1.vector4_f32[0] + V2.vector4_f32[0]; |
||
| 2442 | Result.vector4_f32[1] = V1.vector4_f32[1] + V2.vector4_f32[1]; |
||
| 2443 | Result.vector4_f32[2] = V1.vector4_f32[2] + V2.vector4_f32[2]; |
||
| 2444 | Result.vector4_f32[3] = V1.vector4_f32[3] + V2.vector4_f32[3]; |
||
| 2445 | |||
| 2446 | return Result; |
||
| 2447 | |||
| 2448 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2449 | return _mm_add_ps( V1, V2 ); |
||
| 2450 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2451 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2452 | } |
||
| 2453 | |||
| 2454 | //------------------------------------------------------------------------------ |
||
| 2455 | |||
| 2456 | XMFINLINE XMVECTOR XMVectorAddAngles |
||
| 2457 | ( |
||
| 2458 | FXMVECTOR V1, |
||
| 2459 | FXMVECTOR V2 |
||
| 2460 | ) |
||
| 2461 | { |
||
| 2462 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2463 | |||
| 2464 | XMVECTOR Mask; |
||
| 2465 | XMVECTOR Offset; |
||
| 2466 | XMVECTOR Result; |
||
| 2467 | CONST XMVECTOR Zero = XMVectorZero(); |
||
| 2468 | |||
| 2469 | // Add the given angles together. If the range of V1 is such |
||
| 2470 | // that -Pi <= V1 < Pi and the range of V2 is such that |
||
| 2471 | // -2Pi <= V2 <= 2Pi, then the range of the resulting angle |
||
| 2472 | // will be -Pi <= Result < Pi. |
||
| 2473 | Result = XMVectorAdd(V1, V2); |
||
| 2474 | |||
| 2475 | Mask = XMVectorLess(Result, g_XMNegativePi.v); |
||
| 2476 | Offset = XMVectorSelect(Zero, g_XMTwoPi.v, Mask); |
||
| 2477 | |||
| 2478 | Mask = XMVectorGreaterOrEqual(Result, g_XMPi.v); |
||
| 2479 | Offset = XMVectorSelect(Offset, g_XMNegativeTwoPi.v, Mask); |
||
| 2480 | |||
| 2481 | Result = XMVectorAdd(Result, Offset); |
||
| 2482 | |||
| 2483 | return Result; |
||
| 2484 | |||
| 2485 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2486 | // Adjust the angles |
||
| 2487 | XMVECTOR vResult = _mm_add_ps(V1,V2); |
||
| 2488 | // Less than Pi? |
||
| 2489 | XMVECTOR vOffset = _mm_cmplt_ps(vResult,g_XMNegativePi); |
||
| 2490 | vOffset = _mm_and_ps(vOffset,g_XMTwoPi); |
||
| 2491 | // Add 2Pi to all entries less than -Pi |
||
| 2492 | vResult = _mm_add_ps(vResult,vOffset); |
||
| 2493 | // Greater than or equal to Pi? |
||
| 2494 | vOffset = _mm_cmpge_ps(vResult,g_XMPi); |
||
| 2495 | vOffset = _mm_and_ps(vOffset,g_XMTwoPi); |
||
| 2496 | // Sub 2Pi to all entries greater than Pi |
||
| 2497 | vResult = _mm_sub_ps(vResult,vOffset); |
||
| 2498 | return vResult; |
||
| 2499 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2500 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2501 | } |
||
| 2502 | |||
| 2503 | //------------------------------------------------------------------------------ |
||
| 2504 | |||
| 2505 | XMFINLINE XMVECTOR XMVectorSubtract |
||
| 2506 | ( |
||
| 2507 | FXMVECTOR V1, |
||
| 2508 | FXMVECTOR V2 |
||
| 2509 | ) |
||
| 2510 | { |
||
| 2511 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2512 | |||
| 2513 | XMVECTOR Result; |
||
| 2514 | |||
| 2515 | Result.vector4_f32[0] = V1.vector4_f32[0] - V2.vector4_f32[0]; |
||
| 2516 | Result.vector4_f32[1] = V1.vector4_f32[1] - V2.vector4_f32[1]; |
||
| 2517 | Result.vector4_f32[2] = V1.vector4_f32[2] - V2.vector4_f32[2]; |
||
| 2518 | Result.vector4_f32[3] = V1.vector4_f32[3] - V2.vector4_f32[3]; |
||
| 2519 | |||
| 2520 | return Result; |
||
| 2521 | |||
| 2522 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2523 | return _mm_sub_ps( V1, V2 ); |
||
| 2524 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2525 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2526 | } |
||
| 2527 | |||
| 2528 | //------------------------------------------------------------------------------ |
||
| 2529 | |||
| 2530 | XMFINLINE XMVECTOR XMVectorSubtractAngles |
||
| 2531 | ( |
||
| 2532 | FXMVECTOR V1, |
||
| 2533 | FXMVECTOR V2 |
||
| 2534 | ) |
||
| 2535 | { |
||
| 2536 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2537 | |||
| 2538 | XMVECTOR Mask; |
||
| 2539 | XMVECTOR Offset; |
||
| 2540 | XMVECTOR Result; |
||
| 2541 | CONST XMVECTOR Zero = XMVectorZero(); |
||
| 2542 | |||
| 2543 | // Subtract the given angles. If the range of V1 is such |
||
| 2544 | // that -Pi <= V1 < Pi and the range of V2 is such that |
||
| 2545 | // -2Pi <= V2 <= 2Pi, then the range of the resulting angle |
||
| 2546 | // will be -Pi <= Result < Pi. |
||
| 2547 | Result = XMVectorSubtract(V1, V2); |
||
| 2548 | |||
| 2549 | Mask = XMVectorLess(Result, g_XMNegativePi.v); |
||
| 2550 | Offset = XMVectorSelect(Zero, g_XMTwoPi.v, Mask); |
||
| 2551 | |||
| 2552 | Mask = XMVectorGreaterOrEqual(Result, g_XMPi.v); |
||
| 2553 | Offset = XMVectorSelect(Offset, g_XMNegativeTwoPi.v, Mask); |
||
| 2554 | |||
| 2555 | Result = XMVectorAdd(Result, Offset); |
||
| 2556 | |||
| 2557 | return Result; |
||
| 2558 | |||
| 2559 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2560 | // Adjust the angles |
||
| 2561 | XMVECTOR vResult = _mm_sub_ps(V1,V2); |
||
| 2562 | // Less than Pi? |
||
| 2563 | XMVECTOR vOffset = _mm_cmplt_ps(vResult,g_XMNegativePi); |
||
| 2564 | vOffset = _mm_and_ps(vOffset,g_XMTwoPi); |
||
| 2565 | // Add 2Pi to all entries less than -Pi |
||
| 2566 | vResult = _mm_add_ps(vResult,vOffset); |
||
| 2567 | // Greater than or equal to Pi? |
||
| 2568 | vOffset = _mm_cmpge_ps(vResult,g_XMPi); |
||
| 2569 | vOffset = _mm_and_ps(vOffset,g_XMTwoPi); |
||
| 2570 | // Sub 2Pi to all entries greater than Pi |
||
| 2571 | vResult = _mm_sub_ps(vResult,vOffset); |
||
| 2572 | return vResult; |
||
| 2573 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2574 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2575 | } |
||
| 2576 | |||
| 2577 | //------------------------------------------------------------------------------ |
||
| 2578 | |||
| 2579 | XMFINLINE XMVECTOR XMVectorMultiply |
||
| 2580 | ( |
||
| 2581 | FXMVECTOR V1, |
||
| 2582 | FXMVECTOR V2 |
||
| 2583 | ) |
||
| 2584 | { |
||
| 2585 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2586 | XMVECTOR Result = { |
||
| 2587 | V1.vector4_f32[0] * V2.vector4_f32[0], |
||
| 2588 | V1.vector4_f32[1] * V2.vector4_f32[1], |
||
| 2589 | V1.vector4_f32[2] * V2.vector4_f32[2], |
||
| 2590 | V1.vector4_f32[3] * V2.vector4_f32[3] |
||
| 2591 | }; |
||
| 2592 | return Result; |
||
| 2593 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2594 | return _mm_mul_ps( V1, V2 ); |
||
| 2595 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2596 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2597 | } |
||
| 2598 | |||
| 2599 | //------------------------------------------------------------------------------ |
||
| 2600 | |||
| 2601 | XMFINLINE XMVECTOR XMVectorMultiplyAdd |
||
| 2602 | ( |
||
| 2603 | FXMVECTOR V1, |
||
| 2604 | FXMVECTOR V2, |
||
| 2605 | FXMVECTOR V3 |
||
| 2606 | ) |
||
| 2607 | { |
||
| 2608 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2609 | XMVECTOR vResult = { |
||
| 2610 | (V1.vector4_f32[0] * V2.vector4_f32[0]) + V3.vector4_f32[0], |
||
| 2611 | (V1.vector4_f32[1] * V2.vector4_f32[1]) + V3.vector4_f32[1], |
||
| 2612 | (V1.vector4_f32[2] * V2.vector4_f32[2]) + V3.vector4_f32[2], |
||
| 2613 | (V1.vector4_f32[3] * V2.vector4_f32[3]) + V3.vector4_f32[3] |
||
| 2614 | }; |
||
| 2615 | return vResult; |
||
| 2616 | |||
| 2617 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2618 | XMVECTOR vResult = _mm_mul_ps( V1, V2 ); |
||
| 2619 | return _mm_add_ps(vResult, V3 ); |
||
| 2620 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2621 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2622 | } |
||
| 2623 | |||
| 2624 | //------------------------------------------------------------------------------ |
||
| 2625 | |||
| 2626 | XMFINLINE XMVECTOR XMVectorNegativeMultiplySubtract |
||
| 2627 | ( |
||
| 2628 | FXMVECTOR V1, |
||
| 2629 | FXMVECTOR V2, |
||
| 2630 | FXMVECTOR V3 |
||
| 2631 | ) |
||
| 2632 | { |
||
| 2633 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2634 | |||
| 2635 | XMVECTOR vResult = { |
||
| 2636 | V3.vector4_f32[0] - (V1.vector4_f32[0] * V2.vector4_f32[0]), |
||
| 2637 | V3.vector4_f32[1] - (V1.vector4_f32[1] * V2.vector4_f32[1]), |
||
| 2638 | V3.vector4_f32[2] - (V1.vector4_f32[2] * V2.vector4_f32[2]), |
||
| 2639 | V3.vector4_f32[3] - (V1.vector4_f32[3] * V2.vector4_f32[3]) |
||
| 2640 | }; |
||
| 2641 | return vResult; |
||
| 2642 | |||
| 2643 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2644 | XMVECTOR R = _mm_mul_ps( V1, V2 ); |
||
| 2645 | return _mm_sub_ps( V3, R ); |
||
| 2646 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2647 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2648 | } |
||
| 2649 | |||
| 2650 | //------------------------------------------------------------------------------ |
||
| 2651 | |||
| 2652 | XMFINLINE XMVECTOR XMVectorScale |
||
| 2653 | ( |
||
| 2654 | FXMVECTOR V, |
||
| 2655 | FLOAT ScaleFactor |
||
| 2656 | ) |
||
| 2657 | { |
||
| 2658 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2659 | XMVECTOR vResult = { |
||
| 2660 | V.vector4_f32[0] * ScaleFactor, |
||
| 2661 | V.vector4_f32[1] * ScaleFactor, |
||
| 2662 | V.vector4_f32[2] * ScaleFactor, |
||
| 2663 | V.vector4_f32[3] * ScaleFactor |
||
| 2664 | }; |
||
| 2665 | return vResult; |
||
| 2666 | |||
| 2667 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2668 | XMVECTOR vResult = _mm_set_ps1(ScaleFactor); |
||
| 2669 | return _mm_mul_ps(vResult,V); |
||
| 2670 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 2671 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2672 | } |
||
| 2673 | |||
| 2674 | //------------------------------------------------------------------------------ |
||
| 2675 | |||
| 2676 | XMFINLINE XMVECTOR XMVectorReciprocalEst |
||
| 2677 | ( |
||
| 2678 | FXMVECTOR V |
||
| 2679 | ) |
||
| 2680 | { |
||
| 2681 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2682 | |||
| 2683 | XMVECTOR Result; |
||
| 2684 | UINT i; |
||
| 2685 | |||
| 2686 | // Avoid C4701 |
||
| 2687 | Result.vector4_f32[0] = 0.0f; |
||
| 2688 | |||
| 2689 | for (i = 0; i < 4; i++) |
||
| 2690 | { |
||
| 2691 | if (XMISINF(V.vector4_f32[i])) |
||
| 2692 | { |
||
| 2693 | Result.vector4_f32[i] = (V.vector4_f32[i] < 0.0f) ? -0.0f : 0.0f; |
||
| 2694 | } |
||
| 2695 | else if (V.vector4_f32[i] == -0.0f) |
||
| 2696 | { |
||
| 2697 | Result.vector4_u32[i] = 0xFF800000; |
||
| 2698 | } |
||
| 2699 | else if (V.vector4_f32[i] == 0.0f) |
||
| 2700 | { |
||
| 2701 | Result.vector4_u32[i] = 0x7F800000; |
||
| 2702 | } |
||
| 2703 | else |
||
| 2704 | { |
||
| 2705 | Result.vector4_f32[i] = 1.0f / V.vector4_f32[i]; |
||
| 2706 | } |
||
| 2707 | } |
||
| 2708 | |||
| 2709 | return Result; |
||
| 2710 | |||
| 2711 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2712 | return _mm_rcp_ps(V); |
||
| 2713 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2714 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2715 | } |
||
| 2716 | |||
| 2717 | //------------------------------------------------------------------------------ |
||
| 2718 | |||
| 2719 | XMFINLINE XMVECTOR XMVectorReciprocal |
||
| 2720 | ( |
||
| 2721 | FXMVECTOR V |
||
| 2722 | ) |
||
| 2723 | { |
||
| 2724 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2725 | return XMVectorReciprocalEst(V); |
||
| 2726 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2727 | return _mm_div_ps(g_XMOne,V); |
||
| 2728 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2729 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2730 | } |
||
| 2731 | |||
| 2732 | //------------------------------------------------------------------------------ |
||
| 2733 | // Return an estimated square root |
||
| 2734 | XMFINLINE XMVECTOR XMVectorSqrtEst |
||
| 2735 | ( |
||
| 2736 | FXMVECTOR V |
||
| 2737 | ) |
||
| 2738 | { |
||
| 2739 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2740 | XMVECTOR Select; |
||
| 2741 | |||
| 2742 | // if (x == +Infinity) sqrt(x) = +Infinity |
||
| 2743 | // if (x == +0.0f) sqrt(x) = +0.0f |
||
| 2744 | // if (x == -0.0f) sqrt(x) = -0.0f |
||
| 2745 | // if (x < -0.0f) sqrt(x) = QNaN |
||
| 2746 | |||
| 2747 | XMVECTOR Result = XMVectorReciprocalSqrtEst(V); |
||
| 2748 | XMVECTOR Zero = XMVectorZero(); |
||
| 2749 | XMVECTOR VEqualsInfinity = XMVectorEqualInt(V, g_XMInfinity.v); |
||
| 2750 | XMVECTOR VEqualsZero = XMVectorEqual(V, Zero); |
||
| 2751 | Result = XMVectorMultiply(V, Result); |
||
| 2752 | Select = XMVectorEqualInt(VEqualsInfinity, VEqualsZero); |
||
| 2753 | Result = XMVectorSelect(V, Result, Select); |
||
| 2754 | return Result; |
||
| 2755 | |||
| 2756 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2757 | return _mm_sqrt_ps(V); |
||
| 2758 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2759 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2760 | } |
||
| 2761 | |||
| 2762 | //------------------------------------------------------------------------------ |
||
| 2763 | |||
| 2764 | XMFINLINE XMVECTOR XMVectorSqrt |
||
| 2765 | ( |
||
| 2766 | FXMVECTOR V |
||
| 2767 | ) |
||
| 2768 | { |
||
| 2769 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2770 | |||
| 2771 | XMVECTOR Zero; |
||
| 2772 | XMVECTOR VEqualsInfinity, VEqualsZero; |
||
| 2773 | XMVECTOR Select; |
||
| 2774 | XMVECTOR Result; |
||
| 2775 | |||
| 2776 | // if (x == +Infinity) sqrt(x) = +Infinity |
||
| 2777 | // if (x == +0.0f) sqrt(x) = +0.0f |
||
| 2778 | // if (x == -0.0f) sqrt(x) = -0.0f |
||
| 2779 | // if (x < -0.0f) sqrt(x) = QNaN |
||
| 2780 | |||
| 2781 | Result = XMVectorReciprocalSqrt(V); |
||
| 2782 | Zero = XMVectorZero(); |
||
| 2783 | VEqualsInfinity = XMVectorEqualInt(V, g_XMInfinity.v); |
||
| 2784 | VEqualsZero = XMVectorEqual(V, Zero); |
||
| 2785 | Result = XMVectorMultiply(V, Result); |
||
| 2786 | Select = XMVectorEqualInt(VEqualsInfinity, VEqualsZero); |
||
| 2787 | Result = XMVectorSelect(V, Result, Select); |
||
| 2788 | |||
| 2789 | return Result; |
||
| 2790 | |||
| 2791 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2792 | return _mm_sqrt_ps(V); |
||
| 2793 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2794 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2795 | } |
||
| 2796 | |||
| 2797 | //------------------------------------------------------------------------------ |
||
| 2798 | |||
| 2799 | XMFINLINE XMVECTOR XMVectorReciprocalSqrtEst |
||
| 2800 | ( |
||
| 2801 | FXMVECTOR V |
||
| 2802 | ) |
||
| 2803 | { |
||
| 2804 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2805 | |||
| 2806 | XMVECTOR Result; |
||
| 2807 | UINT i; |
||
| 2808 | |||
| 2809 | // Avoid C4701 |
||
| 2810 | Result.vector4_f32[0] = 0.0f; |
||
| 2811 | |||
| 2812 | for (i = 0; i < 4; i++) |
||
| 2813 | { |
||
| 2814 | if (V.vector4_f32[i] == 0.0f) |
||
| 2815 | { |
||
| 2816 | Result.vector4_u32[i] = 0x7F800000; |
||
| 2817 | } |
||
| 2818 | else if (V.vector4_f32[i] == -0.0f) |
||
| 2819 | { |
||
| 2820 | Result.vector4_u32[i] = 0xFF800000; |
||
| 2821 | } |
||
| 2822 | else if (V.vector4_f32[i] < 0.0f) |
||
| 2823 | { |
||
| 2824 | Result.vector4_u32[i] = 0x7FFFFFFF; |
||
| 2825 | } |
||
| 2826 | else if (XMISINF(V.vector4_f32[i])) |
||
| 2827 | { |
||
| 2828 | Result.vector4_f32[i] = 0.0f; |
||
| 2829 | } |
||
| 2830 | else |
||
| 2831 | { |
||
| 2832 | Result.vector4_f32[i] = 1.0f / sqrtf(V.vector4_f32[i]); |
||
| 2833 | } |
||
| 2834 | } |
||
| 2835 | |||
| 2836 | return Result; |
||
| 2837 | |||
| 2838 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2839 | return _mm_rsqrt_ps(V); |
||
| 2840 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2841 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2842 | } |
||
| 2843 | |||
| 2844 | //------------------------------------------------------------------------------ |
||
| 2845 | |||
| 2846 | XMFINLINE XMVECTOR XMVectorReciprocalSqrt |
||
| 2847 | ( |
||
| 2848 | FXMVECTOR V |
||
| 2849 | ) |
||
| 2850 | { |
||
| 2851 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2852 | |||
| 2853 | return XMVectorReciprocalSqrtEst(V); |
||
| 2854 | |||
| 2855 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2856 | XMVECTOR vResult = _mm_sqrt_ps(V); |
||
| 2857 | vResult = _mm_div_ps(g_XMOne,vResult); |
||
| 2858 | return vResult; |
||
| 2859 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2860 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2861 | } |
||
| 2862 | |||
| 2863 | //------------------------------------------------------------------------------ |
||
| 2864 | |||
| 2865 | XMFINLINE XMVECTOR XMVectorExpEst |
||
| 2866 | ( |
||
| 2867 | FXMVECTOR V |
||
| 2868 | ) |
||
| 2869 | { |
||
| 2870 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2871 | |||
| 2872 | XMVECTOR Result; |
||
| 2873 | Result.vector4_f32[0] = powf(2.0f, V.vector4_f32[0]); |
||
| 2874 | Result.vector4_f32[1] = powf(2.0f, V.vector4_f32[1]); |
||
| 2875 | Result.vector4_f32[2] = powf(2.0f, V.vector4_f32[2]); |
||
| 2876 | Result.vector4_f32[3] = powf(2.0f, V.vector4_f32[3]); |
||
| 2877 | return Result; |
||
| 2878 | |||
| 2879 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2880 | XMVECTOR vResult = _mm_setr_ps( |
||
| 2881 | powf(2.0f,XMVectorGetX(V)), |
||
| 2882 | powf(2.0f,XMVectorGetY(V)), |
||
| 2883 | powf(2.0f,XMVectorGetZ(V)), |
||
| 2884 | powf(2.0f,XMVectorGetW(V))); |
||
| 2885 | return vResult; |
||
| 2886 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2887 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2888 | } |
||
| 2889 | |||
| 2890 | //------------------------------------------------------------------------------ |
||
| 2891 | |||
| 2892 | XMINLINE XMVECTOR XMVectorExp |
||
| 2893 | ( |
||
| 2894 | FXMVECTOR V |
||
| 2895 | ) |
||
| 2896 | { |
||
| 2897 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2898 | |||
| 2899 | XMVECTOR E, S; |
||
| 2900 | XMVECTOR R, R2, R3, R4; |
||
| 2901 | XMVECTOR V0, V1; |
||
| 2902 | XMVECTOR C0X, C0Y, C0Z, C0W; |
||
| 2903 | XMVECTOR C1X, C1Y, C1Z, C1W; |
||
| 2904 | XMVECTOR Result; |
||
| 2905 | static CONST XMVECTOR C0 = {1.0f, -6.93147182e-1f, 2.40226462e-1f, -5.55036440e-2f}; |
||
| 2906 | static CONST XMVECTOR C1 = {9.61597636e-3f, -1.32823968e-3f, 1.47491097e-4f, -1.08635004e-5f}; |
||
| 2907 | |||
| 2908 | R = XMVectorFloor(V); |
||
| 2909 | E = XMVectorExpEst(R); |
||
| 2910 | R = XMVectorSubtract(V, R); |
||
| 2911 | R2 = XMVectorMultiply(R, R); |
||
| 2912 | R3 = XMVectorMultiply(R, R2); |
||
| 2913 | R4 = XMVectorMultiply(R2, R2); |
||
| 2914 | |||
| 2915 | C0X = XMVectorSplatX(C0); |
||
| 2916 | C0Y = XMVectorSplatY(C0); |
||
| 2917 | C0Z = XMVectorSplatZ(C0); |
||
| 2918 | C0W = XMVectorSplatW(C0); |
||
| 2919 | |||
| 2920 | C1X = XMVectorSplatX(C1); |
||
| 2921 | C1Y = XMVectorSplatY(C1); |
||
| 2922 | C1Z = XMVectorSplatZ(C1); |
||
| 2923 | C1W = XMVectorSplatW(C1); |
||
| 2924 | |||
| 2925 | V0 = XMVectorMultiplyAdd(R, C0Y, C0X); |
||
| 2926 | V0 = XMVectorMultiplyAdd(R2, C0Z, V0); |
||
| 2927 | V0 = XMVectorMultiplyAdd(R3, C0W, V0); |
||
| 2928 | |||
| 2929 | V1 = XMVectorMultiplyAdd(R, C1Y, C1X); |
||
| 2930 | V1 = XMVectorMultiplyAdd(R2, C1Z, V1); |
||
| 2931 | V1 = XMVectorMultiplyAdd(R3, C1W, V1); |
||
| 2932 | |||
| 2933 | S = XMVectorMultiplyAdd(R4, V1, V0); |
||
| 2934 | |||
| 2935 | S = XMVectorReciprocal(S); |
||
| 2936 | Result = XMVectorMultiply(E, S); |
||
| 2937 | |||
| 2938 | return Result; |
||
| 2939 | |||
| 2940 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 2941 | static CONST XMVECTORF32 C0 = {1.0f, -6.93147182e-1f, 2.40226462e-1f, -5.55036440e-2f}; |
||
| 2942 | static CONST XMVECTORF32 C1 = {9.61597636e-3f, -1.32823968e-3f, 1.47491097e-4f, -1.08635004e-5f}; |
||
| 2943 | |||
| 2944 | // Get the integer of the input |
||
| 2945 | XMVECTOR R = XMVectorFloor(V); |
||
| 2946 | // Get the exponent estimate |
||
| 2947 | XMVECTOR E = XMVectorExpEst(R); |
||
| 2948 | // Get the fractional only |
||
| 2949 | R = _mm_sub_ps(V,R); |
||
| 2950 | // Get R^2 |
||
| 2951 | XMVECTOR R2 = _mm_mul_ps(R,R); |
||
| 2952 | // And R^3 |
||
| 2953 | XMVECTOR R3 = _mm_mul_ps(R,R2); |
||
| 2954 | |||
| 2955 | XMVECTOR V0 = _mm_load_ps1(&C0.f[1]); |
||
| 2956 | V0 = _mm_mul_ps(V0,R); |
||
| 2957 | XMVECTOR vConstants = _mm_load_ps1(&C0.f[0]); |
||
| 2958 | V0 = _mm_add_ps(V0,vConstants); |
||
| 2959 | vConstants = _mm_load_ps1(&C0.f[2]); |
||
| 2960 | vConstants = _mm_mul_ps(vConstants,R2); |
||
| 2961 | V0 = _mm_add_ps(V0,vConstants); |
||
| 2962 | vConstants = _mm_load_ps1(&C0.f[3]); |
||
| 2963 | vConstants = _mm_mul_ps(vConstants,R3); |
||
| 2964 | V0 = _mm_add_ps(V0,vConstants); |
||
| 2965 | |||
| 2966 | XMVECTOR V1 = _mm_load_ps1(&C1.f[1]); |
||
| 2967 | V1 = _mm_mul_ps(V1,R); |
||
| 2968 | vConstants = _mm_load_ps1(&C1.f[0]); |
||
| 2969 | V1 = _mm_add_ps(V1,vConstants); |
||
| 2970 | vConstants = _mm_load_ps1(&C1.f[2]); |
||
| 2971 | vConstants = _mm_mul_ps(vConstants,R2); |
||
| 2972 | V1 = _mm_add_ps(V1,vConstants); |
||
| 2973 | vConstants = _mm_load_ps1(&C1.f[3]); |
||
| 2974 | vConstants = _mm_mul_ps(vConstants,R3); |
||
| 2975 | V1 = _mm_add_ps(V1,vConstants); |
||
| 2976 | // R2 = R^4 |
||
| 2977 | R2 = _mm_mul_ps(R2,R2); |
||
| 2978 | R2 = _mm_mul_ps(R2,V1); |
||
| 2979 | R2 = _mm_add_ps(R2,V0); |
||
| 2980 | E = _mm_div_ps(E,R2); |
||
| 2981 | return E; |
||
| 2982 | #else // _XM_VMX128_INTRINSICS_ |
||
| 2983 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 2984 | } |
||
| 2985 | |||
| 2986 | //------------------------------------------------------------------------------ |
||
| 2987 | |||
| 2988 | XMFINLINE XMVECTOR XMVectorLogEst |
||
| 2989 | ( |
||
| 2990 | FXMVECTOR V |
||
| 2991 | ) |
||
| 2992 | { |
||
| 2993 | #if defined(_XM_NO_INTRINSICS_) |
||
| 2994 | |||
| 2995 | FLOAT fScale = (1.0f / logf(2.0f)); |
||
| 2996 | XMVECTOR Result; |
||
| 2997 | |||
| 2998 | Result.vector4_f32[0] = logf(V.vector4_f32[0])*fScale; |
||
| 2999 | Result.vector4_f32[1] = logf(V.vector4_f32[1])*fScale; |
||
| 3000 | Result.vector4_f32[2] = logf(V.vector4_f32[2])*fScale; |
||
| 3001 | Result.vector4_f32[3] = logf(V.vector4_f32[3])*fScale; |
||
| 3002 | return Result; |
||
| 3003 | |||
| 3004 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3005 | XMVECTOR vScale = _mm_set_ps1(1.0f / logf(2.0f)); |
||
| 3006 | XMVECTOR vResult = _mm_setr_ps( |
||
| 3007 | logf(XMVectorGetX(V)), |
||
| 3008 | logf(XMVectorGetY(V)), |
||
| 3009 | logf(XMVectorGetZ(V)), |
||
| 3010 | logf(XMVectorGetW(V))); |
||
| 3011 | vResult = _mm_mul_ps(vResult,vScale); |
||
| 3012 | return vResult; |
||
| 3013 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3014 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3015 | } |
||
| 3016 | |||
| 3017 | //------------------------------------------------------------------------------ |
||
| 3018 | |||
| 3019 | XMINLINE XMVECTOR XMVectorLog |
||
| 3020 | ( |
||
| 3021 | FXMVECTOR V |
||
| 3022 | ) |
||
| 3023 | { |
||
| 3024 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3025 | FLOAT fScale = (1.0f / logf(2.0f)); |
||
| 3026 | XMVECTOR Result; |
||
| 3027 | |||
| 3028 | Result.vector4_f32[0] = logf(V.vector4_f32[0])*fScale; |
||
| 3029 | Result.vector4_f32[1] = logf(V.vector4_f32[1])*fScale; |
||
| 3030 | Result.vector4_f32[2] = logf(V.vector4_f32[2])*fScale; |
||
| 3031 | Result.vector4_f32[3] = logf(V.vector4_f32[3])*fScale; |
||
| 3032 | return Result; |
||
| 3033 | |||
| 3034 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3035 | XMVECTOR vScale = _mm_set_ps1(1.0f / logf(2.0f)); |
||
| 3036 | XMVECTOR vResult = _mm_setr_ps( |
||
| 3037 | logf(XMVectorGetX(V)), |
||
| 3038 | logf(XMVectorGetY(V)), |
||
| 3039 | logf(XMVectorGetZ(V)), |
||
| 3040 | logf(XMVectorGetW(V))); |
||
| 3041 | vResult = _mm_mul_ps(vResult,vScale); |
||
| 3042 | return vResult; |
||
| 3043 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3044 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3045 | } |
||
| 3046 | |||
| 3047 | //------------------------------------------------------------------------------ |
||
| 3048 | |||
| 3049 | XMFINLINE XMVECTOR XMVectorPowEst |
||
| 3050 | ( |
||
| 3051 | FXMVECTOR V1, |
||
| 3052 | FXMVECTOR V2 |
||
| 3053 | ) |
||
| 3054 | { |
||
| 3055 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3056 | |||
| 3057 | XMVECTOR Result; |
||
| 3058 | |||
| 3059 | Result.vector4_f32[0] = powf(V1.vector4_f32[0], V2.vector4_f32[0]); |
||
| 3060 | Result.vector4_f32[1] = powf(V1.vector4_f32[1], V2.vector4_f32[1]); |
||
| 3061 | Result.vector4_f32[2] = powf(V1.vector4_f32[2], V2.vector4_f32[2]); |
||
| 3062 | Result.vector4_f32[3] = powf(V1.vector4_f32[3], V2.vector4_f32[3]); |
||
| 3063 | |||
| 3064 | return Result; |
||
| 3065 | |||
| 3066 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3067 | XMVECTOR vResult = _mm_setr_ps( |
||
| 3068 | powf(XMVectorGetX(V1),XMVectorGetX(V2)), |
||
| 3069 | powf(XMVectorGetY(V1),XMVectorGetY(V2)), |
||
| 3070 | powf(XMVectorGetZ(V1),XMVectorGetZ(V2)), |
||
| 3071 | powf(XMVectorGetW(V1),XMVectorGetW(V2))); |
||
| 3072 | return vResult; |
||
| 3073 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3074 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3075 | } |
||
| 3076 | |||
| 3077 | //------------------------------------------------------------------------------ |
||
| 3078 | |||
| 3079 | XMFINLINE XMVECTOR XMVectorPow |
||
| 3080 | ( |
||
| 3081 | FXMVECTOR V1, |
||
| 3082 | FXMVECTOR V2 |
||
| 3083 | ) |
||
| 3084 | { |
||
| 3085 | #if defined(_XM_NO_INTRINSICS_) || defined(_XM_SSE_INTRINSICS_) |
||
| 3086 | |||
| 3087 | return XMVectorPowEst(V1, V2); |
||
| 3088 | |||
| 3089 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3090 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3091 | } |
||
| 3092 | |||
| 3093 | //------------------------------------------------------------------------------ |
||
| 3094 | |||
| 3095 | XMFINLINE XMVECTOR XMVectorAbs |
||
| 3096 | ( |
||
| 3097 | FXMVECTOR V |
||
| 3098 | ) |
||
| 3099 | { |
||
| 3100 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3101 | XMVECTOR vResult = { |
||
| 3102 | fabsf(V.vector4_f32[0]), |
||
| 3103 | fabsf(V.vector4_f32[1]), |
||
| 3104 | fabsf(V.vector4_f32[2]), |
||
| 3105 | fabsf(V.vector4_f32[3]) |
||
| 3106 | }; |
||
| 3107 | return vResult; |
||
| 3108 | |||
| 3109 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3110 | XMVECTOR vResult = _mm_setzero_ps(); |
||
| 3111 | vResult = _mm_sub_ps(vResult,V); |
||
| 3112 | vResult = _mm_max_ps(vResult,V); |
||
| 3113 | return vResult; |
||
| 3114 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3115 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3116 | } |
||
| 3117 | |||
| 3118 | //------------------------------------------------------------------------------ |
||
| 3119 | |||
| 3120 | XMFINLINE XMVECTOR XMVectorMod |
||
| 3121 | ( |
||
| 3122 | FXMVECTOR V1, |
||
| 3123 | FXMVECTOR V2 |
||
| 3124 | ) |
||
| 3125 | { |
||
| 3126 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3127 | |||
| 3128 | XMVECTOR Reciprocal; |
||
| 3129 | XMVECTOR Quotient; |
||
| 3130 | XMVECTOR Result; |
||
| 3131 | |||
| 3132 | // V1 % V2 = V1 - V2 * truncate(V1 / V2) |
||
| 3133 | Reciprocal = XMVectorReciprocal(V2); |
||
| 3134 | Quotient = XMVectorMultiply(V1, Reciprocal); |
||
| 3135 | Quotient = XMVectorTruncate(Quotient); |
||
| 3136 | Result = XMVectorNegativeMultiplySubtract(V2, Quotient, V1); |
||
| 3137 | |||
| 3138 | return Result; |
||
| 3139 | |||
| 3140 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3141 | XMVECTOR vResult = _mm_div_ps(V1, V2); |
||
| 3142 | vResult = XMVectorTruncate(vResult); |
||
| 3143 | vResult = _mm_mul_ps(vResult,V2); |
||
| 3144 | vResult = _mm_sub_ps(V1,vResult); |
||
| 3145 | return vResult; |
||
| 3146 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3147 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3148 | } |
||
| 3149 | |||
| 3150 | //------------------------------------------------------------------------------ |
||
| 3151 | |||
| 3152 | XMFINLINE XMVECTOR XMVectorModAngles |
||
| 3153 | ( |
||
| 3154 | FXMVECTOR Angles |
||
| 3155 | ) |
||
| 3156 | { |
||
| 3157 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3158 | |||
| 3159 | XMVECTOR V; |
||
| 3160 | XMVECTOR Result; |
||
| 3161 | |||
| 3162 | // Modulo the range of the given angles such that -XM_PI <= Angles < XM_PI |
||
| 3163 | V = XMVectorMultiply(Angles, g_XMReciprocalTwoPi.v); |
||
| 3164 | V = XMVectorRound(V); |
||
| 3165 | Result = XMVectorNegativeMultiplySubtract(g_XMTwoPi.v, V, Angles); |
||
| 3166 | |||
| 3167 | return Result; |
||
| 3168 | |||
| 3169 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3170 | // Modulo the range of the given angles such that -XM_PI <= Angles < XM_PI |
||
| 3171 | XMVECTOR vResult = _mm_mul_ps(Angles,g_XMReciprocalTwoPi); |
||
| 3172 | // Use the inline function due to complexity for rounding |
||
| 3173 | vResult = XMVectorRound(vResult); |
||
| 3174 | vResult = _mm_mul_ps(vResult,g_XMTwoPi); |
||
| 3175 | vResult = _mm_sub_ps(Angles,vResult); |
||
| 3176 | return vResult; |
||
| 3177 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3178 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3179 | } |
||
| 3180 | |||
| 3181 | //------------------------------------------------------------------------------ |
||
| 3182 | |||
| 3183 | XMINLINE XMVECTOR XMVectorSin |
||
| 3184 | ( |
||
| 3185 | FXMVECTOR V |
||
| 3186 | ) |
||
| 3187 | { |
||
| 3188 | |||
| 3189 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3190 | |||
| 3191 | XMVECTOR V1, V2, V3, V5, V7, V9, V11, V13, V15, V17, V19, V21, V23; |
||
| 3192 | XMVECTOR S1, S2, S3, S4, S5, S6, S7, S8, S9, S10, S11; |
||
| 3193 | XMVECTOR Result; |
||
| 3194 | |||
| 3195 | V1 = XMVectorModAngles(V); |
||
| 3196 | |||
| 3197 | // sin(V) ~= V - V^3 / 3! + V^5 / 5! - V^7 / 7! + V^9 / 9! - V^11 / 11! + V^13 / 13! - |
||
| 3198 | // V^15 / 15! + V^17 / 17! - V^19 / 19! + V^21 / 21! - V^23 / 23! (for -PI <= V < PI) |
||
| 3199 | V2 = XMVectorMultiply(V1, V1); |
||
| 3200 | V3 = XMVectorMultiply(V2, V1); |
||
| 3201 | V5 = XMVectorMultiply(V3, V2); |
||
| 3202 | V7 = XMVectorMultiply(V5, V2); |
||
| 3203 | V9 = XMVectorMultiply(V7, V2); |
||
| 3204 | V11 = XMVectorMultiply(V9, V2); |
||
| 3205 | V13 = XMVectorMultiply(V11, V2); |
||
| 3206 | V15 = XMVectorMultiply(V13, V2); |
||
| 3207 | V17 = XMVectorMultiply(V15, V2); |
||
| 3208 | V19 = XMVectorMultiply(V17, V2); |
||
| 3209 | V21 = XMVectorMultiply(V19, V2); |
||
| 3210 | V23 = XMVectorMultiply(V21, V2); |
||
| 3211 | |||
| 3212 | S1 = XMVectorSplatY(g_XMSinCoefficients0.v); |
||
| 3213 | S2 = XMVectorSplatZ(g_XMSinCoefficients0.v); |
||
| 3214 | S3 = XMVectorSplatW(g_XMSinCoefficients0.v); |
||
| 3215 | S4 = XMVectorSplatX(g_XMSinCoefficients1.v); |
||
| 3216 | S5 = XMVectorSplatY(g_XMSinCoefficients1.v); |
||
| 3217 | S6 = XMVectorSplatZ(g_XMSinCoefficients1.v); |
||
| 3218 | S7 = XMVectorSplatW(g_XMSinCoefficients1.v); |
||
| 3219 | S8 = XMVectorSplatX(g_XMSinCoefficients2.v); |
||
| 3220 | S9 = XMVectorSplatY(g_XMSinCoefficients2.v); |
||
| 3221 | S10 = XMVectorSplatZ(g_XMSinCoefficients2.v); |
||
| 3222 | S11 = XMVectorSplatW(g_XMSinCoefficients2.v); |
||
| 3223 | |||
| 3224 | Result = XMVectorMultiplyAdd(S1, V3, V1); |
||
| 3225 | Result = XMVectorMultiplyAdd(S2, V5, Result); |
||
| 3226 | Result = XMVectorMultiplyAdd(S3, V7, Result); |
||
| 3227 | Result = XMVectorMultiplyAdd(S4, V9, Result); |
||
| 3228 | Result = XMVectorMultiplyAdd(S5, V11, Result); |
||
| 3229 | Result = XMVectorMultiplyAdd(S6, V13, Result); |
||
| 3230 | Result = XMVectorMultiplyAdd(S7, V15, Result); |
||
| 3231 | Result = XMVectorMultiplyAdd(S8, V17, Result); |
||
| 3232 | Result = XMVectorMultiplyAdd(S9, V19, Result); |
||
| 3233 | Result = XMVectorMultiplyAdd(S10, V21, Result); |
||
| 3234 | Result = XMVectorMultiplyAdd(S11, V23, Result); |
||
| 3235 | |||
| 3236 | return Result; |
||
| 3237 | |||
| 3238 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3239 | // Force the value within the bounds of pi |
||
| 3240 | XMVECTOR vResult = XMVectorModAngles(V); |
||
| 3241 | // Each on is V to the "num" power |
||
| 3242 | // V2 = V1^2 |
||
| 3243 | XMVECTOR V2 = _mm_mul_ps(vResult,vResult); |
||
| 3244 | // V1^3 |
||
| 3245 | XMVECTOR vPower = _mm_mul_ps(vResult,V2); |
||
| 3246 | XMVECTOR vConstants = _mm_load_ps1(&g_XMSinCoefficients0.f[1]); |
||
| 3247 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3248 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3249 | |||
| 3250 | // V^5 |
||
| 3251 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3252 | vConstants = _mm_load_ps1(&g_XMSinCoefficients0.f[2]); |
||
| 3253 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3254 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3255 | |||
| 3256 | // V^7 |
||
| 3257 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3258 | vConstants = _mm_load_ps1(&g_XMSinCoefficients0.f[3]); |
||
| 3259 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3260 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3261 | |||
| 3262 | // V^9 |
||
| 3263 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3264 | vConstants = _mm_load_ps1(&g_XMSinCoefficients1.f[0]); |
||
| 3265 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3266 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3267 | |||
| 3268 | // V^11 |
||
| 3269 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3270 | vConstants = _mm_load_ps1(&g_XMSinCoefficients1.f[1]); |
||
| 3271 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3272 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3273 | |||
| 3274 | // V^13 |
||
| 3275 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3276 | vConstants = _mm_load_ps1(&g_XMSinCoefficients1.f[2]); |
||
| 3277 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3278 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3279 | |||
| 3280 | // V^15 |
||
| 3281 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3282 | vConstants = _mm_load_ps1(&g_XMSinCoefficients1.f[3]); |
||
| 3283 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3284 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3285 | |||
| 3286 | // V^17 |
||
| 3287 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3288 | vConstants = _mm_load_ps1(&g_XMSinCoefficients2.f[0]); |
||
| 3289 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3290 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3291 | |||
| 3292 | // V^19 |
||
| 3293 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3294 | vConstants = _mm_load_ps1(&g_XMSinCoefficients2.f[1]); |
||
| 3295 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3296 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3297 | |||
| 3298 | // V^21 |
||
| 3299 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3300 | vConstants = _mm_load_ps1(&g_XMSinCoefficients2.f[2]); |
||
| 3301 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3302 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3303 | |||
| 3304 | // V^23 |
||
| 3305 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3306 | vConstants = _mm_load_ps1(&g_XMSinCoefficients2.f[3]); |
||
| 3307 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3308 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3309 | return vResult; |
||
| 3310 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3311 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3312 | } |
||
| 3313 | |||
| 3314 | //------------------------------------------------------------------------------ |
||
| 3315 | |||
| 3316 | XMINLINE XMVECTOR XMVectorCos |
||
| 3317 | ( |
||
| 3318 | FXMVECTOR V |
||
| 3319 | ) |
||
| 3320 | { |
||
| 3321 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3322 | |||
| 3323 | XMVECTOR V1, V2, V4, V6, V8, V10, V12, V14, V16, V18, V20, V22; |
||
| 3324 | XMVECTOR C1, C2, C3, C4, C5, C6, C7, C8, C9, C10, C11; |
||
| 3325 | XMVECTOR Result; |
||
| 3326 | |||
| 3327 | V1 = XMVectorModAngles(V); |
||
| 3328 | |||
| 3329 | // cos(V) ~= 1 - V^2 / 2! + V^4 / 4! - V^6 / 6! + V^8 / 8! - V^10 / 10! + V^12 / 12! - |
||
| 3330 | // V^14 / 14! + V^16 / 16! - V^18 / 18! + V^20 / 20! - V^22 / 22! (for -PI <= V < PI) |
||
| 3331 | V2 = XMVectorMultiply(V1, V1); |
||
| 3332 | V4 = XMVectorMultiply(V2, V2); |
||
| 3333 | V6 = XMVectorMultiply(V4, V2); |
||
| 3334 | V8 = XMVectorMultiply(V4, V4); |
||
| 3335 | V10 = XMVectorMultiply(V6, V4); |
||
| 3336 | V12 = XMVectorMultiply(V6, V6); |
||
| 3337 | V14 = XMVectorMultiply(V8, V6); |
||
| 3338 | V16 = XMVectorMultiply(V8, V8); |
||
| 3339 | V18 = XMVectorMultiply(V10, V8); |
||
| 3340 | V20 = XMVectorMultiply(V10, V10); |
||
| 3341 | V22 = XMVectorMultiply(V12, V10); |
||
| 3342 | |||
| 3343 | C1 = XMVectorSplatY(g_XMCosCoefficients0.v); |
||
| 3344 | C2 = XMVectorSplatZ(g_XMCosCoefficients0.v); |
||
| 3345 | C3 = XMVectorSplatW(g_XMCosCoefficients0.v); |
||
| 3346 | C4 = XMVectorSplatX(g_XMCosCoefficients1.v); |
||
| 3347 | C5 = XMVectorSplatY(g_XMCosCoefficients1.v); |
||
| 3348 | C6 = XMVectorSplatZ(g_XMCosCoefficients1.v); |
||
| 3349 | C7 = XMVectorSplatW(g_XMCosCoefficients1.v); |
||
| 3350 | C8 = XMVectorSplatX(g_XMCosCoefficients2.v); |
||
| 3351 | C9 = XMVectorSplatY(g_XMCosCoefficients2.v); |
||
| 3352 | C10 = XMVectorSplatZ(g_XMCosCoefficients2.v); |
||
| 3353 | C11 = XMVectorSplatW(g_XMCosCoefficients2.v); |
||
| 3354 | |||
| 3355 | Result = XMVectorMultiplyAdd(C1, V2, g_XMOne.v); |
||
| 3356 | Result = XMVectorMultiplyAdd(C2, V4, Result); |
||
| 3357 | Result = XMVectorMultiplyAdd(C3, V6, Result); |
||
| 3358 | Result = XMVectorMultiplyAdd(C4, V8, Result); |
||
| 3359 | Result = XMVectorMultiplyAdd(C5, V10, Result); |
||
| 3360 | Result = XMVectorMultiplyAdd(C6, V12, Result); |
||
| 3361 | Result = XMVectorMultiplyAdd(C7, V14, Result); |
||
| 3362 | Result = XMVectorMultiplyAdd(C8, V16, Result); |
||
| 3363 | Result = XMVectorMultiplyAdd(C9, V18, Result); |
||
| 3364 | Result = XMVectorMultiplyAdd(C10, V20, Result); |
||
| 3365 | Result = XMVectorMultiplyAdd(C11, V22, Result); |
||
| 3366 | |||
| 3367 | return Result; |
||
| 3368 | |||
| 3369 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3370 | // Force the value within the bounds of pi |
||
| 3371 | XMVECTOR V2 = XMVectorModAngles(V); |
||
| 3372 | // Each on is V to the "num" power |
||
| 3373 | // V2 = V1^2 |
||
| 3374 | V2 = _mm_mul_ps(V2,V2); |
||
| 3375 | // V^2 |
||
| 3376 | XMVECTOR vConstants = _mm_load_ps1(&g_XMCosCoefficients0.f[1]); |
||
| 3377 | vConstants = _mm_mul_ps(vConstants,V2); |
||
| 3378 | XMVECTOR vResult = _mm_add_ps(vConstants,g_XMOne); |
||
| 3379 | |||
| 3380 | // V^4 |
||
| 3381 | XMVECTOR vPower = _mm_mul_ps(V2,V2); |
||
| 3382 | vConstants = _mm_load_ps1(&g_XMCosCoefficients0.f[2]); |
||
| 3383 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3384 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3385 | |||
| 3386 | // V^6 |
||
| 3387 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3388 | vConstants = _mm_load_ps1(&g_XMCosCoefficients0.f[3]); |
||
| 3389 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3390 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3391 | |||
| 3392 | // V^8 |
||
| 3393 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3394 | vConstants = _mm_load_ps1(&g_XMCosCoefficients1.f[0]); |
||
| 3395 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3396 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3397 | |||
| 3398 | // V^10 |
||
| 3399 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3400 | vConstants = _mm_load_ps1(&g_XMCosCoefficients1.f[1]); |
||
| 3401 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3402 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3403 | |||
| 3404 | // V^12 |
||
| 3405 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3406 | vConstants = _mm_load_ps1(&g_XMCosCoefficients1.f[2]); |
||
| 3407 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3408 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3409 | |||
| 3410 | // V^14 |
||
| 3411 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3412 | vConstants = _mm_load_ps1(&g_XMCosCoefficients1.f[3]); |
||
| 3413 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3414 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3415 | |||
| 3416 | // V^16 |
||
| 3417 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3418 | vConstants = _mm_load_ps1(&g_XMCosCoefficients2.f[0]); |
||
| 3419 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3420 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3421 | |||
| 3422 | // V^18 |
||
| 3423 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3424 | vConstants = _mm_load_ps1(&g_XMCosCoefficients2.f[1]); |
||
| 3425 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3426 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3427 | |||
| 3428 | // V^20 |
||
| 3429 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3430 | vConstants = _mm_load_ps1(&g_XMCosCoefficients2.f[2]); |
||
| 3431 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3432 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3433 | |||
| 3434 | // V^22 |
||
| 3435 | vPower = _mm_mul_ps(vPower,V2); |
||
| 3436 | vConstants = _mm_load_ps1(&g_XMCosCoefficients2.f[3]); |
||
| 3437 | vConstants = _mm_mul_ps(vConstants,vPower); |
||
| 3438 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 3439 | return vResult; |
||
| 3440 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3441 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3442 | } |
||
| 3443 | |||
| 3444 | //------------------------------------------------------------------------------ |
||
| 3445 | |||
| 3446 | XMINLINE VOID XMVectorSinCos |
||
| 3447 | ( |
||
| 3448 | XMVECTOR* pSin, |
||
| 3449 | XMVECTOR* pCos, |
||
| 3450 | FXMVECTOR V |
||
| 3451 | ) |
||
| 3452 | { |
||
| 3453 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3454 | |||
| 3455 | XMVECTOR V1, V2, V3, V4, V5, V6, V7, V8, V9, V10, V11, V12, V13; |
||
| 3456 | XMVECTOR V14, V15, V16, V17, V18, V19, V20, V21, V22, V23; |
||
| 3457 | XMVECTOR S1, S2, S3, S4, S5, S6, S7, S8, S9, S10, S11; |
||
| 3458 | XMVECTOR C1, C2, C3, C4, C5, C6, C7, C8, C9, C10, C11; |
||
| 3459 | XMVECTOR Sin, Cos; |
||
| 3460 | |||
| 3461 | XMASSERT(pSin); |
||
| 3462 | XMASSERT(pCos); |
||
| 3463 | |||
| 3464 | V1 = XMVectorModAngles(V); |
||
| 3465 | |||
| 3466 | // sin(V) ~= V - V^3 / 3! + V^5 / 5! - V^7 / 7! + V^9 / 9! - V^11 / 11! + V^13 / 13! - |
||
| 3467 | // V^15 / 15! + V^17 / 17! - V^19 / 19! + V^21 / 21! - V^23 / 23! (for -PI <= V < PI) |
||
| 3468 | // cos(V) ~= 1 - V^2 / 2! + V^4 / 4! - V^6 / 6! + V^8 / 8! - V^10 / 10! + V^12 / 12! - |
||
| 3469 | // V^14 / 14! + V^16 / 16! - V^18 / 18! + V^20 / 20! - V^22 / 22! (for -PI <= V < PI) |
||
| 3470 | |||
| 3471 | V2 = XMVectorMultiply(V1, V1); |
||
| 3472 | V3 = XMVectorMultiply(V2, V1); |
||
| 3473 | V4 = XMVectorMultiply(V2, V2); |
||
| 3474 | V5 = XMVectorMultiply(V3, V2); |
||
| 3475 | V6 = XMVectorMultiply(V3, V3); |
||
| 3476 | V7 = XMVectorMultiply(V4, V3); |
||
| 3477 | V8 = XMVectorMultiply(V4, V4); |
||
| 3478 | V9 = XMVectorMultiply(V5, V4); |
||
| 3479 | V10 = XMVectorMultiply(V5, V5); |
||
| 3480 | V11 = XMVectorMultiply(V6, V5); |
||
| 3481 | V12 = XMVectorMultiply(V6, V6); |
||
| 3482 | V13 = XMVectorMultiply(V7, V6); |
||
| 3483 | V14 = XMVectorMultiply(V7, V7); |
||
| 3484 | V15 = XMVectorMultiply(V8, V7); |
||
| 3485 | V16 = XMVectorMultiply(V8, V8); |
||
| 3486 | V17 = XMVectorMultiply(V9, V8); |
||
| 3487 | V18 = XMVectorMultiply(V9, V9); |
||
| 3488 | V19 = XMVectorMultiply(V10, V9); |
||
| 3489 | V20 = XMVectorMultiply(V10, V10); |
||
| 3490 | V21 = XMVectorMultiply(V11, V10); |
||
| 3491 | V22 = XMVectorMultiply(V11, V11); |
||
| 3492 | V23 = XMVectorMultiply(V12, V11); |
||
| 3493 | |||
| 3494 | S1 = XMVectorSplatY(g_XMSinCoefficients0.v); |
||
| 3495 | S2 = XMVectorSplatZ(g_XMSinCoefficients0.v); |
||
| 3496 | S3 = XMVectorSplatW(g_XMSinCoefficients0.v); |
||
| 3497 | S4 = XMVectorSplatX(g_XMSinCoefficients1.v); |
||
| 3498 | S5 = XMVectorSplatY(g_XMSinCoefficients1.v); |
||
| 3499 | S6 = XMVectorSplatZ(g_XMSinCoefficients1.v); |
||
| 3500 | S7 = XMVectorSplatW(g_XMSinCoefficients1.v); |
||
| 3501 | S8 = XMVectorSplatX(g_XMSinCoefficients2.v); |
||
| 3502 | S9 = XMVectorSplatY(g_XMSinCoefficients2.v); |
||
| 3503 | S10 = XMVectorSplatZ(g_XMSinCoefficients2.v); |
||
| 3504 | S11 = XMVectorSplatW(g_XMSinCoefficients2.v); |
||
| 3505 | |||
| 3506 | C1 = XMVectorSplatY(g_XMCosCoefficients0.v); |
||
| 3507 | C2 = XMVectorSplatZ(g_XMCosCoefficients0.v); |
||
| 3508 | C3 = XMVectorSplatW(g_XMCosCoefficients0.v); |
||
| 3509 | C4 = XMVectorSplatX(g_XMCosCoefficients1.v); |
||
| 3510 | C5 = XMVectorSplatY(g_XMCosCoefficients1.v); |
||
| 3511 | C6 = XMVectorSplatZ(g_XMCosCoefficients1.v); |
||
| 3512 | C7 = XMVectorSplatW(g_XMCosCoefficients1.v); |
||
| 3513 | C8 = XMVectorSplatX(g_XMCosCoefficients2.v); |
||
| 3514 | C9 = XMVectorSplatY(g_XMCosCoefficients2.v); |
||
| 3515 | C10 = XMVectorSplatZ(g_XMCosCoefficients2.v); |
||
| 3516 | C11 = XMVectorSplatW(g_XMCosCoefficients2.v); |
||
| 3517 | |||
| 3518 | Sin = XMVectorMultiplyAdd(S1, V3, V1); |
||
| 3519 | Sin = XMVectorMultiplyAdd(S2, V5, Sin); |
||
| 3520 | Sin = XMVectorMultiplyAdd(S3, V7, Sin); |
||
| 3521 | Sin = XMVectorMultiplyAdd(S4, V9, Sin); |
||
| 3522 | Sin = XMVectorMultiplyAdd(S5, V11, Sin); |
||
| 3523 | Sin = XMVectorMultiplyAdd(S6, V13, Sin); |
||
| 3524 | Sin = XMVectorMultiplyAdd(S7, V15, Sin); |
||
| 3525 | Sin = XMVectorMultiplyAdd(S8, V17, Sin); |
||
| 3526 | Sin = XMVectorMultiplyAdd(S9, V19, Sin); |
||
| 3527 | Sin = XMVectorMultiplyAdd(S10, V21, Sin); |
||
| 3528 | Sin = XMVectorMultiplyAdd(S11, V23, Sin); |
||
| 3529 | |||
| 3530 | Cos = XMVectorMultiplyAdd(C1, V2, g_XMOne.v); |
||
| 3531 | Cos = XMVectorMultiplyAdd(C2, V4, Cos); |
||
| 3532 | Cos = XMVectorMultiplyAdd(C3, V6, Cos); |
||
| 3533 | Cos = XMVectorMultiplyAdd(C4, V8, Cos); |
||
| 3534 | Cos = XMVectorMultiplyAdd(C5, V10, Cos); |
||
| 3535 | Cos = XMVectorMultiplyAdd(C6, V12, Cos); |
||
| 3536 | Cos = XMVectorMultiplyAdd(C7, V14, Cos); |
||
| 3537 | Cos = XMVectorMultiplyAdd(C8, V16, Cos); |
||
| 3538 | Cos = XMVectorMultiplyAdd(C9, V18, Cos); |
||
| 3539 | Cos = XMVectorMultiplyAdd(C10, V20, Cos); |
||
| 3540 | Cos = XMVectorMultiplyAdd(C11, V22, Cos); |
||
| 3541 | |||
| 3542 | *pSin = Sin; |
||
| 3543 | *pCos = Cos; |
||
| 3544 | |||
| 3545 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3546 | XMASSERT(pSin); |
||
| 3547 | XMASSERT(pCos); |
||
| 3548 | XMVECTOR V1, V2, V3, V4, V5, V6, V7, V8, V9, V10, V11, V12, V13; |
||
| 3549 | XMVECTOR V14, V15, V16, V17, V18, V19, V20, V21, V22, V23; |
||
| 3550 | XMVECTOR S1, S2, S3, S4, S5, S6, S7, S8, S9, S10, S11; |
||
| 3551 | XMVECTOR C1, C2, C3, C4, C5, C6, C7, C8, C9, C10, C11; |
||
| 3552 | XMVECTOR Sin, Cos; |
||
| 3553 | |||
| 3554 | V1 = XMVectorModAngles(V); |
||
| 3555 | |||
| 3556 | // sin(V) ~= V - V^3 / 3! + V^5 / 5! - V^7 / 7! + V^9 / 9! - V^11 / 11! + V^13 / 13! - |
||
| 3557 | // V^15 / 15! + V^17 / 17! - V^19 / 19! + V^21 / 21! - V^23 / 23! (for -PI <= V < PI) |
||
| 3558 | // cos(V) ~= 1 - V^2 / 2! + V^4 / 4! - V^6 / 6! + V^8 / 8! - V^10 / 10! + V^12 / 12! - |
||
| 3559 | // V^14 / 14! + V^16 / 16! - V^18 / 18! + V^20 / 20! - V^22 / 22! (for -PI <= V < PI) |
||
| 3560 | |||
| 3561 | V2 = XMVectorMultiply(V1, V1); |
||
| 3562 | V3 = XMVectorMultiply(V2, V1); |
||
| 3563 | V4 = XMVectorMultiply(V2, V2); |
||
| 3564 | V5 = XMVectorMultiply(V3, V2); |
||
| 3565 | V6 = XMVectorMultiply(V3, V3); |
||
| 3566 | V7 = XMVectorMultiply(V4, V3); |
||
| 3567 | V8 = XMVectorMultiply(V4, V4); |
||
| 3568 | V9 = XMVectorMultiply(V5, V4); |
||
| 3569 | V10 = XMVectorMultiply(V5, V5); |
||
| 3570 | V11 = XMVectorMultiply(V6, V5); |
||
| 3571 | V12 = XMVectorMultiply(V6, V6); |
||
| 3572 | V13 = XMVectorMultiply(V7, V6); |
||
| 3573 | V14 = XMVectorMultiply(V7, V7); |
||
| 3574 | V15 = XMVectorMultiply(V8, V7); |
||
| 3575 | V16 = XMVectorMultiply(V8, V8); |
||
| 3576 | V17 = XMVectorMultiply(V9, V8); |
||
| 3577 | V18 = XMVectorMultiply(V9, V9); |
||
| 3578 | V19 = XMVectorMultiply(V10, V9); |
||
| 3579 | V20 = XMVectorMultiply(V10, V10); |
||
| 3580 | V21 = XMVectorMultiply(V11, V10); |
||
| 3581 | V22 = XMVectorMultiply(V11, V11); |
||
| 3582 | V23 = XMVectorMultiply(V12, V11); |
||
| 3583 | |||
| 3584 | S1 = _mm_load_ps1(&g_XMSinCoefficients0.f[1]); |
||
| 3585 | S2 = _mm_load_ps1(&g_XMSinCoefficients0.f[2]); |
||
| 3586 | S3 = _mm_load_ps1(&g_XMSinCoefficients0.f[3]); |
||
| 3587 | S4 = _mm_load_ps1(&g_XMSinCoefficients1.f[0]); |
||
| 3588 | S5 = _mm_load_ps1(&g_XMSinCoefficients1.f[1]); |
||
| 3589 | S6 = _mm_load_ps1(&g_XMSinCoefficients1.f[2]); |
||
| 3590 | S7 = _mm_load_ps1(&g_XMSinCoefficients1.f[3]); |
||
| 3591 | S8 = _mm_load_ps1(&g_XMSinCoefficients2.f[0]); |
||
| 3592 | S9 = _mm_load_ps1(&g_XMSinCoefficients2.f[1]); |
||
| 3593 | S10 = _mm_load_ps1(&g_XMSinCoefficients2.f[2]); |
||
| 3594 | S11 = _mm_load_ps1(&g_XMSinCoefficients2.f[3]); |
||
| 3595 | |||
| 3596 | C1 = _mm_load_ps1(&g_XMCosCoefficients0.f[1]); |
||
| 3597 | C2 = _mm_load_ps1(&g_XMCosCoefficients0.f[2]); |
||
| 3598 | C3 = _mm_load_ps1(&g_XMCosCoefficients0.f[3]); |
||
| 3599 | C4 = _mm_load_ps1(&g_XMCosCoefficients1.f[0]); |
||
| 3600 | C5 = _mm_load_ps1(&g_XMCosCoefficients1.f[1]); |
||
| 3601 | C6 = _mm_load_ps1(&g_XMCosCoefficients1.f[2]); |
||
| 3602 | C7 = _mm_load_ps1(&g_XMCosCoefficients1.f[3]); |
||
| 3603 | C8 = _mm_load_ps1(&g_XMCosCoefficients2.f[0]); |
||
| 3604 | C9 = _mm_load_ps1(&g_XMCosCoefficients2.f[1]); |
||
| 3605 | C10 = _mm_load_ps1(&g_XMCosCoefficients2.f[2]); |
||
| 3606 | C11 = _mm_load_ps1(&g_XMCosCoefficients2.f[3]); |
||
| 3607 | |||
| 3608 | S1 = _mm_mul_ps(S1,V3); |
||
| 3609 | Sin = _mm_add_ps(S1,V1); |
||
| 3610 | Sin = XMVectorMultiplyAdd(S2, V5, Sin); |
||
| 3611 | Sin = XMVectorMultiplyAdd(S3, V7, Sin); |
||
| 3612 | Sin = XMVectorMultiplyAdd(S4, V9, Sin); |
||
| 3613 | Sin = XMVectorMultiplyAdd(S5, V11, Sin); |
||
| 3614 | Sin = XMVectorMultiplyAdd(S6, V13, Sin); |
||
| 3615 | Sin = XMVectorMultiplyAdd(S7, V15, Sin); |
||
| 3616 | Sin = XMVectorMultiplyAdd(S8, V17, Sin); |
||
| 3617 | Sin = XMVectorMultiplyAdd(S9, V19, Sin); |
||
| 3618 | Sin = XMVectorMultiplyAdd(S10, V21, Sin); |
||
| 3619 | Sin = XMVectorMultiplyAdd(S11, V23, Sin); |
||
| 3620 | |||
| 3621 | Cos = _mm_mul_ps(C1,V2); |
||
| 3622 | Cos = _mm_add_ps(Cos,g_XMOne); |
||
| 3623 | Cos = XMVectorMultiplyAdd(C2, V4, Cos); |
||
| 3624 | Cos = XMVectorMultiplyAdd(C3, V6, Cos); |
||
| 3625 | Cos = XMVectorMultiplyAdd(C4, V8, Cos); |
||
| 3626 | Cos = XMVectorMultiplyAdd(C5, V10, Cos); |
||
| 3627 | Cos = XMVectorMultiplyAdd(C6, V12, Cos); |
||
| 3628 | Cos = XMVectorMultiplyAdd(C7, V14, Cos); |
||
| 3629 | Cos = XMVectorMultiplyAdd(C8, V16, Cos); |
||
| 3630 | Cos = XMVectorMultiplyAdd(C9, V18, Cos); |
||
| 3631 | Cos = XMVectorMultiplyAdd(C10, V20, Cos); |
||
| 3632 | Cos = XMVectorMultiplyAdd(C11, V22, Cos); |
||
| 3633 | |||
| 3634 | *pSin = Sin; |
||
| 3635 | *pCos = Cos; |
||
| 3636 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3637 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3638 | } |
||
| 3639 | |||
| 3640 | //------------------------------------------------------------------------------ |
||
| 3641 | |||
| 3642 | XMINLINE XMVECTOR XMVectorTan |
||
| 3643 | ( |
||
| 3644 | FXMVECTOR V |
||
| 3645 | ) |
||
| 3646 | { |
||
| 3647 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3648 | |||
| 3649 | // Cody and Waite algorithm to compute tangent. |
||
| 3650 | |||
| 3651 | XMVECTOR VA, VB, VC, VC2; |
||
| 3652 | XMVECTOR T0, T1, T2, T3, T4, T5, T6, T7; |
||
| 3653 | XMVECTOR C0, C1, TwoDivPi, Epsilon; |
||
| 3654 | XMVECTOR N, D; |
||
| 3655 | XMVECTOR R0, R1; |
||
| 3656 | XMVECTOR VIsZero, VCNearZero, VBIsEven; |
||
| 3657 | XMVECTOR Zero; |
||
| 3658 | XMVECTOR Result; |
||
| 3659 | UINT i; |
||
| 3660 | static CONST XMVECTOR TanCoefficients0 = {1.0f, -4.667168334e-1f, 2.566383229e-2f, -3.118153191e-4f}; |
||
| 3661 | static CONST XMVECTOR TanCoefficients1 = {4.981943399e-7f, -1.333835001e-1f, 3.424887824e-3f, -1.786170734e-5f}; |
||
| 3662 | static CONST XMVECTOR TanConstants = {1.570796371f, 6.077100628e-11f, 0.000244140625f, 2.0f / XM_PI}; |
||
| 3663 | static CONST XMVECTORU32 Mask = {0x1, 0x1, 0x1, 0x1}; |
||
| 3664 | |||
| 3665 | TwoDivPi = XMVectorSplatW(TanConstants); |
||
| 3666 | |||
| 3667 | Zero = XMVectorZero(); |
||
| 3668 | |||
| 3669 | C0 = XMVectorSplatX(TanConstants); |
||
| 3670 | C1 = XMVectorSplatY(TanConstants); |
||
| 3671 | Epsilon = XMVectorSplatZ(TanConstants); |
||
| 3672 | |||
| 3673 | VA = XMVectorMultiply(V, TwoDivPi); |
||
| 3674 | |||
| 3675 | VA = XMVectorRound(VA); |
||
| 3676 | |||
| 3677 | VC = XMVectorNegativeMultiplySubtract(VA, C0, V); |
||
| 3678 | |||
| 3679 | VB = XMVectorAbs(VA); |
||
| 3680 | |||
| 3681 | VC = XMVectorNegativeMultiplySubtract(VA, C1, VC); |
||
| 3682 | |||
| 3683 | for (i = 0; i < 4; i++) |
||
| 3684 | { |
||
| 3685 | VB.vector4_u32[i] = (UINT)VB.vector4_f32[i]; |
||
| 3686 | } |
||
| 3687 | |||
| 3688 | VC2 = XMVectorMultiply(VC, VC); |
||
| 3689 | |||
| 3690 | T7 = XMVectorSplatW(TanCoefficients1); |
||
| 3691 | T6 = XMVectorSplatZ(TanCoefficients1); |
||
| 3692 | T4 = XMVectorSplatX(TanCoefficients1); |
||
| 3693 | T3 = XMVectorSplatW(TanCoefficients0); |
||
| 3694 | T5 = XMVectorSplatY(TanCoefficients1); |
||
| 3695 | T2 = XMVectorSplatZ(TanCoefficients0); |
||
| 3696 | T1 = XMVectorSplatY(TanCoefficients0); |
||
| 3697 | T0 = XMVectorSplatX(TanCoefficients0); |
||
| 3698 | |||
| 3699 | VBIsEven = XMVectorAndInt(VB, Mask.v); |
||
| 3700 | VBIsEven = XMVectorEqualInt(VBIsEven, Zero); |
||
| 3701 | |||
| 3702 | N = XMVectorMultiplyAdd(VC2, T7, T6); |
||
| 3703 | D = XMVectorMultiplyAdd(VC2, T4, T3); |
||
| 3704 | N = XMVectorMultiplyAdd(VC2, N, T5); |
||
| 3705 | D = XMVectorMultiplyAdd(VC2, D, T2); |
||
| 3706 | N = XMVectorMultiply(VC2, N); |
||
| 3707 | D = XMVectorMultiplyAdd(VC2, D, T1); |
||
| 3708 | N = XMVectorMultiplyAdd(VC, N, VC); |
||
| 3709 | VCNearZero = XMVectorInBounds(VC, Epsilon); |
||
| 3710 | D = XMVectorMultiplyAdd(VC2, D, T0); |
||
| 3711 | |||
| 3712 | N = XMVectorSelect(N, VC, VCNearZero); |
||
| 3713 | D = XMVectorSelect(D, g_XMOne.v, VCNearZero); |
||
| 3714 | |||
| 3715 | R0 = XMVectorNegate(N); |
||
| 3716 | R1 = XMVectorReciprocal(D); |
||
| 3717 | R0 = XMVectorReciprocal(R0); |
||
| 3718 | R1 = XMVectorMultiply(N, R1); |
||
| 3719 | R0 = XMVectorMultiply(D, R0); |
||
| 3720 | |||
| 3721 | VIsZero = XMVectorEqual(V, Zero); |
||
| 3722 | |||
| 3723 | Result = XMVectorSelect(R0, R1, VBIsEven); |
||
| 3724 | |||
| 3725 | Result = XMVectorSelect(Result, Zero, VIsZero); |
||
| 3726 | |||
| 3727 | return Result; |
||
| 3728 | |||
| 3729 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3730 | // Cody and Waite algorithm to compute tangent. |
||
| 3731 | |||
| 3732 | XMVECTOR VA, VB, VC, VC2; |
||
| 3733 | XMVECTOR T0, T1, T2, T3, T4, T5, T6, T7; |
||
| 3734 | XMVECTOR C0, C1, TwoDivPi, Epsilon; |
||
| 3735 | XMVECTOR N, D; |
||
| 3736 | XMVECTOR R0, R1; |
||
| 3737 | XMVECTOR VIsZero, VCNearZero, VBIsEven; |
||
| 3738 | XMVECTOR Zero; |
||
| 3739 | XMVECTOR Result; |
||
| 3740 | static CONST XMVECTORF32 TanCoefficients0 = {1.0f, -4.667168334e-1f, 2.566383229e-2f, -3.118153191e-4f}; |
||
| 3741 | static CONST XMVECTORF32 TanCoefficients1 = {4.981943399e-7f, -1.333835001e-1f, 3.424887824e-3f, -1.786170734e-5f}; |
||
| 3742 | static CONST XMVECTORF32 TanConstants = {1.570796371f, 6.077100628e-11f, 0.000244140625f, 2.0f / XM_PI}; |
||
| 3743 | static CONST XMVECTORI32 Mask = {0x1, 0x1, 0x1, 0x1}; |
||
| 3744 | |||
| 3745 | TwoDivPi = XMVectorSplatW(TanConstants); |
||
| 3746 | |||
| 3747 | Zero = XMVectorZero(); |
||
| 3748 | |||
| 3749 | C0 = XMVectorSplatX(TanConstants); |
||
| 3750 | C1 = XMVectorSplatY(TanConstants); |
||
| 3751 | Epsilon = XMVectorSplatZ(TanConstants); |
||
| 3752 | |||
| 3753 | VA = XMVectorMultiply(V, TwoDivPi); |
||
| 3754 | |||
| 3755 | VA = XMVectorRound(VA); |
||
| 3756 | |||
| 3757 | VC = XMVectorNegativeMultiplySubtract(VA, C0, V); |
||
| 3758 | |||
| 3759 | VB = XMVectorAbs(VA); |
||
| 3760 | |||
| 3761 | VC = XMVectorNegativeMultiplySubtract(VA, C1, VC); |
||
| 3762 | |||
| 3763 | reinterpret_cast<__m128i *>(&VB)[0] = _mm_cvttps_epi32(VB); |
||
| 3764 | |||
| 3765 | VC2 = XMVectorMultiply(VC, VC); |
||
| 3766 | |||
| 3767 | T7 = XMVectorSplatW(TanCoefficients1); |
||
| 3768 | T6 = XMVectorSplatZ(TanCoefficients1); |
||
| 3769 | T4 = XMVectorSplatX(TanCoefficients1); |
||
| 3770 | T3 = XMVectorSplatW(TanCoefficients0); |
||
| 3771 | T5 = XMVectorSplatY(TanCoefficients1); |
||
| 3772 | T2 = XMVectorSplatZ(TanCoefficients0); |
||
| 3773 | T1 = XMVectorSplatY(TanCoefficients0); |
||
| 3774 | T0 = XMVectorSplatX(TanCoefficients0); |
||
| 3775 | |||
| 3776 | VBIsEven = XMVectorAndInt(VB,Mask); |
||
| 3777 | VBIsEven = XMVectorEqualInt(VBIsEven, Zero); |
||
| 3778 | |||
| 3779 | N = XMVectorMultiplyAdd(VC2, T7, T6); |
||
| 3780 | D = XMVectorMultiplyAdd(VC2, T4, T3); |
||
| 3781 | N = XMVectorMultiplyAdd(VC2, N, T5); |
||
| 3782 | D = XMVectorMultiplyAdd(VC2, D, T2); |
||
| 3783 | N = XMVectorMultiply(VC2, N); |
||
| 3784 | D = XMVectorMultiplyAdd(VC2, D, T1); |
||
| 3785 | N = XMVectorMultiplyAdd(VC, N, VC); |
||
| 3786 | VCNearZero = XMVectorInBounds(VC, Epsilon); |
||
| 3787 | D = XMVectorMultiplyAdd(VC2, D, T0); |
||
| 3788 | |||
| 3789 | N = XMVectorSelect(N, VC, VCNearZero); |
||
| 3790 | D = XMVectorSelect(D, g_XMOne, VCNearZero); |
||
| 3791 | R0 = XMVectorNegate(N); |
||
| 3792 | R1 = _mm_div_ps(N,D); |
||
| 3793 | R0 = _mm_div_ps(D,R0); |
||
| 3794 | VIsZero = XMVectorEqual(V, Zero); |
||
| 3795 | Result = XMVectorSelect(R0, R1, VBIsEven); |
||
| 3796 | Result = XMVectorSelect(Result, Zero, VIsZero); |
||
| 3797 | |||
| 3798 | return Result; |
||
| 3799 | |||
| 3800 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3801 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3802 | } |
||
| 3803 | |||
| 3804 | //------------------------------------------------------------------------------ |
||
| 3805 | |||
| 3806 | XMINLINE XMVECTOR XMVectorSinH |
||
| 3807 | ( |
||
| 3808 | FXMVECTOR V |
||
| 3809 | ) |
||
| 3810 | { |
||
| 3811 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3812 | |||
| 3813 | XMVECTOR V1, V2; |
||
| 3814 | XMVECTOR E1, E2; |
||
| 3815 | XMVECTOR Result; |
||
| 3816 | static CONST XMVECTORF32 Scale = {1.442695040888963f, 1.442695040888963f, 1.442695040888963f, 1.442695040888963f}; // 1.0f / ln(2.0f) |
||
| 3817 | |||
| 3818 | V1 = XMVectorMultiplyAdd(V, Scale.v, g_XMNegativeOne.v); |
||
| 3819 | V2 = XMVectorNegativeMultiplySubtract(V, Scale.v, g_XMNegativeOne.v); |
||
| 3820 | |||
| 3821 | E1 = XMVectorExp(V1); |
||
| 3822 | E2 = XMVectorExp(V2); |
||
| 3823 | |||
| 3824 | Result = XMVectorSubtract(E1, E2); |
||
| 3825 | |||
| 3826 | return Result; |
||
| 3827 | |||
| 3828 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3829 | XMVECTOR V1, V2; |
||
| 3830 | XMVECTOR E1, E2; |
||
| 3831 | XMVECTOR Result; |
||
| 3832 | static CONST XMVECTORF32 Scale = {1.442695040888963f, 1.442695040888963f, 1.442695040888963f, 1.442695040888963f}; // 1.0f / ln(2.0f) |
||
| 3833 | |||
| 3834 | V1 = _mm_mul_ps(V, Scale); |
||
| 3835 | V1 = _mm_add_ps(V1,g_XMNegativeOne); |
||
| 3836 | V2 = _mm_mul_ps(V, Scale); |
||
| 3837 | V2 = _mm_sub_ps(g_XMNegativeOne,V2); |
||
| 3838 | E1 = XMVectorExp(V1); |
||
| 3839 | E2 = XMVectorExp(V2); |
||
| 3840 | |||
| 3841 | Result = _mm_sub_ps(E1, E2); |
||
| 3842 | |||
| 3843 | return Result; |
||
| 3844 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3845 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3846 | } |
||
| 3847 | |||
| 3848 | //------------------------------------------------------------------------------ |
||
| 3849 | |||
| 3850 | XMINLINE XMVECTOR XMVectorCosH |
||
| 3851 | ( |
||
| 3852 | FXMVECTOR V |
||
| 3853 | ) |
||
| 3854 | { |
||
| 3855 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3856 | |||
| 3857 | XMVECTOR V1, V2; |
||
| 3858 | XMVECTOR E1, E2; |
||
| 3859 | XMVECTOR Result; |
||
| 3860 | static CONST XMVECTOR Scale = {1.442695040888963f, 1.442695040888963f, 1.442695040888963f, 1.442695040888963f}; // 1.0f / ln(2.0f) |
||
| 3861 | |||
| 3862 | V1 = XMVectorMultiplyAdd(V, Scale, g_XMNegativeOne.v); |
||
| 3863 | V2 = XMVectorNegativeMultiplySubtract(V, Scale, g_XMNegativeOne.v); |
||
| 3864 | |||
| 3865 | E1 = XMVectorExp(V1); |
||
| 3866 | E2 = XMVectorExp(V2); |
||
| 3867 | |||
| 3868 | Result = XMVectorAdd(E1, E2); |
||
| 3869 | |||
| 3870 | return Result; |
||
| 3871 | |||
| 3872 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3873 | XMVECTOR V1, V2; |
||
| 3874 | XMVECTOR E1, E2; |
||
| 3875 | XMVECTOR Result; |
||
| 3876 | static CONST XMVECTORF32 Scale = {1.442695040888963f, 1.442695040888963f, 1.442695040888963f, 1.442695040888963f}; // 1.0f / ln(2.0f) |
||
| 3877 | |||
| 3878 | V1 = _mm_mul_ps(V,Scale); |
||
| 3879 | V1 = _mm_add_ps(V1,g_XMNegativeOne); |
||
| 3880 | V2 = _mm_mul_ps(V, Scale); |
||
| 3881 | V2 = _mm_sub_ps(g_XMNegativeOne,V2); |
||
| 3882 | E1 = XMVectorExp(V1); |
||
| 3883 | E2 = XMVectorExp(V2); |
||
| 3884 | Result = _mm_add_ps(E1, E2); |
||
| 3885 | return Result; |
||
| 3886 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3887 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3888 | } |
||
| 3889 | |||
| 3890 | //------------------------------------------------------------------------------ |
||
| 3891 | |||
| 3892 | XMINLINE XMVECTOR XMVectorTanH |
||
| 3893 | ( |
||
| 3894 | FXMVECTOR V |
||
| 3895 | ) |
||
| 3896 | { |
||
| 3897 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3898 | |||
| 3899 | XMVECTOR E; |
||
| 3900 | XMVECTOR Result; |
||
| 3901 | static CONST XMVECTORF32 Scale = {2.8853900817779268f, 2.8853900817779268f, 2.8853900817779268f, 2.8853900817779268f}; // 2.0f / ln(2.0f) |
||
| 3902 | |||
| 3903 | E = XMVectorMultiply(V, Scale.v); |
||
| 3904 | E = XMVectorExp(E); |
||
| 3905 | E = XMVectorMultiplyAdd(E, g_XMOneHalf.v, g_XMOneHalf.v); |
||
| 3906 | E = XMVectorReciprocal(E); |
||
| 3907 | |||
| 3908 | Result = XMVectorSubtract(g_XMOne.v, E); |
||
| 3909 | |||
| 3910 | return Result; |
||
| 3911 | |||
| 3912 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3913 | static CONST XMVECTORF32 Scale = {2.8853900817779268f, 2.8853900817779268f, 2.8853900817779268f, 2.8853900817779268f}; // 2.0f / ln(2.0f) |
||
| 3914 | |||
| 3915 | XMVECTOR E = _mm_mul_ps(V, Scale); |
||
| 3916 | E = XMVectorExp(E); |
||
| 3917 | E = _mm_mul_ps(E,g_XMOneHalf); |
||
| 3918 | E = _mm_add_ps(E,g_XMOneHalf); |
||
| 3919 | E = XMVectorReciprocal(E); |
||
| 3920 | E = _mm_sub_ps(g_XMOne, E); |
||
| 3921 | return E; |
||
| 3922 | #else // _XM_VMX128_INTRINSICS_ |
||
| 3923 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 3924 | } |
||
| 3925 | |||
| 3926 | //------------------------------------------------------------------------------ |
||
| 3927 | |||
| 3928 | XMINLINE XMVECTOR XMVectorASin |
||
| 3929 | ( |
||
| 3930 | FXMVECTOR V |
||
| 3931 | ) |
||
| 3932 | { |
||
| 3933 | #if defined(_XM_NO_INTRINSICS_) |
||
| 3934 | |||
| 3935 | XMVECTOR V2, V3, AbsV; |
||
| 3936 | XMVECTOR C0, C1, C2, C3, C4, C5, C6, C7, C8, C9, C10, C11; |
||
| 3937 | XMVECTOR R0, R1, R2, R3, R4; |
||
| 3938 | XMVECTOR OneMinusAbsV; |
||
| 3939 | XMVECTOR Rsq; |
||
| 3940 | XMVECTOR Result; |
||
| 3941 | static CONST XMVECTOR OnePlusEpsilon = {1.00000011921f, 1.00000011921f, 1.00000011921f, 1.00000011921f}; |
||
| 3942 | |||
| 3943 | // asin(V) = V * (C0 + C1 * V + C2 * V^2 + C3 * V^3 + C4 * V^4 + C5 * V^5) + (1 - V) * rsq(1 - V) * |
||
| 3944 | // V * (C6 + C7 * V + C8 * V^2 + C9 * V^3 + C10 * V^4 + C11 * V^5) |
||
| 3945 | |||
| 3946 | AbsV = XMVectorAbs(V); |
||
| 3947 | |||
| 3948 | V2 = XMVectorMultiply(V, V); |
||
| 3949 | V3 = XMVectorMultiply(V2, AbsV); |
||
| 3950 | |||
| 3951 | R4 = XMVectorNegativeMultiplySubtract(AbsV, V, V); |
||
| 3952 | |||
| 3953 | OneMinusAbsV = XMVectorSubtract(OnePlusEpsilon, AbsV); |
||
| 3954 | Rsq = XMVectorReciprocalSqrt(OneMinusAbsV); |
||
| 3955 | |||
| 3956 | C0 = XMVectorSplatX(g_XMASinCoefficients0.v); |
||
| 3957 | C1 = XMVectorSplatY(g_XMASinCoefficients0.v); |
||
| 3958 | C2 = XMVectorSplatZ(g_XMASinCoefficients0.v); |
||
| 3959 | C3 = XMVectorSplatW(g_XMASinCoefficients0.v); |
||
| 3960 | |||
| 3961 | C4 = XMVectorSplatX(g_XMASinCoefficients1.v); |
||
| 3962 | C5 = XMVectorSplatY(g_XMASinCoefficients1.v); |
||
| 3963 | C6 = XMVectorSplatZ(g_XMASinCoefficients1.v); |
||
| 3964 | C7 = XMVectorSplatW(g_XMASinCoefficients1.v); |
||
| 3965 | |||
| 3966 | C8 = XMVectorSplatX(g_XMASinCoefficients2.v); |
||
| 3967 | C9 = XMVectorSplatY(g_XMASinCoefficients2.v); |
||
| 3968 | C10 = XMVectorSplatZ(g_XMASinCoefficients2.v); |
||
| 3969 | C11 = XMVectorSplatW(g_XMASinCoefficients2.v); |
||
| 3970 | |||
| 3971 | R0 = XMVectorMultiplyAdd(C3, AbsV, C7); |
||
| 3972 | R1 = XMVectorMultiplyAdd(C1, AbsV, C5); |
||
| 3973 | R2 = XMVectorMultiplyAdd(C2, AbsV, C6); |
||
| 3974 | R3 = XMVectorMultiplyAdd(C0, AbsV, C4); |
||
| 3975 | |||
| 3976 | R0 = XMVectorMultiplyAdd(R0, AbsV, C11); |
||
| 3977 | R1 = XMVectorMultiplyAdd(R1, AbsV, C9); |
||
| 3978 | R2 = XMVectorMultiplyAdd(R2, AbsV, C10); |
||
| 3979 | R3 = XMVectorMultiplyAdd(R3, AbsV, C8); |
||
| 3980 | |||
| 3981 | R0 = XMVectorMultiplyAdd(R2, V3, R0); |
||
| 3982 | R1 = XMVectorMultiplyAdd(R3, V3, R1); |
||
| 3983 | |||
| 3984 | R0 = XMVectorMultiply(V, R0); |
||
| 3985 | R1 = XMVectorMultiply(R4, R1); |
||
| 3986 | |||
| 3987 | Result = XMVectorMultiplyAdd(R1, Rsq, R0); |
||
| 3988 | |||
| 3989 | return Result; |
||
| 3990 | |||
| 3991 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 3992 | static CONST XMVECTORF32 OnePlusEpsilon = {1.00000011921f, 1.00000011921f, 1.00000011921f, 1.00000011921f}; |
||
| 3993 | |||
| 3994 | // asin(V) = V * (C0 + C1 * V + C2 * V^2 + C3 * V^3 + C4 * V^4 + C5 * V^5) + (1 - V) * rsq(1 - V) * |
||
| 3995 | // V * (C6 + C7 * V + C8 * V^2 + C9 * V^3 + C10 * V^4 + C11 * V^5) |
||
| 3996 | // Get abs(V) |
||
| 3997 | XMVECTOR vAbsV = _mm_setzero_ps(); |
||
| 3998 | vAbsV = _mm_sub_ps(vAbsV,V); |
||
| 3999 | vAbsV = _mm_max_ps(vAbsV,V); |
||
| 4000 | |||
| 4001 | XMVECTOR R0 = vAbsV; |
||
| 4002 | XMVECTOR vConstants = _mm_load_ps1(&g_XMASinCoefficients0.f[3]); |
||
| 4003 | R0 = _mm_mul_ps(R0,vConstants); |
||
| 4004 | vConstants = _mm_load_ps1(&g_XMASinCoefficients1.f[3]); |
||
| 4005 | R0 = _mm_add_ps(R0,vConstants); |
||
| 4006 | |||
| 4007 | XMVECTOR R1 = vAbsV; |
||
| 4008 | vConstants = _mm_load_ps1(&g_XMASinCoefficients0.f[1]); |
||
| 4009 | R1 = _mm_mul_ps(R1,vConstants); |
||
| 4010 | vConstants = _mm_load_ps1(&g_XMASinCoefficients1.f[1]); |
||
| 4011 | R1 = _mm_add_ps(R1, vConstants); |
||
| 4012 | |||
| 4013 | XMVECTOR R2 = vAbsV; |
||
| 4014 | vConstants = _mm_load_ps1(&g_XMASinCoefficients0.f[2]); |
||
| 4015 | R2 = _mm_mul_ps(R2,vConstants); |
||
| 4016 | vConstants = _mm_load_ps1(&g_XMASinCoefficients1.f[2]); |
||
| 4017 | R2 = _mm_add_ps(R2, vConstants); |
||
| 4018 | |||
| 4019 | XMVECTOR R3 = vAbsV; |
||
| 4020 | vConstants = _mm_load_ps1(&g_XMASinCoefficients0.f[0]); |
||
| 4021 | R3 = _mm_mul_ps(R3,vConstants); |
||
| 4022 | vConstants = _mm_load_ps1(&g_XMASinCoefficients1.f[0]); |
||
| 4023 | R3 = _mm_add_ps(R3, vConstants); |
||
| 4024 | |||
| 4025 | vConstants = _mm_load_ps1(&g_XMASinCoefficients2.f[3]); |
||
| 4026 | R0 = _mm_mul_ps(R0,vAbsV); |
||
| 4027 | R0 = _mm_add_ps(R0,vConstants); |
||
| 4028 | |||
| 4029 | vConstants = _mm_load_ps1(&g_XMASinCoefficients2.f[1]); |
||
| 4030 | R1 = _mm_mul_ps(R1,vAbsV); |
||
| 4031 | R1 = _mm_add_ps(R1,vConstants); |
||
| 4032 | |||
| 4033 | vConstants = _mm_load_ps1(&g_XMASinCoefficients2.f[2]); |
||
| 4034 | R2 = _mm_mul_ps(R2,vAbsV); |
||
| 4035 | R2 = _mm_add_ps(R2,vConstants); |
||
| 4036 | |||
| 4037 | vConstants = _mm_load_ps1(&g_XMASinCoefficients2.f[0]); |
||
| 4038 | R3 = _mm_mul_ps(R3,vAbsV); |
||
| 4039 | R3 = _mm_add_ps(R3,vConstants); |
||
| 4040 | |||
| 4041 | // V3 = V^3 |
||
| 4042 | vConstants = _mm_mul_ps(V,V); |
||
| 4043 | vConstants = _mm_mul_ps(vConstants, vAbsV); |
||
| 4044 | // Mul by V^3 |
||
| 4045 | R2 = _mm_mul_ps(R2,vConstants); |
||
| 4046 | R3 = _mm_mul_ps(R3,vConstants); |
||
| 4047 | // Merge the results |
||
| 4048 | R0 = _mm_add_ps(R0,R2); |
||
| 4049 | R1 = _mm_add_ps(R1,R3); |
||
| 4050 | |||
| 4051 | R0 = _mm_mul_ps(R0,V); |
||
| 4052 | // vConstants = V-(V^2 retaining sign) |
||
| 4053 | vConstants = _mm_mul_ps(vAbsV, V); |
||
| 4054 | vConstants = _mm_sub_ps(V,vConstants); |
||
| 4055 | R1 = _mm_mul_ps(R1,vConstants); |
||
| 4056 | vConstants = _mm_sub_ps(OnePlusEpsilon,vAbsV); |
||
| 4057 | // Do NOT use rsqrt/mul. This needs the precision |
||
| 4058 | vConstants = _mm_sqrt_ps(vConstants); |
||
| 4059 | R1 = _mm_div_ps(R1,vConstants); |
||
| 4060 | R0 = _mm_add_ps(R0,R1); |
||
| 4061 | return R0; |
||
| 4062 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4063 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4064 | } |
||
| 4065 | |||
| 4066 | //------------------------------------------------------------------------------ |
||
| 4067 | |||
| 4068 | XMINLINE XMVECTOR XMVectorACos |
||
| 4069 | ( |
||
| 4070 | FXMVECTOR V |
||
| 4071 | ) |
||
| 4072 | { |
||
| 4073 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4074 | |||
| 4075 | XMVECTOR V2, V3, AbsV; |
||
| 4076 | XMVECTOR C0, C1, C2, C3, C4, C5, C6, C7, C8, C9, C10, C11; |
||
| 4077 | XMVECTOR R0, R1, R2, R3, R4; |
||
| 4078 | XMVECTOR OneMinusAbsV; |
||
| 4079 | XMVECTOR Rsq; |
||
| 4080 | XMVECTOR Result; |
||
| 4081 | static CONST XMVECTOR OnePlusEpsilon = {1.00000011921f, 1.00000011921f, 1.00000011921f, 1.00000011921f}; |
||
| 4082 | |||
| 4083 | // acos(V) = PI / 2 - asin(V) |
||
| 4084 | |||
| 4085 | AbsV = XMVectorAbs(V); |
||
| 4086 | |||
| 4087 | V2 = XMVectorMultiply(V, V); |
||
| 4088 | V3 = XMVectorMultiply(V2, AbsV); |
||
| 4089 | |||
| 4090 | R4 = XMVectorNegativeMultiplySubtract(AbsV, V, V); |
||
| 4091 | |||
| 4092 | OneMinusAbsV = XMVectorSubtract(OnePlusEpsilon, AbsV); |
||
| 4093 | Rsq = XMVectorReciprocalSqrt(OneMinusAbsV); |
||
| 4094 | |||
| 4095 | C0 = XMVectorSplatX(g_XMASinCoefficients0.v); |
||
| 4096 | C1 = XMVectorSplatY(g_XMASinCoefficients0.v); |
||
| 4097 | C2 = XMVectorSplatZ(g_XMASinCoefficients0.v); |
||
| 4098 | C3 = XMVectorSplatW(g_XMASinCoefficients0.v); |
||
| 4099 | |||
| 4100 | C4 = XMVectorSplatX(g_XMASinCoefficients1.v); |
||
| 4101 | C5 = XMVectorSplatY(g_XMASinCoefficients1.v); |
||
| 4102 | C6 = XMVectorSplatZ(g_XMASinCoefficients1.v); |
||
| 4103 | C7 = XMVectorSplatW(g_XMASinCoefficients1.v); |
||
| 4104 | |||
| 4105 | C8 = XMVectorSplatX(g_XMASinCoefficients2.v); |
||
| 4106 | C9 = XMVectorSplatY(g_XMASinCoefficients2.v); |
||
| 4107 | C10 = XMVectorSplatZ(g_XMASinCoefficients2.v); |
||
| 4108 | C11 = XMVectorSplatW(g_XMASinCoefficients2.v); |
||
| 4109 | |||
| 4110 | R0 = XMVectorMultiplyAdd(C3, AbsV, C7); |
||
| 4111 | R1 = XMVectorMultiplyAdd(C1, AbsV, C5); |
||
| 4112 | R2 = XMVectorMultiplyAdd(C2, AbsV, C6); |
||
| 4113 | R3 = XMVectorMultiplyAdd(C0, AbsV, C4); |
||
| 4114 | |||
| 4115 | R0 = XMVectorMultiplyAdd(R0, AbsV, C11); |
||
| 4116 | R1 = XMVectorMultiplyAdd(R1, AbsV, C9); |
||
| 4117 | R2 = XMVectorMultiplyAdd(R2, AbsV, C10); |
||
| 4118 | R3 = XMVectorMultiplyAdd(R3, AbsV, C8); |
||
| 4119 | |||
| 4120 | R0 = XMVectorMultiplyAdd(R2, V3, R0); |
||
| 4121 | R1 = XMVectorMultiplyAdd(R3, V3, R1); |
||
| 4122 | |||
| 4123 | R0 = XMVectorMultiply(V, R0); |
||
| 4124 | R1 = XMVectorMultiply(R4, R1); |
||
| 4125 | |||
| 4126 | Result = XMVectorMultiplyAdd(R1, Rsq, R0); |
||
| 4127 | |||
| 4128 | Result = XMVectorSubtract(g_XMHalfPi.v, Result); |
||
| 4129 | |||
| 4130 | return Result; |
||
| 4131 | |||
| 4132 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4133 | static CONST XMVECTORF32 OnePlusEpsilon = {1.00000011921f, 1.00000011921f, 1.00000011921f, 1.00000011921f}; |
||
| 4134 | // Uses only 6 registers for good code on x86 targets |
||
| 4135 | // acos(V) = PI / 2 - asin(V) |
||
| 4136 | // Get abs(V) |
||
| 4137 | XMVECTOR vAbsV = _mm_setzero_ps(); |
||
| 4138 | vAbsV = _mm_sub_ps(vAbsV,V); |
||
| 4139 | vAbsV = _mm_max_ps(vAbsV,V); |
||
| 4140 | // Perform the series in precision groups to |
||
| 4141 | // retain precision across 20 bits. (3 bits of imprecision due to operations) |
||
| 4142 | XMVECTOR R0 = vAbsV; |
||
| 4143 | XMVECTOR vConstants = _mm_load_ps1(&g_XMASinCoefficients0.f[3]); |
||
| 4144 | R0 = _mm_mul_ps(R0,vConstants); |
||
| 4145 | vConstants = _mm_load_ps1(&g_XMASinCoefficients1.f[3]); |
||
| 4146 | R0 = _mm_add_ps(R0,vConstants); |
||
| 4147 | R0 = _mm_mul_ps(R0,vAbsV); |
||
| 4148 | vConstants = _mm_load_ps1(&g_XMASinCoefficients2.f[3]); |
||
| 4149 | R0 = _mm_add_ps(R0,vConstants); |
||
| 4150 | |||
| 4151 | XMVECTOR R1 = vAbsV; |
||
| 4152 | vConstants = _mm_load_ps1(&g_XMASinCoefficients0.f[1]); |
||
| 4153 | R1 = _mm_mul_ps(R1,vConstants); |
||
| 4154 | vConstants = _mm_load_ps1(&g_XMASinCoefficients1.f[1]); |
||
| 4155 | R1 = _mm_add_ps(R1,vConstants); |
||
| 4156 | R1 = _mm_mul_ps(R1, vAbsV); |
||
| 4157 | vConstants = _mm_load_ps1(&g_XMASinCoefficients2.f[1]); |
||
| 4158 | R1 = _mm_add_ps(R1,vConstants); |
||
| 4159 | |||
| 4160 | XMVECTOR R2 = vAbsV; |
||
| 4161 | vConstants = _mm_load_ps1(&g_XMASinCoefficients0.f[2]); |
||
| 4162 | R2 = _mm_mul_ps(R2,vConstants); |
||
| 4163 | vConstants = _mm_load_ps1(&g_XMASinCoefficients1.f[2]); |
||
| 4164 | R2 = _mm_add_ps(R2,vConstants); |
||
| 4165 | R2 = _mm_mul_ps(R2, vAbsV); |
||
| 4166 | vConstants = _mm_load_ps1(&g_XMASinCoefficients2.f[2]); |
||
| 4167 | R2 = _mm_add_ps(R2,vConstants); |
||
| 4168 | |||
| 4169 | XMVECTOR R3 = vAbsV; |
||
| 4170 | vConstants = _mm_load_ps1(&g_XMASinCoefficients0.f[0]); |
||
| 4171 | R3 = _mm_mul_ps(R3,vConstants); |
||
| 4172 | vConstants = _mm_load_ps1(&g_XMASinCoefficients1.f[0]); |
||
| 4173 | R3 = _mm_add_ps(R3,vConstants); |
||
| 4174 | R3 = _mm_mul_ps(R3, vAbsV); |
||
| 4175 | vConstants = _mm_load_ps1(&g_XMASinCoefficients2.f[0]); |
||
| 4176 | R3 = _mm_add_ps(R3,vConstants); |
||
| 4177 | |||
| 4178 | // vConstants = V^3 |
||
| 4179 | vConstants = _mm_mul_ps(V,V); |
||
| 4180 | vConstants = _mm_mul_ps(vConstants,vAbsV); |
||
| 4181 | R2 = _mm_mul_ps(R2,vConstants); |
||
| 4182 | R3 = _mm_mul_ps(R3,vConstants); |
||
| 4183 | // Add the pair of values together here to retain |
||
| 4184 | // as much precision as possible |
||
| 4185 | R0 = _mm_add_ps(R0,R2); |
||
| 4186 | R1 = _mm_add_ps(R1,R3); |
||
| 4187 | |||
| 4188 | R0 = _mm_mul_ps(R0,V); |
||
| 4189 | // vConstants = V-(V*abs(V)) |
||
| 4190 | vConstants = _mm_mul_ps(V,vAbsV); |
||
| 4191 | vConstants = _mm_sub_ps(V,vConstants); |
||
| 4192 | R1 = _mm_mul_ps(R1,vConstants); |
||
| 4193 | // Episilon exists to allow 1.0 as an answer |
||
| 4194 | vConstants = _mm_sub_ps(OnePlusEpsilon, vAbsV); |
||
| 4195 | // Use sqrt instead of rsqrt for precision |
||
| 4196 | vConstants = _mm_sqrt_ps(vConstants); |
||
| 4197 | R1 = _mm_div_ps(R1,vConstants); |
||
| 4198 | R1 = _mm_add_ps(R1,R0); |
||
| 4199 | vConstants = _mm_sub_ps(g_XMHalfPi,R1); |
||
| 4200 | return vConstants; |
||
| 4201 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4202 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4203 | } |
||
| 4204 | |||
| 4205 | //------------------------------------------------------------------------------ |
||
| 4206 | |||
| 4207 | XMINLINE XMVECTOR XMVectorATan |
||
| 4208 | ( |
||
| 4209 | FXMVECTOR V |
||
| 4210 | ) |
||
| 4211 | { |
||
| 4212 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4213 | |||
| 4214 | // Cody and Waite algorithm to compute inverse tangent. |
||
| 4215 | |||
| 4216 | XMVECTOR N, D; |
||
| 4217 | XMVECTOR VF, G, ReciprocalF, AbsF, FA, FB; |
||
| 4218 | XMVECTOR Sqrt3, Sqrt3MinusOne, TwoMinusSqrt3; |
||
| 4219 | XMVECTOR HalfPi, OneThirdPi, OneSixthPi, Epsilon, MinV, MaxV; |
||
| 4220 | XMVECTOR Zero; |
||
| 4221 | XMVECTOR NegativeHalfPi; |
||
| 4222 | XMVECTOR Angle1, Angle2; |
||
| 4223 | XMVECTOR F_GT_One, F_GT_TwoMinusSqrt3, AbsF_LT_Epsilon, V_LT_Zero, V_GT_MaxV, V_LT_MinV; |
||
| 4224 | XMVECTOR NegativeResult, Result; |
||
| 4225 | XMVECTOR P0, P1, P2, P3, Q0, Q1, Q2, Q3; |
||
| 4226 | static CONST XMVECTOR ATanConstants0 = {-1.3688768894e+1f, -2.0505855195e+1f, -8.4946240351f, -8.3758299368e-1f}; |
||
| 4227 | static CONST XMVECTOR ATanConstants1 = {4.1066306682e+1f, 8.6157349597e+1f, 5.9578436142e+1f, 1.5024001160e+1f}; |
||
| 4228 | static CONST XMVECTOR ATanConstants2 = {1.732050808f, 7.320508076e-1f, 2.679491924e-1f, 0.000244140625f}; // <sqrt(3), sqrt(3) - 1, 2 - sqrt(3), Epsilon> |
||
| 4229 | static CONST XMVECTOR ATanConstants3 = {XM_PIDIV2, XM_PI / 3.0f, XM_PI / 6.0f, 8.507059173e+37f}; // <Pi / 2, Pi / 3, Pi / 6, MaxV> |
||
| 4230 | |||
| 4231 | Zero = XMVectorZero(); |
||
| 4232 | |||
| 4233 | P0 = XMVectorSplatX(ATanConstants0); |
||
| 4234 | P1 = XMVectorSplatY(ATanConstants0); |
||
| 4235 | P2 = XMVectorSplatZ(ATanConstants0); |
||
| 4236 | P3 = XMVectorSplatW(ATanConstants0); |
||
| 4237 | |||
| 4238 | Q0 = XMVectorSplatX(ATanConstants1); |
||
| 4239 | Q1 = XMVectorSplatY(ATanConstants1); |
||
| 4240 | Q2 = XMVectorSplatZ(ATanConstants1); |
||
| 4241 | Q3 = XMVectorSplatW(ATanConstants1); |
||
| 4242 | |||
| 4243 | Sqrt3 = XMVectorSplatX(ATanConstants2); |
||
| 4244 | Sqrt3MinusOne = XMVectorSplatY(ATanConstants2); |
||
| 4245 | TwoMinusSqrt3 = XMVectorSplatZ(ATanConstants2); |
||
| 4246 | Epsilon = XMVectorSplatW(ATanConstants2); |
||
| 4247 | |||
| 4248 | HalfPi = XMVectorSplatX(ATanConstants3); |
||
| 4249 | OneThirdPi = XMVectorSplatY(ATanConstants3); |
||
| 4250 | OneSixthPi = XMVectorSplatZ(ATanConstants3); |
||
| 4251 | MaxV = XMVectorSplatW(ATanConstants3); |
||
| 4252 | |||
| 4253 | VF = XMVectorAbs(V); |
||
| 4254 | ReciprocalF = XMVectorReciprocal(VF); |
||
| 4255 | |||
| 4256 | F_GT_One = XMVectorGreater(VF, g_XMOne.v); |
||
| 4257 | |||
| 4258 | VF = XMVectorSelect(VF, ReciprocalF, F_GT_One); |
||
| 4259 | Angle1 = XMVectorSelect(Zero, HalfPi, F_GT_One); |
||
| 4260 | Angle2 = XMVectorSelect(OneSixthPi, OneThirdPi, F_GT_One); |
||
| 4261 | |||
| 4262 | F_GT_TwoMinusSqrt3 = XMVectorGreater(VF, TwoMinusSqrt3); |
||
| 4263 | |||
| 4264 | FA = XMVectorMultiplyAdd(Sqrt3MinusOne, VF, VF); |
||
| 4265 | FA = XMVectorAdd(FA, g_XMNegativeOne.v); |
||
| 4266 | FB = XMVectorAdd(VF, Sqrt3); |
||
| 4267 | FB = XMVectorReciprocal(FB); |
||
| 4268 | FA = XMVectorMultiply(FA, FB); |
||
| 4269 | |||
| 4270 | VF = XMVectorSelect(VF, FA, F_GT_TwoMinusSqrt3); |
||
| 4271 | Angle1 = XMVectorSelect(Angle1, Angle2, F_GT_TwoMinusSqrt3); |
||
| 4272 | |||
| 4273 | AbsF = XMVectorAbs(VF); |
||
| 4274 | AbsF_LT_Epsilon = XMVectorLess(AbsF, Epsilon); |
||
| 4275 | |||
| 4276 | G = XMVectorMultiply(VF, VF); |
||
| 4277 | |||
| 4278 | D = XMVectorAdd(G, Q3); |
||
| 4279 | D = XMVectorMultiplyAdd(D, G, Q2); |
||
| 4280 | D = XMVectorMultiplyAdd(D, G, Q1); |
||
| 4281 | D = XMVectorMultiplyAdd(D, G, Q0); |
||
| 4282 | D = XMVectorReciprocal(D); |
||
| 4283 | |||
| 4284 | N = XMVectorMultiplyAdd(P3, G, P2); |
||
| 4285 | N = XMVectorMultiplyAdd(N, G, P1); |
||
| 4286 | N = XMVectorMultiplyAdd(N, G, P0); |
||
| 4287 | N = XMVectorMultiply(N, G); |
||
| 4288 | Result = XMVectorMultiply(N, D); |
||
| 4289 | |||
| 4290 | Result = XMVectorMultiplyAdd(Result, VF, VF); |
||
| 4291 | |||
| 4292 | Result = XMVectorSelect(Result, VF, AbsF_LT_Epsilon); |
||
| 4293 | |||
| 4294 | NegativeResult = XMVectorNegate(Result); |
||
| 4295 | Result = XMVectorSelect(Result, NegativeResult, F_GT_One); |
||
| 4296 | |||
| 4297 | Result = XMVectorAdd(Result, Angle1); |
||
| 4298 | |||
| 4299 | V_LT_Zero = XMVectorLess(V, Zero); |
||
| 4300 | NegativeResult = XMVectorNegate(Result); |
||
| 4301 | Result = XMVectorSelect(Result, NegativeResult, V_LT_Zero); |
||
| 4302 | |||
| 4303 | MinV = XMVectorNegate(MaxV); |
||
| 4304 | NegativeHalfPi = XMVectorNegate(HalfPi); |
||
| 4305 | V_GT_MaxV = XMVectorGreater(V, MaxV); |
||
| 4306 | V_LT_MinV = XMVectorLess(V, MinV); |
||
| 4307 | Result = XMVectorSelect(Result, g_XMHalfPi.v, V_GT_MaxV); |
||
| 4308 | Result = XMVectorSelect(Result, NegativeHalfPi, V_LT_MinV); |
||
| 4309 | |||
| 4310 | return Result; |
||
| 4311 | |||
| 4312 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4313 | static CONST XMVECTORF32 ATanConstants0 = {-1.3688768894e+1f, -2.0505855195e+1f, -8.4946240351f, -8.3758299368e-1f}; |
||
| 4314 | static CONST XMVECTORF32 ATanConstants1 = {4.1066306682e+1f, 8.6157349597e+1f, 5.9578436142e+1f, 1.5024001160e+1f}; |
||
| 4315 | static CONST XMVECTORF32 ATanConstants2 = {1.732050808f, 7.320508076e-1f, 2.679491924e-1f, 0.000244140625f}; // <sqrt(3), sqrt(3) - 1, 2 - sqrt(3), Epsilon> |
||
| 4316 | static CONST XMVECTORF32 ATanConstants3 = {XM_PIDIV2, XM_PI / 3.0f, XM_PI / 6.0f, 8.507059173e+37f}; // <Pi / 2, Pi / 3, Pi / 6, MaxV> |
||
| 4317 | |||
| 4318 | XMVECTOR VF = XMVectorAbs(V); |
||
| 4319 | XMVECTOR F_GT_One = _mm_cmpgt_ps(VF,g_XMOne); |
||
| 4320 | XMVECTOR ReciprocalF = XMVectorReciprocal(VF); |
||
| 4321 | VF = XMVectorSelect(VF, ReciprocalF, F_GT_One); |
||
| 4322 | XMVECTOR Zero = XMVectorZero(); |
||
| 4323 | XMVECTOR HalfPi = _mm_load_ps1(&ATanConstants3.f[0]); |
||
| 4324 | XMVECTOR Angle1 = XMVectorSelect(Zero, HalfPi, F_GT_One); |
||
| 4325 | // Pi/3 |
||
| 4326 | XMVECTOR vConstants = _mm_load_ps1(&ATanConstants3.f[1]); |
||
| 4327 | // Pi/6 |
||
| 4328 | XMVECTOR Angle2 = _mm_load_ps1(&ATanConstants3.f[2]); |
||
| 4329 | Angle2 = XMVectorSelect(Angle2, vConstants, F_GT_One); |
||
| 4330 | |||
| 4331 | // 1-sqrt(3) |
||
| 4332 | XMVECTOR FA = _mm_load_ps1(&ATanConstants2.f[1]); |
||
| 4333 | FA = _mm_mul_ps(FA,VF); |
||
| 4334 | FA = _mm_add_ps(FA,VF); |
||
| 4335 | FA = _mm_add_ps(FA,g_XMNegativeOne); |
||
| 4336 | // sqrt(3) |
||
| 4337 | vConstants = _mm_load_ps1(&ATanConstants2.f[0]); |
||
| 4338 | vConstants = _mm_add_ps(vConstants,VF); |
||
| 4339 | FA = _mm_div_ps(FA,vConstants); |
||
| 4340 | |||
| 4341 | // 2-sqrt(3) |
||
| 4342 | vConstants = _mm_load_ps1(&ATanConstants2.f[2]); |
||
| 4343 | // >2-sqrt(3)? |
||
| 4344 | vConstants = _mm_cmpgt_ps(VF,vConstants); |
||
| 4345 | VF = XMVectorSelect(VF, FA, vConstants); |
||
| 4346 | Angle1 = XMVectorSelect(Angle1, Angle2, vConstants); |
||
| 4347 | |||
| 4348 | XMVECTOR AbsF = XMVectorAbs(VF); |
||
| 4349 | |||
| 4350 | XMVECTOR G = _mm_mul_ps(VF,VF); |
||
| 4351 | XMVECTOR D = _mm_load_ps1(&ATanConstants1.f[3]); |
||
| 4352 | D = _mm_add_ps(D,G); |
||
| 4353 | D = _mm_mul_ps(D,G); |
||
| 4354 | vConstants = _mm_load_ps1(&ATanConstants1.f[2]); |
||
| 4355 | D = _mm_add_ps(D,vConstants); |
||
| 4356 | D = _mm_mul_ps(D,G); |
||
| 4357 | vConstants = _mm_load_ps1(&ATanConstants1.f[1]); |
||
| 4358 | D = _mm_add_ps(D,vConstants); |
||
| 4359 | D = _mm_mul_ps(D,G); |
||
| 4360 | vConstants = _mm_load_ps1(&ATanConstants1.f[0]); |
||
| 4361 | D = _mm_add_ps(D,vConstants); |
||
| 4362 | |||
| 4363 | XMVECTOR N = _mm_load_ps1(&ATanConstants0.f[3]); |
||
| 4364 | N = _mm_mul_ps(N,G); |
||
| 4365 | vConstants = _mm_load_ps1(&ATanConstants0.f[2]); |
||
| 4366 | N = _mm_add_ps(N,vConstants); |
||
| 4367 | N = _mm_mul_ps(N,G); |
||
| 4368 | vConstants = _mm_load_ps1(&ATanConstants0.f[1]); |
||
| 4369 | N = _mm_add_ps(N,vConstants); |
||
| 4370 | N = _mm_mul_ps(N,G); |
||
| 4371 | vConstants = _mm_load_ps1(&ATanConstants0.f[0]); |
||
| 4372 | N = _mm_add_ps(N,vConstants); |
||
| 4373 | N = _mm_mul_ps(N,G); |
||
| 4374 | XMVECTOR Result = _mm_div_ps(N,D); |
||
| 4375 | |||
| 4376 | Result = _mm_mul_ps(Result,VF); |
||
| 4377 | Result = _mm_add_ps(Result,VF); |
||
| 4378 | // Epsilon |
||
| 4379 | vConstants = _mm_load_ps1(&ATanConstants2.f[3]); |
||
| 4380 | vConstants = _mm_cmpge_ps(vConstants,AbsF); |
||
| 4381 | Result = XMVectorSelect(Result,VF,vConstants); |
||
| 4382 | |||
| 4383 | XMVECTOR NegativeResult = _mm_mul_ps(Result,g_XMNegativeOne); |
||
| 4384 | Result = XMVectorSelect(Result,NegativeResult,F_GT_One); |
||
| 4385 | Result = _mm_add_ps(Result,Angle1); |
||
| 4386 | |||
| 4387 | Zero = _mm_cmpge_ps(Zero,V); |
||
| 4388 | NegativeResult = _mm_mul_ps(Result,g_XMNegativeOne); |
||
| 4389 | Result = XMVectorSelect(Result,NegativeResult,Zero); |
||
| 4390 | |||
| 4391 | XMVECTOR MaxV = _mm_load_ps1(&ATanConstants3.f[3]); |
||
| 4392 | XMVECTOR MinV = _mm_mul_ps(MaxV,g_XMNegativeOne); |
||
| 4393 | // Negate HalfPi |
||
| 4394 | HalfPi = _mm_mul_ps(HalfPi,g_XMNegativeOne); |
||
| 4395 | MaxV = _mm_cmple_ps(MaxV,V); |
||
| 4396 | MinV = _mm_cmpge_ps(MinV,V); |
||
| 4397 | Result = XMVectorSelect(Result,g_XMHalfPi,MaxV); |
||
| 4398 | // HalfPi = -HalfPi |
||
| 4399 | Result = XMVectorSelect(Result,HalfPi,MinV); |
||
| 4400 | return Result; |
||
| 4401 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4402 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4403 | } |
||
| 4404 | |||
| 4405 | //------------------------------------------------------------------------------ |
||
| 4406 | |||
| 4407 | XMINLINE XMVECTOR XMVectorATan2 |
||
| 4408 | ( |
||
| 4409 | FXMVECTOR Y, |
||
| 4410 | FXMVECTOR X |
||
| 4411 | ) |
||
| 4412 | { |
||
| 4413 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4414 | |||
| 4415 | // Return the inverse tangent of Y / X in the range of -Pi to Pi with the following exceptions: |
||
| 4416 | |||
| 4417 | // Y == 0 and X is Negative -> Pi with the sign of Y |
||
| 4418 | // Y == 0 and X is Positive -> 0 with the sign of Y |
||
| 4419 | // Y != 0 and X == 0 -> Pi / 2 with the sign of Y |
||
| 4420 | // X == -Infinity and Finite Y > 0 -> Pi with the sign of Y |
||
| 4421 | // X == +Infinity and Finite Y > 0 -> 0 with the sign of Y |
||
| 4422 | // Y == Infinity and X is Finite -> Pi / 2 with the sign of Y |
||
| 4423 | // Y == Infinity and X == -Infinity -> 3Pi / 4 with the sign of Y |
||
| 4424 | // Y == Infinity and X == +Infinity -> Pi / 4 with the sign of Y |
||
| 4425 | // TODO: Return Y / X if the result underflows |
||
| 4426 | |||
| 4427 | XMVECTOR Reciprocal; |
||
| 4428 | XMVECTOR V; |
||
| 4429 | XMVECTOR YSign; |
||
| 4430 | XMVECTOR Pi, PiOverTwo, PiOverFour, ThreePiOverFour; |
||
| 4431 | XMVECTOR YEqualsZero, XEqualsZero, XIsPositive, YEqualsInfinity, XEqualsInfinity, FiniteYGreaterZero; |
||
| 4432 | XMVECTOR ATanResultValid; |
||
| 4433 | XMVECTOR R0, R1, R2, R3, R4, R5, R6, R7; |
||
| 4434 | XMVECTOR Zero; |
||
| 4435 | XMVECTOR Result; |
||
| 4436 | static CONST XMVECTOR ATan2Constants = {XM_PI, XM_PIDIV2, XM_PIDIV4, XM_PI * 3.0f / 4.0f}; |
||
| 4437 | |||
| 4438 | Zero = XMVectorZero(); |
||
| 4439 | ATanResultValid = XMVectorTrueInt(); |
||
| 4440 | |||
| 4441 | Pi = XMVectorSplatX(ATan2Constants); |
||
| 4442 | PiOverTwo = XMVectorSplatY(ATan2Constants); |
||
| 4443 | PiOverFour = XMVectorSplatZ(ATan2Constants); |
||
| 4444 | ThreePiOverFour = XMVectorSplatW(ATan2Constants); |
||
| 4445 | |||
| 4446 | YEqualsZero = XMVectorEqual(Y, Zero); |
||
| 4447 | XEqualsZero = XMVectorEqual(X, Zero); |
||
| 4448 | XIsPositive = XMVectorAndInt(X, g_XMNegativeZero.v); |
||
| 4449 | XIsPositive = XMVectorEqualInt(XIsPositive, Zero); |
||
| 4450 | YEqualsInfinity = XMVectorIsInfinite(Y); |
||
| 4451 | XEqualsInfinity = XMVectorIsInfinite(X); |
||
| 4452 | FiniteYGreaterZero = XMVectorGreater(Y, Zero); |
||
| 4453 | FiniteYGreaterZero = XMVectorSelect(FiniteYGreaterZero, Zero, YEqualsInfinity); |
||
| 4454 | |||
| 4455 | YSign = XMVectorAndInt(Y, g_XMNegativeZero.v); |
||
| 4456 | Pi = XMVectorOrInt(Pi, YSign); |
||
| 4457 | PiOverTwo = XMVectorOrInt(PiOverTwo, YSign); |
||
| 4458 | PiOverFour = XMVectorOrInt(PiOverFour, YSign); |
||
| 4459 | ThreePiOverFour = XMVectorOrInt(ThreePiOverFour, YSign); |
||
| 4460 | |||
| 4461 | R1 = XMVectorSelect(Pi, YSign, XIsPositive); |
||
| 4462 | R2 = XMVectorSelect(ATanResultValid, PiOverTwo, XEqualsZero); |
||
| 4463 | R3 = XMVectorSelect(R2, R1, YEqualsZero); |
||
| 4464 | R4 = XMVectorSelect(ThreePiOverFour, PiOverFour, XIsPositive); |
||
| 4465 | R5 = XMVectorSelect(PiOverTwo, R4, XEqualsInfinity); |
||
| 4466 | R6 = XMVectorSelect(R3, R5, YEqualsInfinity); |
||
| 4467 | R7 = XMVectorSelect(R6, R1, FiniteYGreaterZero); |
||
| 4468 | Result = XMVectorSelect(R6, R7, XEqualsInfinity); |
||
| 4469 | ATanResultValid = XMVectorEqualInt(Result, ATanResultValid); |
||
| 4470 | |||
| 4471 | Reciprocal = XMVectorReciprocal(X); |
||
| 4472 | V = XMVectorMultiply(Y, Reciprocal); |
||
| 4473 | R0 = XMVectorATan(V); |
||
| 4474 | |||
| 4475 | Result = XMVectorSelect(Result, R0, ATanResultValid); |
||
| 4476 | |||
| 4477 | return Result; |
||
| 4478 | |||
| 4479 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4480 | static CONST XMVECTORF32 ATan2Constants = {XM_PI, XM_PIDIV2, XM_PIDIV4, XM_PI * 3.0f / 4.0f}; |
||
| 4481 | // Mask if Y>0 && Y!=INF |
||
| 4482 | XMVECTOR FiniteYGreaterZero = _mm_cmpgt_ps(Y,g_XMZero); |
||
| 4483 | XMVECTOR YEqualsInfinity = XMVectorIsInfinite(Y); |
||
| 4484 | FiniteYGreaterZero = _mm_andnot_ps(YEqualsInfinity,FiniteYGreaterZero); |
||
| 4485 | // Get the sign of (Y&0x80000000) |
||
| 4486 | XMVECTOR YSign = _mm_and_ps(Y, g_XMNegativeZero); |
||
| 4487 | // Get the sign bits of X |
||
| 4488 | XMVECTOR XIsPositive = _mm_and_ps(X,g_XMNegativeZero); |
||
| 4489 | // Change them to masks |
||
| 4490 | XIsPositive = XMVectorEqualInt(XIsPositive,g_XMZero); |
||
| 4491 | // Get Pi |
||
| 4492 | XMVECTOR R1 = _mm_load_ps1(&ATan2Constants.f[0]); |
||
| 4493 | // Copy the sign of Y |
||
| 4494 | R1 = _mm_or_ps(R1,YSign); |
||
| 4495 | R1 = XMVectorSelect(R1,YSign,XIsPositive); |
||
| 4496 | // Mask for X==0 |
||
| 4497 | XMVECTOR vConstants = _mm_cmpeq_ps(X,g_XMZero); |
||
| 4498 | // Get Pi/2 with with sign of Y |
||
| 4499 | XMVECTOR PiOverTwo = _mm_load_ps1(&ATan2Constants.f[1]); |
||
| 4500 | PiOverTwo = _mm_or_ps(PiOverTwo,YSign); |
||
| 4501 | XMVECTOR R2 = XMVectorSelect(g_XMNegOneMask,PiOverTwo,vConstants); |
||
| 4502 | // Mask for Y==0 |
||
| 4503 | vConstants = _mm_cmpeq_ps(Y,g_XMZero); |
||
| 4504 | R2 = XMVectorSelect(R2,R1,vConstants); |
||
| 4505 | // Get Pi/4 with sign of Y |
||
| 4506 | XMVECTOR PiOverFour = _mm_load_ps1(&ATan2Constants.f[2]); |
||
| 4507 | PiOverFour = _mm_or_ps(PiOverFour,YSign); |
||
| 4508 | // Get (Pi*3)/4 with sign of Y |
||
| 4509 | XMVECTOR ThreePiOverFour = _mm_load_ps1(&ATan2Constants.f[3]); |
||
| 4510 | ThreePiOverFour = _mm_or_ps(ThreePiOverFour,YSign); |
||
| 4511 | vConstants = XMVectorSelect(ThreePiOverFour, PiOverFour, XIsPositive); |
||
| 4512 | XMVECTOR XEqualsInfinity = XMVectorIsInfinite(X); |
||
| 4513 | vConstants = XMVectorSelect(PiOverTwo,vConstants,XEqualsInfinity); |
||
| 4514 | |||
| 4515 | XMVECTOR vResult = XMVectorSelect(R2,vConstants,YEqualsInfinity); |
||
| 4516 | vConstants = XMVectorSelect(vResult,R1,FiniteYGreaterZero); |
||
| 4517 | // At this point, any entry that's zero will get the result |
||
| 4518 | // from XMVectorATan(), otherwise, return the failsafe value |
||
| 4519 | vResult = XMVectorSelect(vResult,vConstants,XEqualsInfinity); |
||
| 4520 | // Any entries not 0xFFFFFFFF, are considered precalculated |
||
| 4521 | XMVECTOR ATanResultValid = XMVectorEqualInt(vResult,g_XMNegOneMask); |
||
| 4522 | // Let's do the ATan2 function |
||
| 4523 | vConstants = _mm_div_ps(Y,X); |
||
| 4524 | vConstants = XMVectorATan(vConstants); |
||
| 4525 | // Discard entries that have been declared void |
||
| 4526 | vResult = XMVectorSelect(vResult,vConstants,ATanResultValid); |
||
| 4527 | return vResult; |
||
| 4528 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4529 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4530 | } |
||
| 4531 | |||
| 4532 | //------------------------------------------------------------------------------ |
||
| 4533 | |||
| 4534 | XMFINLINE XMVECTOR XMVectorSinEst |
||
| 4535 | ( |
||
| 4536 | FXMVECTOR V |
||
| 4537 | ) |
||
| 4538 | { |
||
| 4539 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4540 | |||
| 4541 | XMVECTOR V2, V3, V5, V7; |
||
| 4542 | XMVECTOR S1, S2, S3; |
||
| 4543 | XMVECTOR Result; |
||
| 4544 | |||
| 4545 | // sin(V) ~= V - V^3 / 3! + V^5 / 5! - V^7 / 7! (for -PI <= V < PI) |
||
| 4546 | V2 = XMVectorMultiply(V, V); |
||
| 4547 | V3 = XMVectorMultiply(V2, V); |
||
| 4548 | V5 = XMVectorMultiply(V3, V2); |
||
| 4549 | V7 = XMVectorMultiply(V5, V2); |
||
| 4550 | |||
| 4551 | S1 = XMVectorSplatY(g_XMSinEstCoefficients.v); |
||
| 4552 | S2 = XMVectorSplatZ(g_XMSinEstCoefficients.v); |
||
| 4553 | S3 = XMVectorSplatW(g_XMSinEstCoefficients.v); |
||
| 4554 | |||
| 4555 | Result = XMVectorMultiplyAdd(S1, V3, V); |
||
| 4556 | Result = XMVectorMultiplyAdd(S2, V5, Result); |
||
| 4557 | Result = XMVectorMultiplyAdd(S3, V7, Result); |
||
| 4558 | |||
| 4559 | return Result; |
||
| 4560 | |||
| 4561 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4562 | // sin(V) ~= V - V^3 / 3! + V^5 / 5! - V^7 / 7! (for -PI <= V < PI) |
||
| 4563 | XMVECTOR V2 = _mm_mul_ps(V,V); |
||
| 4564 | XMVECTOR V3 = _mm_mul_ps(V2,V); |
||
| 4565 | XMVECTOR vResult = _mm_load_ps1(&g_XMSinEstCoefficients.f[1]); |
||
| 4566 | vResult = _mm_mul_ps(vResult,V3); |
||
| 4567 | vResult = _mm_add_ps(vResult,V); |
||
| 4568 | XMVECTOR vConstants = _mm_load_ps1(&g_XMSinEstCoefficients.f[2]); |
||
| 4569 | // V^5 |
||
| 4570 | V3 = _mm_mul_ps(V3,V2); |
||
| 4571 | vConstants = _mm_mul_ps(vConstants,V3); |
||
| 4572 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 4573 | vConstants = _mm_load_ps1(&g_XMSinEstCoefficients.f[3]); |
||
| 4574 | // V^7 |
||
| 4575 | V3 = _mm_mul_ps(V3,V2); |
||
| 4576 | vConstants = _mm_mul_ps(vConstants,V3); |
||
| 4577 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 4578 | return vResult; |
||
| 4579 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4580 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4581 | } |
||
| 4582 | |||
| 4583 | //------------------------------------------------------------------------------ |
||
| 4584 | |||
| 4585 | XMFINLINE XMVECTOR XMVectorCosEst |
||
| 4586 | ( |
||
| 4587 | FXMVECTOR V |
||
| 4588 | ) |
||
| 4589 | { |
||
| 4590 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4591 | |||
| 4592 | XMVECTOR V2, V4, V6; |
||
| 4593 | XMVECTOR C0, C1, C2, C3; |
||
| 4594 | XMVECTOR Result; |
||
| 4595 | |||
| 4596 | V2 = XMVectorMultiply(V, V); |
||
| 4597 | V4 = XMVectorMultiply(V2, V2); |
||
| 4598 | V6 = XMVectorMultiply(V4, V2); |
||
| 4599 | |||
| 4600 | C0 = XMVectorSplatX(g_XMCosEstCoefficients.v); |
||
| 4601 | C1 = XMVectorSplatY(g_XMCosEstCoefficients.v); |
||
| 4602 | C2 = XMVectorSplatZ(g_XMCosEstCoefficients.v); |
||
| 4603 | C3 = XMVectorSplatW(g_XMCosEstCoefficients.v); |
||
| 4604 | |||
| 4605 | Result = XMVectorMultiplyAdd(C1, V2, C0); |
||
| 4606 | Result = XMVectorMultiplyAdd(C2, V4, Result); |
||
| 4607 | Result = XMVectorMultiplyAdd(C3, V6, Result); |
||
| 4608 | |||
| 4609 | return Result; |
||
| 4610 | |||
| 4611 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4612 | // Get V^2 |
||
| 4613 | XMVECTOR V2 = _mm_mul_ps(V,V); |
||
| 4614 | XMVECTOR vResult = _mm_load_ps1(&g_XMCosEstCoefficients.f[1]); |
||
| 4615 | vResult = _mm_mul_ps(vResult,V2); |
||
| 4616 | XMVECTOR vConstants = _mm_load_ps1(&g_XMCosEstCoefficients.f[0]); |
||
| 4617 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 4618 | vConstants = _mm_load_ps1(&g_XMCosEstCoefficients.f[2]); |
||
| 4619 | // Get V^4 |
||
| 4620 | XMVECTOR V4 = _mm_mul_ps(V2, V2); |
||
| 4621 | vConstants = _mm_mul_ps(vConstants,V4); |
||
| 4622 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 4623 | vConstants = _mm_load_ps1(&g_XMCosEstCoefficients.f[3]); |
||
| 4624 | // It's really V^6 |
||
| 4625 | V4 = _mm_mul_ps(V4,V2); |
||
| 4626 | vConstants = _mm_mul_ps(vConstants,V4); |
||
| 4627 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 4628 | return vResult; |
||
| 4629 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4630 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4631 | } |
||
| 4632 | |||
| 4633 | //------------------------------------------------------------------------------ |
||
| 4634 | |||
| 4635 | XMFINLINE VOID XMVectorSinCosEst |
||
| 4636 | ( |
||
| 4637 | XMVECTOR* pSin, |
||
| 4638 | XMVECTOR* pCos, |
||
| 4639 | FXMVECTOR V |
||
| 4640 | ) |
||
| 4641 | { |
||
| 4642 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4643 | |||
| 4644 | XMVECTOR V2, V3, V4, V5, V6, V7; |
||
| 4645 | XMVECTOR S1, S2, S3; |
||
| 4646 | XMVECTOR C0, C1, C2, C3; |
||
| 4647 | XMVECTOR Sin, Cos; |
||
| 4648 | |||
| 4649 | XMASSERT(pSin); |
||
| 4650 | XMASSERT(pCos); |
||
| 4651 | |||
| 4652 | // sin(V) ~= V - V^3 / 3! + V^5 / 5! - V^7 / 7! (for -PI <= V < PI) |
||
| 4653 | // cos(V) ~= 1 - V^2 / 2! + V^4 / 4! - V^6 / 6! (for -PI <= V < PI) |
||
| 4654 | V2 = XMVectorMultiply(V, V); |
||
| 4655 | V3 = XMVectorMultiply(V2, V); |
||
| 4656 | V4 = XMVectorMultiply(V2, V2); |
||
| 4657 | V5 = XMVectorMultiply(V3, V2); |
||
| 4658 | V6 = XMVectorMultiply(V3, V3); |
||
| 4659 | V7 = XMVectorMultiply(V4, V3); |
||
| 4660 | |||
| 4661 | S1 = XMVectorSplatY(g_XMSinEstCoefficients.v); |
||
| 4662 | S2 = XMVectorSplatZ(g_XMSinEstCoefficients.v); |
||
| 4663 | S3 = XMVectorSplatW(g_XMSinEstCoefficients.v); |
||
| 4664 | |||
| 4665 | C0 = XMVectorSplatX(g_XMCosEstCoefficients.v); |
||
| 4666 | C1 = XMVectorSplatY(g_XMCosEstCoefficients.v); |
||
| 4667 | C2 = XMVectorSplatZ(g_XMCosEstCoefficients.v); |
||
| 4668 | C3 = XMVectorSplatW(g_XMCosEstCoefficients.v); |
||
| 4669 | |||
| 4670 | Sin = XMVectorMultiplyAdd(S1, V3, V); |
||
| 4671 | Sin = XMVectorMultiplyAdd(S2, V5, Sin); |
||
| 4672 | Sin = XMVectorMultiplyAdd(S3, V7, Sin); |
||
| 4673 | |||
| 4674 | Cos = XMVectorMultiplyAdd(C1, V2, C0); |
||
| 4675 | Cos = XMVectorMultiplyAdd(C2, V4, Cos); |
||
| 4676 | Cos = XMVectorMultiplyAdd(C3, V6, Cos); |
||
| 4677 | |||
| 4678 | *pSin = Sin; |
||
| 4679 | *pCos = Cos; |
||
| 4680 | |||
| 4681 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4682 | XMASSERT(pSin); |
||
| 4683 | XMASSERT(pCos); |
||
| 4684 | XMVECTOR V2, V3, V4, V5, V6, V7; |
||
| 4685 | XMVECTOR S1, S2, S3; |
||
| 4686 | XMVECTOR C0, C1, C2, C3; |
||
| 4687 | XMVECTOR Sin, Cos; |
||
| 4688 | |||
| 4689 | // sin(V) ~= V - V^3 / 3! + V^5 / 5! - V^7 / 7! (for -PI <= V < PI) |
||
| 4690 | // cos(V) ~= 1 - V^2 / 2! + V^4 / 4! - V^6 / 6! (for -PI <= V < PI) |
||
| 4691 | V2 = XMVectorMultiply(V, V); |
||
| 4692 | V3 = XMVectorMultiply(V2, V); |
||
| 4693 | V4 = XMVectorMultiply(V2, V2); |
||
| 4694 | V5 = XMVectorMultiply(V3, V2); |
||
| 4695 | V6 = XMVectorMultiply(V3, V3); |
||
| 4696 | V7 = XMVectorMultiply(V4, V3); |
||
| 4697 | |||
| 4698 | S1 = _mm_load_ps1(&g_XMSinEstCoefficients.f[1]); |
||
| 4699 | S2 = _mm_load_ps1(&g_XMSinEstCoefficients.f[2]); |
||
| 4700 | S3 = _mm_load_ps1(&g_XMSinEstCoefficients.f[3]); |
||
| 4701 | |||
| 4702 | C0 = _mm_load_ps1(&g_XMCosEstCoefficients.f[0]); |
||
| 4703 | C1 = _mm_load_ps1(&g_XMCosEstCoefficients.f[1]); |
||
| 4704 | C2 = _mm_load_ps1(&g_XMCosEstCoefficients.f[2]); |
||
| 4705 | C3 = _mm_load_ps1(&g_XMCosEstCoefficients.f[3]); |
||
| 4706 | |||
| 4707 | Sin = XMVectorMultiplyAdd(S1, V3, V); |
||
| 4708 | Sin = XMVectorMultiplyAdd(S2, V5, Sin); |
||
| 4709 | Sin = XMVectorMultiplyAdd(S3, V7, Sin); |
||
| 4710 | |||
| 4711 | Cos = XMVectorMultiplyAdd(C1, V2, C0); |
||
| 4712 | Cos = XMVectorMultiplyAdd(C2, V4, Cos); |
||
| 4713 | Cos = XMVectorMultiplyAdd(C3, V6, Cos); |
||
| 4714 | |||
| 4715 | *pSin = Sin; |
||
| 4716 | *pCos = Cos; |
||
| 4717 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4718 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4719 | } |
||
| 4720 | |||
| 4721 | //------------------------------------------------------------------------------ |
||
| 4722 | |||
| 4723 | XMFINLINE XMVECTOR XMVectorTanEst |
||
| 4724 | ( |
||
| 4725 | FXMVECTOR V |
||
| 4726 | ) |
||
| 4727 | { |
||
| 4728 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4729 | |||
| 4730 | XMVECTOR V1, V2, V1T0, V1T1, V2T2; |
||
| 4731 | XMVECTOR T0, T1, T2; |
||
| 4732 | XMVECTOR N, D; |
||
| 4733 | XMVECTOR OneOverPi; |
||
| 4734 | XMVECTOR Result; |
||
| 4735 | |||
| 4736 | OneOverPi = XMVectorSplatW(g_XMTanEstCoefficients.v); |
||
| 4737 | |||
| 4738 | V1 = XMVectorMultiply(V, OneOverPi); |
||
| 4739 | V1 = XMVectorRound(V1); |
||
| 4740 | |||
| 4741 | V1 = XMVectorNegativeMultiplySubtract(g_XMPi.v, V1, V); |
||
| 4742 | |||
| 4743 | T0 = XMVectorSplatX(g_XMTanEstCoefficients.v); |
||
| 4744 | T1 = XMVectorSplatY(g_XMTanEstCoefficients.v); |
||
| 4745 | T2 = XMVectorSplatZ(g_XMTanEstCoefficients.v); |
||
| 4746 | |||
| 4747 | V2T2 = XMVectorNegativeMultiplySubtract(V1, V1, T2); |
||
| 4748 | V2 = XMVectorMultiply(V1, V1); |
||
| 4749 | V1T0 = XMVectorMultiply(V1, T0); |
||
| 4750 | V1T1 = XMVectorMultiply(V1, T1); |
||
| 4751 | |||
| 4752 | D = XMVectorReciprocalEst(V2T2); |
||
| 4753 | N = XMVectorMultiplyAdd(V2, V1T1, V1T0); |
||
| 4754 | |||
| 4755 | Result = XMVectorMultiply(N, D); |
||
| 4756 | |||
| 4757 | return Result; |
||
| 4758 | |||
| 4759 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4760 | XMVECTOR V1, V2, V1T0, V1T1, V2T2; |
||
| 4761 | XMVECTOR T0, T1, T2; |
||
| 4762 | XMVECTOR N, D; |
||
| 4763 | XMVECTOR OneOverPi; |
||
| 4764 | XMVECTOR Result; |
||
| 4765 | |||
| 4766 | OneOverPi = XMVectorSplatW(g_XMTanEstCoefficients); |
||
| 4767 | |||
| 4768 | V1 = XMVectorMultiply(V, OneOverPi); |
||
| 4769 | V1 = XMVectorRound(V1); |
||
| 4770 | |||
| 4771 | V1 = XMVectorNegativeMultiplySubtract(g_XMPi, V1, V); |
||
| 4772 | |||
| 4773 | T0 = XMVectorSplatX(g_XMTanEstCoefficients); |
||
| 4774 | T1 = XMVectorSplatY(g_XMTanEstCoefficients); |
||
| 4775 | T2 = XMVectorSplatZ(g_XMTanEstCoefficients); |
||
| 4776 | |||
| 4777 | V2T2 = XMVectorNegativeMultiplySubtract(V1, V1, T2); |
||
| 4778 | V2 = XMVectorMultiply(V1, V1); |
||
| 4779 | V1T0 = XMVectorMultiply(V1, T0); |
||
| 4780 | V1T1 = XMVectorMultiply(V1, T1); |
||
| 4781 | |||
| 4782 | D = XMVectorReciprocalEst(V2T2); |
||
| 4783 | N = XMVectorMultiplyAdd(V2, V1T1, V1T0); |
||
| 4784 | |||
| 4785 | Result = XMVectorMultiply(N, D); |
||
| 4786 | |||
| 4787 | return Result; |
||
| 4788 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4789 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4790 | } |
||
| 4791 | |||
| 4792 | //------------------------------------------------------------------------------ |
||
| 4793 | |||
| 4794 | XMFINLINE XMVECTOR XMVectorSinHEst |
||
| 4795 | ( |
||
| 4796 | FXMVECTOR V |
||
| 4797 | ) |
||
| 4798 | { |
||
| 4799 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4800 | |||
| 4801 | XMVECTOR V1, V2; |
||
| 4802 | XMVECTOR E1, E2; |
||
| 4803 | XMVECTOR Result; |
||
| 4804 | static CONST XMVECTORF32 Scale = {1.442695040888963f, 1.442695040888963f, 1.442695040888963f, 1.442695040888963f}; // 1.0f / ln(2.0f) |
||
| 4805 | |||
| 4806 | V1 = XMVectorMultiplyAdd(V, Scale.v, g_XMNegativeOne.v); |
||
| 4807 | V2 = XMVectorNegativeMultiplySubtract(V, Scale.v, g_XMNegativeOne.v); |
||
| 4808 | |||
| 4809 | E1 = XMVectorExpEst(V1); |
||
| 4810 | E2 = XMVectorExpEst(V2); |
||
| 4811 | |||
| 4812 | Result = XMVectorSubtract(E1, E2); |
||
| 4813 | |||
| 4814 | return Result; |
||
| 4815 | |||
| 4816 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4817 | XMVECTOR V1, V2; |
||
| 4818 | XMVECTOR E1, E2; |
||
| 4819 | XMVECTOR Result; |
||
| 4820 | static CONST XMVECTORF32 Scale = {1.442695040888963f, 1.442695040888963f, 1.442695040888963f, 1.442695040888963f}; // 1.0f / ln(2.0f) |
||
| 4821 | |||
| 4822 | V1 = _mm_mul_ps(V,Scale); |
||
| 4823 | V1 = _mm_add_ps(V1,g_XMNegativeOne); |
||
| 4824 | V2 = _mm_mul_ps(V,Scale); |
||
| 4825 | V2 = _mm_sub_ps(g_XMNegativeOne,V2); |
||
| 4826 | E1 = XMVectorExpEst(V1); |
||
| 4827 | E2 = XMVectorExpEst(V2); |
||
| 4828 | Result = _mm_sub_ps(E1, E2); |
||
| 4829 | return Result; |
||
| 4830 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4831 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4832 | } |
||
| 4833 | |||
| 4834 | //------------------------------------------------------------------------------ |
||
| 4835 | |||
| 4836 | XMFINLINE XMVECTOR XMVectorCosHEst |
||
| 4837 | ( |
||
| 4838 | FXMVECTOR V |
||
| 4839 | ) |
||
| 4840 | { |
||
| 4841 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4842 | |||
| 4843 | XMVECTOR V1, V2; |
||
| 4844 | XMVECTOR E1, E2; |
||
| 4845 | XMVECTOR Result; |
||
| 4846 | static CONST XMVECTOR Scale = {1.442695040888963f, 1.442695040888963f, 1.442695040888963f, 1.442695040888963f}; // 1.0f / ln(2.0f) |
||
| 4847 | |||
| 4848 | V1 = XMVectorMultiplyAdd(V, Scale, g_XMNegativeOne.v); |
||
| 4849 | V2 = XMVectorNegativeMultiplySubtract(V, Scale, g_XMNegativeOne.v); |
||
| 4850 | |||
| 4851 | E1 = XMVectorExpEst(V1); |
||
| 4852 | E2 = XMVectorExpEst(V2); |
||
| 4853 | |||
| 4854 | Result = XMVectorAdd(E1, E2); |
||
| 4855 | |||
| 4856 | return Result; |
||
| 4857 | |||
| 4858 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4859 | XMVECTOR V1, V2; |
||
| 4860 | XMVECTOR E1, E2; |
||
| 4861 | XMVECTOR Result; |
||
| 4862 | static CONST XMVECTORF32 Scale = {1.442695040888963f, 1.442695040888963f, 1.442695040888963f, 1.442695040888963f}; // 1.0f / ln(2.0f) |
||
| 4863 | |||
| 4864 | V1 = _mm_mul_ps(V,Scale); |
||
| 4865 | V1 = _mm_add_ps(V1,g_XMNegativeOne); |
||
| 4866 | V2 = _mm_mul_ps(V, Scale); |
||
| 4867 | V2 = _mm_sub_ps(g_XMNegativeOne,V2); |
||
| 4868 | E1 = XMVectorExpEst(V1); |
||
| 4869 | E2 = XMVectorExpEst(V2); |
||
| 4870 | Result = _mm_add_ps(E1, E2); |
||
| 4871 | return Result; |
||
| 4872 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4873 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4874 | } |
||
| 4875 | |||
| 4876 | //------------------------------------------------------------------------------ |
||
| 4877 | |||
| 4878 | XMFINLINE XMVECTOR XMVectorTanHEst |
||
| 4879 | ( |
||
| 4880 | FXMVECTOR V |
||
| 4881 | ) |
||
| 4882 | { |
||
| 4883 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4884 | |||
| 4885 | XMVECTOR E; |
||
| 4886 | XMVECTOR Result; |
||
| 4887 | static CONST XMVECTOR Scale = {2.8853900817779268f, 2.8853900817779268f, 2.8853900817779268f, 2.8853900817779268f}; // 2.0f / ln(2.0f) |
||
| 4888 | |||
| 4889 | E = XMVectorMultiply(V, Scale); |
||
| 4890 | E = XMVectorExpEst(E); |
||
| 4891 | E = XMVectorMultiplyAdd(E, g_XMOneHalf.v, g_XMOneHalf.v); |
||
| 4892 | E = XMVectorReciprocalEst(E); |
||
| 4893 | |||
| 4894 | Result = XMVectorSubtract(g_XMOne.v, E); |
||
| 4895 | |||
| 4896 | return Result; |
||
| 4897 | |||
| 4898 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4899 | static CONST XMVECTORF32 Scale = {2.8853900817779268f, 2.8853900817779268f, 2.8853900817779268f, 2.8853900817779268f}; // 2.0f / ln(2.0f) |
||
| 4900 | |||
| 4901 | XMVECTOR E = _mm_mul_ps(V, Scale); |
||
| 4902 | E = XMVectorExpEst(E); |
||
| 4903 | E = _mm_mul_ps(E,g_XMOneHalf); |
||
| 4904 | E = _mm_add_ps(E,g_XMOneHalf); |
||
| 4905 | E = XMVectorReciprocalEst(E); |
||
| 4906 | E = _mm_sub_ps(g_XMOne, E); |
||
| 4907 | return E; |
||
| 4908 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4909 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4910 | } |
||
| 4911 | |||
| 4912 | //------------------------------------------------------------------------------ |
||
| 4913 | |||
| 4914 | XMFINLINE XMVECTOR XMVectorASinEst |
||
| 4915 | ( |
||
| 4916 | FXMVECTOR V |
||
| 4917 | ) |
||
| 4918 | { |
||
| 4919 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4920 | |||
| 4921 | XMVECTOR AbsV, V2, VD, VC0, V2C3; |
||
| 4922 | XMVECTOR C0, C1, C2, C3; |
||
| 4923 | XMVECTOR D, Rsq, SqrtD; |
||
| 4924 | XMVECTOR OnePlusEps; |
||
| 4925 | XMVECTOR Result; |
||
| 4926 | |||
| 4927 | AbsV = XMVectorAbs(V); |
||
| 4928 | |||
| 4929 | OnePlusEps = XMVectorSplatX(g_XMASinEstConstants.v); |
||
| 4930 | |||
| 4931 | C0 = XMVectorSplatX(g_XMASinEstCoefficients.v); |
||
| 4932 | C1 = XMVectorSplatY(g_XMASinEstCoefficients.v); |
||
| 4933 | C2 = XMVectorSplatZ(g_XMASinEstCoefficients.v); |
||
| 4934 | C3 = XMVectorSplatW(g_XMASinEstCoefficients.v); |
||
| 4935 | |||
| 4936 | D = XMVectorSubtract(OnePlusEps, AbsV); |
||
| 4937 | |||
| 4938 | Rsq = XMVectorReciprocalSqrtEst(D); |
||
| 4939 | SqrtD = XMVectorMultiply(D, Rsq); |
||
| 4940 | |||
| 4941 | V2 = XMVectorMultiply(V, AbsV); |
||
| 4942 | V2C3 = XMVectorMultiply(V2, C3); |
||
| 4943 | VD = XMVectorMultiply(D, AbsV); |
||
| 4944 | VC0 = XMVectorMultiply(V, C0); |
||
| 4945 | |||
| 4946 | Result = XMVectorMultiply(V, C1); |
||
| 4947 | Result = XMVectorMultiplyAdd(V2, C2, Result); |
||
| 4948 | Result = XMVectorMultiplyAdd(V2C3, VD, Result); |
||
| 4949 | Result = XMVectorMultiplyAdd(VC0, SqrtD, Result); |
||
| 4950 | |||
| 4951 | return Result; |
||
| 4952 | |||
| 4953 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 4954 | // Get abs(V) |
||
| 4955 | XMVECTOR vAbsV = _mm_setzero_ps(); |
||
| 4956 | vAbsV = _mm_sub_ps(vAbsV,V); |
||
| 4957 | vAbsV = _mm_max_ps(vAbsV,V); |
||
| 4958 | |||
| 4959 | XMVECTOR D = _mm_load_ps1(&g_XMASinEstConstants.f[0]); |
||
| 4960 | D = _mm_sub_ps(D,vAbsV); |
||
| 4961 | // Since this is an estimate, rqsrt is okay |
||
| 4962 | XMVECTOR vConstants = _mm_rsqrt_ps(D); |
||
| 4963 | XMVECTOR SqrtD = _mm_mul_ps(D,vConstants); |
||
| 4964 | // V2 = V^2 retaining sign |
||
| 4965 | XMVECTOR V2 = _mm_mul_ps(V,vAbsV); |
||
| 4966 | D = _mm_mul_ps(D,vAbsV); |
||
| 4967 | |||
| 4968 | XMVECTOR vResult = _mm_load_ps1(&g_XMASinEstCoefficients.f[1]); |
||
| 4969 | vResult = _mm_mul_ps(vResult,V); |
||
| 4970 | vConstants = _mm_load_ps1(&g_XMASinEstCoefficients.f[2]); |
||
| 4971 | vConstants = _mm_mul_ps(vConstants,V2); |
||
| 4972 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 4973 | |||
| 4974 | vConstants = _mm_load_ps1(&g_XMASinEstCoefficients.f[3]); |
||
| 4975 | vConstants = _mm_mul_ps(vConstants,V2); |
||
| 4976 | vConstants = _mm_mul_ps(vConstants,D); |
||
| 4977 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 4978 | |||
| 4979 | vConstants = _mm_load_ps1(&g_XMASinEstCoefficients.f[0]); |
||
| 4980 | vConstants = _mm_mul_ps(vConstants,V); |
||
| 4981 | vConstants = _mm_mul_ps(vConstants,SqrtD); |
||
| 4982 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 4983 | return vResult; |
||
| 4984 | #else // _XM_VMX128_INTRINSICS_ |
||
| 4985 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 4986 | } |
||
| 4987 | |||
| 4988 | //------------------------------------------------------------------------------ |
||
| 4989 | |||
| 4990 | XMFINLINE XMVECTOR XMVectorACosEst |
||
| 4991 | ( |
||
| 4992 | FXMVECTOR V |
||
| 4993 | ) |
||
| 4994 | { |
||
| 4995 | #if defined(_XM_NO_INTRINSICS_) |
||
| 4996 | |||
| 4997 | XMVECTOR AbsV, V2, VD, VC0, V2C3; |
||
| 4998 | XMVECTOR C0, C1, C2, C3; |
||
| 4999 | XMVECTOR D, Rsq, SqrtD; |
||
| 5000 | XMVECTOR OnePlusEps, HalfPi; |
||
| 5001 | XMVECTOR Result; |
||
| 5002 | |||
| 5003 | // acos(V) = PI / 2 - asin(V) |
||
| 5004 | |||
| 5005 | AbsV = XMVectorAbs(V); |
||
| 5006 | |||
| 5007 | OnePlusEps = XMVectorSplatX(g_XMASinEstConstants.v); |
||
| 5008 | HalfPi = XMVectorSplatY(g_XMASinEstConstants.v); |
||
| 5009 | |||
| 5010 | C0 = XMVectorSplatX(g_XMASinEstCoefficients.v); |
||
| 5011 | C1 = XMVectorSplatY(g_XMASinEstCoefficients.v); |
||
| 5012 | C2 = XMVectorSplatZ(g_XMASinEstCoefficients.v); |
||
| 5013 | C3 = XMVectorSplatW(g_XMASinEstCoefficients.v); |
||
| 5014 | |||
| 5015 | D = XMVectorSubtract(OnePlusEps, AbsV); |
||
| 5016 | |||
| 5017 | Rsq = XMVectorReciprocalSqrtEst(D); |
||
| 5018 | SqrtD = XMVectorMultiply(D, Rsq); |
||
| 5019 | |||
| 5020 | V2 = XMVectorMultiply(V, AbsV); |
||
| 5021 | V2C3 = XMVectorMultiply(V2, C3); |
||
| 5022 | VD = XMVectorMultiply(D, AbsV); |
||
| 5023 | VC0 = XMVectorMultiply(V, C0); |
||
| 5024 | |||
| 5025 | Result = XMVectorMultiply(V, C1); |
||
| 5026 | Result = XMVectorMultiplyAdd(V2, C2, Result); |
||
| 5027 | Result = XMVectorMultiplyAdd(V2C3, VD, Result); |
||
| 5028 | Result = XMVectorMultiplyAdd(VC0, SqrtD, Result); |
||
| 5029 | Result = XMVectorSubtract(HalfPi, Result); |
||
| 5030 | |||
| 5031 | return Result; |
||
| 5032 | |||
| 5033 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5034 | // acos(V) = PI / 2 - asin(V) |
||
| 5035 | // Get abs(V) |
||
| 5036 | XMVECTOR vAbsV = _mm_setzero_ps(); |
||
| 5037 | vAbsV = _mm_sub_ps(vAbsV,V); |
||
| 5038 | vAbsV = _mm_max_ps(vAbsV,V); |
||
| 5039 | // Calc D |
||
| 5040 | XMVECTOR D = _mm_load_ps1(&g_XMASinEstConstants.f[0]); |
||
| 5041 | D = _mm_sub_ps(D,vAbsV); |
||
| 5042 | // SqrtD = sqrt(D-abs(V)) estimated |
||
| 5043 | XMVECTOR vConstants = _mm_rsqrt_ps(D); |
||
| 5044 | XMVECTOR SqrtD = _mm_mul_ps(D,vConstants); |
||
| 5045 | // V2 = V^2 while retaining sign |
||
| 5046 | XMVECTOR V2 = _mm_mul_ps(V, vAbsV); |
||
| 5047 | // Drop vAbsV here. D = (Const-abs(V))*abs(V) |
||
| 5048 | D = _mm_mul_ps(D, vAbsV); |
||
| 5049 | |||
| 5050 | XMVECTOR vResult = _mm_load_ps1(&g_XMASinEstCoefficients.f[1]); |
||
| 5051 | vResult = _mm_mul_ps(vResult,V); |
||
| 5052 | vConstants = _mm_load_ps1(&g_XMASinEstCoefficients.f[2]); |
||
| 5053 | vConstants = _mm_mul_ps(vConstants,V2); |
||
| 5054 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 5055 | |||
| 5056 | vConstants = _mm_load_ps1(&g_XMASinEstCoefficients.f[3]); |
||
| 5057 | vConstants = _mm_mul_ps(vConstants,V2); |
||
| 5058 | vConstants = _mm_mul_ps(vConstants,D); |
||
| 5059 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 5060 | |||
| 5061 | vConstants = _mm_load_ps1(&g_XMASinEstCoefficients.f[0]); |
||
| 5062 | vConstants = _mm_mul_ps(vConstants,V); |
||
| 5063 | vConstants = _mm_mul_ps(vConstants,SqrtD); |
||
| 5064 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 5065 | |||
| 5066 | vConstants = _mm_load_ps1(&g_XMASinEstConstants.f[1]); |
||
| 5067 | vResult = _mm_sub_ps(vConstants,vResult); |
||
| 5068 | return vResult; |
||
| 5069 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5070 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5071 | } |
||
| 5072 | |||
| 5073 | //------------------------------------------------------------------------------ |
||
| 5074 | |||
| 5075 | XMFINLINE XMVECTOR XMVectorATanEst |
||
| 5076 | ( |
||
| 5077 | FXMVECTOR V |
||
| 5078 | ) |
||
| 5079 | { |
||
| 5080 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5081 | |||
| 5082 | XMVECTOR AbsV, V2S2, N, D; |
||
| 5083 | XMVECTOR S0, S1, S2; |
||
| 5084 | XMVECTOR HalfPi; |
||
| 5085 | XMVECTOR Result; |
||
| 5086 | |||
| 5087 | S0 = XMVectorSplatX(g_XMATanEstCoefficients.v); |
||
| 5088 | S1 = XMVectorSplatY(g_XMATanEstCoefficients.v); |
||
| 5089 | S2 = XMVectorSplatZ(g_XMATanEstCoefficients.v); |
||
| 5090 | HalfPi = XMVectorSplatW(g_XMATanEstCoefficients.v); |
||
| 5091 | |||
| 5092 | AbsV = XMVectorAbs(V); |
||
| 5093 | |||
| 5094 | V2S2 = XMVectorMultiplyAdd(V, V, S2); |
||
| 5095 | N = XMVectorMultiplyAdd(AbsV, HalfPi, S0); |
||
| 5096 | D = XMVectorMultiplyAdd(AbsV, S1, V2S2); |
||
| 5097 | N = XMVectorMultiply(N, V); |
||
| 5098 | D = XMVectorReciprocalEst(D); |
||
| 5099 | |||
| 5100 | Result = XMVectorMultiply(N, D); |
||
| 5101 | |||
| 5102 | return Result; |
||
| 5103 | |||
| 5104 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5105 | // Get abs(V) |
||
| 5106 | XMVECTOR vAbsV = _mm_setzero_ps(); |
||
| 5107 | vAbsV = _mm_sub_ps(vAbsV,V); |
||
| 5108 | vAbsV = _mm_max_ps(vAbsV,V); |
||
| 5109 | |||
| 5110 | XMVECTOR vResult = _mm_load_ps1(&g_XMATanEstCoefficients.f[3]); |
||
| 5111 | vResult = _mm_mul_ps(vResult,vAbsV); |
||
| 5112 | XMVECTOR vConstants = _mm_load_ps1(&g_XMATanEstCoefficients.f[0]); |
||
| 5113 | vResult = _mm_add_ps(vResult,vConstants); |
||
| 5114 | vResult = _mm_mul_ps(vResult,V); |
||
| 5115 | |||
| 5116 | XMVECTOR D = _mm_mul_ps(V,V); |
||
| 5117 | vConstants = _mm_load_ps1(&g_XMATanEstCoefficients.f[2]); |
||
| 5118 | D = _mm_add_ps(D,vConstants); |
||
| 5119 | vConstants = _mm_load_ps1(&g_XMATanEstCoefficients.f[1]); |
||
| 5120 | vConstants = _mm_mul_ps(vConstants,vAbsV); |
||
| 5121 | D = _mm_add_ps(D,vConstants); |
||
| 5122 | vResult = _mm_div_ps(vResult,D); |
||
| 5123 | return vResult; |
||
| 5124 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5125 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5126 | } |
||
| 5127 | |||
| 5128 | //------------------------------------------------------------------------------ |
||
| 5129 | |||
| 5130 | XMFINLINE XMVECTOR XMVectorATan2Est |
||
| 5131 | ( |
||
| 5132 | FXMVECTOR Y, |
||
| 5133 | FXMVECTOR X |
||
| 5134 | ) |
||
| 5135 | { |
||
| 5136 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5137 | |||
| 5138 | XMVECTOR Reciprocal; |
||
| 5139 | XMVECTOR V; |
||
| 5140 | XMVECTOR YSign; |
||
| 5141 | XMVECTOR Pi, PiOverTwo, PiOverFour, ThreePiOverFour; |
||
| 5142 | XMVECTOR YEqualsZero, XEqualsZero, XIsPositive, YEqualsInfinity, XEqualsInfinity, FiniteYGreaterZero; |
||
| 5143 | XMVECTOR ATanResultValid; |
||
| 5144 | XMVECTOR R0, R1, R2, R3, R4, R5, R6, R7; |
||
| 5145 | XMVECTOR Zero; |
||
| 5146 | XMVECTOR Result; |
||
| 5147 | static CONST XMVECTOR ATan2Constants = {XM_PI, XM_PIDIV2, XM_PIDIV4, XM_PI * 3.0f / 4.0f}; |
||
| 5148 | |||
| 5149 | Zero = XMVectorZero(); |
||
| 5150 | ATanResultValid = XMVectorTrueInt(); |
||
| 5151 | |||
| 5152 | Pi = XMVectorSplatX(ATan2Constants); |
||
| 5153 | PiOverTwo = XMVectorSplatY(ATan2Constants); |
||
| 5154 | PiOverFour = XMVectorSplatZ(ATan2Constants); |
||
| 5155 | ThreePiOverFour = XMVectorSplatW(ATan2Constants); |
||
| 5156 | |||
| 5157 | YEqualsZero = XMVectorEqual(Y, Zero); |
||
| 5158 | XEqualsZero = XMVectorEqual(X, Zero); |
||
| 5159 | XIsPositive = XMVectorAndInt(X, g_XMNegativeZero.v); |
||
| 5160 | XIsPositive = XMVectorEqualInt(XIsPositive, Zero); |
||
| 5161 | YEqualsInfinity = XMVectorIsInfinite(Y); |
||
| 5162 | XEqualsInfinity = XMVectorIsInfinite(X); |
||
| 5163 | FiniteYGreaterZero = XMVectorGreater(Y, Zero); |
||
| 5164 | FiniteYGreaterZero = XMVectorSelect(FiniteYGreaterZero, Zero, YEqualsInfinity); |
||
| 5165 | |||
| 5166 | YSign = XMVectorAndInt(Y, g_XMNegativeZero.v); |
||
| 5167 | Pi = XMVectorOrInt(Pi, YSign); |
||
| 5168 | PiOverTwo = XMVectorOrInt(PiOverTwo, YSign); |
||
| 5169 | PiOverFour = XMVectorOrInt(PiOverFour, YSign); |
||
| 5170 | ThreePiOverFour = XMVectorOrInt(ThreePiOverFour, YSign); |
||
| 5171 | |||
| 5172 | R1 = XMVectorSelect(Pi, YSign, XIsPositive); |
||
| 5173 | R2 = XMVectorSelect(ATanResultValid, PiOverTwo, XEqualsZero); |
||
| 5174 | R3 = XMVectorSelect(R2, R1, YEqualsZero); |
||
| 5175 | R4 = XMVectorSelect(ThreePiOverFour, PiOverFour, XIsPositive); |
||
| 5176 | R5 = XMVectorSelect(PiOverTwo, R4, XEqualsInfinity); |
||
| 5177 | R6 = XMVectorSelect(R3, R5, YEqualsInfinity); |
||
| 5178 | R7 = XMVectorSelect(R6, R1, FiniteYGreaterZero); |
||
| 5179 | Result = XMVectorSelect(R6, R7, XEqualsInfinity); |
||
| 5180 | ATanResultValid = XMVectorEqualInt(Result, ATanResultValid); |
||
| 5181 | |||
| 5182 | Reciprocal = XMVectorReciprocalEst(X); |
||
| 5183 | V = XMVectorMultiply(Y, Reciprocal); |
||
| 5184 | R0 = XMVectorATanEst(V); |
||
| 5185 | |||
| 5186 | Result = XMVectorSelect(Result, R0, ATanResultValid); |
||
| 5187 | |||
| 5188 | return Result; |
||
| 5189 | |||
| 5190 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5191 | static CONST XMVECTORF32 ATan2Constants = {XM_PI, XM_PIDIV2, XM_PIDIV4, XM_PI * 3.0f / 4.0f}; |
||
| 5192 | // Mask if Y>0 && Y!=INF |
||
| 5193 | XMVECTOR FiniteYGreaterZero = _mm_cmpgt_ps(Y,g_XMZero); |
||
| 5194 | XMVECTOR YEqualsInfinity = XMVectorIsInfinite(Y); |
||
| 5195 | FiniteYGreaterZero = _mm_andnot_ps(YEqualsInfinity,FiniteYGreaterZero); |
||
| 5196 | // Get the sign of (Y&0x80000000) |
||
| 5197 | XMVECTOR YSign = _mm_and_ps(Y, g_XMNegativeZero); |
||
| 5198 | // Get the sign bits of X |
||
| 5199 | XMVECTOR XIsPositive = _mm_and_ps(X,g_XMNegativeZero); |
||
| 5200 | // Change them to masks |
||
| 5201 | XIsPositive = XMVectorEqualInt(XIsPositive,g_XMZero); |
||
| 5202 | // Get Pi |
||
| 5203 | XMVECTOR R1 = _mm_load_ps1(&ATan2Constants.f[0]); |
||
| 5204 | // Copy the sign of Y |
||
| 5205 | R1 = _mm_or_ps(R1,YSign); |
||
| 5206 | R1 = XMVectorSelect(R1,YSign,XIsPositive); |
||
| 5207 | // Mask for X==0 |
||
| 5208 | XMVECTOR vConstants = _mm_cmpeq_ps(X,g_XMZero); |
||
| 5209 | // Get Pi/2 with with sign of Y |
||
| 5210 | XMVECTOR PiOverTwo = _mm_load_ps1(&ATan2Constants.f[1]); |
||
| 5211 | PiOverTwo = _mm_or_ps(PiOverTwo,YSign); |
||
| 5212 | XMVECTOR R2 = XMVectorSelect(g_XMNegOneMask,PiOverTwo,vConstants); |
||
| 5213 | // Mask for Y==0 |
||
| 5214 | vConstants = _mm_cmpeq_ps(Y,g_XMZero); |
||
| 5215 | R2 = XMVectorSelect(R2,R1,vConstants); |
||
| 5216 | // Get Pi/4 with sign of Y |
||
| 5217 | XMVECTOR PiOverFour = _mm_load_ps1(&ATan2Constants.f[2]); |
||
| 5218 | PiOverFour = _mm_or_ps(PiOverFour,YSign); |
||
| 5219 | // Get (Pi*3)/4 with sign of Y |
||
| 5220 | XMVECTOR ThreePiOverFour = _mm_load_ps1(&ATan2Constants.f[3]); |
||
| 5221 | ThreePiOverFour = _mm_or_ps(ThreePiOverFour,YSign); |
||
| 5222 | vConstants = XMVectorSelect(ThreePiOverFour, PiOverFour, XIsPositive); |
||
| 5223 | XMVECTOR XEqualsInfinity = XMVectorIsInfinite(X); |
||
| 5224 | vConstants = XMVectorSelect(PiOverTwo,vConstants,XEqualsInfinity); |
||
| 5225 | |||
| 5226 | XMVECTOR vResult = XMVectorSelect(R2,vConstants,YEqualsInfinity); |
||
| 5227 | vConstants = XMVectorSelect(vResult,R1,FiniteYGreaterZero); |
||
| 5228 | // At this point, any entry that's zero will get the result |
||
| 5229 | // from XMVectorATan(), otherwise, return the failsafe value |
||
| 5230 | vResult = XMVectorSelect(vResult,vConstants,XEqualsInfinity); |
||
| 5231 | // Any entries not 0xFFFFFFFF, are considered precalculated |
||
| 5232 | XMVECTOR ATanResultValid = XMVectorEqualInt(vResult,g_XMNegOneMask); |
||
| 5233 | // Let's do the ATan2 function |
||
| 5234 | vConstants = _mm_div_ps(Y,X); |
||
| 5235 | vConstants = XMVectorATanEst(vConstants); |
||
| 5236 | // Discard entries that have been declared void |
||
| 5237 | vResult = XMVectorSelect(vResult,vConstants,ATanResultValid); |
||
| 5238 | return vResult; |
||
| 5239 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5240 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5241 | } |
||
| 5242 | |||
| 5243 | //------------------------------------------------------------------------------ |
||
| 5244 | |||
| 5245 | XMFINLINE XMVECTOR XMVectorLerp |
||
| 5246 | ( |
||
| 5247 | FXMVECTOR V0, |
||
| 5248 | FXMVECTOR V1, |
||
| 5249 | FLOAT t |
||
| 5250 | ) |
||
| 5251 | { |
||
| 5252 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5253 | |||
| 5254 | XMVECTOR Scale; |
||
| 5255 | XMVECTOR Length; |
||
| 5256 | XMVECTOR Result; |
||
| 5257 | |||
| 5258 | // V0 + t * (V1 - V0) |
||
| 5259 | Scale = XMVectorReplicate(t); |
||
| 5260 | Length = XMVectorSubtract(V1, V0); |
||
| 5261 | Result = XMVectorMultiplyAdd(Length, Scale, V0); |
||
| 5262 | |||
| 5263 | return Result; |
||
| 5264 | |||
| 5265 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5266 | XMVECTOR L, S; |
||
| 5267 | XMVECTOR Result; |
||
| 5268 | |||
| 5269 | L = _mm_sub_ps( V1, V0 ); |
||
| 5270 | |||
| 5271 | S = _mm_set_ps1( t ); |
||
| 5272 | |||
| 5273 | Result = _mm_mul_ps( L, S ); |
||
| 5274 | |||
| 5275 | return _mm_add_ps( Result, V0 ); |
||
| 5276 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 5277 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5278 | } |
||
| 5279 | |||
| 5280 | //------------------------------------------------------------------------------ |
||
| 5281 | |||
| 5282 | XMFINLINE XMVECTOR XMVectorLerpV |
||
| 5283 | ( |
||
| 5284 | FXMVECTOR V0, |
||
| 5285 | FXMVECTOR V1, |
||
| 5286 | FXMVECTOR T |
||
| 5287 | ) |
||
| 5288 | { |
||
| 5289 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5290 | |||
| 5291 | XMVECTOR Length; |
||
| 5292 | XMVECTOR Result; |
||
| 5293 | |||
| 5294 | // V0 + T * (V1 - V0) |
||
| 5295 | Length = XMVectorSubtract(V1, V0); |
||
| 5296 | Result = XMVectorMultiplyAdd(Length, T, V0); |
||
| 5297 | |||
| 5298 | return Result; |
||
| 5299 | |||
| 5300 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5301 | XMVECTOR Length; |
||
| 5302 | XMVECTOR Result; |
||
| 5303 | |||
| 5304 | Length = _mm_sub_ps( V1, V0 ); |
||
| 5305 | |||
| 5306 | Result = _mm_mul_ps( Length, T ); |
||
| 5307 | |||
| 5308 | return _mm_add_ps( Result, V0 ); |
||
| 5309 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5310 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5311 | } |
||
| 5312 | |||
| 5313 | //------------------------------------------------------------------------------ |
||
| 5314 | |||
| 5315 | XMFINLINE XMVECTOR XMVectorHermite |
||
| 5316 | ( |
||
| 5317 | FXMVECTOR Position0, |
||
| 5318 | FXMVECTOR Tangent0, |
||
| 5319 | FXMVECTOR Position1, |
||
| 5320 | CXMVECTOR Tangent1, |
||
| 5321 | FLOAT t |
||
| 5322 | ) |
||
| 5323 | { |
||
| 5324 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5325 | |||
| 5326 | XMVECTOR P0; |
||
| 5327 | XMVECTOR T0; |
||
| 5328 | XMVECTOR P1; |
||
| 5329 | XMVECTOR T1; |
||
| 5330 | XMVECTOR Result; |
||
| 5331 | FLOAT t2; |
||
| 5332 | FLOAT t3; |
||
| 5333 | |||
| 5334 | // Result = (2 * t^3 - 3 * t^2 + 1) * Position0 + |
||
| 5335 | // (t^3 - 2 * t^2 + t) * Tangent0 + |
||
| 5336 | // (-2 * t^3 + 3 * t^2) * Position1 + |
||
| 5337 | // (t^3 - t^2) * Tangent1 |
||
| 5338 | t2 = t * t; |
||
| 5339 | t3 = t * t2; |
||
| 5340 | |||
| 5341 | P0 = XMVectorReplicate(2.0f * t3 - 3.0f * t2 + 1.0f); |
||
| 5342 | T0 = XMVectorReplicate(t3 - 2.0f * t2 + t); |
||
| 5343 | P1 = XMVectorReplicate(-2.0f * t3 + 3.0f * t2); |
||
| 5344 | T1 = XMVectorReplicate(t3 - t2); |
||
| 5345 | |||
| 5346 | Result = XMVectorMultiply(P0, Position0); |
||
| 5347 | Result = XMVectorMultiplyAdd(T0, Tangent0, Result); |
||
| 5348 | Result = XMVectorMultiplyAdd(P1, Position1, Result); |
||
| 5349 | Result = XMVectorMultiplyAdd(T1, Tangent1, Result); |
||
| 5350 | |||
| 5351 | return Result; |
||
| 5352 | |||
| 5353 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5354 | FLOAT t2 = t * t; |
||
| 5355 | FLOAT t3 = t * t2; |
||
| 5356 | |||
| 5357 | XMVECTOR P0 = _mm_set_ps1(2.0f * t3 - 3.0f * t2 + 1.0f); |
||
| 5358 | XMVECTOR T0 = _mm_set_ps1(t3 - 2.0f * t2 + t); |
||
| 5359 | XMVECTOR P1 = _mm_set_ps1(-2.0f * t3 + 3.0f * t2); |
||
| 5360 | XMVECTOR T1 = _mm_set_ps1(t3 - t2); |
||
| 5361 | |||
| 5362 | XMVECTOR vResult = _mm_mul_ps(P0, Position0); |
||
| 5363 | XMVECTOR vTemp = _mm_mul_ps(T0, Tangent0); |
||
| 5364 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 5365 | vTemp = _mm_mul_ps(P1, Position1); |
||
| 5366 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 5367 | vTemp = _mm_mul_ps(T1, Tangent1); |
||
| 5368 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 5369 | return vResult; |
||
| 5370 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 5371 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5372 | } |
||
| 5373 | |||
| 5374 | //------------------------------------------------------------------------------ |
||
| 5375 | |||
| 5376 | XMFINLINE XMVECTOR XMVectorHermiteV |
||
| 5377 | ( |
||
| 5378 | FXMVECTOR Position0, |
||
| 5379 | FXMVECTOR Tangent0, |
||
| 5380 | FXMVECTOR Position1, |
||
| 5381 | CXMVECTOR Tangent1, |
||
| 5382 | CXMVECTOR T |
||
| 5383 | ) |
||
| 5384 | { |
||
| 5385 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5386 | |||
| 5387 | XMVECTOR P0; |
||
| 5388 | XMVECTOR T0; |
||
| 5389 | XMVECTOR P1; |
||
| 5390 | XMVECTOR T1; |
||
| 5391 | XMVECTOR Result; |
||
| 5392 | XMVECTOR T2; |
||
| 5393 | XMVECTOR T3; |
||
| 5394 | |||
| 5395 | // Result = (2 * t^3 - 3 * t^2 + 1) * Position0 + |
||
| 5396 | // (t^3 - 2 * t^2 + t) * Tangent0 + |
||
| 5397 | // (-2 * t^3 + 3 * t^2) * Position1 + |
||
| 5398 | // (t^3 - t^2) * Tangent1 |
||
| 5399 | T2 = XMVectorMultiply(T, T); |
||
| 5400 | T3 = XMVectorMultiply(T , T2); |
||
| 5401 | |||
| 5402 | P0 = XMVectorReplicate(2.0f * T3.vector4_f32[0] - 3.0f * T2.vector4_f32[0] + 1.0f); |
||
| 5403 | T0 = XMVectorReplicate(T3.vector4_f32[1] - 2.0f * T2.vector4_f32[1] + T.vector4_f32[1]); |
||
| 5404 | P1 = XMVectorReplicate(-2.0f * T3.vector4_f32[2] + 3.0f * T2.vector4_f32[2]); |
||
| 5405 | T1 = XMVectorReplicate(T3.vector4_f32[3] - T2.vector4_f32[3]); |
||
| 5406 | |||
| 5407 | Result = XMVectorMultiply(P0, Position0); |
||
| 5408 | Result = XMVectorMultiplyAdd(T0, Tangent0, Result); |
||
| 5409 | Result = XMVectorMultiplyAdd(P1, Position1, Result); |
||
| 5410 | Result = XMVectorMultiplyAdd(T1, Tangent1, Result); |
||
| 5411 | |||
| 5412 | return Result; |
||
| 5413 | |||
| 5414 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5415 | static const XMVECTORF32 CatMulT2 = {-3.0f,-2.0f,3.0f,-1.0f}; |
||
| 5416 | static const XMVECTORF32 CatMulT3 = {2.0f,1.0f,-2.0f,1.0f}; |
||
| 5417 | |||
| 5418 | // Result = (2 * t^3 - 3 * t^2 + 1) * Position0 + |
||
| 5419 | // (t^3 - 2 * t^2 + t) * Tangent0 + |
||
| 5420 | // (-2 * t^3 + 3 * t^2) * Position1 + |
||
| 5421 | // (t^3 - t^2) * Tangent1 |
||
| 5422 | XMVECTOR T2 = _mm_mul_ps(T,T); |
||
| 5423 | XMVECTOR T3 = _mm_mul_ps(T,T2); |
||
| 5424 | // Mul by the constants against t^2 |
||
| 5425 | T2 = _mm_mul_ps(T2,CatMulT2); |
||
| 5426 | // Mul by the constants against t^3 |
||
| 5427 | T3 = _mm_mul_ps(T3,CatMulT3); |
||
| 5428 | // T3 now has the pre-result. |
||
| 5429 | T3 = _mm_add_ps(T3,T2); |
||
| 5430 | // I need to add t.y only |
||
| 5431 | T2 = _mm_and_ps(T,g_XMMaskY); |
||
| 5432 | T3 = _mm_add_ps(T3,T2); |
||
| 5433 | // Add 1.0f to x |
||
| 5434 | T3 = _mm_add_ps(T3,g_XMIdentityR0); |
||
| 5435 | // Now, I have the constants created |
||
| 5436 | // Mul the x constant to Position0 |
||
| 5437 | XMVECTOR vResult = _mm_shuffle_ps(T3,T3,_MM_SHUFFLE(0,0,0,0)); |
||
| 5438 | vResult = _mm_mul_ps(vResult,Position0); |
||
| 5439 | // Mul the y constant to Tangent0 |
||
| 5440 | T2 = _mm_shuffle_ps(T3,T3,_MM_SHUFFLE(1,1,1,1)); |
||
| 5441 | T2 = _mm_mul_ps(T2,Tangent0); |
||
| 5442 | vResult = _mm_add_ps(vResult,T2); |
||
| 5443 | // Mul the z constant to Position1 |
||
| 5444 | T2 = _mm_shuffle_ps(T3,T3,_MM_SHUFFLE(2,2,2,2)); |
||
| 5445 | T2 = _mm_mul_ps(T2,Position1); |
||
| 5446 | vResult = _mm_add_ps(vResult,T2); |
||
| 5447 | // Mul the w constant to Tangent1 |
||
| 5448 | T3 = _mm_shuffle_ps(T3,T3,_MM_SHUFFLE(3,3,3,3)); |
||
| 5449 | T3 = _mm_mul_ps(T3,Tangent1); |
||
| 5450 | vResult = _mm_add_ps(vResult,T3); |
||
| 5451 | return vResult; |
||
| 5452 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5453 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5454 | } |
||
| 5455 | |||
| 5456 | //------------------------------------------------------------------------------ |
||
| 5457 | |||
| 5458 | XMFINLINE XMVECTOR XMVectorCatmullRom |
||
| 5459 | ( |
||
| 5460 | FXMVECTOR Position0, |
||
| 5461 | FXMVECTOR Position1, |
||
| 5462 | FXMVECTOR Position2, |
||
| 5463 | CXMVECTOR Position3, |
||
| 5464 | FLOAT t |
||
| 5465 | ) |
||
| 5466 | { |
||
| 5467 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5468 | |||
| 5469 | XMVECTOR P0; |
||
| 5470 | XMVECTOR P1; |
||
| 5471 | XMVECTOR P2; |
||
| 5472 | XMVECTOR P3; |
||
| 5473 | XMVECTOR Result; |
||
| 5474 | FLOAT t2; |
||
| 5475 | FLOAT t3; |
||
| 5476 | |||
| 5477 | // Result = ((-t^3 + 2 * t^2 - t) * Position0 + |
||
| 5478 | // (3 * t^3 - 5 * t^2 + 2) * Position1 + |
||
| 5479 | // (-3 * t^3 + 4 * t^2 + t) * Position2 + |
||
| 5480 | // (t^3 - t^2) * Position3) * 0.5 |
||
| 5481 | t2 = t * t; |
||
| 5482 | t3 = t * t2; |
||
| 5483 | |||
| 5484 | P0 = XMVectorReplicate((-t3 + 2.0f * t2 - t) * 0.5f); |
||
| 5485 | P1 = XMVectorReplicate((3.0f * t3 - 5.0f * t2 + 2.0f) * 0.5f); |
||
| 5486 | P2 = XMVectorReplicate((-3.0f * t3 + 4.0f * t2 + t) * 0.5f); |
||
| 5487 | P3 = XMVectorReplicate((t3 - t2) * 0.5f); |
||
| 5488 | |||
| 5489 | Result = XMVectorMultiply(P0, Position0); |
||
| 5490 | Result = XMVectorMultiplyAdd(P1, Position1, Result); |
||
| 5491 | Result = XMVectorMultiplyAdd(P2, Position2, Result); |
||
| 5492 | Result = XMVectorMultiplyAdd(P3, Position3, Result); |
||
| 5493 | |||
| 5494 | return Result; |
||
| 5495 | |||
| 5496 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5497 | FLOAT t2 = t * t; |
||
| 5498 | FLOAT t3 = t * t2; |
||
| 5499 | |||
| 5500 | XMVECTOR P0 = _mm_set_ps1((-t3 + 2.0f * t2 - t) * 0.5f); |
||
| 5501 | XMVECTOR P1 = _mm_set_ps1((3.0f * t3 - 5.0f * t2 + 2.0f) * 0.5f); |
||
| 5502 | XMVECTOR P2 = _mm_set_ps1((-3.0f * t3 + 4.0f * t2 + t) * 0.5f); |
||
| 5503 | XMVECTOR P3 = _mm_set_ps1((t3 - t2) * 0.5f); |
||
| 5504 | |||
| 5505 | P0 = _mm_mul_ps(P0, Position0); |
||
| 5506 | P1 = _mm_mul_ps(P1, Position1); |
||
| 5507 | P2 = _mm_mul_ps(P2, Position2); |
||
| 5508 | P3 = _mm_mul_ps(P3, Position3); |
||
| 5509 | P0 = _mm_add_ps(P0,P1); |
||
| 5510 | P2 = _mm_add_ps(P2,P3); |
||
| 5511 | P0 = _mm_add_ps(P0,P2); |
||
| 5512 | return P0; |
||
| 5513 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 5514 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5515 | } |
||
| 5516 | |||
| 5517 | //------------------------------------------------------------------------------ |
||
| 5518 | |||
| 5519 | XMFINLINE XMVECTOR XMVectorCatmullRomV |
||
| 5520 | ( |
||
| 5521 | FXMVECTOR Position0, |
||
| 5522 | FXMVECTOR Position1, |
||
| 5523 | FXMVECTOR Position2, |
||
| 5524 | CXMVECTOR Position3, |
||
| 5525 | CXMVECTOR T |
||
| 5526 | ) |
||
| 5527 | { |
||
| 5528 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5529 | float fx = T.vector4_f32[0]; |
||
| 5530 | float fy = T.vector4_f32[1]; |
||
| 5531 | float fz = T.vector4_f32[2]; |
||
| 5532 | float fw = T.vector4_f32[3]; |
||
| 5533 | XMVECTOR vResult = { |
||
| 5534 | 0.5f*((-fx*fx*fx+2*fx*fx-fx)*Position0.vector4_f32[0]+ |
||
| 5535 | (3*fx*fx*fx-5*fx*fx+2)*Position1.vector4_f32[0]+ |
||
| 5536 | (-3*fx*fx*fx+4*fx*fx+fx)*Position2.vector4_f32[0]+ |
||
| 5537 | (fx*fx*fx-fx*fx)*Position3.vector4_f32[0]), |
||
| 5538 | 0.5f*((-fy*fy*fy+2*fy*fy-fy)*Position0.vector4_f32[1]+ |
||
| 5539 | (3*fy*fy*fy-5*fy*fy+2)*Position1.vector4_f32[1]+ |
||
| 5540 | (-3*fy*fy*fy+4*fy*fy+fy)*Position2.vector4_f32[1]+ |
||
| 5541 | (fy*fy*fy-fy*fy)*Position3.vector4_f32[1]), |
||
| 5542 | 0.5f*((-fz*fz*fz+2*fz*fz-fz)*Position0.vector4_f32[2]+ |
||
| 5543 | (3*fz*fz*fz-5*fz*fz+2)*Position1.vector4_f32[2]+ |
||
| 5544 | (-3*fz*fz*fz+4*fz*fz+fz)*Position2.vector4_f32[2]+ |
||
| 5545 | (fz*fz*fz-fz*fz)*Position3.vector4_f32[2]), |
||
| 5546 | 0.5f*((-fw*fw*fw+2*fw*fw-fw)*Position0.vector4_f32[3]+ |
||
| 5547 | (3*fw*fw*fw-5*fw*fw+2)*Position1.vector4_f32[3]+ |
||
| 5548 | (-3*fw*fw*fw+4*fw*fw+fw)*Position2.vector4_f32[3]+ |
||
| 5549 | (fw*fw*fw-fw*fw)*Position3.vector4_f32[3]) |
||
| 5550 | }; |
||
| 5551 | return vResult; |
||
| 5552 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5553 | static const XMVECTORF32 Catmul2 = {2.0f,2.0f,2.0f,2.0f}; |
||
| 5554 | static const XMVECTORF32 Catmul3 = {3.0f,3.0f,3.0f,3.0f}; |
||
| 5555 | static const XMVECTORF32 Catmul4 = {4.0f,4.0f,4.0f,4.0f}; |
||
| 5556 | static const XMVECTORF32 Catmul5 = {5.0f,5.0f,5.0f,5.0f}; |
||
| 5557 | // Cache T^2 and T^3 |
||
| 5558 | XMVECTOR T2 = _mm_mul_ps(T,T); |
||
| 5559 | XMVECTOR T3 = _mm_mul_ps(T,T2); |
||
| 5560 | // Perform the Position0 term |
||
| 5561 | XMVECTOR vResult = _mm_add_ps(T2,T2); |
||
| 5562 | vResult = _mm_sub_ps(vResult,T); |
||
| 5563 | vResult = _mm_sub_ps(vResult,T3); |
||
| 5564 | vResult = _mm_mul_ps(vResult,Position0); |
||
| 5565 | // Perform the Position1 term and add |
||
| 5566 | XMVECTOR vTemp = _mm_mul_ps(T3,Catmul3); |
||
| 5567 | XMVECTOR vTemp2 = _mm_mul_ps(T2,Catmul5); |
||
| 5568 | vTemp = _mm_sub_ps(vTemp,vTemp2); |
||
| 5569 | vTemp = _mm_add_ps(vTemp,Catmul2); |
||
| 5570 | vTemp = _mm_mul_ps(vTemp,Position1); |
||
| 5571 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 5572 | // Perform the Position2 term and add |
||
| 5573 | vTemp = _mm_mul_ps(T2,Catmul4); |
||
| 5574 | vTemp2 = _mm_mul_ps(T3,Catmul3); |
||
| 5575 | vTemp = _mm_sub_ps(vTemp,vTemp2); |
||
| 5576 | vTemp = _mm_add_ps(vTemp,T); |
||
| 5577 | vTemp = _mm_mul_ps(vTemp,Position2); |
||
| 5578 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 5579 | // Position3 is the last term |
||
| 5580 | T3 = _mm_sub_ps(T3,T2); |
||
| 5581 | T3 = _mm_mul_ps(T3,Position3); |
||
| 5582 | vResult = _mm_add_ps(vResult,T3); |
||
| 5583 | // Multiply by 0.5f and exit |
||
| 5584 | vResult = _mm_mul_ps(vResult,g_XMOneHalf); |
||
| 5585 | return vResult; |
||
| 5586 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5587 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5588 | } |
||
| 5589 | |||
| 5590 | //------------------------------------------------------------------------------ |
||
| 5591 | |||
| 5592 | XMFINLINE XMVECTOR XMVectorBaryCentric |
||
| 5593 | ( |
||
| 5594 | FXMVECTOR Position0, |
||
| 5595 | FXMVECTOR Position1, |
||
| 5596 | FXMVECTOR Position2, |
||
| 5597 | FLOAT f, |
||
| 5598 | FLOAT g |
||
| 5599 | ) |
||
| 5600 | { |
||
| 5601 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5602 | |||
| 5603 | // Result = Position0 + f * (Position1 - Position0) + g * (Position2 - Position0) |
||
| 5604 | XMVECTOR P10; |
||
| 5605 | XMVECTOR P20; |
||
| 5606 | XMVECTOR ScaleF; |
||
| 5607 | XMVECTOR ScaleG; |
||
| 5608 | XMVECTOR Result; |
||
| 5609 | |||
| 5610 | P10 = XMVectorSubtract(Position1, Position0); |
||
| 5611 | ScaleF = XMVectorReplicate(f); |
||
| 5612 | |||
| 5613 | P20 = XMVectorSubtract(Position2, Position0); |
||
| 5614 | ScaleG = XMVectorReplicate(g); |
||
| 5615 | |||
| 5616 | Result = XMVectorMultiplyAdd(P10, ScaleF, Position0); |
||
| 5617 | Result = XMVectorMultiplyAdd(P20, ScaleG, Result); |
||
| 5618 | |||
| 5619 | return Result; |
||
| 5620 | |||
| 5621 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5622 | XMVECTOR R1 = _mm_sub_ps(Position1,Position0); |
||
| 5623 | XMVECTOR SF = _mm_set_ps1(f); |
||
| 5624 | XMVECTOR R2 = _mm_sub_ps(Position2,Position0); |
||
| 5625 | XMVECTOR SG = _mm_set_ps1(g); |
||
| 5626 | R1 = _mm_mul_ps(R1,SF); |
||
| 5627 | R2 = _mm_mul_ps(R2,SG); |
||
| 5628 | R1 = _mm_add_ps(R1,Position0); |
||
| 5629 | R1 = _mm_add_ps(R1,R2); |
||
| 5630 | return R1; |
||
| 5631 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 5632 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5633 | } |
||
| 5634 | |||
| 5635 | //------------------------------------------------------------------------------ |
||
| 5636 | |||
| 5637 | XMFINLINE XMVECTOR XMVectorBaryCentricV |
||
| 5638 | ( |
||
| 5639 | FXMVECTOR Position0, |
||
| 5640 | FXMVECTOR Position1, |
||
| 5641 | FXMVECTOR Position2, |
||
| 5642 | CXMVECTOR F, |
||
| 5643 | CXMVECTOR G |
||
| 5644 | ) |
||
| 5645 | { |
||
| 5646 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5647 | |||
| 5648 | // Result = Position0 + f * (Position1 - Position0) + g * (Position2 - Position0) |
||
| 5649 | XMVECTOR P10; |
||
| 5650 | XMVECTOR P20; |
||
| 5651 | XMVECTOR Result; |
||
| 5652 | |||
| 5653 | P10 = XMVectorSubtract(Position1, Position0); |
||
| 5654 | P20 = XMVectorSubtract(Position2, Position0); |
||
| 5655 | |||
| 5656 | Result = XMVectorMultiplyAdd(P10, F, Position0); |
||
| 5657 | Result = XMVectorMultiplyAdd(P20, G, Result); |
||
| 5658 | |||
| 5659 | return Result; |
||
| 5660 | |||
| 5661 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5662 | XMVECTOR R1 = _mm_sub_ps(Position1,Position0); |
||
| 5663 | XMVECTOR R2 = _mm_sub_ps(Position2,Position0); |
||
| 5664 | R1 = _mm_mul_ps(R1,F); |
||
| 5665 | R2 = _mm_mul_ps(R2,G); |
||
| 5666 | R1 = _mm_add_ps(R1,Position0); |
||
| 5667 | R1 = _mm_add_ps(R1,R2); |
||
| 5668 | return R1; |
||
| 5669 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5670 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5671 | } |
||
| 5672 | |||
| 5673 | /**************************************************************************** |
||
| 5674 | * |
||
| 5675 | * 2D Vector |
||
| 5676 | * |
||
| 5677 | ****************************************************************************/ |
||
| 5678 | |||
| 5679 | //------------------------------------------------------------------------------ |
||
| 5680 | // Comparison operations |
||
| 5681 | //------------------------------------------------------------------------------ |
||
| 5682 | |||
| 5683 | //------------------------------------------------------------------------------ |
||
| 5684 | |||
| 5685 | XMFINLINE BOOL XMVector2Equal |
||
| 5686 | ( |
||
| 5687 | FXMVECTOR V1, |
||
| 5688 | FXMVECTOR V2 |
||
| 5689 | ) |
||
| 5690 | { |
||
| 5691 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5692 | return (((V1.vector4_f32[0] == V2.vector4_f32[0]) && (V1.vector4_f32[1] == V2.vector4_f32[1])) != 0); |
||
| 5693 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5694 | XMVECTOR vTemp = _mm_cmpeq_ps(V1,V2); |
||
| 5695 | // z and w are don't care |
||
| 5696 | return (((_mm_movemask_ps(vTemp)&3)==3) != 0); |
||
| 5697 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5698 | return XMComparisonAllTrue(XMVector2EqualR(V1, V2)); |
||
| 5699 | #endif |
||
| 5700 | } |
||
| 5701 | |||
| 5702 | |||
| 5703 | //------------------------------------------------------------------------------ |
||
| 5704 | |||
| 5705 | XMFINLINE UINT XMVector2EqualR |
||
| 5706 | ( |
||
| 5707 | FXMVECTOR V1, |
||
| 5708 | FXMVECTOR V2 |
||
| 5709 | ) |
||
| 5710 | { |
||
| 5711 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5712 | |||
| 5713 | UINT CR = 0; |
||
| 5714 | |||
| 5715 | if ((V1.vector4_f32[0] == V2.vector4_f32[0]) && |
||
| 5716 | (V1.vector4_f32[1] == V2.vector4_f32[1])) |
||
| 5717 | { |
||
| 5718 | CR = XM_CRMASK_CR6TRUE; |
||
| 5719 | } |
||
| 5720 | else if ((V1.vector4_f32[0] != V2.vector4_f32[0]) && |
||
| 5721 | (V1.vector4_f32[1] != V2.vector4_f32[1])) |
||
| 5722 | { |
||
| 5723 | CR = XM_CRMASK_CR6FALSE; |
||
| 5724 | } |
||
| 5725 | return CR; |
||
| 5726 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5727 | XMVECTOR vTemp = _mm_cmpeq_ps(V1,V2); |
||
| 5728 | // z and w are don't care |
||
| 5729 | int iTest = _mm_movemask_ps(vTemp)&3; |
||
| 5730 | UINT CR = 0; |
||
| 5731 | if (iTest==3) |
||
| 5732 | { |
||
| 5733 | CR = XM_CRMASK_CR6TRUE; |
||
| 5734 | } |
||
| 5735 | else if (!iTest) |
||
| 5736 | { |
||
| 5737 | CR = XM_CRMASK_CR6FALSE; |
||
| 5738 | } |
||
| 5739 | return CR; |
||
| 5740 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5741 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5742 | } |
||
| 5743 | |||
| 5744 | //------------------------------------------------------------------------------ |
||
| 5745 | |||
| 5746 | XMFINLINE BOOL XMVector2EqualInt |
||
| 5747 | ( |
||
| 5748 | FXMVECTOR V1, |
||
| 5749 | FXMVECTOR V2 |
||
| 5750 | ) |
||
| 5751 | { |
||
| 5752 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5753 | return (((V1.vector4_u32[0] == V2.vector4_u32[0]) && (V1.vector4_u32[1] == V2.vector4_u32[1])) != 0); |
||
| 5754 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5755 | __m128i vTemp = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0]); |
||
| 5756 | return (((_mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTemp)[0])&3)==3) != 0); |
||
| 5757 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5758 | return XMComparisonAllTrue(XMVector2EqualIntR(V1, V2)); |
||
| 5759 | #endif |
||
| 5760 | } |
||
| 5761 | |||
| 5762 | //------------------------------------------------------------------------------ |
||
| 5763 | |||
| 5764 | XMFINLINE UINT XMVector2EqualIntR |
||
| 5765 | ( |
||
| 5766 | FXMVECTOR V1, |
||
| 5767 | FXMVECTOR V2 |
||
| 5768 | ) |
||
| 5769 | { |
||
| 5770 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5771 | |||
| 5772 | UINT CR = 0; |
||
| 5773 | if ((V1.vector4_u32[0] == V2.vector4_u32[0]) && |
||
| 5774 | (V1.vector4_u32[1] == V2.vector4_u32[1])) |
||
| 5775 | { |
||
| 5776 | CR = XM_CRMASK_CR6TRUE; |
||
| 5777 | } |
||
| 5778 | else if ((V1.vector4_u32[0] != V2.vector4_u32[0]) && |
||
| 5779 | (V1.vector4_u32[1] != V2.vector4_u32[1])) |
||
| 5780 | { |
||
| 5781 | CR = XM_CRMASK_CR6FALSE; |
||
| 5782 | } |
||
| 5783 | return CR; |
||
| 5784 | |||
| 5785 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5786 | __m128i vTemp = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0]); |
||
| 5787 | int iTest = _mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTemp)[0])&3; |
||
| 5788 | UINT CR = 0; |
||
| 5789 | if (iTest==3) |
||
| 5790 | { |
||
| 5791 | CR = XM_CRMASK_CR6TRUE; |
||
| 5792 | } |
||
| 5793 | else if (!iTest) |
||
| 5794 | { |
||
| 5795 | CR = XM_CRMASK_CR6FALSE; |
||
| 5796 | } |
||
| 5797 | return CR; |
||
| 5798 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5799 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5800 | } |
||
| 5801 | |||
| 5802 | //------------------------------------------------------------------------------ |
||
| 5803 | |||
| 5804 | XMFINLINE BOOL XMVector2NearEqual |
||
| 5805 | ( |
||
| 5806 | FXMVECTOR V1, |
||
| 5807 | FXMVECTOR V2, |
||
| 5808 | FXMVECTOR Epsilon |
||
| 5809 | ) |
||
| 5810 | { |
||
| 5811 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5812 | FLOAT dx, dy; |
||
| 5813 | dx = fabsf(V1.vector4_f32[0]-V2.vector4_f32[0]); |
||
| 5814 | dy = fabsf(V1.vector4_f32[1]-V2.vector4_f32[1]); |
||
| 5815 | return ((dx <= Epsilon.vector4_f32[0]) && |
||
| 5816 | (dy <= Epsilon.vector4_f32[1])); |
||
| 5817 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5818 | // Get the difference |
||
| 5819 | XMVECTOR vDelta = _mm_sub_ps(V1,V2); |
||
| 5820 | // Get the absolute value of the difference |
||
| 5821 | XMVECTOR vTemp = _mm_setzero_ps(); |
||
| 5822 | vTemp = _mm_sub_ps(vTemp,vDelta); |
||
| 5823 | vTemp = _mm_max_ps(vTemp,vDelta); |
||
| 5824 | vTemp = _mm_cmple_ps(vTemp,Epsilon); |
||
| 5825 | // z and w are don't care |
||
| 5826 | return (((_mm_movemask_ps(vTemp)&3)==0x3) != 0); |
||
| 5827 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5828 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5829 | } |
||
| 5830 | |||
| 5831 | //------------------------------------------------------------------------------ |
||
| 5832 | |||
| 5833 | XMFINLINE BOOL XMVector2NotEqual |
||
| 5834 | ( |
||
| 5835 | FXMVECTOR V1, |
||
| 5836 | FXMVECTOR V2 |
||
| 5837 | ) |
||
| 5838 | { |
||
| 5839 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5840 | return (((V1.vector4_f32[0] != V2.vector4_f32[0]) || (V1.vector4_f32[1] != V2.vector4_f32[1])) != 0); |
||
| 5841 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5842 | XMVECTOR vTemp = _mm_cmpeq_ps(V1,V2); |
||
| 5843 | // z and w are don't care |
||
| 5844 | return (((_mm_movemask_ps(vTemp)&3)!=3) != 0); |
||
| 5845 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5846 | return XMComparisonAnyFalse(XMVector2EqualR(V1, V2)); |
||
| 5847 | #endif |
||
| 5848 | } |
||
| 5849 | |||
| 5850 | //------------------------------------------------------------------------------ |
||
| 5851 | |||
| 5852 | XMFINLINE BOOL XMVector2NotEqualInt |
||
| 5853 | ( |
||
| 5854 | FXMVECTOR V1, |
||
| 5855 | FXMVECTOR V2 |
||
| 5856 | ) |
||
| 5857 | { |
||
| 5858 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5859 | return (((V1.vector4_u32[0] != V2.vector4_u32[0]) || (V1.vector4_u32[1] != V2.vector4_u32[1])) != 0); |
||
| 5860 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5861 | __m128i vTemp = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0]); |
||
| 5862 | return (((_mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTemp)[0])&3)!=3) != 0); |
||
| 5863 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5864 | return XMComparisonAnyFalse(XMVector2EqualIntR(V1, V2)); |
||
| 5865 | #endif |
||
| 5866 | } |
||
| 5867 | |||
| 5868 | //------------------------------------------------------------------------------ |
||
| 5869 | |||
| 5870 | XMFINLINE BOOL XMVector2Greater |
||
| 5871 | ( |
||
| 5872 | FXMVECTOR V1, |
||
| 5873 | FXMVECTOR V2 |
||
| 5874 | ) |
||
| 5875 | { |
||
| 5876 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5877 | return (((V1.vector4_f32[0] > V2.vector4_f32[0]) && (V1.vector4_f32[1] > V2.vector4_f32[1])) != 0); |
||
| 5878 | |||
| 5879 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5880 | XMVECTOR vTemp = _mm_cmpgt_ps(V1,V2); |
||
| 5881 | // z and w are don't care |
||
| 5882 | return (((_mm_movemask_ps(vTemp)&3)==3) != 0); |
||
| 5883 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5884 | return XMComparisonAllTrue(XMVector2GreaterR(V1, V2)); |
||
| 5885 | #endif |
||
| 5886 | } |
||
| 5887 | |||
| 5888 | //------------------------------------------------------------------------------ |
||
| 5889 | |||
| 5890 | XMFINLINE UINT XMVector2GreaterR |
||
| 5891 | ( |
||
| 5892 | FXMVECTOR V1, |
||
| 5893 | FXMVECTOR V2 |
||
| 5894 | ) |
||
| 5895 | { |
||
| 5896 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5897 | |||
| 5898 | UINT CR = 0; |
||
| 5899 | if ((V1.vector4_f32[0] > V2.vector4_f32[0]) && |
||
| 5900 | (V1.vector4_f32[1] > V2.vector4_f32[1])) |
||
| 5901 | { |
||
| 5902 | CR = XM_CRMASK_CR6TRUE; |
||
| 5903 | } |
||
| 5904 | else if ((V1.vector4_f32[0] <= V2.vector4_f32[0]) && |
||
| 5905 | (V1.vector4_f32[1] <= V2.vector4_f32[1])) |
||
| 5906 | { |
||
| 5907 | CR = XM_CRMASK_CR6FALSE; |
||
| 5908 | } |
||
| 5909 | return CR; |
||
| 5910 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5911 | XMVECTOR vTemp = _mm_cmpgt_ps(V1,V2); |
||
| 5912 | int iTest = _mm_movemask_ps(vTemp)&3; |
||
| 5913 | UINT CR = 0; |
||
| 5914 | if (iTest==3) |
||
| 5915 | { |
||
| 5916 | CR = XM_CRMASK_CR6TRUE; |
||
| 5917 | } |
||
| 5918 | else if (!iTest) |
||
| 5919 | { |
||
| 5920 | CR = XM_CRMASK_CR6FALSE; |
||
| 5921 | } |
||
| 5922 | return CR; |
||
| 5923 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5924 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5925 | } |
||
| 5926 | |||
| 5927 | //------------------------------------------------------------------------------ |
||
| 5928 | |||
| 5929 | XMFINLINE BOOL XMVector2GreaterOrEqual |
||
| 5930 | ( |
||
| 5931 | FXMVECTOR V1, |
||
| 5932 | FXMVECTOR V2 |
||
| 5933 | ) |
||
| 5934 | { |
||
| 5935 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5936 | return (((V1.vector4_f32[0] >= V2.vector4_f32[0]) && (V1.vector4_f32[1] >= V2.vector4_f32[1])) != 0); |
||
| 5937 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5938 | XMVECTOR vTemp = _mm_cmpge_ps(V1,V2); |
||
| 5939 | return (((_mm_movemask_ps(vTemp)&3)==3) != 0); |
||
| 5940 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5941 | return XMComparisonAllTrue(XMVector2GreaterOrEqualR(V1, V2)); |
||
| 5942 | #endif |
||
| 5943 | } |
||
| 5944 | |||
| 5945 | //------------------------------------------------------------------------------ |
||
| 5946 | |||
| 5947 | XMFINLINE UINT XMVector2GreaterOrEqualR |
||
| 5948 | ( |
||
| 5949 | FXMVECTOR V1, |
||
| 5950 | FXMVECTOR V2 |
||
| 5951 | ) |
||
| 5952 | { |
||
| 5953 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5954 | UINT CR = 0; |
||
| 5955 | if ((V1.vector4_f32[0] >= V2.vector4_f32[0]) && |
||
| 5956 | (V1.vector4_f32[1] >= V2.vector4_f32[1])) |
||
| 5957 | { |
||
| 5958 | CR = XM_CRMASK_CR6TRUE; |
||
| 5959 | } |
||
| 5960 | else if ((V1.vector4_f32[0] < V2.vector4_f32[0]) && |
||
| 5961 | (V1.vector4_f32[1] < V2.vector4_f32[1])) |
||
| 5962 | { |
||
| 5963 | CR = XM_CRMASK_CR6FALSE; |
||
| 5964 | } |
||
| 5965 | return CR; |
||
| 5966 | |||
| 5967 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5968 | XMVECTOR vTemp = _mm_cmpge_ps(V1,V2); |
||
| 5969 | int iTest = _mm_movemask_ps(vTemp)&3; |
||
| 5970 | UINT CR = 0; |
||
| 5971 | if (iTest == 3) |
||
| 5972 | { |
||
| 5973 | CR = XM_CRMASK_CR6TRUE; |
||
| 5974 | } |
||
| 5975 | else if (!iTest) |
||
| 5976 | { |
||
| 5977 | CR = XM_CRMASK_CR6FALSE; |
||
| 5978 | } |
||
| 5979 | return CR; |
||
| 5980 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5981 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 5982 | } |
||
| 5983 | |||
| 5984 | //------------------------------------------------------------------------------ |
||
| 5985 | |||
| 5986 | XMFINLINE BOOL XMVector2Less |
||
| 5987 | ( |
||
| 5988 | FXMVECTOR V1, |
||
| 5989 | FXMVECTOR V2 |
||
| 5990 | ) |
||
| 5991 | { |
||
| 5992 | #if defined(_XM_NO_INTRINSICS_) |
||
| 5993 | return (((V1.vector4_f32[0] < V2.vector4_f32[0]) && (V1.vector4_f32[1] < V2.vector4_f32[1])) != 0); |
||
| 5994 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 5995 | XMVECTOR vTemp = _mm_cmplt_ps(V1,V2); |
||
| 5996 | return (((_mm_movemask_ps(vTemp)&3)==3) != 0); |
||
| 5997 | #else // _XM_VMX128_INTRINSICS_ |
||
| 5998 | return XMComparisonAllTrue(XMVector2GreaterR(V2, V1)); |
||
| 5999 | #endif |
||
| 6000 | } |
||
| 6001 | |||
| 6002 | //------------------------------------------------------------------------------ |
||
| 6003 | |||
| 6004 | XMFINLINE BOOL XMVector2LessOrEqual |
||
| 6005 | ( |
||
| 6006 | FXMVECTOR V1, |
||
| 6007 | FXMVECTOR V2 |
||
| 6008 | ) |
||
| 6009 | { |
||
| 6010 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6011 | return (((V1.vector4_f32[0] <= V2.vector4_f32[0]) && (V1.vector4_f32[1] <= V2.vector4_f32[1])) != 0); |
||
| 6012 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6013 | XMVECTOR vTemp = _mm_cmple_ps(V1,V2); |
||
| 6014 | return (((_mm_movemask_ps(vTemp)&3)==3) != 0); |
||
| 6015 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6016 | return XMComparisonAllTrue(XMVector2GreaterOrEqualR(V2, V1)); |
||
| 6017 | #endif |
||
| 6018 | } |
||
| 6019 | |||
| 6020 | //------------------------------------------------------------------------------ |
||
| 6021 | |||
| 6022 | XMFINLINE BOOL XMVector2InBounds |
||
| 6023 | ( |
||
| 6024 | FXMVECTOR V, |
||
| 6025 | FXMVECTOR Bounds |
||
| 6026 | ) |
||
| 6027 | { |
||
| 6028 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6029 | return (((V.vector4_f32[0] <= Bounds.vector4_f32[0] && V.vector4_f32[0] >= -Bounds.vector4_f32[0]) && |
||
| 6030 | (V.vector4_f32[1] <= Bounds.vector4_f32[1] && V.vector4_f32[1] >= -Bounds.vector4_f32[1])) != 0); |
||
| 6031 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6032 | // Test if less than or equal |
||
| 6033 | XMVECTOR vTemp1 = _mm_cmple_ps(V,Bounds); |
||
| 6034 | // Negate the bounds |
||
| 6035 | XMVECTOR vTemp2 = _mm_mul_ps(Bounds,g_XMNegativeOne); |
||
| 6036 | // Test if greater or equal (Reversed) |
||
| 6037 | vTemp2 = _mm_cmple_ps(vTemp2,V); |
||
| 6038 | // Blend answers |
||
| 6039 | vTemp1 = _mm_and_ps(vTemp1,vTemp2); |
||
| 6040 | // x and y in bounds? (z and w are don't care) |
||
| 6041 | return (((_mm_movemask_ps(vTemp1)&0x3)==0x3) != 0); |
||
| 6042 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6043 | return XMComparisonAllInBounds(XMVector2InBoundsR(V, Bounds)); |
||
| 6044 | #endif |
||
| 6045 | } |
||
| 6046 | |||
| 6047 | //------------------------------------------------------------------------------ |
||
| 6048 | |||
| 6049 | XMFINLINE UINT XMVector2InBoundsR |
||
| 6050 | ( |
||
| 6051 | FXMVECTOR V, |
||
| 6052 | FXMVECTOR Bounds |
||
| 6053 | ) |
||
| 6054 | { |
||
| 6055 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6056 | UINT CR = 0; |
||
| 6057 | if ((V.vector4_f32[0] <= Bounds.vector4_f32[0] && V.vector4_f32[0] >= -Bounds.vector4_f32[0]) && |
||
| 6058 | (V.vector4_f32[1] <= Bounds.vector4_f32[1] && V.vector4_f32[1] >= -Bounds.vector4_f32[1])) |
||
| 6059 | { |
||
| 6060 | CR = XM_CRMASK_CR6BOUNDS; |
||
| 6061 | } |
||
| 6062 | return CR; |
||
| 6063 | |||
| 6064 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6065 | // Test if less than or equal |
||
| 6066 | XMVECTOR vTemp1 = _mm_cmple_ps(V,Bounds); |
||
| 6067 | // Negate the bounds |
||
| 6068 | XMVECTOR vTemp2 = _mm_mul_ps(Bounds,g_XMNegativeOne); |
||
| 6069 | // Test if greater or equal (Reversed) |
||
| 6070 | vTemp2 = _mm_cmple_ps(vTemp2,V); |
||
| 6071 | // Blend answers |
||
| 6072 | vTemp1 = _mm_and_ps(vTemp1,vTemp2); |
||
| 6073 | // x and y in bounds? (z and w are don't care) |
||
| 6074 | return ((_mm_movemask_ps(vTemp1)&0x3)==0x3) ? XM_CRMASK_CR6BOUNDS : 0; |
||
| 6075 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6076 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6077 | } |
||
| 6078 | |||
| 6079 | //------------------------------------------------------------------------------ |
||
| 6080 | |||
| 6081 | XMFINLINE BOOL XMVector2IsNaN |
||
| 6082 | ( |
||
| 6083 | FXMVECTOR V |
||
| 6084 | ) |
||
| 6085 | { |
||
| 6086 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6087 | return (XMISNAN(V.vector4_f32[0]) || |
||
| 6088 | XMISNAN(V.vector4_f32[1])); |
||
| 6089 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6090 | // Mask off the exponent |
||
| 6091 | __m128i vTempInf = _mm_and_si128(reinterpret_cast<const __m128i *>(&V)[0],g_XMInfinity); |
||
| 6092 | // Mask off the mantissa |
||
| 6093 | __m128i vTempNan = _mm_and_si128(reinterpret_cast<const __m128i *>(&V)[0],g_XMQNaNTest); |
||
| 6094 | // Are any of the exponents == 0x7F800000? |
||
| 6095 | vTempInf = _mm_cmpeq_epi32(vTempInf,g_XMInfinity); |
||
| 6096 | // Are any of the mantissa's zero? (SSE2 doesn't have a neq test) |
||
| 6097 | vTempNan = _mm_cmpeq_epi32(vTempNan,g_XMZero); |
||
| 6098 | // Perform a not on the NaN test to be true on NON-zero mantissas |
||
| 6099 | vTempNan = _mm_andnot_si128(vTempNan,vTempInf); |
||
| 6100 | // If x or y are NaN, the signs are true after the merge above |
||
| 6101 | return ((_mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTempNan)[0])&3) != 0); |
||
| 6102 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6103 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6104 | } |
||
| 6105 | |||
| 6106 | //------------------------------------------------------------------------------ |
||
| 6107 | |||
| 6108 | XMFINLINE BOOL XMVector2IsInfinite |
||
| 6109 | ( |
||
| 6110 | FXMVECTOR V |
||
| 6111 | ) |
||
| 6112 | { |
||
| 6113 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6114 | |||
| 6115 | return (XMISINF(V.vector4_f32[0]) || |
||
| 6116 | XMISINF(V.vector4_f32[1])); |
||
| 6117 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6118 | // Mask off the sign bit |
||
| 6119 | __m128 vTemp = _mm_and_ps(V,g_XMAbsMask); |
||
| 6120 | // Compare to infinity |
||
| 6121 | vTemp = _mm_cmpeq_ps(vTemp,g_XMInfinity); |
||
| 6122 | // If x or z are infinity, the signs are true. |
||
| 6123 | return ((_mm_movemask_ps(vTemp)&3) != 0); |
||
| 6124 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6125 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6126 | } |
||
| 6127 | |||
| 6128 | //------------------------------------------------------------------------------ |
||
| 6129 | // Computation operations |
||
| 6130 | //------------------------------------------------------------------------------ |
||
| 6131 | |||
| 6132 | //------------------------------------------------------------------------------ |
||
| 6133 | |||
| 6134 | XMFINLINE XMVECTOR XMVector2Dot |
||
| 6135 | ( |
||
| 6136 | FXMVECTOR V1, |
||
| 6137 | FXMVECTOR V2 |
||
| 6138 | ) |
||
| 6139 | { |
||
| 6140 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6141 | |||
| 6142 | XMVECTOR Result; |
||
| 6143 | |||
| 6144 | Result.vector4_f32[0] = |
||
| 6145 | Result.vector4_f32[1] = |
||
| 6146 | Result.vector4_f32[2] = |
||
| 6147 | Result.vector4_f32[3] = V1.vector4_f32[0] * V2.vector4_f32[0] + V1.vector4_f32[1] * V2.vector4_f32[1]; |
||
| 6148 | |||
| 6149 | return Result; |
||
| 6150 | |||
| 6151 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6152 | // Perform the dot product on x and y |
||
| 6153 | XMVECTOR vLengthSq = _mm_mul_ps(V1,V2); |
||
| 6154 | // vTemp has y splatted |
||
| 6155 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,1,1,1)); |
||
| 6156 | // x+y |
||
| 6157 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 6158 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 6159 | return vLengthSq; |
||
| 6160 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6161 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6162 | } |
||
| 6163 | |||
| 6164 | //------------------------------------------------------------------------------ |
||
| 6165 | |||
| 6166 | XMFINLINE XMVECTOR XMVector2Cross |
||
| 6167 | ( |
||
| 6168 | FXMVECTOR V1, |
||
| 6169 | FXMVECTOR V2 |
||
| 6170 | ) |
||
| 6171 | { |
||
| 6172 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6173 | FLOAT fCross = (V1.vector4_f32[0] * V2.vector4_f32[1]) - (V1.vector4_f32[1] * V2.vector4_f32[0]); |
||
| 6174 | XMVECTOR vResult = { |
||
| 6175 | fCross, |
||
| 6176 | fCross, |
||
| 6177 | fCross, |
||
| 6178 | fCross |
||
| 6179 | }; |
||
| 6180 | return vResult; |
||
| 6181 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6182 | // Swap x and y |
||
| 6183 | XMVECTOR vResult = _mm_shuffle_ps(V2,V2,_MM_SHUFFLE(0,1,0,1)); |
||
| 6184 | // Perform the muls |
||
| 6185 | vResult = _mm_mul_ps(vResult,V1); |
||
| 6186 | // Splat y |
||
| 6187 | XMVECTOR vTemp = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(1,1,1,1)); |
||
| 6188 | // Sub the values |
||
| 6189 | vResult = _mm_sub_ss(vResult,vTemp); |
||
| 6190 | // Splat the cross product |
||
| 6191 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(0,0,0,0)); |
||
| 6192 | return vResult; |
||
| 6193 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6194 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6195 | } |
||
| 6196 | |||
| 6197 | //------------------------------------------------------------------------------ |
||
| 6198 | |||
| 6199 | XMFINLINE XMVECTOR XMVector2LengthSq |
||
| 6200 | ( |
||
| 6201 | FXMVECTOR V |
||
| 6202 | ) |
||
| 6203 | { |
||
| 6204 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6205 | return XMVector2Dot(V, V); |
||
| 6206 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6207 | // Perform the dot product on x and y |
||
| 6208 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 6209 | // vTemp has y splatted |
||
| 6210 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,1,1,1)); |
||
| 6211 | // x+y |
||
| 6212 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 6213 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 6214 | return vLengthSq; |
||
| 6215 | #else |
||
| 6216 | return XMVector2Dot(V, V); |
||
| 6217 | #endif |
||
| 6218 | } |
||
| 6219 | |||
| 6220 | //------------------------------------------------------------------------------ |
||
| 6221 | |||
| 6222 | XMFINLINE XMVECTOR XMVector2ReciprocalLengthEst |
||
| 6223 | ( |
||
| 6224 | FXMVECTOR V |
||
| 6225 | ) |
||
| 6226 | { |
||
| 6227 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6228 | |||
| 6229 | XMVECTOR Result; |
||
| 6230 | |||
| 6231 | Result = XMVector2LengthSq(V); |
||
| 6232 | Result = XMVectorReciprocalSqrtEst(Result); |
||
| 6233 | |||
| 6234 | return Result; |
||
| 6235 | |||
| 6236 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6237 | // Perform the dot product on x and y |
||
| 6238 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 6239 | // vTemp has y splatted |
||
| 6240 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,1,1,1)); |
||
| 6241 | // x+y |
||
| 6242 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 6243 | vLengthSq = _mm_rsqrt_ss(vLengthSq); |
||
| 6244 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 6245 | return vLengthSq; |
||
| 6246 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6247 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6248 | } |
||
| 6249 | |||
| 6250 | //------------------------------------------------------------------------------ |
||
| 6251 | |||
| 6252 | XMFINLINE XMVECTOR XMVector2ReciprocalLength |
||
| 6253 | ( |
||
| 6254 | FXMVECTOR V |
||
| 6255 | ) |
||
| 6256 | { |
||
| 6257 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6258 | |||
| 6259 | XMVECTOR Result; |
||
| 6260 | |||
| 6261 | Result = XMVector2LengthSq(V); |
||
| 6262 | Result = XMVectorReciprocalSqrt(Result); |
||
| 6263 | |||
| 6264 | return Result; |
||
| 6265 | |||
| 6266 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6267 | // Perform the dot product on x and y |
||
| 6268 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 6269 | // vTemp has y splatted |
||
| 6270 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,1,1,1)); |
||
| 6271 | // x+y |
||
| 6272 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 6273 | vLengthSq = _mm_sqrt_ss(vLengthSq); |
||
| 6274 | vLengthSq = _mm_div_ss(g_XMOne,vLengthSq); |
||
| 6275 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 6276 | return vLengthSq; |
||
| 6277 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6278 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6279 | } |
||
| 6280 | |||
| 6281 | //------------------------------------------------------------------------------ |
||
| 6282 | |||
| 6283 | XMFINLINE XMVECTOR XMVector2LengthEst |
||
| 6284 | ( |
||
| 6285 | FXMVECTOR V |
||
| 6286 | ) |
||
| 6287 | { |
||
| 6288 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6289 | XMVECTOR Result; |
||
| 6290 | Result = XMVector2LengthSq(V); |
||
| 6291 | Result = XMVectorSqrtEst(Result); |
||
| 6292 | return Result; |
||
| 6293 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6294 | // Perform the dot product on x and y |
||
| 6295 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 6296 | // vTemp has y splatted |
||
| 6297 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,1,1,1)); |
||
| 6298 | // x+y |
||
| 6299 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 6300 | vLengthSq = _mm_sqrt_ss(vLengthSq); |
||
| 6301 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 6302 | return vLengthSq; |
||
| 6303 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6304 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6305 | } |
||
| 6306 | |||
| 6307 | //------------------------------------------------------------------------------ |
||
| 6308 | |||
| 6309 | XMFINLINE XMVECTOR XMVector2Length |
||
| 6310 | ( |
||
| 6311 | FXMVECTOR V |
||
| 6312 | ) |
||
| 6313 | { |
||
| 6314 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6315 | |||
| 6316 | XMVECTOR Result; |
||
| 6317 | Result = XMVector2LengthSq(V); |
||
| 6318 | Result = XMVectorSqrt(Result); |
||
| 6319 | return Result; |
||
| 6320 | |||
| 6321 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6322 | // Perform the dot product on x and y |
||
| 6323 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 6324 | // vTemp has y splatted |
||
| 6325 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,1,1,1)); |
||
| 6326 | // x+y |
||
| 6327 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 6328 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 6329 | vLengthSq = _mm_sqrt_ps(vLengthSq); |
||
| 6330 | return vLengthSq; |
||
| 6331 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6332 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6333 | } |
||
| 6334 | |||
| 6335 | //------------------------------------------------------------------------------ |
||
| 6336 | // XMVector2NormalizeEst uses a reciprocal estimate and |
||
| 6337 | // returns QNaN on zero and infinite vectors. |
||
| 6338 | |||
| 6339 | XMFINLINE XMVECTOR XMVector2NormalizeEst |
||
| 6340 | ( |
||
| 6341 | FXMVECTOR V |
||
| 6342 | ) |
||
| 6343 | { |
||
| 6344 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6345 | |||
| 6346 | XMVECTOR Result; |
||
| 6347 | Result = XMVector2ReciprocalLength(V); |
||
| 6348 | Result = XMVectorMultiply(V, Result); |
||
| 6349 | return Result; |
||
| 6350 | |||
| 6351 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6352 | // Perform the dot product on x and y |
||
| 6353 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 6354 | // vTemp has y splatted |
||
| 6355 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,1,1,1)); |
||
| 6356 | // x+y |
||
| 6357 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 6358 | vLengthSq = _mm_rsqrt_ss(vLengthSq); |
||
| 6359 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 6360 | vLengthSq = _mm_mul_ps(vLengthSq,V); |
||
| 6361 | return vLengthSq; |
||
| 6362 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6363 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6364 | } |
||
| 6365 | |||
| 6366 | //------------------------------------------------------------------------------ |
||
| 6367 | |||
| 6368 | XMFINLINE XMVECTOR XMVector2Normalize |
||
| 6369 | ( |
||
| 6370 | FXMVECTOR V |
||
| 6371 | ) |
||
| 6372 | { |
||
| 6373 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6374 | |||
| 6375 | XMVECTOR LengthSq; |
||
| 6376 | XMVECTOR Zero; |
||
| 6377 | XMVECTOR InfiniteLength; |
||
| 6378 | XMVECTOR ZeroLength; |
||
| 6379 | XMVECTOR Select; |
||
| 6380 | XMVECTOR Result; |
||
| 6381 | |||
| 6382 | LengthSq = XMVector2LengthSq(V); |
||
| 6383 | Zero = XMVectorZero(); |
||
| 6384 | Result = XMVectorReciprocalSqrt(LengthSq); |
||
| 6385 | InfiniteLength = XMVectorEqualInt(LengthSq, g_XMInfinity.v); |
||
| 6386 | ZeroLength = XMVectorEqual(LengthSq, Zero); |
||
| 6387 | Result = XMVectorMultiply(V, Result); |
||
| 6388 | Select = XMVectorEqualInt(InfiniteLength, ZeroLength); |
||
| 6389 | Result = XMVectorSelect(LengthSq, Result, Select); |
||
| 6390 | |||
| 6391 | return Result; |
||
| 6392 | |||
| 6393 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6394 | // Perform the dot product on x and y only |
||
| 6395 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 6396 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,1,1,1)); |
||
| 6397 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 6398 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 6399 | // Prepare for the division |
||
| 6400 | XMVECTOR vResult = _mm_sqrt_ps(vLengthSq); |
||
| 6401 | // Failsafe on zero (Or epsilon) length planes |
||
| 6402 | // If the length is infinity, set the elements to zero |
||
| 6403 | vLengthSq = _mm_cmpneq_ps(vLengthSq,g_XMInfinity); |
||
| 6404 | // Reciprocal mul to perform the normalization |
||
| 6405 | vResult = _mm_div_ps(V,vResult); |
||
| 6406 | // Any that are infinity, set to zero |
||
| 6407 | vResult = _mm_and_ps(vResult,vLengthSq); |
||
| 6408 | return vResult; |
||
| 6409 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6410 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6411 | } |
||
| 6412 | |||
| 6413 | //------------------------------------------------------------------------------ |
||
| 6414 | |||
| 6415 | XMFINLINE XMVECTOR XMVector2ClampLength |
||
| 6416 | ( |
||
| 6417 | FXMVECTOR V, |
||
| 6418 | FLOAT LengthMin, |
||
| 6419 | FLOAT LengthMax |
||
| 6420 | ) |
||
| 6421 | { |
||
| 6422 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6423 | |||
| 6424 | XMVECTOR ClampMax; |
||
| 6425 | XMVECTOR ClampMin; |
||
| 6426 | |||
| 6427 | ClampMax = XMVectorReplicate(LengthMax); |
||
| 6428 | ClampMin = XMVectorReplicate(LengthMin); |
||
| 6429 | |||
| 6430 | return XMVector2ClampLengthV(V, ClampMin, ClampMax); |
||
| 6431 | |||
| 6432 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6433 | XMVECTOR ClampMax = _mm_set_ps1(LengthMax); |
||
| 6434 | XMVECTOR ClampMin = _mm_set_ps1(LengthMin); |
||
| 6435 | return XMVector2ClampLengthV(V, ClampMin, ClampMax); |
||
| 6436 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 6437 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6438 | } |
||
| 6439 | |||
| 6440 | //------------------------------------------------------------------------------ |
||
| 6441 | |||
| 6442 | XMFINLINE XMVECTOR XMVector2ClampLengthV |
||
| 6443 | ( |
||
| 6444 | FXMVECTOR V, |
||
| 6445 | FXMVECTOR LengthMin, |
||
| 6446 | FXMVECTOR LengthMax |
||
| 6447 | ) |
||
| 6448 | { |
||
| 6449 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6450 | |||
| 6451 | XMVECTOR ClampLength; |
||
| 6452 | XMVECTOR LengthSq; |
||
| 6453 | XMVECTOR RcpLength; |
||
| 6454 | XMVECTOR Length; |
||
| 6455 | XMVECTOR Normal; |
||
| 6456 | XMVECTOR Zero; |
||
| 6457 | XMVECTOR InfiniteLength; |
||
| 6458 | XMVECTOR ZeroLength; |
||
| 6459 | XMVECTOR Select; |
||
| 6460 | XMVECTOR ControlMax; |
||
| 6461 | XMVECTOR ControlMin; |
||
| 6462 | XMVECTOR Control; |
||
| 6463 | XMVECTOR Result; |
||
| 6464 | |||
| 6465 | XMASSERT((LengthMin.vector4_f32[1] == LengthMin.vector4_f32[0])); |
||
| 6466 | XMASSERT((LengthMax.vector4_f32[1] == LengthMax.vector4_f32[0])); |
||
| 6467 | XMASSERT(XMVector2GreaterOrEqual(LengthMin, XMVectorZero())); |
||
| 6468 | XMASSERT(XMVector2GreaterOrEqual(LengthMax, XMVectorZero())); |
||
| 6469 | XMASSERT(XMVector2GreaterOrEqual(LengthMax, LengthMin)); |
||
| 6470 | |||
| 6471 | LengthSq = XMVector2LengthSq(V); |
||
| 6472 | |||
| 6473 | Zero = XMVectorZero(); |
||
| 6474 | |||
| 6475 | RcpLength = XMVectorReciprocalSqrt(LengthSq); |
||
| 6476 | |||
| 6477 | InfiniteLength = XMVectorEqualInt(LengthSq, g_XMInfinity.v); |
||
| 6478 | ZeroLength = XMVectorEqual(LengthSq, Zero); |
||
| 6479 | |||
| 6480 | Length = XMVectorMultiply(LengthSq, RcpLength); |
||
| 6481 | |||
| 6482 | Normal = XMVectorMultiply(V, RcpLength); |
||
| 6483 | |||
| 6484 | Select = XMVectorEqualInt(InfiniteLength, ZeroLength); |
||
| 6485 | Length = XMVectorSelect(LengthSq, Length, Select); |
||
| 6486 | Normal = XMVectorSelect(LengthSq, Normal, Select); |
||
| 6487 | |||
| 6488 | ControlMax = XMVectorGreater(Length, LengthMax); |
||
| 6489 | ControlMin = XMVectorLess(Length, LengthMin); |
||
| 6490 | |||
| 6491 | ClampLength = XMVectorSelect(Length, LengthMax, ControlMax); |
||
| 6492 | ClampLength = XMVectorSelect(ClampLength, LengthMin, ControlMin); |
||
| 6493 | |||
| 6494 | Result = XMVectorMultiply(Normal, ClampLength); |
||
| 6495 | |||
| 6496 | // Preserve the original vector (with no precision loss) if the length falls within the given range |
||
| 6497 | Control = XMVectorEqualInt(ControlMax, ControlMin); |
||
| 6498 | Result = XMVectorSelect(Result, V, Control); |
||
| 6499 | |||
| 6500 | return Result; |
||
| 6501 | |||
| 6502 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6503 | XMVECTOR ClampLength; |
||
| 6504 | XMVECTOR LengthSq; |
||
| 6505 | XMVECTOR RcpLength; |
||
| 6506 | XMVECTOR Length; |
||
| 6507 | XMVECTOR Normal; |
||
| 6508 | XMVECTOR InfiniteLength; |
||
| 6509 | XMVECTOR ZeroLength; |
||
| 6510 | XMVECTOR Select; |
||
| 6511 | XMVECTOR ControlMax; |
||
| 6512 | XMVECTOR ControlMin; |
||
| 6513 | XMVECTOR Control; |
||
| 6514 | XMVECTOR Result; |
||
| 6515 | |||
| 6516 | XMASSERT((XMVectorGetY(LengthMin) == XMVectorGetX(LengthMin))); |
||
| 6517 | XMASSERT((XMVectorGetY(LengthMax) == XMVectorGetX(LengthMax))); |
||
| 6518 | XMASSERT(XMVector2GreaterOrEqual(LengthMin, g_XMZero)); |
||
| 6519 | XMASSERT(XMVector2GreaterOrEqual(LengthMax, g_XMZero)); |
||
| 6520 | XMASSERT(XMVector2GreaterOrEqual(LengthMax, LengthMin)); |
||
| 6521 | LengthSq = XMVector2LengthSq(V); |
||
| 6522 | RcpLength = XMVectorReciprocalSqrt(LengthSq); |
||
| 6523 | InfiniteLength = XMVectorEqualInt(LengthSq, g_XMInfinity); |
||
| 6524 | ZeroLength = XMVectorEqual(LengthSq, g_XMZero); |
||
| 6525 | Length = _mm_mul_ps(LengthSq, RcpLength); |
||
| 6526 | Normal = _mm_mul_ps(V, RcpLength); |
||
| 6527 | Select = XMVectorEqualInt(InfiniteLength, ZeroLength); |
||
| 6528 | Length = XMVectorSelect(LengthSq, Length, Select); |
||
| 6529 | Normal = XMVectorSelect(LengthSq, Normal, Select); |
||
| 6530 | ControlMax = XMVectorGreater(Length, LengthMax); |
||
| 6531 | ControlMin = XMVectorLess(Length, LengthMin); |
||
| 6532 | ClampLength = XMVectorSelect(Length, LengthMax, ControlMax); |
||
| 6533 | ClampLength = XMVectorSelect(ClampLength, LengthMin, ControlMin); |
||
| 6534 | Result = _mm_mul_ps(Normal, ClampLength); |
||
| 6535 | // Preserve the original vector (with no precision loss) if the length falls within the given range |
||
| 6536 | Control = XMVectorEqualInt(ControlMax, ControlMin); |
||
| 6537 | Result = XMVectorSelect(Result, V, Control); |
||
| 6538 | return Result; |
||
| 6539 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6540 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6541 | } |
||
| 6542 | |||
| 6543 | //------------------------------------------------------------------------------ |
||
| 6544 | |||
| 6545 | XMFINLINE XMVECTOR XMVector2Reflect |
||
| 6546 | ( |
||
| 6547 | FXMVECTOR Incident, |
||
| 6548 | FXMVECTOR Normal |
||
| 6549 | ) |
||
| 6550 | { |
||
| 6551 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6552 | |||
| 6553 | XMVECTOR Result; |
||
| 6554 | |||
| 6555 | // Result = Incident - (2 * dot(Incident, Normal)) * Normal |
||
| 6556 | Result = XMVector2Dot(Incident, Normal); |
||
| 6557 | Result = XMVectorAdd(Result, Result); |
||
| 6558 | Result = XMVectorNegativeMultiplySubtract(Result, Normal, Incident); |
||
| 6559 | |||
| 6560 | return Result; |
||
| 6561 | |||
| 6562 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6563 | // Result = Incident - (2 * dot(Incident, Normal)) * Normal |
||
| 6564 | XMVECTOR Result = XMVector2Dot(Incident,Normal); |
||
| 6565 | Result = _mm_add_ps(Result, Result); |
||
| 6566 | Result = _mm_mul_ps(Result, Normal); |
||
| 6567 | Result = _mm_sub_ps(Incident,Result); |
||
| 6568 | return Result; |
||
| 6569 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6570 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6571 | } |
||
| 6572 | |||
| 6573 | //------------------------------------------------------------------------------ |
||
| 6574 | |||
| 6575 | XMFINLINE XMVECTOR XMVector2Refract |
||
| 6576 | ( |
||
| 6577 | FXMVECTOR Incident, |
||
| 6578 | FXMVECTOR Normal, |
||
| 6579 | FLOAT RefractionIndex |
||
| 6580 | ) |
||
| 6581 | { |
||
| 6582 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6583 | XMVECTOR Index; |
||
| 6584 | Index = XMVectorReplicate(RefractionIndex); |
||
| 6585 | return XMVector2RefractV(Incident, Normal, Index); |
||
| 6586 | |||
| 6587 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6588 | XMVECTOR Index = _mm_set_ps1(RefractionIndex); |
||
| 6589 | return XMVector2RefractV(Incident,Normal,Index); |
||
| 6590 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 6591 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6592 | } |
||
| 6593 | |||
| 6594 | //------------------------------------------------------------------------------ |
||
| 6595 | |||
| 6596 | // Return the refraction of a 2D vector |
||
| 6597 | XMFINLINE XMVECTOR XMVector2RefractV |
||
| 6598 | ( |
||
| 6599 | FXMVECTOR Incident, |
||
| 6600 | FXMVECTOR Normal, |
||
| 6601 | FXMVECTOR RefractionIndex |
||
| 6602 | ) |
||
| 6603 | { |
||
| 6604 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6605 | float IDotN; |
||
| 6606 | float RX,RY; |
||
| 6607 | XMVECTOR vResult; |
||
| 6608 | // Result = RefractionIndex * Incident - Normal * (RefractionIndex * dot(Incident, Normal) + |
||
| 6609 | // sqrt(1 - RefractionIndex * RefractionIndex * (1 - dot(Incident, Normal) * dot(Incident, Normal)))) |
||
| 6610 | IDotN = (Incident.vector4_f32[0]*Normal.vector4_f32[0])+(Incident.vector4_f32[1]*Normal.vector4_f32[1]); |
||
| 6611 | // R = 1.0f - RefractionIndex * RefractionIndex * (1.0f - IDotN * IDotN) |
||
| 6612 | RY = 1.0f-(IDotN*IDotN); |
||
| 6613 | RX = 1.0f-(RY*RefractionIndex.vector4_f32[0]*RefractionIndex.vector4_f32[0]); |
||
| 6614 | RY = 1.0f-(RY*RefractionIndex.vector4_f32[1]*RefractionIndex.vector4_f32[1]); |
||
| 6615 | if (RX>=0.0f) { |
||
| 6616 | RX = (RefractionIndex.vector4_f32[0]*Incident.vector4_f32[0])-(Normal.vector4_f32[0]*((RefractionIndex.vector4_f32[0]*IDotN)+sqrtf(RX))); |
||
| 6617 | } else { |
||
| 6618 | RX = 0.0f; |
||
| 6619 | } |
||
| 6620 | if (RY>=0.0f) { |
||
| 6621 | RY = (RefractionIndex.vector4_f32[1]*Incident.vector4_f32[1])-(Normal.vector4_f32[1]*((RefractionIndex.vector4_f32[1]*IDotN)+sqrtf(RY))); |
||
| 6622 | } else { |
||
| 6623 | RY = 0.0f; |
||
| 6624 | } |
||
| 6625 | vResult.vector4_f32[0] = RX; |
||
| 6626 | vResult.vector4_f32[1] = RY; |
||
| 6627 | vResult.vector4_f32[2] = 0.0f; |
||
| 6628 | vResult.vector4_f32[3] = 0.0f; |
||
| 6629 | return vResult; |
||
| 6630 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6631 | // Result = RefractionIndex * Incident - Normal * (RefractionIndex * dot(Incident, Normal) + |
||
| 6632 | // sqrt(1 - RefractionIndex * RefractionIndex * (1 - dot(Incident, Normal) * dot(Incident, Normal)))) |
||
| 6633 | // Get the 2D Dot product of Incident-Normal |
||
| 6634 | XMVECTOR IDotN = _mm_mul_ps(Incident,Normal); |
||
| 6635 | XMVECTOR vTemp = _mm_shuffle_ps(IDotN,IDotN,_MM_SHUFFLE(1,1,1,1)); |
||
| 6636 | IDotN = _mm_add_ss(IDotN,vTemp); |
||
| 6637 | IDotN = _mm_shuffle_ps(IDotN,IDotN,_MM_SHUFFLE(0,0,0,0)); |
||
| 6638 | // vTemp = 1.0f - RefractionIndex * RefractionIndex * (1.0f - IDotN * IDotN) |
||
| 6639 | vTemp = _mm_mul_ps(IDotN,IDotN); |
||
| 6640 | vTemp = _mm_sub_ps(g_XMOne,vTemp); |
||
| 6641 | vTemp = _mm_mul_ps(vTemp,RefractionIndex); |
||
| 6642 | vTemp = _mm_mul_ps(vTemp,RefractionIndex); |
||
| 6643 | vTemp = _mm_sub_ps(g_XMOne,vTemp); |
||
| 6644 | // If any terms are <=0, sqrt() will fail, punt to zero |
||
| 6645 | XMVECTOR vMask = _mm_cmpgt_ps(vTemp,g_XMZero); |
||
| 6646 | // R = RefractionIndex * IDotN + sqrt(R) |
||
| 6647 | vTemp = _mm_sqrt_ps(vTemp); |
||
| 6648 | XMVECTOR vResult = _mm_mul_ps(RefractionIndex,IDotN); |
||
| 6649 | vTemp = _mm_add_ps(vTemp,vResult); |
||
| 6650 | // Result = RefractionIndex * Incident - Normal * R |
||
| 6651 | vResult = _mm_mul_ps(RefractionIndex,Incident); |
||
| 6652 | vTemp = _mm_mul_ps(vTemp,Normal); |
||
| 6653 | vResult = _mm_sub_ps(vResult,vTemp); |
||
| 6654 | vResult = _mm_and_ps(vResult,vMask); |
||
| 6655 | return vResult; |
||
| 6656 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6657 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6658 | } |
||
| 6659 | |||
| 6660 | //------------------------------------------------------------------------------ |
||
| 6661 | |||
| 6662 | XMFINLINE XMVECTOR XMVector2Orthogonal |
||
| 6663 | ( |
||
| 6664 | FXMVECTOR V |
||
| 6665 | ) |
||
| 6666 | { |
||
| 6667 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6668 | |||
| 6669 | XMVECTOR Result; |
||
| 6670 | |||
| 6671 | Result.vector4_f32[0] = -V.vector4_f32[1]; |
||
| 6672 | Result.vector4_f32[1] = V.vector4_f32[0]; |
||
| 6673 | |||
| 6674 | return Result; |
||
| 6675 | |||
| 6676 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6677 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,2,0,1)); |
||
| 6678 | vResult = _mm_mul_ps(vResult,g_XMNegateX); |
||
| 6679 | return vResult; |
||
| 6680 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6681 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6682 | } |
||
| 6683 | |||
| 6684 | //------------------------------------------------------------------------------ |
||
| 6685 | |||
| 6686 | XMFINLINE XMVECTOR XMVector2AngleBetweenNormalsEst |
||
| 6687 | ( |
||
| 6688 | FXMVECTOR N1, |
||
| 6689 | FXMVECTOR N2 |
||
| 6690 | ) |
||
| 6691 | { |
||
| 6692 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6693 | |||
| 6694 | XMVECTOR NegativeOne; |
||
| 6695 | XMVECTOR One; |
||
| 6696 | XMVECTOR Result; |
||
| 6697 | |||
| 6698 | Result = XMVector2Dot(N1, N2); |
||
| 6699 | NegativeOne = XMVectorSplatConstant(-1, 0); |
||
| 6700 | One = XMVectorSplatOne(); |
||
| 6701 | Result = XMVectorClamp(Result, NegativeOne, One); |
||
| 6702 | Result = XMVectorACosEst(Result); |
||
| 6703 | |||
| 6704 | return Result; |
||
| 6705 | |||
| 6706 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6707 | XMVECTOR vResult = XMVector2Dot(N1,N2); |
||
| 6708 | // Clamp to -1.0f to 1.0f |
||
| 6709 | vResult = _mm_max_ps(vResult,g_XMNegativeOne); |
||
| 6710 | vResult = _mm_min_ps(vResult,g_XMOne);; |
||
| 6711 | vResult = XMVectorACosEst(vResult); |
||
| 6712 | return vResult; |
||
| 6713 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6714 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6715 | } |
||
| 6716 | |||
| 6717 | //------------------------------------------------------------------------------ |
||
| 6718 | |||
| 6719 | XMFINLINE XMVECTOR XMVector2AngleBetweenNormals |
||
| 6720 | ( |
||
| 6721 | FXMVECTOR N1, |
||
| 6722 | FXMVECTOR N2 |
||
| 6723 | ) |
||
| 6724 | { |
||
| 6725 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6726 | |||
| 6727 | XMVECTOR NegativeOne; |
||
| 6728 | XMVECTOR One; |
||
| 6729 | XMVECTOR Result; |
||
| 6730 | |||
| 6731 | Result = XMVector2Dot(N1, N2); |
||
| 6732 | NegativeOne = XMVectorSplatConstant(-1, 0); |
||
| 6733 | One = XMVectorSplatOne(); |
||
| 6734 | Result = XMVectorClamp(Result, NegativeOne, One); |
||
| 6735 | Result = XMVectorACos(Result); |
||
| 6736 | |||
| 6737 | return Result; |
||
| 6738 | |||
| 6739 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6740 | XMVECTOR vResult = XMVector2Dot(N1,N2); |
||
| 6741 | // Clamp to -1.0f to 1.0f |
||
| 6742 | vResult = _mm_max_ps(vResult,g_XMNegativeOne); |
||
| 6743 | vResult = _mm_min_ps(vResult,g_XMOne);; |
||
| 6744 | vResult = XMVectorACos(vResult); |
||
| 6745 | return vResult; |
||
| 6746 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6747 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6748 | } |
||
| 6749 | |||
| 6750 | //------------------------------------------------------------------------------ |
||
| 6751 | |||
| 6752 | XMFINLINE XMVECTOR XMVector2AngleBetweenVectors |
||
| 6753 | ( |
||
| 6754 | FXMVECTOR V1, |
||
| 6755 | FXMVECTOR V2 |
||
| 6756 | ) |
||
| 6757 | { |
||
| 6758 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6759 | |||
| 6760 | XMVECTOR L1; |
||
| 6761 | XMVECTOR L2; |
||
| 6762 | XMVECTOR Dot; |
||
| 6763 | XMVECTOR CosAngle; |
||
| 6764 | XMVECTOR NegativeOne; |
||
| 6765 | XMVECTOR One; |
||
| 6766 | XMVECTOR Result; |
||
| 6767 | |||
| 6768 | L1 = XMVector2ReciprocalLength(V1); |
||
| 6769 | L2 = XMVector2ReciprocalLength(V2); |
||
| 6770 | |||
| 6771 | Dot = XMVector2Dot(V1, V2); |
||
| 6772 | |||
| 6773 | L1 = XMVectorMultiply(L1, L2); |
||
| 6774 | |||
| 6775 | CosAngle = XMVectorMultiply(Dot, L1); |
||
| 6776 | NegativeOne = XMVectorSplatConstant(-1, 0); |
||
| 6777 | One = XMVectorSplatOne(); |
||
| 6778 | CosAngle = XMVectorClamp(CosAngle, NegativeOne, One); |
||
| 6779 | |||
| 6780 | Result = XMVectorACos(CosAngle); |
||
| 6781 | |||
| 6782 | return Result; |
||
| 6783 | |||
| 6784 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6785 | XMVECTOR L1; |
||
| 6786 | XMVECTOR L2; |
||
| 6787 | XMVECTOR Dot; |
||
| 6788 | XMVECTOR CosAngle; |
||
| 6789 | XMVECTOR Result; |
||
| 6790 | L1 = XMVector2ReciprocalLength(V1); |
||
| 6791 | L2 = XMVector2ReciprocalLength(V2); |
||
| 6792 | Dot = XMVector2Dot(V1, V2); |
||
| 6793 | L1 = _mm_mul_ps(L1, L2); |
||
| 6794 | CosAngle = _mm_mul_ps(Dot, L1); |
||
| 6795 | CosAngle = XMVectorClamp(CosAngle, g_XMNegativeOne,g_XMOne); |
||
| 6796 | Result = XMVectorACos(CosAngle); |
||
| 6797 | return Result; |
||
| 6798 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6799 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6800 | } |
||
| 6801 | |||
| 6802 | //------------------------------------------------------------------------------ |
||
| 6803 | |||
| 6804 | XMFINLINE XMVECTOR XMVector2LinePointDistance |
||
| 6805 | ( |
||
| 6806 | FXMVECTOR LinePoint1, |
||
| 6807 | FXMVECTOR LinePoint2, |
||
| 6808 | FXMVECTOR Point |
||
| 6809 | ) |
||
| 6810 | { |
||
| 6811 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6812 | |||
| 6813 | XMVECTOR PointVector; |
||
| 6814 | XMVECTOR LineVector; |
||
| 6815 | XMVECTOR ReciprocalLengthSq; |
||
| 6816 | XMVECTOR PointProjectionScale; |
||
| 6817 | XMVECTOR DistanceVector; |
||
| 6818 | XMVECTOR Result; |
||
| 6819 | |||
| 6820 | // Given a vector PointVector from LinePoint1 to Point and a vector |
||
| 6821 | // LineVector from LinePoint1 to LinePoint2, the scaled distance |
||
| 6822 | // PointProjectionScale from LinePoint1 to the perpendicular projection |
||
| 6823 | // of PointVector onto the line is defined as: |
||
| 6824 | // |
||
| 6825 | // PointProjectionScale = dot(PointVector, LineVector) / LengthSq(LineVector) |
||
| 6826 | |||
| 6827 | PointVector = XMVectorSubtract(Point, LinePoint1); |
||
| 6828 | LineVector = XMVectorSubtract(LinePoint2, LinePoint1); |
||
| 6829 | |||
| 6830 | ReciprocalLengthSq = XMVector2LengthSq(LineVector); |
||
| 6831 | ReciprocalLengthSq = XMVectorReciprocal(ReciprocalLengthSq); |
||
| 6832 | |||
| 6833 | PointProjectionScale = XMVector2Dot(PointVector, LineVector); |
||
| 6834 | PointProjectionScale = XMVectorMultiply(PointProjectionScale, ReciprocalLengthSq); |
||
| 6835 | |||
| 6836 | DistanceVector = XMVectorMultiply(LineVector, PointProjectionScale); |
||
| 6837 | DistanceVector = XMVectorSubtract(PointVector, DistanceVector); |
||
| 6838 | |||
| 6839 | Result = XMVector2Length(DistanceVector); |
||
| 6840 | |||
| 6841 | return Result; |
||
| 6842 | |||
| 6843 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6844 | XMVECTOR PointVector = _mm_sub_ps(Point,LinePoint1); |
||
| 6845 | XMVECTOR LineVector = _mm_sub_ps(LinePoint2,LinePoint1); |
||
| 6846 | XMVECTOR ReciprocalLengthSq = XMVector2LengthSq(LineVector); |
||
| 6847 | XMVECTOR vResult = XMVector2Dot(PointVector,LineVector); |
||
| 6848 | vResult = _mm_div_ps(vResult,ReciprocalLengthSq); |
||
| 6849 | vResult = _mm_mul_ps(vResult,LineVector); |
||
| 6850 | vResult = _mm_sub_ps(PointVector,vResult); |
||
| 6851 | vResult = XMVector2Length(vResult); |
||
| 6852 | return vResult; |
||
| 6853 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6854 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6855 | } |
||
| 6856 | |||
| 6857 | //------------------------------------------------------------------------------ |
||
| 6858 | |||
| 6859 | XMFINLINE XMVECTOR XMVector2IntersectLine |
||
| 6860 | ( |
||
| 6861 | FXMVECTOR Line1Point1, |
||
| 6862 | FXMVECTOR Line1Point2, |
||
| 6863 | FXMVECTOR Line2Point1, |
||
| 6864 | CXMVECTOR Line2Point2 |
||
| 6865 | ) |
||
| 6866 | { |
||
| 6867 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6868 | |||
| 6869 | XMVECTOR V1; |
||
| 6870 | XMVECTOR V2; |
||
| 6871 | XMVECTOR V3; |
||
| 6872 | XMVECTOR C1; |
||
| 6873 | XMVECTOR C2; |
||
| 6874 | XMVECTOR Result; |
||
| 6875 | CONST XMVECTOR Zero = XMVectorZero(); |
||
| 6876 | |||
| 6877 | V1 = XMVectorSubtract(Line1Point2, Line1Point1); |
||
| 6878 | V2 = XMVectorSubtract(Line2Point2, Line2Point1); |
||
| 6879 | V3 = XMVectorSubtract(Line1Point1, Line2Point1); |
||
| 6880 | |||
| 6881 | C1 = XMVector2Cross(V1, V2); |
||
| 6882 | C2 = XMVector2Cross(V2, V3); |
||
| 6883 | |||
| 6884 | if (XMVector2NearEqual(C1, Zero, g_XMEpsilon.v)) |
||
| 6885 | { |
||
| 6886 | if (XMVector2NearEqual(C2, Zero, g_XMEpsilon.v)) |
||
| 6887 | { |
||
| 6888 | // Coincident |
||
| 6889 | Result = g_XMInfinity.v; |
||
| 6890 | } |
||
| 6891 | else |
||
| 6892 | { |
||
| 6893 | // Parallel |
||
| 6894 | Result = g_XMQNaN.v; |
||
| 6895 | } |
||
| 6896 | } |
||
| 6897 | else |
||
| 6898 | { |
||
| 6899 | // Intersection point = Line1Point1 + V1 * (C2 / C1) |
||
| 6900 | XMVECTOR Scale; |
||
| 6901 | Scale = XMVectorReciprocal(C1); |
||
| 6902 | Scale = XMVectorMultiply(C2, Scale); |
||
| 6903 | Result = XMVectorMultiplyAdd(V1, Scale, Line1Point1); |
||
| 6904 | } |
||
| 6905 | |||
| 6906 | return Result; |
||
| 6907 | |||
| 6908 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6909 | XMVECTOR V1 = _mm_sub_ps(Line1Point2, Line1Point1); |
||
| 6910 | XMVECTOR V2 = _mm_sub_ps(Line2Point2, Line2Point1); |
||
| 6911 | XMVECTOR V3 = _mm_sub_ps(Line1Point1, Line2Point1); |
||
| 6912 | // Generate the cross products |
||
| 6913 | XMVECTOR C1 = XMVector2Cross(V1, V2); |
||
| 6914 | XMVECTOR C2 = XMVector2Cross(V2, V3); |
||
| 6915 | // If C1 is not close to epsilon, use the calculated value |
||
| 6916 | XMVECTOR vResultMask = _mm_setzero_ps(); |
||
| 6917 | vResultMask = _mm_sub_ps(vResultMask,C1); |
||
| 6918 | vResultMask = _mm_max_ps(vResultMask,C1); |
||
| 6919 | // 0xFFFFFFFF if the calculated value is to be used |
||
| 6920 | vResultMask = _mm_cmpgt_ps(vResultMask,g_XMEpsilon); |
||
| 6921 | // If C1 is close to epsilon, which fail type is it? INFINITY or NAN? |
||
| 6922 | XMVECTOR vFailMask = _mm_setzero_ps(); |
||
| 6923 | vFailMask = _mm_sub_ps(vFailMask,C2); |
||
| 6924 | vFailMask = _mm_max_ps(vFailMask,C2); |
||
| 6925 | vFailMask = _mm_cmple_ps(vFailMask,g_XMEpsilon); |
||
| 6926 | XMVECTOR vFail = _mm_and_ps(vFailMask,g_XMInfinity); |
||
| 6927 | vFailMask = _mm_andnot_ps(vFailMask,g_XMQNaN); |
||
| 6928 | // vFail is NAN or INF |
||
| 6929 | vFail = _mm_or_ps(vFail,vFailMask); |
||
| 6930 | // Intersection point = Line1Point1 + V1 * (C2 / C1) |
||
| 6931 | XMVECTOR vResult = _mm_div_ps(C2,C1); |
||
| 6932 | vResult = _mm_mul_ps(vResult,V1); |
||
| 6933 | vResult = _mm_add_ps(vResult,Line1Point1); |
||
| 6934 | // Use result, or failure value |
||
| 6935 | vResult = _mm_and_ps(vResult,vResultMask); |
||
| 6936 | vResultMask = _mm_andnot_ps(vResultMask,vFail); |
||
| 6937 | vResult = _mm_or_ps(vResult,vResultMask); |
||
| 6938 | return vResult; |
||
| 6939 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6940 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6941 | } |
||
| 6942 | |||
| 6943 | //------------------------------------------------------------------------------ |
||
| 6944 | |||
| 6945 | XMFINLINE XMVECTOR XMVector2Transform |
||
| 6946 | ( |
||
| 6947 | FXMVECTOR V, |
||
| 6948 | CXMMATRIX M |
||
| 6949 | ) |
||
| 6950 | { |
||
| 6951 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6952 | |||
| 6953 | XMVECTOR X; |
||
| 6954 | XMVECTOR Y; |
||
| 6955 | XMVECTOR Result; |
||
| 6956 | |||
| 6957 | Y = XMVectorSplatY(V); |
||
| 6958 | X = XMVectorSplatX(V); |
||
| 6959 | |||
| 6960 | Result = XMVectorMultiplyAdd(Y, M.r[1], M.r[3]); |
||
| 6961 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 6962 | |||
| 6963 | return Result; |
||
| 6964 | |||
| 6965 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 6966 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,0,0,0)); |
||
| 6967 | vResult = _mm_mul_ps(vResult,M.r[0]); |
||
| 6968 | XMVECTOR vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 6969 | vTemp = _mm_mul_ps(vTemp,M.r[1]); |
||
| 6970 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 6971 | vResult = _mm_add_ps(vResult,M.r[3]); |
||
| 6972 | return vResult; |
||
| 6973 | #else // _XM_VMX128_INTRINSICS_ |
||
| 6974 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 6975 | } |
||
| 6976 | |||
| 6977 | //------------------------------------------------------------------------------ |
||
| 6978 | |||
| 6979 | XMINLINE XMFLOAT4* XMVector2TransformStream |
||
| 6980 | ( |
||
| 6981 | XMFLOAT4* pOutputStream, |
||
| 6982 | UINT OutputStride, |
||
| 6983 | CONST XMFLOAT2* pInputStream, |
||
| 6984 | UINT InputStride, |
||
| 6985 | UINT VectorCount, |
||
| 6986 | CXMMATRIX M |
||
| 6987 | ) |
||
| 6988 | { |
||
| 6989 | #if defined(_XM_NO_INTRINSICS_) |
||
| 6990 | |||
| 6991 | XMVECTOR V; |
||
| 6992 | XMVECTOR X; |
||
| 6993 | XMVECTOR Y; |
||
| 6994 | XMVECTOR Result; |
||
| 6995 | UINT i; |
||
| 6996 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 6997 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 6998 | |||
| 6999 | XMASSERT(pOutputStream); |
||
| 7000 | XMASSERT(pInputStream); |
||
| 7001 | |||
| 7002 | for (i = 0; i < VectorCount; i++) |
||
| 7003 | { |
||
| 7004 | V = XMLoadFloat2((XMFLOAT2*)pInputVector); |
||
| 7005 | Y = XMVectorSplatY(V); |
||
| 7006 | X = XMVectorSplatX(V); |
||
| 7007 | // Y = XMVectorReplicate(((XMFLOAT2*)pInputVector)->y); |
||
| 7008 | // X = XMVectorReplicate(((XMFLOAT2*)pInputVector)->x); |
||
| 7009 | |||
| 7010 | Result = XMVectorMultiplyAdd(Y, M.r[1], M.r[3]); |
||
| 7011 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 7012 | |||
| 7013 | XMStoreFloat4((XMFLOAT4*)pOutputVector, Result); |
||
| 7014 | |||
| 7015 | pInputVector += InputStride; |
||
| 7016 | pOutputVector += OutputStride; |
||
| 7017 | } |
||
| 7018 | |||
| 7019 | return pOutputStream; |
||
| 7020 | |||
| 7021 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7022 | XMASSERT(pOutputStream); |
||
| 7023 | XMASSERT(pInputStream); |
||
| 7024 | UINT i; |
||
| 7025 | const BYTE* pInputVector = (const BYTE*)pInputStream; |
||
| 7026 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 7027 | |||
| 7028 | for (i = 0; i < VectorCount; i++) |
||
| 7029 | { |
||
| 7030 | XMVECTOR X = _mm_load_ps1(&reinterpret_cast<const XMFLOAT2*>(pInputVector)->x); |
||
| 7031 | XMVECTOR vResult = _mm_load_ps1(&reinterpret_cast<const XMFLOAT2*>(pInputVector)->y); |
||
| 7032 | vResult = _mm_mul_ps(vResult,M.r[1]); |
||
| 7033 | vResult = _mm_add_ps(vResult,M.r[3]); |
||
| 7034 | X = _mm_mul_ps(X,M.r[0]); |
||
| 7035 | vResult = _mm_add_ps(vResult,X); |
||
| 7036 | _mm_storeu_ps(reinterpret_cast<float*>(pOutputVector),vResult); |
||
| 7037 | pInputVector += InputStride; |
||
| 7038 | pOutputVector += OutputStride; |
||
| 7039 | } |
||
| 7040 | return pOutputStream; |
||
| 7041 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 7042 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7043 | } |
||
| 7044 | |||
| 7045 | //------------------------------------------------------------------------------ |
||
| 7046 | |||
| 7047 | XMINLINE XMFLOAT4* XMVector2TransformStreamNC |
||
| 7048 | ( |
||
| 7049 | XMFLOAT4* pOutputStream, |
||
| 7050 | UINT OutputStride, |
||
| 7051 | CONST XMFLOAT2* pInputStream, |
||
| 7052 | UINT InputStride, |
||
| 7053 | UINT VectorCount, |
||
| 7054 | CXMMATRIX M |
||
| 7055 | ) |
||
| 7056 | { |
||
| 7057 | #if defined(_XM_NO_INTRINSICS_) || defined(XM_NO_MISALIGNED_VECTOR_ACCESS) || defined(_XM_SSE_INTRINSICS_) |
||
| 7058 | return XMVector2TransformStream( pOutputStream, OutputStride, pInputStream, InputStride, VectorCount, M ); |
||
| 7059 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7060 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7061 | } |
||
| 7062 | |||
| 7063 | //------------------------------------------------------------------------------ |
||
| 7064 | |||
| 7065 | XMFINLINE XMVECTOR XMVector2TransformCoord |
||
| 7066 | ( |
||
| 7067 | FXMVECTOR V, |
||
| 7068 | CXMMATRIX M |
||
| 7069 | ) |
||
| 7070 | { |
||
| 7071 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7072 | |||
| 7073 | XMVECTOR X; |
||
| 7074 | XMVECTOR Y; |
||
| 7075 | XMVECTOR InverseW; |
||
| 7076 | XMVECTOR Result; |
||
| 7077 | |||
| 7078 | Y = XMVectorSplatY(V); |
||
| 7079 | X = XMVectorSplatX(V); |
||
| 7080 | |||
| 7081 | Result = XMVectorMultiplyAdd(Y, M.r[1], M.r[3]); |
||
| 7082 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 7083 | |||
| 7084 | InverseW = XMVectorSplatW(Result); |
||
| 7085 | InverseW = XMVectorReciprocal(InverseW); |
||
| 7086 | |||
| 7087 | Result = XMVectorMultiply(Result, InverseW); |
||
| 7088 | |||
| 7089 | return Result; |
||
| 7090 | |||
| 7091 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7092 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,0,0,0)); |
||
| 7093 | vResult = _mm_mul_ps(vResult,M.r[0]); |
||
| 7094 | XMVECTOR vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 7095 | vTemp = _mm_mul_ps(vTemp,M.r[1]); |
||
| 7096 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 7097 | vResult = _mm_add_ps(vResult,M.r[3]); |
||
| 7098 | vTemp = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,3,3,3)); |
||
| 7099 | vResult = _mm_div_ps(vResult,vTemp); |
||
| 7100 | return vResult; |
||
| 7101 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7102 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7103 | } |
||
| 7104 | |||
| 7105 | //------------------------------------------------------------------------------ |
||
| 7106 | |||
| 7107 | XMINLINE XMFLOAT2* XMVector2TransformCoordStream |
||
| 7108 | ( |
||
| 7109 | XMFLOAT2* pOutputStream, |
||
| 7110 | UINT OutputStride, |
||
| 7111 | CONST XMFLOAT2* pInputStream, |
||
| 7112 | UINT InputStride, |
||
| 7113 | UINT VectorCount, |
||
| 7114 | CXMMATRIX M |
||
| 7115 | ) |
||
| 7116 | { |
||
| 7117 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7118 | |||
| 7119 | XMVECTOR V; |
||
| 7120 | XMVECTOR X; |
||
| 7121 | XMVECTOR Y; |
||
| 7122 | XMVECTOR InverseW; |
||
| 7123 | XMVECTOR Result; |
||
| 7124 | UINT i; |
||
| 7125 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 7126 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 7127 | |||
| 7128 | XMASSERT(pOutputStream); |
||
| 7129 | XMASSERT(pInputStream); |
||
| 7130 | |||
| 7131 | for (i = 0; i < VectorCount; i++) |
||
| 7132 | { |
||
| 7133 | V = XMLoadFloat2((XMFLOAT2*)pInputVector); |
||
| 7134 | Y = XMVectorSplatY(V); |
||
| 7135 | X = XMVectorSplatX(V); |
||
| 7136 | // Y = XMVectorReplicate(((XMFLOAT2*)pInputVector)->y); |
||
| 7137 | // X = XMVectorReplicate(((XMFLOAT2*)pInputVector)->x); |
||
| 7138 | |||
| 7139 | Result = XMVectorMultiplyAdd(Y, M.r[1], M.r[3]); |
||
| 7140 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 7141 | |||
| 7142 | InverseW = XMVectorSplatW(Result); |
||
| 7143 | InverseW = XMVectorReciprocal(InverseW); |
||
| 7144 | |||
| 7145 | Result = XMVectorMultiply(Result, InverseW); |
||
| 7146 | |||
| 7147 | XMStoreFloat2((XMFLOAT2*)pOutputVector, Result); |
||
| 7148 | |||
| 7149 | pInputVector += InputStride; |
||
| 7150 | pOutputVector += OutputStride; |
||
| 7151 | } |
||
| 7152 | |||
| 7153 | return pOutputStream; |
||
| 7154 | |||
| 7155 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7156 | XMASSERT(pOutputStream); |
||
| 7157 | XMASSERT(pInputStream); |
||
| 7158 | UINT i; |
||
| 7159 | const BYTE *pInputVector = (BYTE*)pInputStream; |
||
| 7160 | BYTE *pOutputVector = (BYTE*)pOutputStream; |
||
| 7161 | |||
| 7162 | for (i = 0; i < VectorCount; i++) |
||
| 7163 | { |
||
| 7164 | XMVECTOR X = _mm_load_ps1(&reinterpret_cast<const XMFLOAT2*>(pInputVector)->x); |
||
| 7165 | XMVECTOR vResult = _mm_load_ps1(&reinterpret_cast<const XMFLOAT2*>(pInputVector)->y); |
||
| 7166 | vResult = _mm_mul_ps(vResult,M.r[1]); |
||
| 7167 | vResult = _mm_add_ps(vResult,M.r[3]); |
||
| 7168 | X = _mm_mul_ps(X,M.r[0]); |
||
| 7169 | vResult = _mm_add_ps(vResult,X); |
||
| 7170 | X = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,3,3,3)); |
||
| 7171 | vResult = _mm_div_ps(vResult,X); |
||
| 7172 | _mm_store_sd(reinterpret_cast<double *>(pOutputVector),reinterpret_cast<__m128d *>(&vResult)[0]); |
||
| 7173 | pInputVector += InputStride; |
||
| 7174 | pOutputVector += OutputStride; |
||
| 7175 | } |
||
| 7176 | return pOutputStream; |
||
| 7177 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 7178 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7179 | } |
||
| 7180 | |||
| 7181 | //------------------------------------------------------------------------------ |
||
| 7182 | |||
| 7183 | XMFINLINE XMVECTOR XMVector2TransformNormal |
||
| 7184 | ( |
||
| 7185 | FXMVECTOR V, |
||
| 7186 | CXMMATRIX M |
||
| 7187 | ) |
||
| 7188 | { |
||
| 7189 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7190 | |||
| 7191 | XMVECTOR X; |
||
| 7192 | XMVECTOR Y; |
||
| 7193 | XMVECTOR Result; |
||
| 7194 | |||
| 7195 | Y = XMVectorSplatY(V); |
||
| 7196 | X = XMVectorSplatX(V); |
||
| 7197 | |||
| 7198 | Result = XMVectorMultiply(Y, M.r[1]); |
||
| 7199 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 7200 | |||
| 7201 | return Result; |
||
| 7202 | |||
| 7203 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7204 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,0,0,0)); |
||
| 7205 | vResult = _mm_mul_ps(vResult,M.r[0]); |
||
| 7206 | XMVECTOR vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 7207 | vTemp = _mm_mul_ps(vTemp,M.r[1]); |
||
| 7208 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 7209 | return vResult; |
||
| 7210 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7211 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7212 | } |
||
| 7213 | |||
| 7214 | //------------------------------------------------------------------------------ |
||
| 7215 | |||
| 7216 | XMINLINE XMFLOAT2* XMVector2TransformNormalStream |
||
| 7217 | ( |
||
| 7218 | XMFLOAT2* pOutputStream, |
||
| 7219 | UINT OutputStride, |
||
| 7220 | CONST XMFLOAT2* pInputStream, |
||
| 7221 | UINT InputStride, |
||
| 7222 | UINT VectorCount, |
||
| 7223 | CXMMATRIX M |
||
| 7224 | ) |
||
| 7225 | { |
||
| 7226 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7227 | |||
| 7228 | XMVECTOR V; |
||
| 7229 | XMVECTOR X; |
||
| 7230 | XMVECTOR Y; |
||
| 7231 | XMVECTOR Result; |
||
| 7232 | UINT i; |
||
| 7233 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 7234 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 7235 | |||
| 7236 | XMASSERT(pOutputStream); |
||
| 7237 | XMASSERT(pInputStream); |
||
| 7238 | |||
| 7239 | for (i = 0; i < VectorCount; i++) |
||
| 7240 | { |
||
| 7241 | V = XMLoadFloat2((XMFLOAT2*)pInputVector); |
||
| 7242 | Y = XMVectorSplatY(V); |
||
| 7243 | X = XMVectorSplatX(V); |
||
| 7244 | // Y = XMVectorReplicate(((XMFLOAT2*)pInputVector)->y); |
||
| 7245 | // X = XMVectorReplicate(((XMFLOAT2*)pInputVector)->x); |
||
| 7246 | |||
| 7247 | Result = XMVectorMultiply(Y, M.r[1]); |
||
| 7248 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 7249 | |||
| 7250 | XMStoreFloat2((XMFLOAT2*)pOutputVector, Result); |
||
| 7251 | |||
| 7252 | pInputVector += InputStride; |
||
| 7253 | pOutputVector += OutputStride; |
||
| 7254 | } |
||
| 7255 | |||
| 7256 | return pOutputStream; |
||
| 7257 | |||
| 7258 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7259 | XMASSERT(pOutputStream); |
||
| 7260 | XMASSERT(pInputStream); |
||
| 7261 | UINT i; |
||
| 7262 | const BYTE*pInputVector = (const BYTE*)pInputStream; |
||
| 7263 | BYTE *pOutputVector = (BYTE*)pOutputStream; |
||
| 7264 | for (i = 0; i < VectorCount; i++) |
||
| 7265 | { |
||
| 7266 | XMVECTOR X = _mm_load_ps1(&reinterpret_cast<const XMFLOAT2 *>(pInputVector)->x); |
||
| 7267 | XMVECTOR vResult = _mm_load_ps1(&reinterpret_cast<const XMFLOAT2 *>(pInputVector)->y); |
||
| 7268 | vResult = _mm_mul_ps(vResult,M.r[1]); |
||
| 7269 | X = _mm_mul_ps(X,M.r[0]); |
||
| 7270 | vResult = _mm_add_ps(vResult,X); |
||
| 7271 | _mm_store_sd(reinterpret_cast<double*>(pOutputVector),reinterpret_cast<const __m128d *>(&vResult)[0]); |
||
| 7272 | |||
| 7273 | pInputVector += InputStride; |
||
| 7274 | pOutputVector += OutputStride; |
||
| 7275 | } |
||
| 7276 | |||
| 7277 | return pOutputStream; |
||
| 7278 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 7279 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7280 | } |
||
| 7281 | |||
| 7282 | /**************************************************************************** |
||
| 7283 | * |
||
| 7284 | * 3D Vector |
||
| 7285 | * |
||
| 7286 | ****************************************************************************/ |
||
| 7287 | |||
| 7288 | //------------------------------------------------------------------------------ |
||
| 7289 | // Comparison operations |
||
| 7290 | //------------------------------------------------------------------------------ |
||
| 7291 | |||
| 7292 | //------------------------------------------------------------------------------ |
||
| 7293 | |||
| 7294 | XMFINLINE BOOL XMVector3Equal |
||
| 7295 | ( |
||
| 7296 | FXMVECTOR V1, |
||
| 7297 | FXMVECTOR V2 |
||
| 7298 | ) |
||
| 7299 | { |
||
| 7300 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7301 | return (((V1.vector4_f32[0] == V2.vector4_f32[0]) && (V1.vector4_f32[1] == V2.vector4_f32[1]) && (V1.vector4_f32[2] == V2.vector4_f32[2])) != 0); |
||
| 7302 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7303 | XMVECTOR vTemp = _mm_cmpeq_ps(V1,V2); |
||
| 7304 | return (((_mm_movemask_ps(vTemp)&7)==7) != 0); |
||
| 7305 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7306 | return XMComparisonAllTrue(XMVector3EqualR(V1, V2)); |
||
| 7307 | #endif |
||
| 7308 | } |
||
| 7309 | |||
| 7310 | //------------------------------------------------------------------------------ |
||
| 7311 | |||
| 7312 | XMFINLINE UINT XMVector3EqualR |
||
| 7313 | ( |
||
| 7314 | FXMVECTOR V1, |
||
| 7315 | FXMVECTOR V2 |
||
| 7316 | ) |
||
| 7317 | { |
||
| 7318 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7319 | UINT CR = 0; |
||
| 7320 | if ((V1.vector4_f32[0] == V2.vector4_f32[0]) && |
||
| 7321 | (V1.vector4_f32[1] == V2.vector4_f32[1]) && |
||
| 7322 | (V1.vector4_f32[2] == V2.vector4_f32[2])) |
||
| 7323 | { |
||
| 7324 | CR = XM_CRMASK_CR6TRUE; |
||
| 7325 | } |
||
| 7326 | else if ((V1.vector4_f32[0] != V2.vector4_f32[0]) && |
||
| 7327 | (V1.vector4_f32[1] != V2.vector4_f32[1]) && |
||
| 7328 | (V1.vector4_f32[2] != V2.vector4_f32[2])) |
||
| 7329 | { |
||
| 7330 | CR = XM_CRMASK_CR6FALSE; |
||
| 7331 | } |
||
| 7332 | return CR; |
||
| 7333 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7334 | XMVECTOR vTemp = _mm_cmpeq_ps(V1,V2); |
||
| 7335 | int iTest = _mm_movemask_ps(vTemp)&7; |
||
| 7336 | UINT CR = 0; |
||
| 7337 | if (iTest==7) |
||
| 7338 | { |
||
| 7339 | CR = XM_CRMASK_CR6TRUE; |
||
| 7340 | } |
||
| 7341 | else if (!iTest) |
||
| 7342 | { |
||
| 7343 | CR = XM_CRMASK_CR6FALSE; |
||
| 7344 | } |
||
| 7345 | return CR; |
||
| 7346 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7347 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7348 | } |
||
| 7349 | |||
| 7350 | //------------------------------------------------------------------------------ |
||
| 7351 | |||
| 7352 | XMFINLINE BOOL XMVector3EqualInt |
||
| 7353 | ( |
||
| 7354 | FXMVECTOR V1, |
||
| 7355 | FXMVECTOR V2 |
||
| 7356 | ) |
||
| 7357 | { |
||
| 7358 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7359 | return (((V1.vector4_u32[0] == V2.vector4_u32[0]) && (V1.vector4_u32[1] == V2.vector4_u32[1]) && (V1.vector4_u32[2] == V2.vector4_u32[2])) != 0); |
||
| 7360 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7361 | __m128i vTemp = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0]); |
||
| 7362 | return (((_mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTemp)[0])&7)==7) != 0); |
||
| 7363 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7364 | return XMComparisonAllTrue(XMVector3EqualIntR(V1, V2)); |
||
| 7365 | #endif |
||
| 7366 | } |
||
| 7367 | |||
| 7368 | //------------------------------------------------------------------------------ |
||
| 7369 | |||
| 7370 | XMFINLINE UINT XMVector3EqualIntR |
||
| 7371 | ( |
||
| 7372 | FXMVECTOR V1, |
||
| 7373 | FXMVECTOR V2 |
||
| 7374 | ) |
||
| 7375 | { |
||
| 7376 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7377 | UINT CR = 0; |
||
| 7378 | if ((V1.vector4_u32[0] == V2.vector4_u32[0]) && |
||
| 7379 | (V1.vector4_u32[1] == V2.vector4_u32[1]) && |
||
| 7380 | (V1.vector4_u32[2] == V2.vector4_u32[2])) |
||
| 7381 | { |
||
| 7382 | CR = XM_CRMASK_CR6TRUE; |
||
| 7383 | } |
||
| 7384 | else if ((V1.vector4_u32[0] != V2.vector4_u32[0]) && |
||
| 7385 | (V1.vector4_u32[1] != V2.vector4_u32[1]) && |
||
| 7386 | (V1.vector4_u32[2] != V2.vector4_u32[2])) |
||
| 7387 | { |
||
| 7388 | CR = XM_CRMASK_CR6FALSE; |
||
| 7389 | } |
||
| 7390 | return CR; |
||
| 7391 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7392 | __m128i vTemp = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0]); |
||
| 7393 | int iTemp = _mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTemp)[0])&7; |
||
| 7394 | UINT CR = 0; |
||
| 7395 | if (iTemp==7) |
||
| 7396 | { |
||
| 7397 | CR = XM_CRMASK_CR6TRUE; |
||
| 7398 | } |
||
| 7399 | else if (!iTemp) |
||
| 7400 | { |
||
| 7401 | CR = XM_CRMASK_CR6FALSE; |
||
| 7402 | } |
||
| 7403 | return CR; |
||
| 7404 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7405 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7406 | } |
||
| 7407 | |||
| 7408 | //------------------------------------------------------------------------------ |
||
| 7409 | |||
| 7410 | XMFINLINE BOOL XMVector3NearEqual |
||
| 7411 | ( |
||
| 7412 | FXMVECTOR V1, |
||
| 7413 | FXMVECTOR V2, |
||
| 7414 | FXMVECTOR Epsilon |
||
| 7415 | ) |
||
| 7416 | { |
||
| 7417 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7418 | FLOAT dx, dy, dz; |
||
| 7419 | |||
| 7420 | dx = fabsf(V1.vector4_f32[0]-V2.vector4_f32[0]); |
||
| 7421 | dy = fabsf(V1.vector4_f32[1]-V2.vector4_f32[1]); |
||
| 7422 | dz = fabsf(V1.vector4_f32[2]-V2.vector4_f32[2]); |
||
| 7423 | return (((dx <= Epsilon.vector4_f32[0]) && |
||
| 7424 | (dy <= Epsilon.vector4_f32[1]) && |
||
| 7425 | (dz <= Epsilon.vector4_f32[2])) != 0); |
||
| 7426 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7427 | // Get the difference |
||
| 7428 | XMVECTOR vDelta = _mm_sub_ps(V1,V2); |
||
| 7429 | // Get the absolute value of the difference |
||
| 7430 | XMVECTOR vTemp = _mm_setzero_ps(); |
||
| 7431 | vTemp = _mm_sub_ps(vTemp,vDelta); |
||
| 7432 | vTemp = _mm_max_ps(vTemp,vDelta); |
||
| 7433 | vTemp = _mm_cmple_ps(vTemp,Epsilon); |
||
| 7434 | // w is don't care |
||
| 7435 | return (((_mm_movemask_ps(vTemp)&7)==0x7) != 0); |
||
| 7436 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7437 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7438 | } |
||
| 7439 | |||
| 7440 | //------------------------------------------------------------------------------ |
||
| 7441 | |||
| 7442 | XMFINLINE BOOL XMVector3NotEqual |
||
| 7443 | ( |
||
| 7444 | FXMVECTOR V1, |
||
| 7445 | FXMVECTOR V2 |
||
| 7446 | ) |
||
| 7447 | { |
||
| 7448 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7449 | return (((V1.vector4_f32[0] != V2.vector4_f32[0]) || (V1.vector4_f32[1] != V2.vector4_f32[1]) || (V1.vector4_f32[2] != V2.vector4_f32[2])) != 0); |
||
| 7450 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7451 | XMVECTOR vTemp = _mm_cmpeq_ps(V1,V2); |
||
| 7452 | return (((_mm_movemask_ps(vTemp)&7)!=7) != 0); |
||
| 7453 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7454 | return XMComparisonAnyFalse(XMVector3EqualR(V1, V2)); |
||
| 7455 | #endif |
||
| 7456 | } |
||
| 7457 | |||
| 7458 | //------------------------------------------------------------------------------ |
||
| 7459 | |||
| 7460 | XMFINLINE BOOL XMVector3NotEqualInt |
||
| 7461 | ( |
||
| 7462 | FXMVECTOR V1, |
||
| 7463 | FXMVECTOR V2 |
||
| 7464 | ) |
||
| 7465 | { |
||
| 7466 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7467 | return (((V1.vector4_u32[0] != V2.vector4_u32[0]) || (V1.vector4_u32[1] != V2.vector4_u32[1]) || (V1.vector4_u32[2] != V2.vector4_u32[2])) != 0); |
||
| 7468 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7469 | __m128i vTemp = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0]); |
||
| 7470 | return (((_mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTemp)[0])&7)!=7) != 0); |
||
| 7471 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7472 | return XMComparisonAnyFalse(XMVector3EqualIntR(V1, V2)); |
||
| 7473 | #endif |
||
| 7474 | } |
||
| 7475 | |||
| 7476 | //------------------------------------------------------------------------------ |
||
| 7477 | |||
| 7478 | XMFINLINE BOOL XMVector3Greater |
||
| 7479 | ( |
||
| 7480 | FXMVECTOR V1, |
||
| 7481 | FXMVECTOR V2 |
||
| 7482 | ) |
||
| 7483 | { |
||
| 7484 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7485 | return (((V1.vector4_f32[0] > V2.vector4_f32[0]) && (V1.vector4_f32[1] > V2.vector4_f32[1]) && (V1.vector4_f32[2] > V2.vector4_f32[2])) != 0); |
||
| 7486 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7487 | XMVECTOR vTemp = _mm_cmpgt_ps(V1,V2); |
||
| 7488 | return (((_mm_movemask_ps(vTemp)&7)==7) != 0); |
||
| 7489 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7490 | return XMComparisonAllTrue(XMVector3GreaterR(V1, V2)); |
||
| 7491 | #endif |
||
| 7492 | } |
||
| 7493 | |||
| 7494 | //------------------------------------------------------------------------------ |
||
| 7495 | |||
| 7496 | XMFINLINE UINT XMVector3GreaterR |
||
| 7497 | ( |
||
| 7498 | FXMVECTOR V1, |
||
| 7499 | FXMVECTOR V2 |
||
| 7500 | ) |
||
| 7501 | { |
||
| 7502 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7503 | UINT CR = 0; |
||
| 7504 | if ((V1.vector4_f32[0] > V2.vector4_f32[0]) && |
||
| 7505 | (V1.vector4_f32[1] > V2.vector4_f32[1]) && |
||
| 7506 | (V1.vector4_f32[2] > V2.vector4_f32[2])) |
||
| 7507 | { |
||
| 7508 | CR = XM_CRMASK_CR6TRUE; |
||
| 7509 | } |
||
| 7510 | else if ((V1.vector4_f32[0] <= V2.vector4_f32[0]) && |
||
| 7511 | (V1.vector4_f32[1] <= V2.vector4_f32[1]) && |
||
| 7512 | (V1.vector4_f32[2] <= V2.vector4_f32[2])) |
||
| 7513 | { |
||
| 7514 | CR = XM_CRMASK_CR6FALSE; |
||
| 7515 | } |
||
| 7516 | return CR; |
||
| 7517 | |||
| 7518 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7519 | XMVECTOR vTemp = _mm_cmpgt_ps(V1,V2); |
||
| 7520 | UINT CR = 0; |
||
| 7521 | int iTest = _mm_movemask_ps(vTemp)&7; |
||
| 7522 | if (iTest==7) |
||
| 7523 | { |
||
| 7524 | CR = XM_CRMASK_CR6TRUE; |
||
| 7525 | } |
||
| 7526 | else if (!iTest) |
||
| 7527 | { |
||
| 7528 | CR = XM_CRMASK_CR6FALSE; |
||
| 7529 | } |
||
| 7530 | return CR; |
||
| 7531 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7532 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7533 | } |
||
| 7534 | |||
| 7535 | //------------------------------------------------------------------------------ |
||
| 7536 | |||
| 7537 | XMFINLINE BOOL XMVector3GreaterOrEqual |
||
| 7538 | ( |
||
| 7539 | FXMVECTOR V1, |
||
| 7540 | FXMVECTOR V2 |
||
| 7541 | ) |
||
| 7542 | { |
||
| 7543 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7544 | return (((V1.vector4_f32[0] >= V2.vector4_f32[0]) && (V1.vector4_f32[1] >= V2.vector4_f32[1]) && (V1.vector4_f32[2] >= V2.vector4_f32[2])) != 0); |
||
| 7545 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7546 | XMVECTOR vTemp = _mm_cmpge_ps(V1,V2); |
||
| 7547 | return (((_mm_movemask_ps(vTemp)&7)==7) != 0); |
||
| 7548 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7549 | return XMComparisonAllTrue(XMVector3GreaterOrEqualR(V1, V2)); |
||
| 7550 | #endif |
||
| 7551 | } |
||
| 7552 | |||
| 7553 | //------------------------------------------------------------------------------ |
||
| 7554 | |||
| 7555 | XMFINLINE UINT XMVector3GreaterOrEqualR |
||
| 7556 | ( |
||
| 7557 | FXMVECTOR V1, |
||
| 7558 | FXMVECTOR V2 |
||
| 7559 | ) |
||
| 7560 | { |
||
| 7561 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7562 | |||
| 7563 | UINT CR = 0; |
||
| 7564 | if ((V1.vector4_f32[0] >= V2.vector4_f32[0]) && |
||
| 7565 | (V1.vector4_f32[1] >= V2.vector4_f32[1]) && |
||
| 7566 | (V1.vector4_f32[2] >= V2.vector4_f32[2])) |
||
| 7567 | { |
||
| 7568 | CR = XM_CRMASK_CR6TRUE; |
||
| 7569 | } |
||
| 7570 | else if ((V1.vector4_f32[0] < V2.vector4_f32[0]) && |
||
| 7571 | (V1.vector4_f32[1] < V2.vector4_f32[1]) && |
||
| 7572 | (V1.vector4_f32[2] < V2.vector4_f32[2])) |
||
| 7573 | { |
||
| 7574 | CR = XM_CRMASK_CR6FALSE; |
||
| 7575 | } |
||
| 7576 | return CR; |
||
| 7577 | |||
| 7578 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7579 | XMVECTOR vTemp = _mm_cmpge_ps(V1,V2); |
||
| 7580 | UINT CR = 0; |
||
| 7581 | int iTest = _mm_movemask_ps(vTemp)&7; |
||
| 7582 | if (iTest==7) |
||
| 7583 | { |
||
| 7584 | CR = XM_CRMASK_CR6TRUE; |
||
| 7585 | } |
||
| 7586 | else if (!iTest) |
||
| 7587 | { |
||
| 7588 | CR = XM_CRMASK_CR6FALSE; |
||
| 7589 | } |
||
| 7590 | return CR; |
||
| 7591 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7592 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7593 | } |
||
| 7594 | |||
| 7595 | //------------------------------------------------------------------------------ |
||
| 7596 | |||
| 7597 | XMFINLINE BOOL XMVector3Less |
||
| 7598 | ( |
||
| 7599 | FXMVECTOR V1, |
||
| 7600 | FXMVECTOR V2 |
||
| 7601 | ) |
||
| 7602 | { |
||
| 7603 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7604 | return (((V1.vector4_f32[0] < V2.vector4_f32[0]) && (V1.vector4_f32[1] < V2.vector4_f32[1]) && (V1.vector4_f32[2] < V2.vector4_f32[2])) != 0); |
||
| 7605 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7606 | XMVECTOR vTemp = _mm_cmplt_ps(V1,V2); |
||
| 7607 | return (((_mm_movemask_ps(vTemp)&7)==7) != 0); |
||
| 7608 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7609 | return XMComparisonAllTrue(XMVector3GreaterR(V2, V1)); |
||
| 7610 | #endif |
||
| 7611 | } |
||
| 7612 | |||
| 7613 | //------------------------------------------------------------------------------ |
||
| 7614 | |||
| 7615 | XMFINLINE BOOL XMVector3LessOrEqual |
||
| 7616 | ( |
||
| 7617 | FXMVECTOR V1, |
||
| 7618 | FXMVECTOR V2 |
||
| 7619 | ) |
||
| 7620 | { |
||
| 7621 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7622 | return (((V1.vector4_f32[0] <= V2.vector4_f32[0]) && (V1.vector4_f32[1] <= V2.vector4_f32[1]) && (V1.vector4_f32[2] <= V2.vector4_f32[2])) != 0); |
||
| 7623 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7624 | XMVECTOR vTemp = _mm_cmple_ps(V1,V2); |
||
| 7625 | return (((_mm_movemask_ps(vTemp)&7)==7) != 0); |
||
| 7626 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7627 | return XMComparisonAllTrue(XMVector3GreaterOrEqualR(V2, V1)); |
||
| 7628 | #endif |
||
| 7629 | } |
||
| 7630 | |||
| 7631 | //------------------------------------------------------------------------------ |
||
| 7632 | |||
| 7633 | XMFINLINE BOOL XMVector3InBounds |
||
| 7634 | ( |
||
| 7635 | FXMVECTOR V, |
||
| 7636 | FXMVECTOR Bounds |
||
| 7637 | ) |
||
| 7638 | { |
||
| 7639 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7640 | return (((V.vector4_f32[0] <= Bounds.vector4_f32[0] && V.vector4_f32[0] >= -Bounds.vector4_f32[0]) && |
||
| 7641 | (V.vector4_f32[1] <= Bounds.vector4_f32[1] && V.vector4_f32[1] >= -Bounds.vector4_f32[1]) && |
||
| 7642 | (V.vector4_f32[2] <= Bounds.vector4_f32[2] && V.vector4_f32[2] >= -Bounds.vector4_f32[2])) != 0); |
||
| 7643 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7644 | // Test if less than or equal |
||
| 7645 | XMVECTOR vTemp1 = _mm_cmple_ps(V,Bounds); |
||
| 7646 | // Negate the bounds |
||
| 7647 | XMVECTOR vTemp2 = _mm_mul_ps(Bounds,g_XMNegativeOne); |
||
| 7648 | // Test if greater or equal (Reversed) |
||
| 7649 | vTemp2 = _mm_cmple_ps(vTemp2,V); |
||
| 7650 | // Blend answers |
||
| 7651 | vTemp1 = _mm_and_ps(vTemp1,vTemp2); |
||
| 7652 | // x,y and z in bounds? (w is don't care) |
||
| 7653 | return (((_mm_movemask_ps(vTemp1)&0x7)==0x7) != 0); |
||
| 7654 | #else |
||
| 7655 | return XMComparisonAllInBounds(XMVector3InBoundsR(V, Bounds)); |
||
| 7656 | #endif |
||
| 7657 | } |
||
| 7658 | |||
| 7659 | //------------------------------------------------------------------------------ |
||
| 7660 | |||
| 7661 | XMFINLINE UINT XMVector3InBoundsR |
||
| 7662 | ( |
||
| 7663 | FXMVECTOR V, |
||
| 7664 | FXMVECTOR Bounds |
||
| 7665 | ) |
||
| 7666 | { |
||
| 7667 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7668 | UINT CR = 0; |
||
| 7669 | if ((V.vector4_f32[0] <= Bounds.vector4_f32[0] && V.vector4_f32[0] >= -Bounds.vector4_f32[0]) && |
||
| 7670 | (V.vector4_f32[1] <= Bounds.vector4_f32[1] && V.vector4_f32[1] >= -Bounds.vector4_f32[1]) && |
||
| 7671 | (V.vector4_f32[2] <= Bounds.vector4_f32[2] && V.vector4_f32[2] >= -Bounds.vector4_f32[2])) |
||
| 7672 | { |
||
| 7673 | CR = XM_CRMASK_CR6BOUNDS; |
||
| 7674 | } |
||
| 7675 | return CR; |
||
| 7676 | |||
| 7677 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7678 | // Test if less than or equal |
||
| 7679 | XMVECTOR vTemp1 = _mm_cmple_ps(V,Bounds); |
||
| 7680 | // Negate the bounds |
||
| 7681 | XMVECTOR vTemp2 = _mm_mul_ps(Bounds,g_XMNegativeOne); |
||
| 7682 | // Test if greater or equal (Reversed) |
||
| 7683 | vTemp2 = _mm_cmple_ps(vTemp2,V); |
||
| 7684 | // Blend answers |
||
| 7685 | vTemp1 = _mm_and_ps(vTemp1,vTemp2); |
||
| 7686 | // x,y and z in bounds? (w is don't care) |
||
| 7687 | return ((_mm_movemask_ps(vTemp1)&0x7)==0x7) ? XM_CRMASK_CR6BOUNDS : 0; |
||
| 7688 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7689 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7690 | } |
||
| 7691 | |||
| 7692 | //------------------------------------------------------------------------------ |
||
| 7693 | |||
| 7694 | XMFINLINE BOOL XMVector3IsNaN |
||
| 7695 | ( |
||
| 7696 | FXMVECTOR V |
||
| 7697 | ) |
||
| 7698 | { |
||
| 7699 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7700 | |||
| 7701 | return (XMISNAN(V.vector4_f32[0]) || |
||
| 7702 | XMISNAN(V.vector4_f32[1]) || |
||
| 7703 | XMISNAN(V.vector4_f32[2])); |
||
| 7704 | |||
| 7705 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7706 | // Mask off the exponent |
||
| 7707 | __m128i vTempInf = _mm_and_si128(reinterpret_cast<const __m128i *>(&V)[0],g_XMInfinity); |
||
| 7708 | // Mask off the mantissa |
||
| 7709 | __m128i vTempNan = _mm_and_si128(reinterpret_cast<const __m128i *>(&V)[0],g_XMQNaNTest); |
||
| 7710 | // Are any of the exponents == 0x7F800000? |
||
| 7711 | vTempInf = _mm_cmpeq_epi32(vTempInf,g_XMInfinity); |
||
| 7712 | // Are any of the mantissa's zero? (SSE2 doesn't have a neq test) |
||
| 7713 | vTempNan = _mm_cmpeq_epi32(vTempNan,g_XMZero); |
||
| 7714 | // Perform a not on the NaN test to be true on NON-zero mantissas |
||
| 7715 | vTempNan = _mm_andnot_si128(vTempNan,vTempInf); |
||
| 7716 | // If x, y or z are NaN, the signs are true after the merge above |
||
| 7717 | return ((_mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTempNan)[0])&7) != 0); |
||
| 7718 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7719 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7720 | } |
||
| 7721 | |||
| 7722 | //------------------------------------------------------------------------------ |
||
| 7723 | |||
| 7724 | XMFINLINE BOOL XMVector3IsInfinite |
||
| 7725 | ( |
||
| 7726 | FXMVECTOR V |
||
| 7727 | ) |
||
| 7728 | { |
||
| 7729 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7730 | return (XMISINF(V.vector4_f32[0]) || |
||
| 7731 | XMISINF(V.vector4_f32[1]) || |
||
| 7732 | XMISINF(V.vector4_f32[2])); |
||
| 7733 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7734 | // Mask off the sign bit |
||
| 7735 | __m128 vTemp = _mm_and_ps(V,g_XMAbsMask); |
||
| 7736 | // Compare to infinity |
||
| 7737 | vTemp = _mm_cmpeq_ps(vTemp,g_XMInfinity); |
||
| 7738 | // If x,y or z are infinity, the signs are true. |
||
| 7739 | return ((_mm_movemask_ps(vTemp)&7) != 0); |
||
| 7740 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7741 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7742 | } |
||
| 7743 | |||
| 7744 | //------------------------------------------------------------------------------ |
||
| 7745 | // Computation operations |
||
| 7746 | //------------------------------------------------------------------------------ |
||
| 7747 | |||
| 7748 | //------------------------------------------------------------------------------ |
||
| 7749 | |||
| 7750 | XMFINLINE XMVECTOR XMVector3Dot |
||
| 7751 | ( |
||
| 7752 | FXMVECTOR V1, |
||
| 7753 | FXMVECTOR V2 |
||
| 7754 | ) |
||
| 7755 | { |
||
| 7756 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7757 | FLOAT fValue = V1.vector4_f32[0] * V2.vector4_f32[0] + V1.vector4_f32[1] * V2.vector4_f32[1] + V1.vector4_f32[2] * V2.vector4_f32[2]; |
||
| 7758 | XMVECTOR vResult = { |
||
| 7759 | fValue, |
||
| 7760 | fValue, |
||
| 7761 | fValue, |
||
| 7762 | fValue |
||
| 7763 | }; |
||
| 7764 | return vResult; |
||
| 7765 | |||
| 7766 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7767 | // Perform the dot product |
||
| 7768 | XMVECTOR vDot = _mm_mul_ps(V1,V2); |
||
| 7769 | // x=Dot.vector4_f32[1], y=Dot.vector4_f32[2] |
||
| 7770 | XMVECTOR vTemp = _mm_shuffle_ps(vDot,vDot,_MM_SHUFFLE(2,1,2,1)); |
||
| 7771 | // Result.vector4_f32[0] = x+y |
||
| 7772 | vDot = _mm_add_ss(vDot,vTemp); |
||
| 7773 | // x=Dot.vector4_f32[2] |
||
| 7774 | vTemp = _mm_shuffle_ps(vTemp,vTemp,_MM_SHUFFLE(1,1,1,1)); |
||
| 7775 | // Result.vector4_f32[0] = (x+y)+z |
||
| 7776 | vDot = _mm_add_ss(vDot,vTemp); |
||
| 7777 | // Splat x |
||
| 7778 | return _mm_shuffle_ps(vDot,vDot,_MM_SHUFFLE(0,0,0,0)); |
||
| 7779 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7780 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7781 | } |
||
| 7782 | |||
| 7783 | //------------------------------------------------------------------------------ |
||
| 7784 | |||
| 7785 | XMFINLINE XMVECTOR XMVector3Cross |
||
| 7786 | ( |
||
| 7787 | FXMVECTOR V1, |
||
| 7788 | FXMVECTOR V2 |
||
| 7789 | ) |
||
| 7790 | { |
||
| 7791 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7792 | XMVECTOR vResult = { |
||
| 7793 | (V1.vector4_f32[1] * V2.vector4_f32[2]) - (V1.vector4_f32[2] * V2.vector4_f32[1]), |
||
| 7794 | (V1.vector4_f32[2] * V2.vector4_f32[0]) - (V1.vector4_f32[0] * V2.vector4_f32[2]), |
||
| 7795 | (V1.vector4_f32[0] * V2.vector4_f32[1]) - (V1.vector4_f32[1] * V2.vector4_f32[0]), |
||
| 7796 | 0.0f |
||
| 7797 | }; |
||
| 7798 | return vResult; |
||
| 7799 | |||
| 7800 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7801 | // y1,z1,x1,w1 |
||
| 7802 | XMVECTOR vTemp1 = _mm_shuffle_ps(V1,V1,_MM_SHUFFLE(3,0,2,1)); |
||
| 7803 | // z2,x2,y2,w2 |
||
| 7804 | XMVECTOR vTemp2 = _mm_shuffle_ps(V2,V2,_MM_SHUFFLE(3,1,0,2)); |
||
| 7805 | // Perform the left operation |
||
| 7806 | XMVECTOR vResult = _mm_mul_ps(vTemp1,vTemp2); |
||
| 7807 | // z1,x1,y1,w1 |
||
| 7808 | vTemp1 = _mm_shuffle_ps(vTemp1,vTemp1,_MM_SHUFFLE(3,0,2,1)); |
||
| 7809 | // y2,z2,x2,w2 |
||
| 7810 | vTemp2 = _mm_shuffle_ps(vTemp2,vTemp2,_MM_SHUFFLE(3,1,0,2)); |
||
| 7811 | // Perform the right operation |
||
| 7812 | vTemp1 = _mm_mul_ps(vTemp1,vTemp2); |
||
| 7813 | // Subract the right from left, and return answer |
||
| 7814 | vResult = _mm_sub_ps(vResult,vTemp1); |
||
| 7815 | // Set w to zero |
||
| 7816 | return _mm_and_ps(vResult,g_XMMask3); |
||
| 7817 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7818 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7819 | } |
||
| 7820 | |||
| 7821 | //------------------------------------------------------------------------------ |
||
| 7822 | |||
| 7823 | XMFINLINE XMVECTOR XMVector3LengthSq |
||
| 7824 | ( |
||
| 7825 | FXMVECTOR V |
||
| 7826 | ) |
||
| 7827 | { |
||
| 7828 | return XMVector3Dot(V, V); |
||
| 7829 | } |
||
| 7830 | |||
| 7831 | //------------------------------------------------------------------------------ |
||
| 7832 | |||
| 7833 | XMFINLINE XMVECTOR XMVector3ReciprocalLengthEst |
||
| 7834 | ( |
||
| 7835 | FXMVECTOR V |
||
| 7836 | ) |
||
| 7837 | { |
||
| 7838 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7839 | |||
| 7840 | XMVECTOR Result; |
||
| 7841 | |||
| 7842 | Result = XMVector3LengthSq(V); |
||
| 7843 | Result = XMVectorReciprocalSqrtEst(Result); |
||
| 7844 | |||
| 7845 | return Result; |
||
| 7846 | |||
| 7847 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7848 | // Perform the dot product on x,y and z |
||
| 7849 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 7850 | // vTemp has z and y |
||
| 7851 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,2,1,2)); |
||
| 7852 | // x+z, y |
||
| 7853 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 7854 | // y,y,y,y |
||
| 7855 | vTemp = _mm_shuffle_ps(vTemp,vTemp,_MM_SHUFFLE(1,1,1,1)); |
||
| 7856 | // x+z+y,??,??,?? |
||
| 7857 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 7858 | // Splat the length squared |
||
| 7859 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 7860 | // Get the reciprocal |
||
| 7861 | vLengthSq = _mm_rsqrt_ps(vLengthSq); |
||
| 7862 | return vLengthSq; |
||
| 7863 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7864 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7865 | } |
||
| 7866 | |||
| 7867 | //------------------------------------------------------------------------------ |
||
| 7868 | |||
| 7869 | XMFINLINE XMVECTOR XMVector3ReciprocalLength |
||
| 7870 | ( |
||
| 7871 | FXMVECTOR V |
||
| 7872 | ) |
||
| 7873 | { |
||
| 7874 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7875 | |||
| 7876 | XMVECTOR Result; |
||
| 7877 | |||
| 7878 | Result = XMVector3LengthSq(V); |
||
| 7879 | Result = XMVectorReciprocalSqrt(Result); |
||
| 7880 | |||
| 7881 | return Result; |
||
| 7882 | |||
| 7883 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7884 | // Perform the dot product |
||
| 7885 | XMVECTOR vDot = _mm_mul_ps(V,V); |
||
| 7886 | // x=Dot.y, y=Dot.z |
||
| 7887 | XMVECTOR vTemp = _mm_shuffle_ps(vDot,vDot,_MM_SHUFFLE(2,1,2,1)); |
||
| 7888 | // Result.x = x+y |
||
| 7889 | vDot = _mm_add_ss(vDot,vTemp); |
||
| 7890 | // x=Dot.z |
||
| 7891 | vTemp = _mm_shuffle_ps(vTemp,vTemp,_MM_SHUFFLE(1,1,1,1)); |
||
| 7892 | // Result.x = (x+y)+z |
||
| 7893 | vDot = _mm_add_ss(vDot,vTemp); |
||
| 7894 | // Splat x |
||
| 7895 | vDot = _mm_shuffle_ps(vDot,vDot,_MM_SHUFFLE(0,0,0,0)); |
||
| 7896 | // Get the reciprocal |
||
| 7897 | vDot = _mm_sqrt_ps(vDot); |
||
| 7898 | // Get the reciprocal |
||
| 7899 | vDot = _mm_div_ps(g_XMOne,vDot); |
||
| 7900 | return vDot; |
||
| 7901 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7902 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7903 | } |
||
| 7904 | |||
| 7905 | //------------------------------------------------------------------------------ |
||
| 7906 | |||
| 7907 | XMFINLINE XMVECTOR XMVector3LengthEst |
||
| 7908 | ( |
||
| 7909 | FXMVECTOR V |
||
| 7910 | ) |
||
| 7911 | { |
||
| 7912 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7913 | |||
| 7914 | XMVECTOR Result; |
||
| 7915 | |||
| 7916 | Result = XMVector3LengthSq(V); |
||
| 7917 | Result = XMVectorSqrtEst(Result); |
||
| 7918 | |||
| 7919 | return Result; |
||
| 7920 | |||
| 7921 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7922 | // Perform the dot product on x,y and z |
||
| 7923 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 7924 | // vTemp has z and y |
||
| 7925 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,2,1,2)); |
||
| 7926 | // x+z, y |
||
| 7927 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 7928 | // y,y,y,y |
||
| 7929 | vTemp = _mm_shuffle_ps(vTemp,vTemp,_MM_SHUFFLE(1,1,1,1)); |
||
| 7930 | // x+z+y,??,??,?? |
||
| 7931 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 7932 | // Splat the length squared |
||
| 7933 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 7934 | // Get the length |
||
| 7935 | vLengthSq = _mm_sqrt_ps(vLengthSq); |
||
| 7936 | return vLengthSq; |
||
| 7937 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7938 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7939 | } |
||
| 7940 | |||
| 7941 | //------------------------------------------------------------------------------ |
||
| 7942 | |||
| 7943 | XMFINLINE XMVECTOR XMVector3Length |
||
| 7944 | ( |
||
| 7945 | FXMVECTOR V |
||
| 7946 | ) |
||
| 7947 | { |
||
| 7948 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7949 | |||
| 7950 | XMVECTOR Result; |
||
| 7951 | |||
| 7952 | Result = XMVector3LengthSq(V); |
||
| 7953 | Result = XMVectorSqrt(Result); |
||
| 7954 | |||
| 7955 | return Result; |
||
| 7956 | |||
| 7957 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7958 | // Perform the dot product on x,y and z |
||
| 7959 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 7960 | // vTemp has z and y |
||
| 7961 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,2,1,2)); |
||
| 7962 | // x+z, y |
||
| 7963 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 7964 | // y,y,y,y |
||
| 7965 | vTemp = _mm_shuffle_ps(vTemp,vTemp,_MM_SHUFFLE(1,1,1,1)); |
||
| 7966 | // x+z+y,??,??,?? |
||
| 7967 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 7968 | // Splat the length squared |
||
| 7969 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 7970 | // Get the length |
||
| 7971 | vLengthSq = _mm_sqrt_ps(vLengthSq); |
||
| 7972 | return vLengthSq; |
||
| 7973 | #else // _XM_VMX128_INTRINSICS_ |
||
| 7974 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 7975 | } |
||
| 7976 | |||
| 7977 | //------------------------------------------------------------------------------ |
||
| 7978 | // XMVector3NormalizeEst uses a reciprocal estimate and |
||
| 7979 | // returns QNaN on zero and infinite vectors. |
||
| 7980 | |||
| 7981 | XMFINLINE XMVECTOR XMVector3NormalizeEst |
||
| 7982 | ( |
||
| 7983 | FXMVECTOR V |
||
| 7984 | ) |
||
| 7985 | { |
||
| 7986 | #if defined(_XM_NO_INTRINSICS_) |
||
| 7987 | |||
| 7988 | XMVECTOR Result; |
||
| 7989 | Result = XMVector3ReciprocalLength(V); |
||
| 7990 | Result = XMVectorMultiply(V, Result); |
||
| 7991 | return Result; |
||
| 7992 | |||
| 7993 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 7994 | // Perform the dot product |
||
| 7995 | XMVECTOR vDot = _mm_mul_ps(V,V); |
||
| 7996 | // x=Dot.y, y=Dot.z |
||
| 7997 | XMVECTOR vTemp = _mm_shuffle_ps(vDot,vDot,_MM_SHUFFLE(2,1,2,1)); |
||
| 7998 | // Result.x = x+y |
||
| 7999 | vDot = _mm_add_ss(vDot,vTemp); |
||
| 8000 | // x=Dot.z |
||
| 8001 | vTemp = _mm_shuffle_ps(vTemp,vTemp,_MM_SHUFFLE(1,1,1,1)); |
||
| 8002 | // Result.x = (x+y)+z |
||
| 8003 | vDot = _mm_add_ss(vDot,vTemp); |
||
| 8004 | // Splat x |
||
| 8005 | vDot = _mm_shuffle_ps(vDot,vDot,_MM_SHUFFLE(0,0,0,0)); |
||
| 8006 | // Get the reciprocal |
||
| 8007 | vDot = _mm_rsqrt_ps(vDot); |
||
| 8008 | // Perform the normalization |
||
| 8009 | vDot = _mm_mul_ps(vDot,V); |
||
| 8010 | return vDot; |
||
| 8011 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8012 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8013 | } |
||
| 8014 | |||
| 8015 | //------------------------------------------------------------------------------ |
||
| 8016 | |||
| 8017 | XMFINLINE XMVECTOR XMVector3Normalize |
||
| 8018 | ( |
||
| 8019 | FXMVECTOR V |
||
| 8020 | ) |
||
| 8021 | { |
||
| 8022 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8023 | FLOAT fLengthSq; |
||
| 8024 | XMVECTOR vResult; |
||
| 8025 | |||
| 8026 | fLengthSq = sqrtf((V.vector4_f32[0]*V.vector4_f32[0])+(V.vector4_f32[1]*V.vector4_f32[1])+(V.vector4_f32[2]*V.vector4_f32[2])); |
||
| 8027 | // Prevent divide by zero |
||
| 8028 | if (fLengthSq) { |
||
| 8029 | fLengthSq = 1.0f/fLengthSq; |
||
| 8030 | } |
||
| 8031 | |||
| 8032 | vResult.vector4_f32[0] = V.vector4_f32[0]*fLengthSq; |
||
| 8033 | vResult.vector4_f32[1] = V.vector4_f32[1]*fLengthSq; |
||
| 8034 | vResult.vector4_f32[2] = V.vector4_f32[2]*fLengthSq; |
||
| 8035 | vResult.vector4_f32[3] = V.vector4_f32[3]*fLengthSq; |
||
| 8036 | return vResult; |
||
| 8037 | |||
| 8038 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8039 | // Perform the dot product on x,y and z only |
||
| 8040 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 8041 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(2,1,2,1)); |
||
| 8042 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 8043 | vTemp = _mm_shuffle_ps(vTemp,vTemp,_MM_SHUFFLE(1,1,1,1)); |
||
| 8044 | vLengthSq = _mm_add_ss(vLengthSq,vTemp); |
||
| 8045 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(0,0,0,0)); |
||
| 8046 | // Prepare for the division |
||
| 8047 | XMVECTOR vResult = _mm_sqrt_ps(vLengthSq); |
||
| 8048 | // Failsafe on zero (Or epsilon) length planes |
||
| 8049 | // If the length is infinity, set the elements to zero |
||
| 8050 | vLengthSq = _mm_cmpneq_ps(vLengthSq,g_XMInfinity); |
||
| 8051 | // Divide to perform the normalization |
||
| 8052 | vResult = _mm_div_ps(V,vResult); |
||
| 8053 | // Any that are infinity, set to zero |
||
| 8054 | vResult = _mm_and_ps(vResult,vLengthSq); |
||
| 8055 | return vResult; |
||
| 8056 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8057 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8058 | } |
||
| 8059 | |||
| 8060 | //------------------------------------------------------------------------------ |
||
| 8061 | |||
| 8062 | XMFINLINE XMVECTOR XMVector3ClampLength |
||
| 8063 | ( |
||
| 8064 | FXMVECTOR V, |
||
| 8065 | FLOAT LengthMin, |
||
| 8066 | FLOAT LengthMax |
||
| 8067 | ) |
||
| 8068 | { |
||
| 8069 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8070 | |||
| 8071 | XMVECTOR ClampMax; |
||
| 8072 | XMVECTOR ClampMin; |
||
| 8073 | |||
| 8074 | ClampMax = XMVectorReplicate(LengthMax); |
||
| 8075 | ClampMin = XMVectorReplicate(LengthMin); |
||
| 8076 | |||
| 8077 | return XMVector3ClampLengthV(V, ClampMin, ClampMax); |
||
| 8078 | |||
| 8079 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8080 | XMVECTOR ClampMax = _mm_set_ps1(LengthMax); |
||
| 8081 | XMVECTOR ClampMin = _mm_set_ps1(LengthMin); |
||
| 8082 | return XMVector3ClampLengthV(V,ClampMin,ClampMax); |
||
| 8083 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 8084 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8085 | } |
||
| 8086 | |||
| 8087 | //------------------------------------------------------------------------------ |
||
| 8088 | |||
| 8089 | XMFINLINE XMVECTOR XMVector3ClampLengthV |
||
| 8090 | ( |
||
| 8091 | FXMVECTOR V, |
||
| 8092 | FXMVECTOR LengthMin, |
||
| 8093 | FXMVECTOR LengthMax |
||
| 8094 | ) |
||
| 8095 | { |
||
| 8096 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8097 | |||
| 8098 | XMVECTOR ClampLength; |
||
| 8099 | XMVECTOR LengthSq; |
||
| 8100 | XMVECTOR RcpLength; |
||
| 8101 | XMVECTOR Length; |
||
| 8102 | XMVECTOR Normal; |
||
| 8103 | XMVECTOR Zero; |
||
| 8104 | XMVECTOR InfiniteLength; |
||
| 8105 | XMVECTOR ZeroLength; |
||
| 8106 | XMVECTOR Select; |
||
| 8107 | XMVECTOR ControlMax; |
||
| 8108 | XMVECTOR ControlMin; |
||
| 8109 | XMVECTOR Control; |
||
| 8110 | XMVECTOR Result; |
||
| 8111 | |||
| 8112 | XMASSERT((LengthMin.vector4_f32[1] == LengthMin.vector4_f32[0]) && (LengthMin.vector4_f32[2] == LengthMin.vector4_f32[0])); |
||
| 8113 | XMASSERT((LengthMax.vector4_f32[1] == LengthMax.vector4_f32[0]) && (LengthMax.vector4_f32[2] == LengthMax.vector4_f32[0])); |
||
| 8114 | XMASSERT(XMVector3GreaterOrEqual(LengthMin, XMVectorZero())); |
||
| 8115 | XMASSERT(XMVector3GreaterOrEqual(LengthMax, XMVectorZero())); |
||
| 8116 | XMASSERT(XMVector3GreaterOrEqual(LengthMax, LengthMin)); |
||
| 8117 | |||
| 8118 | LengthSq = XMVector3LengthSq(V); |
||
| 8119 | |||
| 8120 | Zero = XMVectorZero(); |
||
| 8121 | |||
| 8122 | RcpLength = XMVectorReciprocalSqrt(LengthSq); |
||
| 8123 | |||
| 8124 | InfiniteLength = XMVectorEqualInt(LengthSq, g_XMInfinity.v); |
||
| 8125 | ZeroLength = XMVectorEqual(LengthSq, Zero); |
||
| 8126 | |||
| 8127 | Normal = XMVectorMultiply(V, RcpLength); |
||
| 8128 | |||
| 8129 | Length = XMVectorMultiply(LengthSq, RcpLength); |
||
| 8130 | |||
| 8131 | Select = XMVectorEqualInt(InfiniteLength, ZeroLength); |
||
| 8132 | Length = XMVectorSelect(LengthSq, Length, Select); |
||
| 8133 | Normal = XMVectorSelect(LengthSq, Normal, Select); |
||
| 8134 | |||
| 8135 | ControlMax = XMVectorGreater(Length, LengthMax); |
||
| 8136 | ControlMin = XMVectorLess(Length, LengthMin); |
||
| 8137 | |||
| 8138 | ClampLength = XMVectorSelect(Length, LengthMax, ControlMax); |
||
| 8139 | ClampLength = XMVectorSelect(ClampLength, LengthMin, ControlMin); |
||
| 8140 | |||
| 8141 | Result = XMVectorMultiply(Normal, ClampLength); |
||
| 8142 | |||
| 8143 | // Preserve the original vector (with no precision loss) if the length falls within the given range |
||
| 8144 | Control = XMVectorEqualInt(ControlMax, ControlMin); |
||
| 8145 | Result = XMVectorSelect(Result, V, Control); |
||
| 8146 | |||
| 8147 | return Result; |
||
| 8148 | |||
| 8149 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8150 | XMVECTOR ClampLength; |
||
| 8151 | XMVECTOR LengthSq; |
||
| 8152 | XMVECTOR RcpLength; |
||
| 8153 | XMVECTOR Length; |
||
| 8154 | XMVECTOR Normal; |
||
| 8155 | XMVECTOR InfiniteLength; |
||
| 8156 | XMVECTOR ZeroLength; |
||
| 8157 | XMVECTOR Select; |
||
| 8158 | XMVECTOR ControlMax; |
||
| 8159 | XMVECTOR ControlMin; |
||
| 8160 | XMVECTOR Control; |
||
| 8161 | XMVECTOR Result; |
||
| 8162 | |||
| 8163 | XMASSERT((XMVectorGetY(LengthMin) == XMVectorGetX(LengthMin)) && (XMVectorGetZ(LengthMin) == XMVectorGetX(LengthMin))); |
||
| 8164 | XMASSERT((XMVectorGetY(LengthMax) == XMVectorGetX(LengthMax)) && (XMVectorGetZ(LengthMax) == XMVectorGetX(LengthMax))); |
||
| 8165 | XMASSERT(XMVector3GreaterOrEqual(LengthMin, g_XMZero)); |
||
| 8166 | XMASSERT(XMVector3GreaterOrEqual(LengthMax, g_XMZero)); |
||
| 8167 | XMASSERT(XMVector3GreaterOrEqual(LengthMax, LengthMin)); |
||
| 8168 | |||
| 8169 | LengthSq = XMVector3LengthSq(V); |
||
| 8170 | RcpLength = XMVectorReciprocalSqrt(LengthSq); |
||
| 8171 | InfiniteLength = XMVectorEqualInt(LengthSq, g_XMInfinity); |
||
| 8172 | ZeroLength = XMVectorEqual(LengthSq,g_XMZero); |
||
| 8173 | Normal = _mm_mul_ps(V, RcpLength); |
||
| 8174 | Length = _mm_mul_ps(LengthSq, RcpLength); |
||
| 8175 | Select = XMVectorEqualInt(InfiniteLength, ZeroLength); |
||
| 8176 | Length = XMVectorSelect(LengthSq, Length, Select); |
||
| 8177 | Normal = XMVectorSelect(LengthSq, Normal, Select); |
||
| 8178 | ControlMax = XMVectorGreater(Length, LengthMax); |
||
| 8179 | ControlMin = XMVectorLess(Length, LengthMin); |
||
| 8180 | ClampLength = XMVectorSelect(Length, LengthMax, ControlMax); |
||
| 8181 | ClampLength = XMVectorSelect(ClampLength, LengthMin, ControlMin); |
||
| 8182 | Result = _mm_mul_ps(Normal, ClampLength); |
||
| 8183 | // Preserve the original vector (with no precision loss) if the length falls within the given range |
||
| 8184 | Control = XMVectorEqualInt(ControlMax, ControlMin); |
||
| 8185 | Result = XMVectorSelect(Result, V, Control); |
||
| 8186 | return Result; |
||
| 8187 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8188 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8189 | } |
||
| 8190 | |||
| 8191 | //------------------------------------------------------------------------------ |
||
| 8192 | |||
| 8193 | XMFINLINE XMVECTOR XMVector3Reflect |
||
| 8194 | ( |
||
| 8195 | FXMVECTOR Incident, |
||
| 8196 | FXMVECTOR Normal |
||
| 8197 | ) |
||
| 8198 | { |
||
| 8199 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8200 | |||
| 8201 | XMVECTOR Result; |
||
| 8202 | |||
| 8203 | // Result = Incident - (2 * dot(Incident, Normal)) * Normal |
||
| 8204 | Result = XMVector3Dot(Incident, Normal); |
||
| 8205 | Result = XMVectorAdd(Result, Result); |
||
| 8206 | Result = XMVectorNegativeMultiplySubtract(Result, Normal, Incident); |
||
| 8207 | |||
| 8208 | return Result; |
||
| 8209 | |||
| 8210 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8211 | // Result = Incident - (2 * dot(Incident, Normal)) * Normal |
||
| 8212 | XMVECTOR Result = XMVector3Dot(Incident, Normal); |
||
| 8213 | Result = _mm_add_ps(Result, Result); |
||
| 8214 | Result = _mm_mul_ps(Result, Normal); |
||
| 8215 | Result = _mm_sub_ps(Incident,Result); |
||
| 8216 | return Result; |
||
| 8217 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8218 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8219 | } |
||
| 8220 | |||
| 8221 | //------------------------------------------------------------------------------ |
||
| 8222 | |||
| 8223 | XMFINLINE XMVECTOR XMVector3Refract |
||
| 8224 | ( |
||
| 8225 | FXMVECTOR Incident, |
||
| 8226 | FXMVECTOR Normal, |
||
| 8227 | FLOAT RefractionIndex |
||
| 8228 | ) |
||
| 8229 | { |
||
| 8230 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8231 | |||
| 8232 | XMVECTOR Index; |
||
| 8233 | Index = XMVectorReplicate(RefractionIndex); |
||
| 8234 | return XMVector3RefractV(Incident, Normal, Index); |
||
| 8235 | |||
| 8236 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8237 | XMVECTOR Index = _mm_set_ps1(RefractionIndex); |
||
| 8238 | return XMVector3RefractV(Incident,Normal,Index); |
||
| 8239 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 8240 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8241 | } |
||
| 8242 | |||
| 8243 | //------------------------------------------------------------------------------ |
||
| 8244 | |||
| 8245 | XMFINLINE XMVECTOR XMVector3RefractV |
||
| 8246 | ( |
||
| 8247 | FXMVECTOR Incident, |
||
| 8248 | FXMVECTOR Normal, |
||
| 8249 | FXMVECTOR RefractionIndex |
||
| 8250 | ) |
||
| 8251 | { |
||
| 8252 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8253 | |||
| 8254 | XMVECTOR IDotN; |
||
| 8255 | XMVECTOR R; |
||
| 8256 | CONST XMVECTOR Zero = XMVectorZero(); |
||
| 8257 | |||
| 8258 | // Result = RefractionIndex * Incident - Normal * (RefractionIndex * dot(Incident, Normal) + |
||
| 8259 | // sqrt(1 - RefractionIndex * RefractionIndex * (1 - dot(Incident, Normal) * dot(Incident, Normal)))) |
||
| 8260 | |||
| 8261 | IDotN = XMVector3Dot(Incident, Normal); |
||
| 8262 | |||
| 8263 | // R = 1.0f - RefractionIndex * RefractionIndex * (1.0f - IDotN * IDotN) |
||
| 8264 | R = XMVectorNegativeMultiplySubtract(IDotN, IDotN, g_XMOne.v); |
||
| 8265 | R = XMVectorMultiply(R, RefractionIndex); |
||
| 8266 | R = XMVectorNegativeMultiplySubtract(R, RefractionIndex, g_XMOne.v); |
||
| 8267 | |||
| 8268 | if (XMVector4LessOrEqual(R, Zero)) |
||
| 8269 | { |
||
| 8270 | // Total internal reflection |
||
| 8271 | return Zero; |
||
| 8272 | } |
||
| 8273 | else |
||
| 8274 | { |
||
| 8275 | XMVECTOR Result; |
||
| 8276 | |||
| 8277 | // R = RefractionIndex * IDotN + sqrt(R) |
||
| 8278 | R = XMVectorSqrt(R); |
||
| 8279 | R = XMVectorMultiplyAdd(RefractionIndex, IDotN, R); |
||
| 8280 | |||
| 8281 | // Result = RefractionIndex * Incident - Normal * R |
||
| 8282 | Result = XMVectorMultiply(RefractionIndex, Incident); |
||
| 8283 | Result = XMVectorNegativeMultiplySubtract(Normal, R, Result); |
||
| 8284 | |||
| 8285 | return Result; |
||
| 8286 | } |
||
| 8287 | |||
| 8288 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8289 | // Result = RefractionIndex * Incident - Normal * (RefractionIndex * dot(Incident, Normal) + |
||
| 8290 | // sqrt(1 - RefractionIndex * RefractionIndex * (1 - dot(Incident, Normal) * dot(Incident, Normal)))) |
||
| 8291 | XMVECTOR IDotN = XMVector3Dot(Incident, Normal); |
||
| 8292 | // R = 1.0f - RefractionIndex * RefractionIndex * (1.0f - IDotN * IDotN) |
||
| 8293 | XMVECTOR R = _mm_mul_ps(IDotN, IDotN); |
||
| 8294 | R = _mm_sub_ps(g_XMOne,R); |
||
| 8295 | R = _mm_mul_ps(R, RefractionIndex); |
||
| 8296 | R = _mm_mul_ps(R, RefractionIndex); |
||
| 8297 | R = _mm_sub_ps(g_XMOne,R); |
||
| 8298 | |||
| 8299 | XMVECTOR vResult = _mm_cmple_ps(R,g_XMZero); |
||
| 8300 | if (_mm_movemask_ps(vResult)==0x0f) |
||
| 8301 | { |
||
| 8302 | // Total internal reflection |
||
| 8303 | vResult = g_XMZero; |
||
| 8304 | } |
||
| 8305 | else |
||
| 8306 | { |
||
| 8307 | // R = RefractionIndex * IDotN + sqrt(R) |
||
| 8308 | R = _mm_sqrt_ps(R); |
||
| 8309 | vResult = _mm_mul_ps(RefractionIndex,IDotN); |
||
| 8310 | R = _mm_add_ps(R,vResult); |
||
| 8311 | // Result = RefractionIndex * Incident - Normal * R |
||
| 8312 | vResult = _mm_mul_ps(RefractionIndex, Incident); |
||
| 8313 | R = _mm_mul_ps(R,Normal); |
||
| 8314 | vResult = _mm_sub_ps(vResult,R); |
||
| 8315 | } |
||
| 8316 | return vResult; |
||
| 8317 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8318 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8319 | } |
||
| 8320 | |||
| 8321 | //------------------------------------------------------------------------------ |
||
| 8322 | |||
| 8323 | XMFINLINE XMVECTOR XMVector3Orthogonal |
||
| 8324 | ( |
||
| 8325 | FXMVECTOR V |
||
| 8326 | ) |
||
| 8327 | { |
||
| 8328 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8329 | |||
| 8330 | XMVECTOR NegativeV; |
||
| 8331 | XMVECTOR Z, YZYY; |
||
| 8332 | XMVECTOR ZIsNegative, YZYYIsNegative; |
||
| 8333 | XMVECTOR S, D; |
||
| 8334 | XMVECTOR R0, R1; |
||
| 8335 | XMVECTOR Select; |
||
| 8336 | XMVECTOR Zero; |
||
| 8337 | XMVECTOR Result; |
||
| 8338 | static CONST XMVECTORU32 Permute1X0X0X0X = {XM_PERMUTE_1X, XM_PERMUTE_0X, XM_PERMUTE_0X, XM_PERMUTE_0X}; |
||
| 8339 | static CONST XMVECTORU32 Permute0Y0Z0Y0Y= {XM_PERMUTE_0Y, XM_PERMUTE_0Z, XM_PERMUTE_0Y, XM_PERMUTE_0Y}; |
||
| 8340 | |||
| 8341 | Zero = XMVectorZero(); |
||
| 8342 | Z = XMVectorSplatZ(V); |
||
| 8343 | YZYY = XMVectorPermute(V, V, Permute0Y0Z0Y0Y.v); |
||
| 8344 | |||
| 8345 | NegativeV = XMVectorSubtract(Zero, V); |
||
| 8346 | |||
| 8347 | ZIsNegative = XMVectorLess(Z, Zero); |
||
| 8348 | YZYYIsNegative = XMVectorLess(YZYY, Zero); |
||
| 8349 | |||
| 8350 | S = XMVectorAdd(YZYY, Z); |
||
| 8351 | D = XMVectorSubtract(YZYY, Z); |
||
| 8352 | |||
| 8353 | Select = XMVectorEqualInt(ZIsNegative, YZYYIsNegative); |
||
| 8354 | |||
| 8355 | R0 = XMVectorPermute(NegativeV, S, Permute1X0X0X0X.v); |
||
| 8356 | R1 = XMVectorPermute(V, D, Permute1X0X0X0X.v); |
||
| 8357 | |||
| 8358 | Result = XMVectorSelect(R1, R0, Select); |
||
| 8359 | |||
| 8360 | return Result; |
||
| 8361 | |||
| 8362 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8363 | XMVECTOR NegativeV; |
||
| 8364 | XMVECTOR Z, YZYY; |
||
| 8365 | XMVECTOR ZIsNegative, YZYYIsNegative; |
||
| 8366 | XMVECTOR S, D; |
||
| 8367 | XMVECTOR R0, R1; |
||
| 8368 | XMVECTOR Select; |
||
| 8369 | XMVECTOR Zero; |
||
| 8370 | XMVECTOR Result; |
||
| 8371 | static CONST XMVECTORI32 Permute1X0X0X0X = {XM_PERMUTE_1X, XM_PERMUTE_0X, XM_PERMUTE_0X, XM_PERMUTE_0X}; |
||
| 8372 | static CONST XMVECTORI32 Permute0Y0Z0Y0Y= {XM_PERMUTE_0Y, XM_PERMUTE_0Z, XM_PERMUTE_0Y, XM_PERMUTE_0Y}; |
||
| 8373 | |||
| 8374 | Zero = XMVectorZero(); |
||
| 8375 | Z = XMVectorSplatZ(V); |
||
| 8376 | YZYY = XMVectorPermute(V, V, Permute0Y0Z0Y0Y); |
||
| 8377 | |||
| 8378 | NegativeV = _mm_sub_ps(Zero, V); |
||
| 8379 | |||
| 8380 | ZIsNegative = XMVectorLess(Z, Zero); |
||
| 8381 | YZYYIsNegative = XMVectorLess(YZYY, Zero); |
||
| 8382 | |||
| 8383 | S = _mm_add_ps(YZYY, Z); |
||
| 8384 | D = _mm_sub_ps(YZYY, Z); |
||
| 8385 | |||
| 8386 | Select = XMVectorEqualInt(ZIsNegative, YZYYIsNegative); |
||
| 8387 | |||
| 8388 | R0 = XMVectorPermute(NegativeV, S, Permute1X0X0X0X); |
||
| 8389 | R1 = XMVectorPermute(V, D,Permute1X0X0X0X); |
||
| 8390 | Result = XMVectorSelect(R1, R0, Select); |
||
| 8391 | return Result; |
||
| 8392 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8393 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8394 | } |
||
| 8395 | |||
| 8396 | //------------------------------------------------------------------------------ |
||
| 8397 | |||
| 8398 | XMFINLINE XMVECTOR XMVector3AngleBetweenNormalsEst |
||
| 8399 | ( |
||
| 8400 | FXMVECTOR N1, |
||
| 8401 | FXMVECTOR N2 |
||
| 8402 | ) |
||
| 8403 | { |
||
| 8404 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8405 | |||
| 8406 | XMVECTOR Result; |
||
| 8407 | XMVECTOR NegativeOne; |
||
| 8408 | XMVECTOR One; |
||
| 8409 | |||
| 8410 | Result = XMVector3Dot(N1, N2); |
||
| 8411 | NegativeOne = XMVectorSplatConstant(-1, 0); |
||
| 8412 | One = XMVectorSplatOne(); |
||
| 8413 | Result = XMVectorClamp(Result, NegativeOne, One); |
||
| 8414 | Result = XMVectorACosEst(Result); |
||
| 8415 | |||
| 8416 | return Result; |
||
| 8417 | |||
| 8418 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8419 | XMVECTOR vResult = XMVector3Dot(N1,N2); |
||
| 8420 | // Clamp to -1.0f to 1.0f |
||
| 8421 | vResult = _mm_max_ps(vResult,g_XMNegativeOne); |
||
| 8422 | vResult = _mm_min_ps(vResult,g_XMOne); |
||
| 8423 | vResult = XMVectorACosEst(vResult); |
||
| 8424 | return vResult; |
||
| 8425 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8426 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8427 | } |
||
| 8428 | |||
| 8429 | //------------------------------------------------------------------------------ |
||
| 8430 | |||
| 8431 | XMFINLINE XMVECTOR XMVector3AngleBetweenNormals |
||
| 8432 | ( |
||
| 8433 | FXMVECTOR N1, |
||
| 8434 | FXMVECTOR N2 |
||
| 8435 | ) |
||
| 8436 | { |
||
| 8437 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8438 | |||
| 8439 | XMVECTOR Result; |
||
| 8440 | XMVECTOR NegativeOne; |
||
| 8441 | XMVECTOR One; |
||
| 8442 | |||
| 8443 | Result = XMVector3Dot(N1, N2); |
||
| 8444 | NegativeOne = XMVectorSplatConstant(-1, 0); |
||
| 8445 | One = XMVectorSplatOne(); |
||
| 8446 | Result = XMVectorClamp(Result, NegativeOne, One); |
||
| 8447 | Result = XMVectorACos(Result); |
||
| 8448 | |||
| 8449 | return Result; |
||
| 8450 | |||
| 8451 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8452 | XMVECTOR vResult = XMVector3Dot(N1,N2); |
||
| 8453 | // Clamp to -1.0f to 1.0f |
||
| 8454 | vResult = _mm_max_ps(vResult,g_XMNegativeOne); |
||
| 8455 | vResult = _mm_min_ps(vResult,g_XMOne); |
||
| 8456 | vResult = XMVectorACos(vResult); |
||
| 8457 | return vResult; |
||
| 8458 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8459 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8460 | } |
||
| 8461 | |||
| 8462 | //------------------------------------------------------------------------------ |
||
| 8463 | |||
| 8464 | XMFINLINE XMVECTOR XMVector3AngleBetweenVectors |
||
| 8465 | ( |
||
| 8466 | FXMVECTOR V1, |
||
| 8467 | FXMVECTOR V2 |
||
| 8468 | ) |
||
| 8469 | { |
||
| 8470 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8471 | |||
| 8472 | XMVECTOR L1; |
||
| 8473 | XMVECTOR L2; |
||
| 8474 | XMVECTOR Dot; |
||
| 8475 | XMVECTOR CosAngle; |
||
| 8476 | XMVECTOR NegativeOne; |
||
| 8477 | XMVECTOR One; |
||
| 8478 | XMVECTOR Result; |
||
| 8479 | |||
| 8480 | L1 = XMVector3ReciprocalLength(V1); |
||
| 8481 | L2 = XMVector3ReciprocalLength(V2); |
||
| 8482 | |||
| 8483 | Dot = XMVector3Dot(V1, V2); |
||
| 8484 | |||
| 8485 | L1 = XMVectorMultiply(L1, L2); |
||
| 8486 | |||
| 8487 | NegativeOne = XMVectorSplatConstant(-1, 0); |
||
| 8488 | One = XMVectorSplatOne(); |
||
| 8489 | |||
| 8490 | CosAngle = XMVectorMultiply(Dot, L1); |
||
| 8491 | |||
| 8492 | CosAngle = XMVectorClamp(CosAngle, NegativeOne, One); |
||
| 8493 | |||
| 8494 | Result = XMVectorACos(CosAngle); |
||
| 8495 | |||
| 8496 | return Result; |
||
| 8497 | |||
| 8498 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8499 | XMVECTOR L1; |
||
| 8500 | XMVECTOR L2; |
||
| 8501 | XMVECTOR Dot; |
||
| 8502 | XMVECTOR CosAngle; |
||
| 8503 | XMVECTOR Result; |
||
| 8504 | |||
| 8505 | L1 = XMVector3ReciprocalLength(V1); |
||
| 8506 | L2 = XMVector3ReciprocalLength(V2); |
||
| 8507 | Dot = XMVector3Dot(V1, V2); |
||
| 8508 | L1 = _mm_mul_ps(L1, L2); |
||
| 8509 | CosAngle = _mm_mul_ps(Dot, L1); |
||
| 8510 | CosAngle = XMVectorClamp(CosAngle,g_XMNegativeOne,g_XMOne); |
||
| 8511 | Result = XMVectorACos(CosAngle); |
||
| 8512 | return Result; |
||
| 8513 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8514 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8515 | } |
||
| 8516 | |||
| 8517 | //------------------------------------------------------------------------------ |
||
| 8518 | |||
| 8519 | XMFINLINE XMVECTOR XMVector3LinePointDistance |
||
| 8520 | ( |
||
| 8521 | FXMVECTOR LinePoint1, |
||
| 8522 | FXMVECTOR LinePoint2, |
||
| 8523 | FXMVECTOR Point |
||
| 8524 | ) |
||
| 8525 | { |
||
| 8526 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8527 | |||
| 8528 | XMVECTOR PointVector; |
||
| 8529 | XMVECTOR LineVector; |
||
| 8530 | XMVECTOR ReciprocalLengthSq; |
||
| 8531 | XMVECTOR PointProjectionScale; |
||
| 8532 | XMVECTOR DistanceVector; |
||
| 8533 | XMVECTOR Result; |
||
| 8534 | |||
| 8535 | // Given a vector PointVector from LinePoint1 to Point and a vector |
||
| 8536 | // LineVector from LinePoint1 to LinePoint2, the scaled distance |
||
| 8537 | // PointProjectionScale from LinePoint1 to the perpendicular projection |
||
| 8538 | // of PointVector onto the line is defined as: |
||
| 8539 | // |
||
| 8540 | // PointProjectionScale = dot(PointVector, LineVector) / LengthSq(LineVector) |
||
| 8541 | |||
| 8542 | PointVector = XMVectorSubtract(Point, LinePoint1); |
||
| 8543 | LineVector = XMVectorSubtract(LinePoint2, LinePoint1); |
||
| 8544 | |||
| 8545 | ReciprocalLengthSq = XMVector3LengthSq(LineVector); |
||
| 8546 | ReciprocalLengthSq = XMVectorReciprocal(ReciprocalLengthSq); |
||
| 8547 | |||
| 8548 | PointProjectionScale = XMVector3Dot(PointVector, LineVector); |
||
| 8549 | PointProjectionScale = XMVectorMultiply(PointProjectionScale, ReciprocalLengthSq); |
||
| 8550 | |||
| 8551 | DistanceVector = XMVectorMultiply(LineVector, PointProjectionScale); |
||
| 8552 | DistanceVector = XMVectorSubtract(PointVector, DistanceVector); |
||
| 8553 | |||
| 8554 | Result = XMVector3Length(DistanceVector); |
||
| 8555 | |||
| 8556 | return Result; |
||
| 8557 | |||
| 8558 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8559 | XMVECTOR PointVector = _mm_sub_ps(Point,LinePoint1); |
||
| 8560 | XMVECTOR LineVector = _mm_sub_ps(LinePoint2,LinePoint1); |
||
| 8561 | XMVECTOR ReciprocalLengthSq = XMVector3LengthSq(LineVector); |
||
| 8562 | XMVECTOR vResult = XMVector3Dot(PointVector,LineVector); |
||
| 8563 | vResult = _mm_div_ps(vResult,ReciprocalLengthSq); |
||
| 8564 | vResult = _mm_mul_ps(vResult,LineVector); |
||
| 8565 | vResult = _mm_sub_ps(PointVector,vResult); |
||
| 8566 | vResult = XMVector3Length(vResult); |
||
| 8567 | return vResult; |
||
| 8568 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8569 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8570 | } |
||
| 8571 | |||
| 8572 | //------------------------------------------------------------------------------ |
||
| 8573 | |||
| 8574 | XMFINLINE VOID XMVector3ComponentsFromNormal |
||
| 8575 | ( |
||
| 8576 | XMVECTOR* pParallel, |
||
| 8577 | XMVECTOR* pPerpendicular, |
||
| 8578 | FXMVECTOR V, |
||
| 8579 | FXMVECTOR Normal |
||
| 8580 | ) |
||
| 8581 | { |
||
| 8582 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8583 | |||
| 8584 | XMVECTOR Parallel; |
||
| 8585 | XMVECTOR Scale; |
||
| 8586 | |||
| 8587 | XMASSERT(pParallel); |
||
| 8588 | XMASSERT(pPerpendicular); |
||
| 8589 | |||
| 8590 | Scale = XMVector3Dot(V, Normal); |
||
| 8591 | |||
| 8592 | Parallel = XMVectorMultiply(Normal, Scale); |
||
| 8593 | |||
| 8594 | *pParallel = Parallel; |
||
| 8595 | *pPerpendicular = XMVectorSubtract(V, Parallel); |
||
| 8596 | |||
| 8597 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8598 | XMASSERT(pParallel); |
||
| 8599 | XMASSERT(pPerpendicular); |
||
| 8600 | XMVECTOR Scale = XMVector3Dot(V, Normal); |
||
| 8601 | XMVECTOR Parallel = _mm_mul_ps(Normal,Scale); |
||
| 8602 | *pParallel = Parallel; |
||
| 8603 | *pPerpendicular = _mm_sub_ps(V,Parallel); |
||
| 8604 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8605 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8606 | } |
||
| 8607 | |||
| 8608 | //------------------------------------------------------------------------------ |
||
| 8609 | // Transform a vector using a rotation expressed as a unit quaternion |
||
| 8610 | |||
| 8611 | XMFINLINE XMVECTOR XMVector3Rotate |
||
| 8612 | ( |
||
| 8613 | FXMVECTOR V, |
||
| 8614 | FXMVECTOR RotationQuaternion |
||
| 8615 | ) |
||
| 8616 | { |
||
| 8617 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8618 | |||
| 8619 | XMVECTOR A; |
||
| 8620 | XMVECTOR Q; |
||
| 8621 | XMVECTOR Result; |
||
| 8622 | |||
| 8623 | A = XMVectorSelect(g_XMSelect1110.v, V, g_XMSelect1110.v); |
||
| 8624 | Q = XMQuaternionConjugate(RotationQuaternion); |
||
| 8625 | Result = XMQuaternionMultiply(Q, A); |
||
| 8626 | Result = XMQuaternionMultiply(Result, RotationQuaternion); |
||
| 8627 | |||
| 8628 | return Result; |
||
| 8629 | |||
| 8630 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8631 | XMVECTOR A; |
||
| 8632 | XMVECTOR Q; |
||
| 8633 | XMVECTOR Result; |
||
| 8634 | |||
| 8635 | A = _mm_and_ps(V,g_XMMask3); |
||
| 8636 | Q = XMQuaternionConjugate(RotationQuaternion); |
||
| 8637 | Result = XMQuaternionMultiply(Q, A); |
||
| 8638 | Result = XMQuaternionMultiply(Result, RotationQuaternion); |
||
| 8639 | return Result; |
||
| 8640 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8641 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8642 | } |
||
| 8643 | |||
| 8644 | //------------------------------------------------------------------------------ |
||
| 8645 | // Transform a vector using the inverse of a rotation expressed as a unit quaternion |
||
| 8646 | |||
| 8647 | XMFINLINE XMVECTOR XMVector3InverseRotate |
||
| 8648 | ( |
||
| 8649 | FXMVECTOR V, |
||
| 8650 | FXMVECTOR RotationQuaternion |
||
| 8651 | ) |
||
| 8652 | { |
||
| 8653 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8654 | |||
| 8655 | XMVECTOR A; |
||
| 8656 | XMVECTOR Q; |
||
| 8657 | XMVECTOR Result; |
||
| 8658 | |||
| 8659 | A = XMVectorSelect(g_XMSelect1110.v, V, g_XMSelect1110.v); |
||
| 8660 | Result = XMQuaternionMultiply(RotationQuaternion, A); |
||
| 8661 | Q = XMQuaternionConjugate(RotationQuaternion); |
||
| 8662 | Result = XMQuaternionMultiply(Result, Q); |
||
| 8663 | |||
| 8664 | return Result; |
||
| 8665 | |||
| 8666 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8667 | XMVECTOR A; |
||
| 8668 | XMVECTOR Q; |
||
| 8669 | XMVECTOR Result; |
||
| 8670 | A = _mm_and_ps(V,g_XMMask3); |
||
| 8671 | Result = XMQuaternionMultiply(RotationQuaternion, A); |
||
| 8672 | Q = XMQuaternionConjugate(RotationQuaternion); |
||
| 8673 | Result = XMQuaternionMultiply(Result, Q); |
||
| 8674 | return Result; |
||
| 8675 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8676 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8677 | } |
||
| 8678 | |||
| 8679 | //------------------------------------------------------------------------------ |
||
| 8680 | |||
| 8681 | XMFINLINE XMVECTOR XMVector3Transform |
||
| 8682 | ( |
||
| 8683 | FXMVECTOR V, |
||
| 8684 | CXMMATRIX M |
||
| 8685 | ) |
||
| 8686 | { |
||
| 8687 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8688 | |||
| 8689 | XMVECTOR X; |
||
| 8690 | XMVECTOR Y; |
||
| 8691 | XMVECTOR Z; |
||
| 8692 | XMVECTOR Result; |
||
| 8693 | |||
| 8694 | Z = XMVectorSplatZ(V); |
||
| 8695 | Y = XMVectorSplatY(V); |
||
| 8696 | X = XMVectorSplatX(V); |
||
| 8697 | |||
| 8698 | Result = XMVectorMultiplyAdd(Z, M.r[2], M.r[3]); |
||
| 8699 | Result = XMVectorMultiplyAdd(Y, M.r[1], Result); |
||
| 8700 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 8701 | |||
| 8702 | return Result; |
||
| 8703 | |||
| 8704 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8705 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,0,0,0)); |
||
| 8706 | vResult = _mm_mul_ps(vResult,M.r[0]); |
||
| 8707 | XMVECTOR vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 8708 | vTemp = _mm_mul_ps(vTemp,M.r[1]); |
||
| 8709 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 8710 | vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(2,2,2,2)); |
||
| 8711 | vTemp = _mm_mul_ps(vTemp,M.r[2]); |
||
| 8712 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 8713 | vResult = _mm_add_ps(vResult,M.r[3]); |
||
| 8714 | return vResult; |
||
| 8715 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8716 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8717 | } |
||
| 8718 | |||
| 8719 | //------------------------------------------------------------------------------ |
||
| 8720 | |||
| 8721 | XMINLINE XMFLOAT4* XMVector3TransformStream |
||
| 8722 | ( |
||
| 8723 | XMFLOAT4* pOutputStream, |
||
| 8724 | UINT OutputStride, |
||
| 8725 | CONST XMFLOAT3* pInputStream, |
||
| 8726 | UINT InputStride, |
||
| 8727 | UINT VectorCount, |
||
| 8728 | CXMMATRIX M |
||
| 8729 | ) |
||
| 8730 | { |
||
| 8731 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8732 | |||
| 8733 | XMVECTOR V; |
||
| 8734 | XMVECTOR X; |
||
| 8735 | XMVECTOR Y; |
||
| 8736 | XMVECTOR Z; |
||
| 8737 | XMVECTOR Result; |
||
| 8738 | UINT i; |
||
| 8739 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 8740 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 8741 | |||
| 8742 | XMASSERT(pOutputStream); |
||
| 8743 | XMASSERT(pInputStream); |
||
| 8744 | |||
| 8745 | for (i = 0; i < VectorCount; i++) |
||
| 8746 | { |
||
| 8747 | V = XMLoadFloat3((XMFLOAT3*)pInputVector); |
||
| 8748 | Z = XMVectorSplatZ(V); |
||
| 8749 | Y = XMVectorSplatY(V); |
||
| 8750 | X = XMVectorSplatX(V); |
||
| 8751 | |||
| 8752 | Result = XMVectorMultiplyAdd(Z, M.r[2], M.r[3]); |
||
| 8753 | Result = XMVectorMultiplyAdd(Y, M.r[1], Result); |
||
| 8754 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 8755 | |||
| 8756 | XMStoreFloat4((XMFLOAT4*)pOutputVector, Result); |
||
| 8757 | |||
| 8758 | pInputVector += InputStride; |
||
| 8759 | pOutputVector += OutputStride; |
||
| 8760 | } |
||
| 8761 | |||
| 8762 | return pOutputStream; |
||
| 8763 | |||
| 8764 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8765 | XMASSERT(pOutputStream); |
||
| 8766 | XMASSERT(pInputStream); |
||
| 8767 | UINT i; |
||
| 8768 | const BYTE* pInputVector = (const BYTE*)pInputStream; |
||
| 8769 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 8770 | |||
| 8771 | for (i = 0; i < VectorCount; i++) |
||
| 8772 | { |
||
| 8773 | XMVECTOR X = _mm_load_ps1(&reinterpret_cast<const XMFLOAT3 *>(pInputVector)->x); |
||
| 8774 | XMVECTOR Y = _mm_load_ps1(&reinterpret_cast<const XMFLOAT3 *>(pInputVector)->y); |
||
| 8775 | XMVECTOR vResult = _mm_load_ps1(&reinterpret_cast<const XMFLOAT3 *>(pInputVector)->z); |
||
| 8776 | vResult = _mm_mul_ps(vResult,M.r[2]); |
||
| 8777 | vResult = _mm_add_ps(vResult,M.r[3]); |
||
| 8778 | Y = _mm_mul_ps(Y,M.r[1]); |
||
| 8779 | vResult = _mm_add_ps(vResult,Y); |
||
| 8780 | X = _mm_mul_ps(X,M.r[0]); |
||
| 8781 | vResult = _mm_add_ps(vResult,X); |
||
| 8782 | _mm_storeu_ps(reinterpret_cast<float *>(pOutputVector),vResult); |
||
| 8783 | pInputVector += InputStride; |
||
| 8784 | pOutputVector += OutputStride; |
||
| 8785 | } |
||
| 8786 | |||
| 8787 | return pOutputStream; |
||
| 8788 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 8789 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8790 | } |
||
| 8791 | |||
| 8792 | //------------------------------------------------------------------------------ |
||
| 8793 | |||
| 8794 | XMINLINE XMFLOAT4* XMVector3TransformStreamNC |
||
| 8795 | ( |
||
| 8796 | XMFLOAT4* pOutputStream, |
||
| 8797 | UINT OutputStride, |
||
| 8798 | CONST XMFLOAT3* pInputStream, |
||
| 8799 | UINT InputStride, |
||
| 8800 | UINT VectorCount, |
||
| 8801 | CXMMATRIX M |
||
| 8802 | ) |
||
| 8803 | { |
||
| 8804 | #if defined(_XM_NO_INTRINSICS_) || defined(XM_NO_MISALIGNED_VECTOR_ACCESS) || defined(_XM_SSE_INTRINSICS_) |
||
| 8805 | return XMVector3TransformStream( pOutputStream, OutputStride, pInputStream, InputStride, VectorCount, M ); |
||
| 8806 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8807 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8808 | } |
||
| 8809 | |||
| 8810 | //------------------------------------------------------------------------------ |
||
| 8811 | |||
| 8812 | XMFINLINE XMVECTOR XMVector3TransformCoord |
||
| 8813 | ( |
||
| 8814 | FXMVECTOR V, |
||
| 8815 | CXMMATRIX M |
||
| 8816 | ) |
||
| 8817 | { |
||
| 8818 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8819 | |||
| 8820 | XMVECTOR X; |
||
| 8821 | XMVECTOR Y; |
||
| 8822 | XMVECTOR Z; |
||
| 8823 | XMVECTOR InverseW; |
||
| 8824 | XMVECTOR Result; |
||
| 8825 | |||
| 8826 | Z = XMVectorSplatZ(V); |
||
| 8827 | Y = XMVectorSplatY(V); |
||
| 8828 | X = XMVectorSplatX(V); |
||
| 8829 | |||
| 8830 | Result = XMVectorMultiplyAdd(Z, M.r[2], M.r[3]); |
||
| 8831 | Result = XMVectorMultiplyAdd(Y, M.r[1], Result); |
||
| 8832 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 8833 | |||
| 8834 | InverseW = XMVectorSplatW(Result); |
||
| 8835 | InverseW = XMVectorReciprocal(InverseW); |
||
| 8836 | |||
| 8837 | Result = XMVectorMultiply(Result, InverseW); |
||
| 8838 | |||
| 8839 | return Result; |
||
| 8840 | |||
| 8841 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8842 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,0,0,0)); |
||
| 8843 | vResult = _mm_mul_ps(vResult,M.r[0]); |
||
| 8844 | XMVECTOR vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 8845 | vTemp = _mm_mul_ps(vTemp,M.r[1]); |
||
| 8846 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 8847 | vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(2,2,2,2)); |
||
| 8848 | vTemp = _mm_mul_ps(vTemp,M.r[2]); |
||
| 8849 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 8850 | vResult = _mm_add_ps(vResult,M.r[3]); |
||
| 8851 | vTemp = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,3,3,3)); |
||
| 8852 | vResult = _mm_div_ps(vResult,vTemp); |
||
| 8853 | return vResult; |
||
| 8854 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8855 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8856 | } |
||
| 8857 | |||
| 8858 | //------------------------------------------------------------------------------ |
||
| 8859 | |||
| 8860 | XMINLINE XMFLOAT3* XMVector3TransformCoordStream |
||
| 8861 | ( |
||
| 8862 | XMFLOAT3* pOutputStream, |
||
| 8863 | UINT OutputStride, |
||
| 8864 | CONST XMFLOAT3* pInputStream, |
||
| 8865 | UINT InputStride, |
||
| 8866 | UINT VectorCount, |
||
| 8867 | CXMMATRIX M |
||
| 8868 | ) |
||
| 8869 | { |
||
| 8870 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8871 | |||
| 8872 | XMVECTOR V; |
||
| 8873 | XMVECTOR X; |
||
| 8874 | XMVECTOR Y; |
||
| 8875 | XMVECTOR Z; |
||
| 8876 | XMVECTOR InverseW; |
||
| 8877 | XMVECTOR Result; |
||
| 8878 | UINT i; |
||
| 8879 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 8880 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 8881 | |||
| 8882 | XMASSERT(pOutputStream); |
||
| 8883 | XMASSERT(pInputStream); |
||
| 8884 | |||
| 8885 | for (i = 0; i < VectorCount; i++) |
||
| 8886 | { |
||
| 8887 | V = XMLoadFloat3((XMFLOAT3*)pInputVector); |
||
| 8888 | Z = XMVectorSplatZ(V); |
||
| 8889 | Y = XMVectorSplatY(V); |
||
| 8890 | X = XMVectorSplatX(V); |
||
| 8891 | // Z = XMVectorReplicate(((XMFLOAT3*)pInputVector)->z); |
||
| 8892 | // Y = XMVectorReplicate(((XMFLOAT3*)pInputVector)->y); |
||
| 8893 | // X = XMVectorReplicate(((XMFLOAT3*)pInputVector)->x); |
||
| 8894 | |||
| 8895 | Result = XMVectorMultiplyAdd(Z, M.r[2], M.r[3]); |
||
| 8896 | Result = XMVectorMultiplyAdd(Y, M.r[1], Result); |
||
| 8897 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 8898 | |||
| 8899 | InverseW = XMVectorSplatW(Result); |
||
| 8900 | InverseW = XMVectorReciprocal(InverseW); |
||
| 8901 | |||
| 8902 | Result = XMVectorMultiply(Result, InverseW); |
||
| 8903 | |||
| 8904 | XMStoreFloat3((XMFLOAT3*)pOutputVector, Result); |
||
| 8905 | |||
| 8906 | pInputVector += InputStride; |
||
| 8907 | pOutputVector += OutputStride; |
||
| 8908 | } |
||
| 8909 | |||
| 8910 | return pOutputStream; |
||
| 8911 | |||
| 8912 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8913 | XMASSERT(pOutputStream); |
||
| 8914 | XMASSERT(pInputStream); |
||
| 8915 | |||
| 8916 | UINT i; |
||
| 8917 | const BYTE *pInputVector = (BYTE*)pInputStream; |
||
| 8918 | BYTE *pOutputVector = (BYTE*)pOutputStream; |
||
| 8919 | |||
| 8920 | for (i = 0; i < VectorCount; i++) |
||
| 8921 | { |
||
| 8922 | XMVECTOR X = _mm_load_ps1(&reinterpret_cast<const XMFLOAT3 *>(pInputVector)->x); |
||
| 8923 | XMVECTOR Y = _mm_load_ps1(&reinterpret_cast<const XMFLOAT3 *>(pInputVector)->y); |
||
| 8924 | XMVECTOR vResult = _mm_load_ps1(&reinterpret_cast<const XMFLOAT3 *>(pInputVector)->z); |
||
| 8925 | vResult = _mm_mul_ps(vResult,M.r[2]); |
||
| 8926 | vResult = _mm_add_ps(vResult,M.r[3]); |
||
| 8927 | Y = _mm_mul_ps(Y,M.r[1]); |
||
| 8928 | vResult = _mm_add_ps(vResult,Y); |
||
| 8929 | X = _mm_mul_ps(X,M.r[0]); |
||
| 8930 | vResult = _mm_add_ps(vResult,X); |
||
| 8931 | |||
| 8932 | X = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(3,3,3,3)); |
||
| 8933 | vResult = _mm_div_ps(vResult,X); |
||
| 8934 | _mm_store_ss(&reinterpret_cast<XMFLOAT3 *>(pOutputVector)->x,vResult); |
||
| 8935 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(0,3,2,1)); |
||
| 8936 | _mm_store_ss(&reinterpret_cast<XMFLOAT3 *>(pOutputVector)->y,vResult); |
||
| 8937 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(0,3,2,1)); |
||
| 8938 | _mm_store_ss(&reinterpret_cast<XMFLOAT3 *>(pOutputVector)->z,vResult); |
||
| 8939 | pInputVector += InputStride; |
||
| 8940 | pOutputVector += OutputStride; |
||
| 8941 | } |
||
| 8942 | |||
| 8943 | return pOutputStream; |
||
| 8944 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 8945 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8946 | } |
||
| 8947 | |||
| 8948 | //------------------------------------------------------------------------------ |
||
| 8949 | |||
| 8950 | XMFINLINE XMVECTOR XMVector3TransformNormal |
||
| 8951 | ( |
||
| 8952 | FXMVECTOR V, |
||
| 8953 | CXMMATRIX M |
||
| 8954 | ) |
||
| 8955 | { |
||
| 8956 | #if defined(_XM_NO_INTRINSICS_) |
||
| 8957 | |||
| 8958 | XMVECTOR X; |
||
| 8959 | XMVECTOR Y; |
||
| 8960 | XMVECTOR Z; |
||
| 8961 | XMVECTOR Result; |
||
| 8962 | |||
| 8963 | Z = XMVectorSplatZ(V); |
||
| 8964 | Y = XMVectorSplatY(V); |
||
| 8965 | X = XMVectorSplatX(V); |
||
| 8966 | |||
| 8967 | Result = XMVectorMultiply(Z, M.r[2]); |
||
| 8968 | Result = XMVectorMultiplyAdd(Y, M.r[1], Result); |
||
| 8969 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 8970 | |||
| 8971 | return Result; |
||
| 8972 | |||
| 8973 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 8974 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,0,0,0)); |
||
| 8975 | vResult = _mm_mul_ps(vResult,M.r[0]); |
||
| 8976 | XMVECTOR vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 8977 | vTemp = _mm_mul_ps(vTemp,M.r[1]); |
||
| 8978 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 8979 | vTemp = _mm_shuffle_ps(V,V,_MM_SHUFFLE(2,2,2,2)); |
||
| 8980 | vTemp = _mm_mul_ps(vTemp,M.r[2]); |
||
| 8981 | vResult = _mm_add_ps(vResult,vTemp); |
||
| 8982 | return vResult; |
||
| 8983 | #else // _XM_VMX128_INTRINSICS_ |
||
| 8984 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 8985 | } |
||
| 8986 | |||
| 8987 | //------------------------------------------------------------------------------ |
||
| 8988 | |||
| 8989 | XMINLINE XMFLOAT3* XMVector3TransformNormalStream |
||
| 8990 | ( |
||
| 8991 | XMFLOAT3* pOutputStream, |
||
| 8992 | UINT OutputStride, |
||
| 8993 | CONST XMFLOAT3* pInputStream, |
||
| 8994 | UINT InputStride, |
||
| 8995 | UINT VectorCount, |
||
| 8996 | CXMMATRIX M |
||
| 8997 | ) |
||
| 8998 | { |
||
| 8999 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9000 | |||
| 9001 | XMVECTOR V; |
||
| 9002 | XMVECTOR X; |
||
| 9003 | XMVECTOR Y; |
||
| 9004 | XMVECTOR Z; |
||
| 9005 | XMVECTOR Result; |
||
| 9006 | UINT i; |
||
| 9007 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 9008 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 9009 | |||
| 9010 | XMASSERT(pOutputStream); |
||
| 9011 | XMASSERT(pInputStream); |
||
| 9012 | |||
| 9013 | for (i = 0; i < VectorCount; i++) |
||
| 9014 | { |
||
| 9015 | V = XMLoadFloat3((XMFLOAT3*)pInputVector); |
||
| 9016 | Z = XMVectorSplatZ(V); |
||
| 9017 | Y = XMVectorSplatY(V); |
||
| 9018 | X = XMVectorSplatX(V); |
||
| 9019 | // Z = XMVectorReplicate(((XMFLOAT3*)pInputVector)->z); |
||
| 9020 | // Y = XMVectorReplicate(((XMFLOAT3*)pInputVector)->y); |
||
| 9021 | // X = XMVectorReplicate(((XMFLOAT3*)pInputVector)->x); |
||
| 9022 | |||
| 9023 | Result = XMVectorMultiply(Z, M.r[2]); |
||
| 9024 | Result = XMVectorMultiplyAdd(Y, M.r[1], Result); |
||
| 9025 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 9026 | |||
| 9027 | XMStoreFloat3((XMFLOAT3*)pOutputVector, Result); |
||
| 9028 | |||
| 9029 | pInputVector += InputStride; |
||
| 9030 | pOutputVector += OutputStride; |
||
| 9031 | } |
||
| 9032 | |||
| 9033 | return pOutputStream; |
||
| 9034 | |||
| 9035 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9036 | XMASSERT(pOutputStream); |
||
| 9037 | XMASSERT(pInputStream); |
||
| 9038 | |||
| 9039 | UINT i; |
||
| 9040 | const BYTE *pInputVector = (BYTE*)pInputStream; |
||
| 9041 | BYTE *pOutputVector = (BYTE*)pOutputStream; |
||
| 9042 | |||
| 9043 | for (i = 0; i < VectorCount; i++) |
||
| 9044 | { |
||
| 9045 | XMVECTOR X = _mm_load_ps1(&reinterpret_cast<const XMFLOAT3 *>(pInputVector)->x); |
||
| 9046 | XMVECTOR Y = _mm_load_ps1(&reinterpret_cast<const XMFLOAT3 *>(pInputVector)->y); |
||
| 9047 | XMVECTOR vResult = _mm_load_ps1(&reinterpret_cast<const XMFLOAT3 *>(pInputVector)->z); |
||
| 9048 | vResult = _mm_mul_ps(vResult,M.r[2]); |
||
| 9049 | Y = _mm_mul_ps(Y,M.r[1]); |
||
| 9050 | vResult = _mm_add_ps(vResult,Y); |
||
| 9051 | X = _mm_mul_ps(X,M.r[0]); |
||
| 9052 | vResult = _mm_add_ps(vResult,X); |
||
| 9053 | _mm_store_ss(&reinterpret_cast<XMFLOAT3 *>(pOutputVector)->x,vResult); |
||
| 9054 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(0,3,2,1)); |
||
| 9055 | _mm_store_ss(&reinterpret_cast<XMFLOAT3 *>(pOutputVector)->y,vResult); |
||
| 9056 | vResult = _mm_shuffle_ps(vResult,vResult,_MM_SHUFFLE(0,3,2,1)); |
||
| 9057 | _mm_store_ss(&reinterpret_cast<XMFLOAT3 *>(pOutputVector)->z,vResult); |
||
| 9058 | pInputVector += InputStride; |
||
| 9059 | pOutputVector += OutputStride; |
||
| 9060 | } |
||
| 9061 | |||
| 9062 | return pOutputStream; |
||
| 9063 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 9064 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9065 | } |
||
| 9066 | |||
| 9067 | //------------------------------------------------------------------------------ |
||
| 9068 | |||
| 9069 | XMINLINE XMVECTOR XMVector3Project |
||
| 9070 | ( |
||
| 9071 | FXMVECTOR V, |
||
| 9072 | FLOAT ViewportX, |
||
| 9073 | FLOAT ViewportY, |
||
| 9074 | FLOAT ViewportWidth, |
||
| 9075 | FLOAT ViewportHeight, |
||
| 9076 | FLOAT ViewportMinZ, |
||
| 9077 | FLOAT ViewportMaxZ, |
||
| 9078 | CXMMATRIX Projection, |
||
| 9079 | CXMMATRIX View, |
||
| 9080 | CXMMATRIX World |
||
| 9081 | ) |
||
| 9082 | { |
||
| 9083 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9084 | |||
| 9085 | XMMATRIX Transform; |
||
| 9086 | XMVECTOR Scale; |
||
| 9087 | XMVECTOR Offset; |
||
| 9088 | XMVECTOR Result; |
||
| 9089 | FLOAT HalfViewportWidth = ViewportWidth * 0.5f; |
||
| 9090 | FLOAT HalfViewportHeight = ViewportHeight * 0.5f; |
||
| 9091 | |||
| 9092 | Scale = XMVectorSet(HalfViewportWidth, |
||
| 9093 | -HalfViewportHeight, |
||
| 9094 | ViewportMaxZ - ViewportMinZ, |
||
| 9095 | 0.0f); |
||
| 9096 | |||
| 9097 | Offset = XMVectorSet(ViewportX + HalfViewportWidth, |
||
| 9098 | ViewportY + HalfViewportHeight, |
||
| 9099 | ViewportMinZ, |
||
| 9100 | 0.0f); |
||
| 9101 | |||
| 9102 | Transform = XMMatrixMultiply(World, View); |
||
| 9103 | Transform = XMMatrixMultiply(Transform, Projection); |
||
| 9104 | |||
| 9105 | Result = XMVector3TransformCoord(V, Transform); |
||
| 9106 | |||
| 9107 | Result = XMVectorMultiplyAdd(Result, Scale, Offset); |
||
| 9108 | |||
| 9109 | return Result; |
||
| 9110 | |||
| 9111 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9112 | XMMATRIX Transform; |
||
| 9113 | XMVECTOR Scale; |
||
| 9114 | XMVECTOR Offset; |
||
| 9115 | XMVECTOR Result; |
||
| 9116 | FLOAT HalfViewportWidth = ViewportWidth * 0.5f; |
||
| 9117 | FLOAT HalfViewportHeight = ViewportHeight * 0.5f; |
||
| 9118 | |||
| 9119 | Scale = XMVectorSet(HalfViewportWidth, |
||
| 9120 | -HalfViewportHeight, |
||
| 9121 | ViewportMaxZ - ViewportMinZ, |
||
| 9122 | 0.0f); |
||
| 9123 | |||
| 9124 | Offset = XMVectorSet(ViewportX + HalfViewportWidth, |
||
| 9125 | ViewportY + HalfViewportHeight, |
||
| 9126 | ViewportMinZ, |
||
| 9127 | 0.0f); |
||
| 9128 | Transform = XMMatrixMultiply(World, View); |
||
| 9129 | Transform = XMMatrixMultiply(Transform, Projection); |
||
| 9130 | Result = XMVector3TransformCoord(V, Transform); |
||
| 9131 | Result = _mm_mul_ps(Result,Scale); |
||
| 9132 | Result = _mm_add_ps(Result,Offset); |
||
| 9133 | return Result; |
||
| 9134 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9135 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9136 | } |
||
| 9137 | |||
| 9138 | //------------------------------------------------------------------------------ |
||
| 9139 | |||
| 9140 | XMINLINE XMFLOAT3* XMVector3ProjectStream |
||
| 9141 | ( |
||
| 9142 | XMFLOAT3* pOutputStream, |
||
| 9143 | UINT OutputStride, |
||
| 9144 | CONST XMFLOAT3* pInputStream, |
||
| 9145 | UINT InputStride, |
||
| 9146 | UINT VectorCount, |
||
| 9147 | FLOAT ViewportX, |
||
| 9148 | FLOAT ViewportY, |
||
| 9149 | FLOAT ViewportWidth, |
||
| 9150 | FLOAT ViewportHeight, |
||
| 9151 | FLOAT ViewportMinZ, |
||
| 9152 | FLOAT ViewportMaxZ, |
||
| 9153 | CXMMATRIX Projection, |
||
| 9154 | CXMMATRIX View, |
||
| 9155 | CXMMATRIX World |
||
| 9156 | ) |
||
| 9157 | { |
||
| 9158 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9159 | |||
| 9160 | XMMATRIX Transform; |
||
| 9161 | XMVECTOR V; |
||
| 9162 | XMVECTOR Scale; |
||
| 9163 | XMVECTOR Offset; |
||
| 9164 | XMVECTOR Result; |
||
| 9165 | UINT i; |
||
| 9166 | FLOAT HalfViewportWidth = ViewportWidth * 0.5f; |
||
| 9167 | FLOAT HalfViewportHeight = ViewportHeight * 0.5f; |
||
| 9168 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 9169 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 9170 | |||
| 9171 | XMASSERT(pOutputStream); |
||
| 9172 | XMASSERT(pInputStream); |
||
| 9173 | |||
| 9174 | Scale = XMVectorSet(HalfViewportWidth, |
||
| 9175 | -HalfViewportHeight, |
||
| 9176 | ViewportMaxZ - ViewportMinZ, |
||
| 9177 | 1.0f); |
||
| 9178 | |||
| 9179 | Offset = XMVectorSet(ViewportX + HalfViewportWidth, |
||
| 9180 | ViewportY + HalfViewportHeight, |
||
| 9181 | ViewportMinZ, |
||
| 9182 | 0.0f); |
||
| 9183 | |||
| 9184 | Transform = XMMatrixMultiply(World, View); |
||
| 9185 | Transform = XMMatrixMultiply(Transform, Projection); |
||
| 9186 | |||
| 9187 | for (i = 0; i < VectorCount; i++) |
||
| 9188 | { |
||
| 9189 | V = XMLoadFloat3((XMFLOAT3*)pInputVector); |
||
| 9190 | |||
| 9191 | Result = XMVector3TransformCoord(V, Transform); |
||
| 9192 | |||
| 9193 | Result = XMVectorMultiplyAdd(Result, Scale, Offset); |
||
| 9194 | |||
| 9195 | XMStoreFloat3((XMFLOAT3*)pOutputVector, Result); |
||
| 9196 | |||
| 9197 | pInputVector += InputStride; |
||
| 9198 | pOutputVector += OutputStride; |
||
| 9199 | } |
||
| 9200 | |||
| 9201 | return pOutputStream; |
||
| 9202 | |||
| 9203 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9204 | XMASSERT(pOutputStream); |
||
| 9205 | XMASSERT(pInputStream); |
||
| 9206 | XMMATRIX Transform; |
||
| 9207 | XMVECTOR V; |
||
| 9208 | XMVECTOR Scale; |
||
| 9209 | XMVECTOR Offset; |
||
| 9210 | XMVECTOR Result; |
||
| 9211 | UINT i; |
||
| 9212 | FLOAT HalfViewportWidth = ViewportWidth * 0.5f; |
||
| 9213 | FLOAT HalfViewportHeight = ViewportHeight * 0.5f; |
||
| 9214 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 9215 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 9216 | |||
| 9217 | Scale = XMVectorSet(HalfViewportWidth, |
||
| 9218 | -HalfViewportHeight, |
||
| 9219 | ViewportMaxZ - ViewportMinZ, |
||
| 9220 | 1.0f); |
||
| 9221 | |||
| 9222 | Offset = XMVectorSet(ViewportX + HalfViewportWidth, |
||
| 9223 | ViewportY + HalfViewportHeight, |
||
| 9224 | ViewportMinZ, |
||
| 9225 | 0.0f); |
||
| 9226 | |||
| 9227 | Transform = XMMatrixMultiply(World, View); |
||
| 9228 | Transform = XMMatrixMultiply(Transform, Projection); |
||
| 9229 | |||
| 9230 | for (i = 0; i < VectorCount; i++) |
||
| 9231 | { |
||
| 9232 | V = XMLoadFloat3((XMFLOAT3*)pInputVector); |
||
| 9233 | |||
| 9234 | Result = XMVector3TransformCoord(V, Transform); |
||
| 9235 | |||
| 9236 | Result = _mm_mul_ps(Result,Scale); |
||
| 9237 | Result = _mm_add_ps(Result,Offset); |
||
| 9238 | XMStoreFloat3((XMFLOAT3*)pOutputVector, Result); |
||
| 9239 | pInputVector += InputStride; |
||
| 9240 | pOutputVector += OutputStride; |
||
| 9241 | } |
||
| 9242 | return pOutputStream; |
||
| 9243 | |||
| 9244 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 9245 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9246 | } |
||
| 9247 | |||
| 9248 | //------------------------------------------------------------------------------ |
||
| 9249 | |||
| 9250 | XMFINLINE XMVECTOR XMVector3Unproject |
||
| 9251 | ( |
||
| 9252 | FXMVECTOR V, |
||
| 9253 | FLOAT ViewportX, |
||
| 9254 | FLOAT ViewportY, |
||
| 9255 | FLOAT ViewportWidth, |
||
| 9256 | FLOAT ViewportHeight, |
||
| 9257 | FLOAT ViewportMinZ, |
||
| 9258 | FLOAT ViewportMaxZ, |
||
| 9259 | CXMMATRIX Projection, |
||
| 9260 | CXMMATRIX View, |
||
| 9261 | CXMMATRIX World |
||
| 9262 | ) |
||
| 9263 | { |
||
| 9264 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9265 | |||
| 9266 | XMMATRIX Transform; |
||
| 9267 | XMVECTOR Scale; |
||
| 9268 | XMVECTOR Offset; |
||
| 9269 | XMVECTOR Determinant; |
||
| 9270 | XMVECTOR Result; |
||
| 9271 | CONST XMVECTOR D = XMVectorSet(-1.0f, 1.0f, 0.0f, 0.0f); |
||
| 9272 | |||
| 9273 | Scale = XMVectorSet(ViewportWidth * 0.5f, |
||
| 9274 | -ViewportHeight * 0.5f, |
||
| 9275 | ViewportMaxZ - ViewportMinZ, |
||
| 9276 | 1.0f); |
||
| 9277 | Scale = XMVectorReciprocal(Scale); |
||
| 9278 | |||
| 9279 | Offset = XMVectorSet(-ViewportX, |
||
| 9280 | -ViewportY, |
||
| 9281 | -ViewportMinZ, |
||
| 9282 | 0.0f); |
||
| 9283 | Offset = XMVectorMultiplyAdd(Scale, Offset, D); |
||
| 9284 | |||
| 9285 | Transform = XMMatrixMultiply(World, View); |
||
| 9286 | Transform = XMMatrixMultiply(Transform, Projection); |
||
| 9287 | Transform = XMMatrixInverse(&Determinant, Transform); |
||
| 9288 | |||
| 9289 | Result = XMVectorMultiplyAdd(V, Scale, Offset); |
||
| 9290 | |||
| 9291 | Result = XMVector3TransformCoord(Result, Transform); |
||
| 9292 | |||
| 9293 | return Result; |
||
| 9294 | |||
| 9295 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9296 | XMMATRIX Transform; |
||
| 9297 | XMVECTOR Scale; |
||
| 9298 | XMVECTOR Offset; |
||
| 9299 | XMVECTOR Determinant; |
||
| 9300 | XMVECTOR Result; |
||
| 9301 | CONST XMVECTORF32 D = {-1.0f, 1.0f, 0.0f, 0.0f}; |
||
| 9302 | |||
| 9303 | Scale = XMVectorSet(ViewportWidth * 0.5f, |
||
| 9304 | -ViewportHeight * 0.5f, |
||
| 9305 | ViewportMaxZ - ViewportMinZ, |
||
| 9306 | 1.0f); |
||
| 9307 | Scale = XMVectorReciprocal(Scale); |
||
| 9308 | |||
| 9309 | Offset = XMVectorSet(-ViewportX, |
||
| 9310 | -ViewportY, |
||
| 9311 | -ViewportMinZ, |
||
| 9312 | 0.0f); |
||
| 9313 | Offset = _mm_mul_ps(Offset,Scale); |
||
| 9314 | Offset = _mm_add_ps(Offset,D); |
||
| 9315 | |||
| 9316 | Transform = XMMatrixMultiply(World, View); |
||
| 9317 | Transform = XMMatrixMultiply(Transform, Projection); |
||
| 9318 | Transform = XMMatrixInverse(&Determinant, Transform); |
||
| 9319 | |||
| 9320 | Result = _mm_mul_ps(V,Scale); |
||
| 9321 | Result = _mm_add_ps(Result,Offset); |
||
| 9322 | |||
| 9323 | Result = XMVector3TransformCoord(Result, Transform); |
||
| 9324 | |||
| 9325 | return Result; |
||
| 9326 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9327 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9328 | } |
||
| 9329 | |||
| 9330 | //------------------------------------------------------------------------------ |
||
| 9331 | |||
| 9332 | XMINLINE XMFLOAT3* XMVector3UnprojectStream |
||
| 9333 | ( |
||
| 9334 | XMFLOAT3* pOutputStream, |
||
| 9335 | UINT OutputStride, |
||
| 9336 | CONST XMFLOAT3* pInputStream, |
||
| 9337 | UINT InputStride, |
||
| 9338 | UINT VectorCount, |
||
| 9339 | FLOAT ViewportX, |
||
| 9340 | FLOAT ViewportY, |
||
| 9341 | FLOAT ViewportWidth, |
||
| 9342 | FLOAT ViewportHeight, |
||
| 9343 | FLOAT ViewportMinZ, |
||
| 9344 | FLOAT ViewportMaxZ, |
||
| 9345 | CXMMATRIX Projection, |
||
| 9346 | CXMMATRIX View, |
||
| 9347 | CXMMATRIX World) |
||
| 9348 | { |
||
| 9349 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9350 | |||
| 9351 | XMMATRIX Transform; |
||
| 9352 | XMVECTOR Scale; |
||
| 9353 | XMVECTOR Offset; |
||
| 9354 | XMVECTOR V; |
||
| 9355 | XMVECTOR Determinant; |
||
| 9356 | XMVECTOR Result; |
||
| 9357 | UINT i; |
||
| 9358 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 9359 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 9360 | CONST XMVECTOR D = XMVectorSet(-1.0f, 1.0f, 0.0f, 0.0f); |
||
| 9361 | |||
| 9362 | XMASSERT(pOutputStream); |
||
| 9363 | XMASSERT(pInputStream); |
||
| 9364 | |||
| 9365 | Scale = XMVectorSet(ViewportWidth * 0.5f, |
||
| 9366 | -ViewportHeight * 0.5f, |
||
| 9367 | ViewportMaxZ - ViewportMinZ, |
||
| 9368 | 1.0f); |
||
| 9369 | Scale = XMVectorReciprocal(Scale); |
||
| 9370 | |||
| 9371 | Offset = XMVectorSet(-ViewportX, |
||
| 9372 | -ViewportY, |
||
| 9373 | -ViewportMinZ, |
||
| 9374 | 0.0f); |
||
| 9375 | Offset = XMVectorMultiplyAdd(Scale, Offset, D); |
||
| 9376 | |||
| 9377 | Transform = XMMatrixMultiply(World, View); |
||
| 9378 | Transform = XMMatrixMultiply(Transform, Projection); |
||
| 9379 | Transform = XMMatrixInverse(&Determinant, Transform); |
||
| 9380 | |||
| 9381 | for (i = 0; i < VectorCount; i++) |
||
| 9382 | { |
||
| 9383 | V = XMLoadFloat3((XMFLOAT3*)pInputVector); |
||
| 9384 | |||
| 9385 | Result = XMVectorMultiplyAdd(V, Scale, Offset); |
||
| 9386 | |||
| 9387 | Result = XMVector3TransformCoord(Result, Transform); |
||
| 9388 | |||
| 9389 | XMStoreFloat3((XMFLOAT3*)pOutputVector, Result); |
||
| 9390 | |||
| 9391 | pInputVector += InputStride; |
||
| 9392 | pOutputVector += OutputStride; |
||
| 9393 | } |
||
| 9394 | |||
| 9395 | return pOutputStream; |
||
| 9396 | |||
| 9397 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9398 | XMASSERT(pOutputStream); |
||
| 9399 | XMASSERT(pInputStream); |
||
| 9400 | XMMATRIX Transform; |
||
| 9401 | XMVECTOR Scale; |
||
| 9402 | XMVECTOR Offset; |
||
| 9403 | XMVECTOR V; |
||
| 9404 | XMVECTOR Determinant; |
||
| 9405 | XMVECTOR Result; |
||
| 9406 | UINT i; |
||
| 9407 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 9408 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 9409 | CONST XMVECTORF32 D = {-1.0f, 1.0f, 0.0f, 0.0f}; |
||
| 9410 | |||
| 9411 | Scale = XMVectorSet(ViewportWidth * 0.5f, |
||
| 9412 | -ViewportHeight * 0.5f, |
||
| 9413 | ViewportMaxZ - ViewportMinZ, |
||
| 9414 | 1.0f); |
||
| 9415 | Scale = XMVectorReciprocal(Scale); |
||
| 9416 | |||
| 9417 | Offset = XMVectorSet(-ViewportX, |
||
| 9418 | -ViewportY, |
||
| 9419 | -ViewportMinZ, |
||
| 9420 | 0.0f); |
||
| 9421 | Offset = _mm_mul_ps(Offset,Scale); |
||
| 9422 | Offset = _mm_add_ps(Offset,D); |
||
| 9423 | |||
| 9424 | Transform = XMMatrixMultiply(World, View); |
||
| 9425 | Transform = XMMatrixMultiply(Transform, Projection); |
||
| 9426 | Transform = XMMatrixInverse(&Determinant, Transform); |
||
| 9427 | |||
| 9428 | for (i = 0; i < VectorCount; i++) |
||
| 9429 | { |
||
| 9430 | V = XMLoadFloat3((XMFLOAT3*)pInputVector); |
||
| 9431 | |||
| 9432 | Result = XMVectorMultiplyAdd(V, Scale, Offset); |
||
| 9433 | |||
| 9434 | Result = XMVector3TransformCoord(Result, Transform); |
||
| 9435 | |||
| 9436 | XMStoreFloat3((XMFLOAT3*)pOutputVector, Result); |
||
| 9437 | |||
| 9438 | pInputVector += InputStride; |
||
| 9439 | pOutputVector += OutputStride; |
||
| 9440 | } |
||
| 9441 | |||
| 9442 | return pOutputStream; |
||
| 9443 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 9444 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9445 | } |
||
| 9446 | |||
| 9447 | /**************************************************************************** |
||
| 9448 | * |
||
| 9449 | * 4D Vector |
||
| 9450 | * |
||
| 9451 | ****************************************************************************/ |
||
| 9452 | |||
| 9453 | //------------------------------------------------------------------------------ |
||
| 9454 | // Comparison operations |
||
| 9455 | //------------------------------------------------------------------------------ |
||
| 9456 | |||
| 9457 | //------------------------------------------------------------------------------ |
||
| 9458 | |||
| 9459 | XMFINLINE BOOL XMVector4Equal |
||
| 9460 | ( |
||
| 9461 | FXMVECTOR V1, |
||
| 9462 | FXMVECTOR V2 |
||
| 9463 | ) |
||
| 9464 | { |
||
| 9465 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9466 | return (((V1.vector4_f32[0] == V2.vector4_f32[0]) && (V1.vector4_f32[1] == V2.vector4_f32[1]) && (V1.vector4_f32[2] == V2.vector4_f32[2]) && (V1.vector4_f32[3] == V2.vector4_f32[3])) != 0); |
||
| 9467 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9468 | XMVECTOR vTemp = _mm_cmpeq_ps(V1,V2); |
||
| 9469 | return ((_mm_movemask_ps(vTemp)==0x0f) != 0); |
||
| 9470 | #else |
||
| 9471 | return XMComparisonAllTrue(XMVector4EqualR(V1, V2)); |
||
| 9472 | #endif |
||
| 9473 | } |
||
| 9474 | |||
| 9475 | //------------------------------------------------------------------------------ |
||
| 9476 | |||
| 9477 | XMFINLINE UINT XMVector4EqualR |
||
| 9478 | ( |
||
| 9479 | FXMVECTOR V1, |
||
| 9480 | FXMVECTOR V2 |
||
| 9481 | ) |
||
| 9482 | { |
||
| 9483 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9484 | |||
| 9485 | UINT CR = 0; |
||
| 9486 | |||
| 9487 | if ((V1.vector4_f32[0] == V2.vector4_f32[0]) && |
||
| 9488 | (V1.vector4_f32[1] == V2.vector4_f32[1]) && |
||
| 9489 | (V1.vector4_f32[2] == V2.vector4_f32[2]) && |
||
| 9490 | (V1.vector4_f32[3] == V2.vector4_f32[3])) |
||
| 9491 | { |
||
| 9492 | CR = XM_CRMASK_CR6TRUE; |
||
| 9493 | } |
||
| 9494 | else if ((V1.vector4_f32[0] != V2.vector4_f32[0]) && |
||
| 9495 | (V1.vector4_f32[1] != V2.vector4_f32[1]) && |
||
| 9496 | (V1.vector4_f32[2] != V2.vector4_f32[2]) && |
||
| 9497 | (V1.vector4_f32[3] != V2.vector4_f32[3])) |
||
| 9498 | { |
||
| 9499 | CR = XM_CRMASK_CR6FALSE; |
||
| 9500 | } |
||
| 9501 | return CR; |
||
| 9502 | |||
| 9503 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9504 | XMVECTOR vTemp = _mm_cmpeq_ps(V1,V2); |
||
| 9505 | int iTest = _mm_movemask_ps(vTemp); |
||
| 9506 | UINT CR = 0; |
||
| 9507 | if (iTest==0xf) // All equal? |
||
| 9508 | { |
||
| 9509 | CR = XM_CRMASK_CR6TRUE; |
||
| 9510 | } |
||
| 9511 | else if (iTest==0) // All not equal? |
||
| 9512 | { |
||
| 9513 | CR = XM_CRMASK_CR6FALSE; |
||
| 9514 | } |
||
| 9515 | return CR; |
||
| 9516 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9517 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9518 | } |
||
| 9519 | |||
| 9520 | //------------------------------------------------------------------------------ |
||
| 9521 | |||
| 9522 | XMFINLINE BOOL XMVector4EqualInt |
||
| 9523 | ( |
||
| 9524 | FXMVECTOR V1, |
||
| 9525 | FXMVECTOR V2 |
||
| 9526 | ) |
||
| 9527 | { |
||
| 9528 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9529 | return (((V1.vector4_u32[0] == V2.vector4_u32[0]) && (V1.vector4_u32[1] == V2.vector4_u32[1]) && (V1.vector4_u32[2] == V2.vector4_u32[2]) && (V1.vector4_u32[3] == V2.vector4_u32[3])) != 0); |
||
| 9530 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9531 | __m128i vTemp = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0]); |
||
| 9532 | return ((_mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTemp)[0])==0xf) != 0); |
||
| 9533 | #else |
||
| 9534 | return XMComparisonAllTrue(XMVector4EqualIntR(V1, V2)); |
||
| 9535 | #endif |
||
| 9536 | } |
||
| 9537 | |||
| 9538 | //------------------------------------------------------------------------------ |
||
| 9539 | |||
| 9540 | XMFINLINE UINT XMVector4EqualIntR |
||
| 9541 | ( |
||
| 9542 | FXMVECTOR V1, |
||
| 9543 | FXMVECTOR V2 |
||
| 9544 | ) |
||
| 9545 | { |
||
| 9546 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9547 | UINT CR = 0; |
||
| 9548 | if (V1.vector4_u32[0] == V2.vector4_u32[0] && |
||
| 9549 | V1.vector4_u32[1] == V2.vector4_u32[1] && |
||
| 9550 | V1.vector4_u32[2] == V2.vector4_u32[2] && |
||
| 9551 | V1.vector4_u32[3] == V2.vector4_u32[3]) |
||
| 9552 | { |
||
| 9553 | CR = XM_CRMASK_CR6TRUE; |
||
| 9554 | } |
||
| 9555 | else if (V1.vector4_u32[0] != V2.vector4_u32[0] && |
||
| 9556 | V1.vector4_u32[1] != V2.vector4_u32[1] && |
||
| 9557 | V1.vector4_u32[2] != V2.vector4_u32[2] && |
||
| 9558 | V1.vector4_u32[3] != V2.vector4_u32[3]) |
||
| 9559 | { |
||
| 9560 | CR = XM_CRMASK_CR6FALSE; |
||
| 9561 | } |
||
| 9562 | return CR; |
||
| 9563 | |||
| 9564 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9565 | __m128i vTemp = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0]); |
||
| 9566 | int iTest = _mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTemp)[0]); |
||
| 9567 | UINT CR = 0; |
||
| 9568 | if (iTest==0xf) // All equal? |
||
| 9569 | { |
||
| 9570 | CR = XM_CRMASK_CR6TRUE; |
||
| 9571 | } |
||
| 9572 | else if (iTest==0) // All not equal? |
||
| 9573 | { |
||
| 9574 | CR = XM_CRMASK_CR6FALSE; |
||
| 9575 | } |
||
| 9576 | return CR; |
||
| 9577 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9578 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9579 | } |
||
| 9580 | |||
| 9581 | XMFINLINE BOOL XMVector4NearEqual |
||
| 9582 | ( |
||
| 9583 | FXMVECTOR V1, |
||
| 9584 | FXMVECTOR V2, |
||
| 9585 | FXMVECTOR Epsilon |
||
| 9586 | ) |
||
| 9587 | { |
||
| 9588 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9589 | FLOAT dx, dy, dz, dw; |
||
| 9590 | |||
| 9591 | dx = fabsf(V1.vector4_f32[0]-V2.vector4_f32[0]); |
||
| 9592 | dy = fabsf(V1.vector4_f32[1]-V2.vector4_f32[1]); |
||
| 9593 | dz = fabsf(V1.vector4_f32[2]-V2.vector4_f32[2]); |
||
| 9594 | dw = fabsf(V1.vector4_f32[3]-V2.vector4_f32[3]); |
||
| 9595 | return (((dx <= Epsilon.vector4_f32[0]) && |
||
| 9596 | (dy <= Epsilon.vector4_f32[1]) && |
||
| 9597 | (dz <= Epsilon.vector4_f32[2]) && |
||
| 9598 | (dw <= Epsilon.vector4_f32[3])) != 0); |
||
| 9599 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9600 | // Get the difference |
||
| 9601 | XMVECTOR vDelta = _mm_sub_ps(V1,V2); |
||
| 9602 | // Get the absolute value of the difference |
||
| 9603 | XMVECTOR vTemp = _mm_setzero_ps(); |
||
| 9604 | vTemp = _mm_sub_ps(vTemp,vDelta); |
||
| 9605 | vTemp = _mm_max_ps(vTemp,vDelta); |
||
| 9606 | vTemp = _mm_cmple_ps(vTemp,Epsilon); |
||
| 9607 | return ((_mm_movemask_ps(vTemp)==0xf) != 0); |
||
| 9608 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9609 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9610 | } |
||
| 9611 | |||
| 9612 | //------------------------------------------------------------------------------ |
||
| 9613 | |||
| 9614 | XMFINLINE BOOL XMVector4NotEqual |
||
| 9615 | ( |
||
| 9616 | FXMVECTOR V1, |
||
| 9617 | FXMVECTOR V2 |
||
| 9618 | ) |
||
| 9619 | { |
||
| 9620 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9621 | return (((V1.vector4_f32[0] != V2.vector4_f32[0]) || (V1.vector4_f32[1] != V2.vector4_f32[1]) || (V1.vector4_f32[2] != V2.vector4_f32[2]) || (V1.vector4_f32[3] != V2.vector4_f32[3])) != 0); |
||
| 9622 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9623 | XMVECTOR vTemp = _mm_cmpneq_ps(V1,V2); |
||
| 9624 | return ((_mm_movemask_ps(vTemp)) != 0); |
||
| 9625 | #else |
||
| 9626 | return XMComparisonAnyFalse(XMVector4EqualR(V1, V2)); |
||
| 9627 | #endif |
||
| 9628 | } |
||
| 9629 | |||
| 9630 | //------------------------------------------------------------------------------ |
||
| 9631 | |||
| 9632 | XMFINLINE BOOL XMVector4NotEqualInt |
||
| 9633 | ( |
||
| 9634 | FXMVECTOR V1, |
||
| 9635 | FXMVECTOR V2 |
||
| 9636 | ) |
||
| 9637 | { |
||
| 9638 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9639 | return (((V1.vector4_u32[0] != V2.vector4_u32[0]) || (V1.vector4_u32[1] != V2.vector4_u32[1]) || (V1.vector4_u32[2] != V2.vector4_u32[2]) || (V1.vector4_u32[3] != V2.vector4_u32[3])) != 0); |
||
| 9640 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9641 | __m128i vTemp = _mm_cmpeq_epi32(reinterpret_cast<const __m128i *>(&V1)[0],reinterpret_cast<const __m128i *>(&V2)[0]); |
||
| 9642 | return ((_mm_movemask_ps(reinterpret_cast<const __m128 *>(&vTemp)[0])!=0xF) != 0); |
||
| 9643 | #else |
||
| 9644 | return XMComparisonAnyFalse(XMVector4EqualIntR(V1, V2)); |
||
| 9645 | #endif |
||
| 9646 | } |
||
| 9647 | |||
| 9648 | //------------------------------------------------------------------------------ |
||
| 9649 | |||
| 9650 | XMFINLINE BOOL XMVector4Greater |
||
| 9651 | ( |
||
| 9652 | FXMVECTOR V1, |
||
| 9653 | FXMVECTOR V2 |
||
| 9654 | ) |
||
| 9655 | { |
||
| 9656 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9657 | return (((V1.vector4_f32[0] > V2.vector4_f32[0]) && (V1.vector4_f32[1] > V2.vector4_f32[1]) && (V1.vector4_f32[2] > V2.vector4_f32[2]) && (V1.vector4_f32[3] > V2.vector4_f32[3])) != 0); |
||
| 9658 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9659 | XMVECTOR vTemp = _mm_cmpgt_ps(V1,V2); |
||
| 9660 | return ((_mm_movemask_ps(vTemp)==0x0f) != 0); |
||
| 9661 | #else |
||
| 9662 | return XMComparisonAllTrue(XMVector4GreaterR(V1, V2)); |
||
| 9663 | #endif |
||
| 9664 | } |
||
| 9665 | |||
| 9666 | //------------------------------------------------------------------------------ |
||
| 9667 | |||
| 9668 | XMFINLINE UINT XMVector4GreaterR |
||
| 9669 | ( |
||
| 9670 | FXMVECTOR V1, |
||
| 9671 | FXMVECTOR V2 |
||
| 9672 | ) |
||
| 9673 | { |
||
| 9674 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9675 | UINT CR = 0; |
||
| 9676 | if (V1.vector4_f32[0] > V2.vector4_f32[0] && |
||
| 9677 | V1.vector4_f32[1] > V2.vector4_f32[1] && |
||
| 9678 | V1.vector4_f32[2] > V2.vector4_f32[2] && |
||
| 9679 | V1.vector4_f32[3] > V2.vector4_f32[3]) |
||
| 9680 | { |
||
| 9681 | CR = XM_CRMASK_CR6TRUE; |
||
| 9682 | } |
||
| 9683 | else if (V1.vector4_f32[0] <= V2.vector4_f32[0] && |
||
| 9684 | V1.vector4_f32[1] <= V2.vector4_f32[1] && |
||
| 9685 | V1.vector4_f32[2] <= V2.vector4_f32[2] && |
||
| 9686 | V1.vector4_f32[3] <= V2.vector4_f32[3]) |
||
| 9687 | { |
||
| 9688 | CR = XM_CRMASK_CR6FALSE; |
||
| 9689 | } |
||
| 9690 | return CR; |
||
| 9691 | |||
| 9692 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9693 | UINT CR = 0; |
||
| 9694 | XMVECTOR vTemp = _mm_cmpgt_ps(V1,V2); |
||
| 9695 | int iTest = _mm_movemask_ps(vTemp); |
||
| 9696 | if (iTest==0xf) { |
||
| 9697 | CR = XM_CRMASK_CR6TRUE; |
||
| 9698 | } |
||
| 9699 | else if (!iTest) |
||
| 9700 | { |
||
| 9701 | CR = XM_CRMASK_CR6FALSE; |
||
| 9702 | } |
||
| 9703 | return CR; |
||
| 9704 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9705 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9706 | } |
||
| 9707 | |||
| 9708 | //------------------------------------------------------------------------------ |
||
| 9709 | |||
| 9710 | XMFINLINE BOOL XMVector4GreaterOrEqual |
||
| 9711 | ( |
||
| 9712 | FXMVECTOR V1, |
||
| 9713 | FXMVECTOR V2 |
||
| 9714 | ) |
||
| 9715 | { |
||
| 9716 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9717 | return (((V1.vector4_f32[0] >= V2.vector4_f32[0]) && (V1.vector4_f32[1] >= V2.vector4_f32[1]) && (V1.vector4_f32[2] >= V2.vector4_f32[2]) && (V1.vector4_f32[3] >= V2.vector4_f32[3])) != 0); |
||
| 9718 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9719 | XMVECTOR vTemp = _mm_cmpge_ps(V1,V2); |
||
| 9720 | return ((_mm_movemask_ps(vTemp)==0x0f) != 0); |
||
| 9721 | #else |
||
| 9722 | return XMComparisonAllTrue(XMVector4GreaterOrEqualR(V1, V2)); |
||
| 9723 | #endif |
||
| 9724 | } |
||
| 9725 | |||
| 9726 | //------------------------------------------------------------------------------ |
||
| 9727 | |||
| 9728 | XMFINLINE UINT XMVector4GreaterOrEqualR |
||
| 9729 | ( |
||
| 9730 | FXMVECTOR V1, |
||
| 9731 | FXMVECTOR V2 |
||
| 9732 | ) |
||
| 9733 | { |
||
| 9734 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9735 | UINT CR = 0; |
||
| 9736 | if ((V1.vector4_f32[0] >= V2.vector4_f32[0]) && |
||
| 9737 | (V1.vector4_f32[1] >= V2.vector4_f32[1]) && |
||
| 9738 | (V1.vector4_f32[2] >= V2.vector4_f32[2]) && |
||
| 9739 | (V1.vector4_f32[3] >= V2.vector4_f32[3])) |
||
| 9740 | { |
||
| 9741 | CR = XM_CRMASK_CR6TRUE; |
||
| 9742 | } |
||
| 9743 | else if ((V1.vector4_f32[0] < V2.vector4_f32[0]) && |
||
| 9744 | (V1.vector4_f32[1] < V2.vector4_f32[1]) && |
||
| 9745 | (V1.vector4_f32[2] < V2.vector4_f32[2]) && |
||
| 9746 | (V1.vector4_f32[3] < V2.vector4_f32[3])) |
||
| 9747 | { |
||
| 9748 | CR = XM_CRMASK_CR6FALSE; |
||
| 9749 | } |
||
| 9750 | return CR; |
||
| 9751 | |||
| 9752 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9753 | UINT CR = 0; |
||
| 9754 | XMVECTOR vTemp = _mm_cmpge_ps(V1,V2); |
||
| 9755 | int iTest = _mm_movemask_ps(vTemp); |
||
| 9756 | if (iTest==0x0f) |
||
| 9757 | { |
||
| 9758 | CR = XM_CRMASK_CR6TRUE; |
||
| 9759 | } |
||
| 9760 | else if (!iTest) |
||
| 9761 | { |
||
| 9762 | CR = XM_CRMASK_CR6FALSE; |
||
| 9763 | } |
||
| 9764 | return CR; |
||
| 9765 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9766 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9767 | } |
||
| 9768 | |||
| 9769 | //------------------------------------------------------------------------------ |
||
| 9770 | |||
| 9771 | XMFINLINE BOOL XMVector4Less |
||
| 9772 | ( |
||
| 9773 | FXMVECTOR V1, |
||
| 9774 | FXMVECTOR V2 |
||
| 9775 | ) |
||
| 9776 | { |
||
| 9777 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9778 | return (((V1.vector4_f32[0] < V2.vector4_f32[0]) && (V1.vector4_f32[1] < V2.vector4_f32[1]) && (V1.vector4_f32[2] < V2.vector4_f32[2]) && (V1.vector4_f32[3] < V2.vector4_f32[3])) != 0); |
||
| 9779 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9780 | XMVECTOR vTemp = _mm_cmplt_ps(V1,V2); |
||
| 9781 | return ((_mm_movemask_ps(vTemp)==0x0f) != 0); |
||
| 9782 | #else |
||
| 9783 | return XMComparisonAllTrue(XMVector4GreaterR(V2, V1)); |
||
| 9784 | #endif |
||
| 9785 | } |
||
| 9786 | |||
| 9787 | //------------------------------------------------------------------------------ |
||
| 9788 | |||
| 9789 | XMFINLINE BOOL XMVector4LessOrEqual |
||
| 9790 | ( |
||
| 9791 | FXMVECTOR V1, |
||
| 9792 | FXMVECTOR V2 |
||
| 9793 | ) |
||
| 9794 | { |
||
| 9795 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9796 | return (((V1.vector4_f32[0] <= V2.vector4_f32[0]) && (V1.vector4_f32[1] <= V2.vector4_f32[1]) && (V1.vector4_f32[2] <= V2.vector4_f32[2]) && (V1.vector4_f32[3] <= V2.vector4_f32[3])) != 0); |
||
| 9797 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9798 | XMVECTOR vTemp = _mm_cmple_ps(V1,V2); |
||
| 9799 | return ((_mm_movemask_ps(vTemp)==0x0f) != 0); |
||
| 9800 | #else |
||
| 9801 | return XMComparisonAllTrue(XMVector4GreaterOrEqualR(V2, V1)); |
||
| 9802 | #endif |
||
| 9803 | } |
||
| 9804 | |||
| 9805 | //------------------------------------------------------------------------------ |
||
| 9806 | |||
| 9807 | XMFINLINE BOOL XMVector4InBounds |
||
| 9808 | ( |
||
| 9809 | FXMVECTOR V, |
||
| 9810 | FXMVECTOR Bounds |
||
| 9811 | ) |
||
| 9812 | { |
||
| 9813 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9814 | return (((V.vector4_f32[0] <= Bounds.vector4_f32[0] && V.vector4_f32[0] >= -Bounds.vector4_f32[0]) && |
||
| 9815 | (V.vector4_f32[1] <= Bounds.vector4_f32[1] && V.vector4_f32[1] >= -Bounds.vector4_f32[1]) && |
||
| 9816 | (V.vector4_f32[2] <= Bounds.vector4_f32[2] && V.vector4_f32[2] >= -Bounds.vector4_f32[2]) && |
||
| 9817 | (V.vector4_f32[3] <= Bounds.vector4_f32[3] && V.vector4_f32[3] >= -Bounds.vector4_f32[3])) != 0); |
||
| 9818 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9819 | // Test if less than or equal |
||
| 9820 | XMVECTOR vTemp1 = _mm_cmple_ps(V,Bounds); |
||
| 9821 | // Negate the bounds |
||
| 9822 | XMVECTOR vTemp2 = _mm_mul_ps(Bounds,g_XMNegativeOne); |
||
| 9823 | // Test if greater or equal (Reversed) |
||
| 9824 | vTemp2 = _mm_cmple_ps(vTemp2,V); |
||
| 9825 | // Blend answers |
||
| 9826 | vTemp1 = _mm_and_ps(vTemp1,vTemp2); |
||
| 9827 | // All in bounds? |
||
| 9828 | return ((_mm_movemask_ps(vTemp1)==0x0f) != 0); |
||
| 9829 | #else |
||
| 9830 | return XMComparisonAllInBounds(XMVector4InBoundsR(V, Bounds)); |
||
| 9831 | #endif |
||
| 9832 | } |
||
| 9833 | |||
| 9834 | //------------------------------------------------------------------------------ |
||
| 9835 | |||
| 9836 | XMFINLINE UINT XMVector4InBoundsR |
||
| 9837 | ( |
||
| 9838 | FXMVECTOR V, |
||
| 9839 | FXMVECTOR Bounds |
||
| 9840 | ) |
||
| 9841 | { |
||
| 9842 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9843 | |||
| 9844 | UINT CR = 0; |
||
| 9845 | if ((V.vector4_f32[0] <= Bounds.vector4_f32[0] && V.vector4_f32[0] >= -Bounds.vector4_f32[0]) && |
||
| 9846 | (V.vector4_f32[1] <= Bounds.vector4_f32[1] && V.vector4_f32[1] >= -Bounds.vector4_f32[1]) && |
||
| 9847 | (V.vector4_f32[2] <= Bounds.vector4_f32[2] && V.vector4_f32[2] >= -Bounds.vector4_f32[2]) && |
||
| 9848 | (V.vector4_f32[3] <= Bounds.vector4_f32[3] && V.vector4_f32[3] >= -Bounds.vector4_f32[3])) |
||
| 9849 | { |
||
| 9850 | CR = XM_CRMASK_CR6BOUNDS; |
||
| 9851 | } |
||
| 9852 | return CR; |
||
| 9853 | |||
| 9854 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9855 | // Test if less than or equal |
||
| 9856 | XMVECTOR vTemp1 = _mm_cmple_ps(V,Bounds); |
||
| 9857 | // Negate the bounds |
||
| 9858 | XMVECTOR vTemp2 = _mm_mul_ps(Bounds,g_XMNegativeOne); |
||
| 9859 | // Test if greater or equal (Reversed) |
||
| 9860 | vTemp2 = _mm_cmple_ps(vTemp2,V); |
||
| 9861 | // Blend answers |
||
| 9862 | vTemp1 = _mm_and_ps(vTemp1,vTemp2); |
||
| 9863 | // All in bounds? |
||
| 9864 | return (_mm_movemask_ps(vTemp1)==0x0f) ? XM_CRMASK_CR6BOUNDS : 0; |
||
| 9865 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9866 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9867 | } |
||
| 9868 | |||
| 9869 | //------------------------------------------------------------------------------ |
||
| 9870 | |||
| 9871 | XMFINLINE BOOL XMVector4IsNaN |
||
| 9872 | ( |
||
| 9873 | FXMVECTOR V |
||
| 9874 | ) |
||
| 9875 | { |
||
| 9876 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9877 | return (XMISNAN(V.vector4_f32[0]) || |
||
| 9878 | XMISNAN(V.vector4_f32[1]) || |
||
| 9879 | XMISNAN(V.vector4_f32[2]) || |
||
| 9880 | XMISNAN(V.vector4_f32[3])); |
||
| 9881 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9882 | // Test against itself. NaN is always not equal |
||
| 9883 | XMVECTOR vTempNan = _mm_cmpneq_ps(V,V); |
||
| 9884 | // If any are NaN, the mask is non-zero |
||
| 9885 | return (_mm_movemask_ps(vTempNan)!=0); |
||
| 9886 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9887 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9888 | } |
||
| 9889 | |||
| 9890 | //------------------------------------------------------------------------------ |
||
| 9891 | |||
| 9892 | XMFINLINE BOOL XMVector4IsInfinite |
||
| 9893 | ( |
||
| 9894 | FXMVECTOR V |
||
| 9895 | ) |
||
| 9896 | { |
||
| 9897 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9898 | |||
| 9899 | return (XMISINF(V.vector4_f32[0]) || |
||
| 9900 | XMISINF(V.vector4_f32[1]) || |
||
| 9901 | XMISINF(V.vector4_f32[2]) || |
||
| 9902 | XMISINF(V.vector4_f32[3])); |
||
| 9903 | |||
| 9904 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9905 | // Mask off the sign bit |
||
| 9906 | XMVECTOR vTemp = _mm_and_ps(V,g_XMAbsMask); |
||
| 9907 | // Compare to infinity |
||
| 9908 | vTemp = _mm_cmpeq_ps(vTemp,g_XMInfinity); |
||
| 9909 | // If any are infinity, the signs are true. |
||
| 9910 | return (_mm_movemask_ps(vTemp) != 0); |
||
| 9911 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9912 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9913 | } |
||
| 9914 | |||
| 9915 | //------------------------------------------------------------------------------ |
||
| 9916 | // Computation operations |
||
| 9917 | //------------------------------------------------------------------------------ |
||
| 9918 | |||
| 9919 | //------------------------------------------------------------------------------ |
||
| 9920 | |||
| 9921 | XMFINLINE XMVECTOR XMVector4Dot |
||
| 9922 | ( |
||
| 9923 | FXMVECTOR V1, |
||
| 9924 | FXMVECTOR V2 |
||
| 9925 | ) |
||
| 9926 | { |
||
| 9927 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9928 | |||
| 9929 | XMVECTOR Result; |
||
| 9930 | |||
| 9931 | Result.vector4_f32[0] = |
||
| 9932 | Result.vector4_f32[1] = |
||
| 9933 | Result.vector4_f32[2] = |
||
| 9934 | Result.vector4_f32[3] = V1.vector4_f32[0] * V2.vector4_f32[0] + V1.vector4_f32[1] * V2.vector4_f32[1] + V1.vector4_f32[2] * V2.vector4_f32[2] + V1.vector4_f32[3] * V2.vector4_f32[3]; |
||
| 9935 | |||
| 9936 | return Result; |
||
| 9937 | |||
| 9938 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9939 | XMVECTOR vTemp2 = V2; |
||
| 9940 | XMVECTOR vTemp = _mm_mul_ps(V1,vTemp2); |
||
| 9941 | vTemp2 = _mm_shuffle_ps(vTemp2,vTemp,_MM_SHUFFLE(1,0,0,0)); // Copy X to the Z position and Y to the W position |
||
| 9942 | vTemp2 = _mm_add_ps(vTemp2,vTemp); // Add Z = X+Z; W = Y+W; |
||
| 9943 | vTemp = _mm_shuffle_ps(vTemp,vTemp2,_MM_SHUFFLE(0,3,0,0)); // Copy W to the Z position |
||
| 9944 | vTemp = _mm_add_ps(vTemp,vTemp2); // Add Z and W together |
||
| 9945 | return _mm_shuffle_ps(vTemp,vTemp,_MM_SHUFFLE(2,2,2,2)); // Splat Z and return |
||
| 9946 | #else // _XM_VMX128_INTRINSICS_ |
||
| 9947 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 9948 | } |
||
| 9949 | |||
| 9950 | //------------------------------------------------------------------------------ |
||
| 9951 | |||
| 9952 | XMFINLINE XMVECTOR XMVector4Cross |
||
| 9953 | ( |
||
| 9954 | FXMVECTOR V1, |
||
| 9955 | FXMVECTOR V2, |
||
| 9956 | FXMVECTOR V3 |
||
| 9957 | ) |
||
| 9958 | { |
||
| 9959 | #if defined(_XM_NO_INTRINSICS_) |
||
| 9960 | XMVECTOR Result; |
||
| 9961 | |||
| 9962 | Result.vector4_f32[0] = (((V2.vector4_f32[2]*V3.vector4_f32[3])-(V2.vector4_f32[3]*V3.vector4_f32[2]))*V1.vector4_f32[1])-(((V2.vector4_f32[1]*V3.vector4_f32[3])-(V2.vector4_f32[3]*V3.vector4_f32[1]))*V1.vector4_f32[2])+(((V2.vector4_f32[1]*V3.vector4_f32[2])-(V2.vector4_f32[2]*V3.vector4_f32[1]))*V1.vector4_f32[3]); |
||
| 9963 | Result.vector4_f32[1] = (((V2.vector4_f32[3]*V3.vector4_f32[2])-(V2.vector4_f32[2]*V3.vector4_f32[3]))*V1.vector4_f32[0])-(((V2.vector4_f32[3]*V3.vector4_f32[0])-(V2.vector4_f32[0]*V3.vector4_f32[3]))*V1.vector4_f32[2])+(((V2.vector4_f32[2]*V3.vector4_f32[0])-(V2.vector4_f32[0]*V3.vector4_f32[2]))*V1.vector4_f32[3]); |
||
| 9964 | Result.vector4_f32[2] = (((V2.vector4_f32[1]*V3.vector4_f32[3])-(V2.vector4_f32[3]*V3.vector4_f32[1]))*V1.vector4_f32[0])-(((V2.vector4_f32[0]*V3.vector4_f32[3])-(V2.vector4_f32[3]*V3.vector4_f32[0]))*V1.vector4_f32[1])+(((V2.vector4_f32[0]*V3.vector4_f32[1])-(V2.vector4_f32[1]*V3.vector4_f32[0]))*V1.vector4_f32[3]); |
||
| 9965 | Result.vector4_f32[3] = (((V2.vector4_f32[2]*V3.vector4_f32[1])-(V2.vector4_f32[1]*V3.vector4_f32[2]))*V1.vector4_f32[0])-(((V2.vector4_f32[2]*V3.vector4_f32[0])-(V2.vector4_f32[0]*V3.vector4_f32[2]))*V1.vector4_f32[1])+(((V2.vector4_f32[1]*V3.vector4_f32[0])-(V2.vector4_f32[0]*V3.vector4_f32[1]))*V1.vector4_f32[2]); |
||
| 9966 | return Result; |
||
| 9967 | |||
| 9968 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 9969 | // V2zwyz * V3wzwy |
||
| 9970 | XMVECTOR vResult = _mm_shuffle_ps(V2,V2,_MM_SHUFFLE(2,1,3,2)); |
||
| 9971 | XMVECTOR vTemp3 = _mm_shuffle_ps(V3,V3,_MM_SHUFFLE(1,3,2,3)); |
||
| 9972 | vResult = _mm_mul_ps(vResult,vTemp3); |
||
| 9973 | // - V2wzwy * V3zwyz |
||
| 9974 | XMVECTOR vTemp2 = _mm_shuffle_ps(V2,V2,_MM_SHUFFLE(1,3,2,3)); |
||
| 9975 | vTemp3 = _mm_shuffle_ps(vTemp3,vTemp3,_MM_SHUFFLE(1,3,0,1)); |
||
| 9976 | vTemp2 = _mm_mul_ps(vTemp2,vTemp3); |
||
| 9977 | vResult = _mm_sub_ps(vResult,vTemp2); |
||
| 9978 | // term1 * V1yxxx |
||
| 9979 | XMVECTOR vTemp1 = _mm_shuffle_ps(V1,V1,_MM_SHUFFLE(0,0,0,1)); |
||
| 9980 | vResult = _mm_mul_ps(vResult,vTemp1); |
||
| 9981 | |||
| 9982 | // V2ywxz * V3wxwx |
||
| 9983 | vTemp2 = _mm_shuffle_ps(V2,V2,_MM_SHUFFLE(2,0,3,1)); |
||
| 9984 | vTemp3 = _mm_shuffle_ps(V3,V3,_MM_SHUFFLE(0,3,0,3)); |
||
| 9985 | vTemp3 = _mm_mul_ps(vTemp3,vTemp2); |
||
| 9986 | // - V2wxwx * V3ywxz |
||
| 9987 | vTemp2 = _mm_shuffle_ps(vTemp2,vTemp2,_MM_SHUFFLE(2,1,2,1)); |
||
| 9988 | vTemp1 = _mm_shuffle_ps(V3,V3,_MM_SHUFFLE(2,0,3,1)); |
||
| 9989 | vTemp2 = _mm_mul_ps(vTemp2,vTemp1); |
||
| 9990 | vTemp3 = _mm_sub_ps(vTemp3,vTemp2); |
||
| 9991 | // vResult - temp * V1zzyy |
||
| 9992 | vTemp1 = _mm_shuffle_ps(V1,V1,_MM_SHUFFLE(1,1,2,2)); |
||
| 9993 | vTemp1 = _mm_mul_ps(vTemp1,vTemp3); |
||
| 9994 | vResult = _mm_sub_ps(vResult,vTemp1); |
||
| 9995 | |||
| 9996 | // V2yzxy * V3zxyx |
||
| 9997 | vTemp2 = _mm_shuffle_ps(V2,V2,_MM_SHUFFLE(1,0,2,1)); |
||
| 9998 | vTemp3 = _mm_shuffle_ps(V3,V3,_MM_SHUFFLE(0,1,0,2)); |
||
| 9999 | vTemp3 = _mm_mul_ps(vTemp3,vTemp2); |
||
| 10000 | // - V2zxyx * V3yzxy |
||
| 10001 | vTemp2 = _mm_shuffle_ps(vTemp2,vTemp2,_MM_SHUFFLE(2,0,2,1)); |
||
| 10002 | vTemp1 = _mm_shuffle_ps(V3,V3,_MM_SHUFFLE(1,0,2,1)); |
||
| 10003 | vTemp1 = _mm_mul_ps(vTemp1,vTemp2); |
||
| 10004 | vTemp3 = _mm_sub_ps(vTemp3,vTemp1); |
||
| 10005 | // vResult + term * V1wwwz |
||
| 10006 | vTemp1 = _mm_shuffle_ps(V1,V1,_MM_SHUFFLE(2,3,3,3)); |
||
| 10007 | vTemp3 = _mm_mul_ps(vTemp3,vTemp1); |
||
| 10008 | vResult = _mm_add_ps(vResult,vTemp3); |
||
| 10009 | return vResult; |
||
| 10010 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10011 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10012 | } |
||
| 10013 | |||
| 10014 | //------------------------------------------------------------------------------ |
||
| 10015 | |||
| 10016 | XMFINLINE XMVECTOR XMVector4LengthSq |
||
| 10017 | ( |
||
| 10018 | FXMVECTOR V |
||
| 10019 | ) |
||
| 10020 | { |
||
| 10021 | return XMVector4Dot(V, V); |
||
| 10022 | } |
||
| 10023 | |||
| 10024 | //------------------------------------------------------------------------------ |
||
| 10025 | |||
| 10026 | XMFINLINE XMVECTOR XMVector4ReciprocalLengthEst |
||
| 10027 | ( |
||
| 10028 | FXMVECTOR V |
||
| 10029 | ) |
||
| 10030 | { |
||
| 10031 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10032 | |||
| 10033 | XMVECTOR Result; |
||
| 10034 | |||
| 10035 | Result = XMVector4LengthSq(V); |
||
| 10036 | Result = XMVectorReciprocalSqrtEst(Result); |
||
| 10037 | |||
| 10038 | return Result; |
||
| 10039 | |||
| 10040 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10041 | // Perform the dot product on x,y,z and w |
||
| 10042 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 10043 | // vTemp has z and w |
||
| 10044 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(3,2,3,2)); |
||
| 10045 | // x+z, y+w |
||
| 10046 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10047 | // x+z,x+z,x+z,y+w |
||
| 10048 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,0,0,0)); |
||
| 10049 | // ??,??,y+w,y+w |
||
| 10050 | vTemp = _mm_shuffle_ps(vTemp,vLengthSq,_MM_SHUFFLE(3,3,0,0)); |
||
| 10051 | // ??,??,x+z+y+w,?? |
||
| 10052 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10053 | // Splat the length |
||
| 10054 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(2,2,2,2)); |
||
| 10055 | // Get the reciprocal |
||
| 10056 | vLengthSq = _mm_rsqrt_ps(vLengthSq); |
||
| 10057 | return vLengthSq; |
||
| 10058 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10059 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10060 | } |
||
| 10061 | |||
| 10062 | //------------------------------------------------------------------------------ |
||
| 10063 | |||
| 10064 | XMFINLINE XMVECTOR XMVector4ReciprocalLength |
||
| 10065 | ( |
||
| 10066 | FXMVECTOR V |
||
| 10067 | ) |
||
| 10068 | { |
||
| 10069 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10070 | |||
| 10071 | XMVECTOR Result; |
||
| 10072 | |||
| 10073 | Result = XMVector4LengthSq(V); |
||
| 10074 | Result = XMVectorReciprocalSqrt(Result); |
||
| 10075 | |||
| 10076 | return Result; |
||
| 10077 | |||
| 10078 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10079 | // Perform the dot product on x,y,z and w |
||
| 10080 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 10081 | // vTemp has z and w |
||
| 10082 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(3,2,3,2)); |
||
| 10083 | // x+z, y+w |
||
| 10084 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10085 | // x+z,x+z,x+z,y+w |
||
| 10086 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,0,0,0)); |
||
| 10087 | // ??,??,y+w,y+w |
||
| 10088 | vTemp = _mm_shuffle_ps(vTemp,vLengthSq,_MM_SHUFFLE(3,3,0,0)); |
||
| 10089 | // ??,??,x+z+y+w,?? |
||
| 10090 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10091 | // Splat the length |
||
| 10092 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(2,2,2,2)); |
||
| 10093 | // Get the reciprocal |
||
| 10094 | vLengthSq = _mm_sqrt_ps(vLengthSq); |
||
| 10095 | // Accurate! |
||
| 10096 | vLengthSq = _mm_div_ps(g_XMOne,vLengthSq); |
||
| 10097 | return vLengthSq; |
||
| 10098 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10099 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10100 | } |
||
| 10101 | |||
| 10102 | //------------------------------------------------------------------------------ |
||
| 10103 | |||
| 10104 | XMFINLINE XMVECTOR XMVector4LengthEst |
||
| 10105 | ( |
||
| 10106 | FXMVECTOR V |
||
| 10107 | ) |
||
| 10108 | { |
||
| 10109 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10110 | |||
| 10111 | XMVECTOR Result; |
||
| 10112 | |||
| 10113 | Result = XMVector4LengthSq(V); |
||
| 10114 | Result = XMVectorSqrtEst(Result); |
||
| 10115 | |||
| 10116 | return Result; |
||
| 10117 | |||
| 10118 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10119 | // Perform the dot product on x,y,z and w |
||
| 10120 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 10121 | // vTemp has z and w |
||
| 10122 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(3,2,3,2)); |
||
| 10123 | // x+z, y+w |
||
| 10124 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10125 | // x+z,x+z,x+z,y+w |
||
| 10126 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,0,0,0)); |
||
| 10127 | // ??,??,y+w,y+w |
||
| 10128 | vTemp = _mm_shuffle_ps(vTemp,vLengthSq,_MM_SHUFFLE(3,3,0,0)); |
||
| 10129 | // ??,??,x+z+y+w,?? |
||
| 10130 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10131 | // Splat the length |
||
| 10132 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(2,2,2,2)); |
||
| 10133 | // Prepare for the division |
||
| 10134 | vLengthSq = _mm_sqrt_ps(vLengthSq); |
||
| 10135 | return vLengthSq; |
||
| 10136 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10137 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10138 | } |
||
| 10139 | |||
| 10140 | //------------------------------------------------------------------------------ |
||
| 10141 | |||
| 10142 | XMFINLINE XMVECTOR XMVector4Length |
||
| 10143 | ( |
||
| 10144 | FXMVECTOR V |
||
| 10145 | ) |
||
| 10146 | { |
||
| 10147 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10148 | |||
| 10149 | XMVECTOR Result; |
||
| 10150 | |||
| 10151 | Result = XMVector4LengthSq(V); |
||
| 10152 | Result = XMVectorSqrt(Result); |
||
| 10153 | |||
| 10154 | return Result; |
||
| 10155 | |||
| 10156 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10157 | // Perform the dot product on x,y,z and w |
||
| 10158 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 10159 | // vTemp has z and w |
||
| 10160 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(3,2,3,2)); |
||
| 10161 | // x+z, y+w |
||
| 10162 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10163 | // x+z,x+z,x+z,y+w |
||
| 10164 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,0,0,0)); |
||
| 10165 | // ??,??,y+w,y+w |
||
| 10166 | vTemp = _mm_shuffle_ps(vTemp,vLengthSq,_MM_SHUFFLE(3,3,0,0)); |
||
| 10167 | // ??,??,x+z+y+w,?? |
||
| 10168 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10169 | // Splat the length |
||
| 10170 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(2,2,2,2)); |
||
| 10171 | // Prepare for the division |
||
| 10172 | vLengthSq = _mm_sqrt_ps(vLengthSq); |
||
| 10173 | return vLengthSq; |
||
| 10174 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10175 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10176 | } |
||
| 10177 | |||
| 10178 | //------------------------------------------------------------------------------ |
||
| 10179 | // XMVector4NormalizeEst uses a reciprocal estimate and |
||
| 10180 | // returns QNaN on zero and infinite vectors. |
||
| 10181 | |||
| 10182 | XMFINLINE XMVECTOR XMVector4NormalizeEst |
||
| 10183 | ( |
||
| 10184 | FXMVECTOR V |
||
| 10185 | ) |
||
| 10186 | { |
||
| 10187 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10188 | |||
| 10189 | XMVECTOR Result; |
||
| 10190 | Result = XMVector4ReciprocalLength(V); |
||
| 10191 | Result = XMVectorMultiply(V, Result); |
||
| 10192 | return Result; |
||
| 10193 | |||
| 10194 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10195 | // Perform the dot product on x,y,z and w |
||
| 10196 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 10197 | // vTemp has z and w |
||
| 10198 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(3,2,3,2)); |
||
| 10199 | // x+z, y+w |
||
| 10200 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10201 | // x+z,x+z,x+z,y+w |
||
| 10202 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,0,0,0)); |
||
| 10203 | // ??,??,y+w,y+w |
||
| 10204 | vTemp = _mm_shuffle_ps(vTemp,vLengthSq,_MM_SHUFFLE(3,3,0,0)); |
||
| 10205 | // ??,??,x+z+y+w,?? |
||
| 10206 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10207 | // Splat the length |
||
| 10208 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(2,2,2,2)); |
||
| 10209 | // Prepare for the division |
||
| 10210 | XMVECTOR vResult = _mm_rsqrt_ps(vLengthSq); |
||
| 10211 | // Failsafe on zero (Or epsilon) length planes |
||
| 10212 | // If the length is infinity, set the elements to zero |
||
| 10213 | vLengthSq = _mm_cmpneq_ps(vLengthSq,g_XMInfinity); |
||
| 10214 | // Reciprocal mul to perform the normalization |
||
| 10215 | vResult = _mm_mul_ps(vResult,V); |
||
| 10216 | // Any that are infinity, set to zero |
||
| 10217 | vResult = _mm_and_ps(vResult,vLengthSq); |
||
| 10218 | return vResult; |
||
| 10219 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10220 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10221 | } |
||
| 10222 | |||
| 10223 | //------------------------------------------------------------------------------ |
||
| 10224 | |||
| 10225 | XMFINLINE XMVECTOR XMVector4Normalize |
||
| 10226 | ( |
||
| 10227 | FXMVECTOR V |
||
| 10228 | ) |
||
| 10229 | { |
||
| 10230 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10231 | |||
| 10232 | XMVECTOR LengthSq; |
||
| 10233 | XMVECTOR Zero; |
||
| 10234 | XMVECTOR InfiniteLength; |
||
| 10235 | XMVECTOR ZeroLength; |
||
| 10236 | XMVECTOR Select; |
||
| 10237 | XMVECTOR Result; |
||
| 10238 | |||
| 10239 | LengthSq = XMVector4LengthSq(V); |
||
| 10240 | Zero = XMVectorZero(); |
||
| 10241 | Result = XMVectorReciprocalSqrt(LengthSq); |
||
| 10242 | InfiniteLength = XMVectorEqualInt(LengthSq, g_XMInfinity.v); |
||
| 10243 | ZeroLength = XMVectorEqual(LengthSq, Zero); |
||
| 10244 | Result = XMVectorMultiply(V, Result); |
||
| 10245 | Select = XMVectorEqualInt(InfiniteLength, ZeroLength); |
||
| 10246 | Result = XMVectorSelect(LengthSq, Result, Select); |
||
| 10247 | |||
| 10248 | return Result; |
||
| 10249 | |||
| 10250 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10251 | // Perform the dot product on x,y,z and w |
||
| 10252 | XMVECTOR vLengthSq = _mm_mul_ps(V,V); |
||
| 10253 | // vTemp has z and w |
||
| 10254 | XMVECTOR vTemp = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(3,2,3,2)); |
||
| 10255 | // x+z, y+w |
||
| 10256 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10257 | // x+z,x+z,x+z,y+w |
||
| 10258 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(1,0,0,0)); |
||
| 10259 | // ??,??,y+w,y+w |
||
| 10260 | vTemp = _mm_shuffle_ps(vTemp,vLengthSq,_MM_SHUFFLE(3,3,0,0)); |
||
| 10261 | // ??,??,x+z+y+w,?? |
||
| 10262 | vLengthSq = _mm_add_ps(vLengthSq,vTemp); |
||
| 10263 | // Splat the length |
||
| 10264 | vLengthSq = _mm_shuffle_ps(vLengthSq,vLengthSq,_MM_SHUFFLE(2,2,2,2)); |
||
| 10265 | // Prepare for the division |
||
| 10266 | XMVECTOR vResult = _mm_sqrt_ps(vLengthSq); |
||
| 10267 | // Failsafe on zero (Or epsilon) length planes |
||
| 10268 | // If the length is infinity, set the elements to zero |
||
| 10269 | vLengthSq = _mm_cmpneq_ps(vLengthSq,g_XMInfinity); |
||
| 10270 | // Divide to perform the normalization |
||
| 10271 | vResult = _mm_div_ps(V,vResult); |
||
| 10272 | // Any that are infinity, set to zero |
||
| 10273 | vResult = _mm_and_ps(vResult,vLengthSq); |
||
| 10274 | return vResult; |
||
| 10275 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10276 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10277 | } |
||
| 10278 | |||
| 10279 | //------------------------------------------------------------------------------ |
||
| 10280 | |||
| 10281 | XMFINLINE XMVECTOR XMVector4ClampLength |
||
| 10282 | ( |
||
| 10283 | FXMVECTOR V, |
||
| 10284 | FLOAT LengthMin, |
||
| 10285 | FLOAT LengthMax |
||
| 10286 | ) |
||
| 10287 | { |
||
| 10288 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10289 | |||
| 10290 | XMVECTOR ClampMax; |
||
| 10291 | XMVECTOR ClampMin; |
||
| 10292 | |||
| 10293 | ClampMax = XMVectorReplicate(LengthMax); |
||
| 10294 | ClampMin = XMVectorReplicate(LengthMin); |
||
| 10295 | |||
| 10296 | return XMVector4ClampLengthV(V, ClampMin, ClampMax); |
||
| 10297 | |||
| 10298 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10299 | XMVECTOR ClampMax = _mm_set_ps1(LengthMax); |
||
| 10300 | XMVECTOR ClampMin = _mm_set_ps1(LengthMin); |
||
| 10301 | return XMVector4ClampLengthV(V, ClampMin, ClampMax); |
||
| 10302 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 10303 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10304 | } |
||
| 10305 | |||
| 10306 | //------------------------------------------------------------------------------ |
||
| 10307 | |||
| 10308 | XMFINLINE XMVECTOR XMVector4ClampLengthV |
||
| 10309 | ( |
||
| 10310 | FXMVECTOR V, |
||
| 10311 | FXMVECTOR LengthMin, |
||
| 10312 | FXMVECTOR LengthMax |
||
| 10313 | ) |
||
| 10314 | { |
||
| 10315 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10316 | |||
| 10317 | XMVECTOR ClampLength; |
||
| 10318 | XMVECTOR LengthSq; |
||
| 10319 | XMVECTOR RcpLength; |
||
| 10320 | XMVECTOR Length; |
||
| 10321 | XMVECTOR Normal; |
||
| 10322 | XMVECTOR Zero; |
||
| 10323 | XMVECTOR InfiniteLength; |
||
| 10324 | XMVECTOR ZeroLength; |
||
| 10325 | XMVECTOR Select; |
||
| 10326 | XMVECTOR ControlMax; |
||
| 10327 | XMVECTOR ControlMin; |
||
| 10328 | XMVECTOR Control; |
||
| 10329 | XMVECTOR Result; |
||
| 10330 | |||
| 10331 | XMASSERT((LengthMin.vector4_f32[1] == LengthMin.vector4_f32[0]) && (LengthMin.vector4_f32[2] == LengthMin.vector4_f32[0]) && (LengthMin.vector4_f32[3] == LengthMin.vector4_f32[0])); |
||
| 10332 | XMASSERT((LengthMax.vector4_f32[1] == LengthMax.vector4_f32[0]) && (LengthMax.vector4_f32[2] == LengthMax.vector4_f32[0]) && (LengthMax.vector4_f32[3] == LengthMax.vector4_f32[0])); |
||
| 10333 | XMASSERT(XMVector4GreaterOrEqual(LengthMin, XMVectorZero())); |
||
| 10334 | XMASSERT(XMVector4GreaterOrEqual(LengthMax, XMVectorZero())); |
||
| 10335 | XMASSERT(XMVector4GreaterOrEqual(LengthMax, LengthMin)); |
||
| 10336 | |||
| 10337 | LengthSq = XMVector4LengthSq(V); |
||
| 10338 | |||
| 10339 | Zero = XMVectorZero(); |
||
| 10340 | |||
| 10341 | RcpLength = XMVectorReciprocalSqrt(LengthSq); |
||
| 10342 | |||
| 10343 | InfiniteLength = XMVectorEqualInt(LengthSq, g_XMInfinity.v); |
||
| 10344 | ZeroLength = XMVectorEqual(LengthSq, Zero); |
||
| 10345 | |||
| 10346 | Normal = XMVectorMultiply(V, RcpLength); |
||
| 10347 | |||
| 10348 | Length = XMVectorMultiply(LengthSq, RcpLength); |
||
| 10349 | |||
| 10350 | Select = XMVectorEqualInt(InfiniteLength, ZeroLength); |
||
| 10351 | Length = XMVectorSelect(LengthSq, Length, Select); |
||
| 10352 | Normal = XMVectorSelect(LengthSq, Normal, Select); |
||
| 10353 | |||
| 10354 | ControlMax = XMVectorGreater(Length, LengthMax); |
||
| 10355 | ControlMin = XMVectorLess(Length, LengthMin); |
||
| 10356 | |||
| 10357 | ClampLength = XMVectorSelect(Length, LengthMax, ControlMax); |
||
| 10358 | ClampLength = XMVectorSelect(ClampLength, LengthMin, ControlMin); |
||
| 10359 | |||
| 10360 | Result = XMVectorMultiply(Normal, ClampLength); |
||
| 10361 | |||
| 10362 | // Preserve the original vector (with no precision loss) if the length falls within the given range |
||
| 10363 | Control = XMVectorEqualInt(ControlMax, ControlMin); |
||
| 10364 | Result = XMVectorSelect(Result, V, Control); |
||
| 10365 | |||
| 10366 | return Result; |
||
| 10367 | |||
| 10368 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10369 | XMVECTOR ClampLength; |
||
| 10370 | XMVECTOR LengthSq; |
||
| 10371 | XMVECTOR RcpLength; |
||
| 10372 | XMVECTOR Length; |
||
| 10373 | XMVECTOR Normal; |
||
| 10374 | XMVECTOR Zero; |
||
| 10375 | XMVECTOR InfiniteLength; |
||
| 10376 | XMVECTOR ZeroLength; |
||
| 10377 | XMVECTOR Select; |
||
| 10378 | XMVECTOR ControlMax; |
||
| 10379 | XMVECTOR ControlMin; |
||
| 10380 | XMVECTOR Control; |
||
| 10381 | XMVECTOR Result; |
||
| 10382 | |||
| 10383 | XMASSERT((XMVectorGetY(LengthMin) == XMVectorGetX(LengthMin)) && (XMVectorGetZ(LengthMin) == XMVectorGetX(LengthMin)) && (XMVectorGetW(LengthMin) == XMVectorGetX(LengthMin))); |
||
| 10384 | XMASSERT((XMVectorGetY(LengthMax) == XMVectorGetX(LengthMax)) && (XMVectorGetZ(LengthMax) == XMVectorGetX(LengthMax)) && (XMVectorGetW(LengthMax) == XMVectorGetX(LengthMax))); |
||
| 10385 | XMASSERT(XMVector4GreaterOrEqual(LengthMin, g_XMZero)); |
||
| 10386 | XMASSERT(XMVector4GreaterOrEqual(LengthMax, g_XMZero)); |
||
| 10387 | XMASSERT(XMVector4GreaterOrEqual(LengthMax, LengthMin)); |
||
| 10388 | |||
| 10389 | LengthSq = XMVector4LengthSq(V); |
||
| 10390 | Zero = XMVectorZero(); |
||
| 10391 | RcpLength = XMVectorReciprocalSqrt(LengthSq); |
||
| 10392 | InfiniteLength = XMVectorEqualInt(LengthSq, g_XMInfinity); |
||
| 10393 | ZeroLength = XMVectorEqual(LengthSq, Zero); |
||
| 10394 | Normal = _mm_mul_ps(V, RcpLength); |
||
| 10395 | Length = _mm_mul_ps(LengthSq, RcpLength); |
||
| 10396 | Select = XMVectorEqualInt(InfiniteLength, ZeroLength); |
||
| 10397 | Length = XMVectorSelect(LengthSq, Length, Select); |
||
| 10398 | Normal = XMVectorSelect(LengthSq, Normal, Select); |
||
| 10399 | ControlMax = XMVectorGreater(Length, LengthMax); |
||
| 10400 | ControlMin = XMVectorLess(Length, LengthMin); |
||
| 10401 | ClampLength = XMVectorSelect(Length, LengthMax, ControlMax); |
||
| 10402 | ClampLength = XMVectorSelect(ClampLength, LengthMin, ControlMin); |
||
| 10403 | Result = _mm_mul_ps(Normal, ClampLength); |
||
| 10404 | // Preserve the original vector (with no precision loss) if the length falls within the given range |
||
| 10405 | Control = XMVectorEqualInt(ControlMax,ControlMin); |
||
| 10406 | Result = XMVectorSelect(Result,V,Control); |
||
| 10407 | return Result; |
||
| 10408 | |||
| 10409 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10410 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10411 | } |
||
| 10412 | |||
| 10413 | //------------------------------------------------------------------------------ |
||
| 10414 | |||
| 10415 | XMFINLINE XMVECTOR XMVector4Reflect |
||
| 10416 | ( |
||
| 10417 | FXMVECTOR Incident, |
||
| 10418 | FXMVECTOR Normal |
||
| 10419 | ) |
||
| 10420 | { |
||
| 10421 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10422 | |||
| 10423 | XMVECTOR Result; |
||
| 10424 | |||
| 10425 | // Result = Incident - (2 * dot(Incident, Normal)) * Normal |
||
| 10426 | Result = XMVector4Dot(Incident, Normal); |
||
| 10427 | Result = XMVectorAdd(Result, Result); |
||
| 10428 | Result = XMVectorNegativeMultiplySubtract(Result, Normal, Incident); |
||
| 10429 | |||
| 10430 | return Result; |
||
| 10431 | |||
| 10432 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10433 | // Result = Incident - (2 * dot(Incident, Normal)) * Normal |
||
| 10434 | XMVECTOR Result = XMVector4Dot(Incident,Normal); |
||
| 10435 | Result = _mm_add_ps(Result,Result); |
||
| 10436 | Result = _mm_mul_ps(Result,Normal); |
||
| 10437 | Result = _mm_sub_ps(Incident,Result); |
||
| 10438 | return Result; |
||
| 10439 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10440 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10441 | } |
||
| 10442 | |||
| 10443 | //------------------------------------------------------------------------------ |
||
| 10444 | |||
| 10445 | XMFINLINE XMVECTOR XMVector4Refract |
||
| 10446 | ( |
||
| 10447 | FXMVECTOR Incident, |
||
| 10448 | FXMVECTOR Normal, |
||
| 10449 | FLOAT RefractionIndex |
||
| 10450 | ) |
||
| 10451 | { |
||
| 10452 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10453 | |||
| 10454 | XMVECTOR Index; |
||
| 10455 | Index = XMVectorReplicate(RefractionIndex); |
||
| 10456 | return XMVector4RefractV(Incident, Normal, Index); |
||
| 10457 | |||
| 10458 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10459 | XMVECTOR Index = _mm_set_ps1(RefractionIndex); |
||
| 10460 | return XMVector4RefractV(Incident,Normal,Index); |
||
| 10461 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 10462 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10463 | } |
||
| 10464 | |||
| 10465 | //------------------------------------------------------------------------------ |
||
| 10466 | |||
| 10467 | XMFINLINE XMVECTOR XMVector4RefractV |
||
| 10468 | ( |
||
| 10469 | FXMVECTOR Incident, |
||
| 10470 | FXMVECTOR Normal, |
||
| 10471 | FXMVECTOR RefractionIndex |
||
| 10472 | ) |
||
| 10473 | { |
||
| 10474 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10475 | |||
| 10476 | XMVECTOR IDotN; |
||
| 10477 | XMVECTOR R; |
||
| 10478 | CONST XMVECTOR Zero = XMVectorZero(); |
||
| 10479 | |||
| 10480 | // Result = RefractionIndex * Incident - Normal * (RefractionIndex * dot(Incident, Normal) + |
||
| 10481 | // sqrt(1 - RefractionIndex * RefractionIndex * (1 - dot(Incident, Normal) * dot(Incident, Normal)))) |
||
| 10482 | |||
| 10483 | IDotN = XMVector4Dot(Incident, Normal); |
||
| 10484 | |||
| 10485 | // R = 1.0f - RefractionIndex * RefractionIndex * (1.0f - IDotN * IDotN) |
||
| 10486 | R = XMVectorNegativeMultiplySubtract(IDotN, IDotN, g_XMOne.v); |
||
| 10487 | R = XMVectorMultiply(R, RefractionIndex); |
||
| 10488 | R = XMVectorNegativeMultiplySubtract(R, RefractionIndex, g_XMOne.v); |
||
| 10489 | |||
| 10490 | if (XMVector4LessOrEqual(R, Zero)) |
||
| 10491 | { |
||
| 10492 | // Total internal reflection |
||
| 10493 | return Zero; |
||
| 10494 | } |
||
| 10495 | else |
||
| 10496 | { |
||
| 10497 | XMVECTOR Result; |
||
| 10498 | |||
| 10499 | // R = RefractionIndex * IDotN + sqrt(R) |
||
| 10500 | R = XMVectorSqrt(R); |
||
| 10501 | R = XMVectorMultiplyAdd(RefractionIndex, IDotN, R); |
||
| 10502 | |||
| 10503 | // Result = RefractionIndex * Incident - Normal * R |
||
| 10504 | Result = XMVectorMultiply(RefractionIndex, Incident); |
||
| 10505 | Result = XMVectorNegativeMultiplySubtract(Normal, R, Result); |
||
| 10506 | |||
| 10507 | return Result; |
||
| 10508 | } |
||
| 10509 | |||
| 10510 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10511 | // Result = RefractionIndex * Incident - Normal * (RefractionIndex * dot(Incident, Normal) + |
||
| 10512 | // sqrt(1 - RefractionIndex * RefractionIndex * (1 - dot(Incident, Normal) * dot(Incident, Normal)))) |
||
| 10513 | |||
| 10514 | XMVECTOR IDotN = XMVector4Dot(Incident,Normal); |
||
| 10515 | |||
| 10516 | // R = 1.0f - RefractionIndex * RefractionIndex * (1.0f - IDotN * IDotN) |
||
| 10517 | XMVECTOR R = _mm_mul_ps(IDotN,IDotN); |
||
| 10518 | R = _mm_sub_ps(g_XMOne,R); |
||
| 10519 | R = _mm_mul_ps(R, RefractionIndex); |
||
| 10520 | R = _mm_mul_ps(R, RefractionIndex); |
||
| 10521 | R = _mm_sub_ps(g_XMOne,R); |
||
| 10522 | |||
| 10523 | XMVECTOR vResult = _mm_cmple_ps(R,g_XMZero); |
||
| 10524 | if (_mm_movemask_ps(vResult)==0x0f) |
||
| 10525 | { |
||
| 10526 | // Total internal reflection |
||
| 10527 | vResult = g_XMZero; |
||
| 10528 | } |
||
| 10529 | else |
||
| 10530 | { |
||
| 10531 | // R = RefractionIndex * IDotN + sqrt(R) |
||
| 10532 | R = _mm_sqrt_ps(R); |
||
| 10533 | vResult = _mm_mul_ps(RefractionIndex, IDotN); |
||
| 10534 | R = _mm_add_ps(R,vResult); |
||
| 10535 | // Result = RefractionIndex * Incident - Normal * R |
||
| 10536 | vResult = _mm_mul_ps(RefractionIndex, Incident); |
||
| 10537 | R = _mm_mul_ps(R,Normal); |
||
| 10538 | vResult = _mm_sub_ps(vResult,R); |
||
| 10539 | } |
||
| 10540 | return vResult; |
||
| 10541 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10542 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10543 | } |
||
| 10544 | |||
| 10545 | //------------------------------------------------------------------------------ |
||
| 10546 | |||
| 10547 | XMFINLINE XMVECTOR XMVector4Orthogonal |
||
| 10548 | ( |
||
| 10549 | FXMVECTOR V |
||
| 10550 | ) |
||
| 10551 | { |
||
| 10552 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10553 | |||
| 10554 | XMVECTOR Result; |
||
| 10555 | Result.vector4_f32[0] = V.vector4_f32[2]; |
||
| 10556 | Result.vector4_f32[1] = V.vector4_f32[3]; |
||
| 10557 | Result.vector4_f32[2] = -V.vector4_f32[0]; |
||
| 10558 | Result.vector4_f32[3] = -V.vector4_f32[1]; |
||
| 10559 | return Result; |
||
| 10560 | |||
| 10561 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10562 | static const XMVECTORF32 FlipZW = {1.0f,1.0f,-1.0f,-1.0f}; |
||
| 10563 | XMVECTOR vResult = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,0,3,2)); |
||
| 10564 | vResult = _mm_mul_ps(vResult,FlipZW); |
||
| 10565 | return vResult; |
||
| 10566 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10567 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10568 | } |
||
| 10569 | |||
| 10570 | //------------------------------------------------------------------------------ |
||
| 10571 | |||
| 10572 | XMFINLINE XMVECTOR XMVector4AngleBetweenNormalsEst |
||
| 10573 | ( |
||
| 10574 | FXMVECTOR N1, |
||
| 10575 | FXMVECTOR N2 |
||
| 10576 | ) |
||
| 10577 | { |
||
| 10578 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10579 | |||
| 10580 | XMVECTOR NegativeOne; |
||
| 10581 | XMVECTOR One; |
||
| 10582 | XMVECTOR Result; |
||
| 10583 | |||
| 10584 | Result = XMVector4Dot(N1, N2); |
||
| 10585 | NegativeOne = XMVectorSplatConstant(-1, 0); |
||
| 10586 | One = XMVectorSplatOne(); |
||
| 10587 | Result = XMVectorClamp(Result, NegativeOne, One); |
||
| 10588 | Result = XMVectorACosEst(Result); |
||
| 10589 | |||
| 10590 | return Result; |
||
| 10591 | |||
| 10592 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10593 | XMVECTOR vResult = XMVector4Dot(N1,N2); |
||
| 10594 | // Clamp to -1.0f to 1.0f |
||
| 10595 | vResult = _mm_max_ps(vResult,g_XMNegativeOne); |
||
| 10596 | vResult = _mm_min_ps(vResult,g_XMOne);; |
||
| 10597 | vResult = XMVectorACosEst(vResult); |
||
| 10598 | return vResult; |
||
| 10599 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10600 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10601 | } |
||
| 10602 | |||
| 10603 | //------------------------------------------------------------------------------ |
||
| 10604 | |||
| 10605 | XMFINLINE XMVECTOR XMVector4AngleBetweenNormals |
||
| 10606 | ( |
||
| 10607 | FXMVECTOR N1, |
||
| 10608 | FXMVECTOR N2 |
||
| 10609 | ) |
||
| 10610 | { |
||
| 10611 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10612 | |||
| 10613 | XMVECTOR NegativeOne; |
||
| 10614 | XMVECTOR One; |
||
| 10615 | XMVECTOR Result; |
||
| 10616 | |||
| 10617 | Result = XMVector4Dot(N1, N2); |
||
| 10618 | NegativeOne = XMVectorSplatConstant(-1, 0); |
||
| 10619 | One = XMVectorSplatOne(); |
||
| 10620 | Result = XMVectorClamp(Result, NegativeOne, One); |
||
| 10621 | Result = XMVectorACos(Result); |
||
| 10622 | |||
| 10623 | return Result; |
||
| 10624 | |||
| 10625 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10626 | XMVECTOR vResult = XMVector4Dot(N1,N2); |
||
| 10627 | // Clamp to -1.0f to 1.0f |
||
| 10628 | vResult = _mm_max_ps(vResult,g_XMNegativeOne); |
||
| 10629 | vResult = _mm_min_ps(vResult,g_XMOne);; |
||
| 10630 | vResult = XMVectorACos(vResult); |
||
| 10631 | return vResult; |
||
| 10632 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10633 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10634 | } |
||
| 10635 | |||
| 10636 | //------------------------------------------------------------------------------ |
||
| 10637 | |||
| 10638 | XMFINLINE XMVECTOR XMVector4AngleBetweenVectors |
||
| 10639 | ( |
||
| 10640 | FXMVECTOR V1, |
||
| 10641 | FXMVECTOR V2 |
||
| 10642 | ) |
||
| 10643 | { |
||
| 10644 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10645 | |||
| 10646 | XMVECTOR L1; |
||
| 10647 | XMVECTOR L2; |
||
| 10648 | XMVECTOR Dot; |
||
| 10649 | XMVECTOR CosAngle; |
||
| 10650 | XMVECTOR NegativeOne; |
||
| 10651 | XMVECTOR One; |
||
| 10652 | XMVECTOR Result; |
||
| 10653 | |||
| 10654 | L1 = XMVector4ReciprocalLength(V1); |
||
| 10655 | L2 = XMVector4ReciprocalLength(V2); |
||
| 10656 | |||
| 10657 | Dot = XMVector4Dot(V1, V2); |
||
| 10658 | |||
| 10659 | L1 = XMVectorMultiply(L1, L2); |
||
| 10660 | |||
| 10661 | CosAngle = XMVectorMultiply(Dot, L1); |
||
| 10662 | NegativeOne = XMVectorSplatConstant(-1, 0); |
||
| 10663 | One = XMVectorSplatOne(); |
||
| 10664 | CosAngle = XMVectorClamp(CosAngle, NegativeOne, One); |
||
| 10665 | |||
| 10666 | Result = XMVectorACos(CosAngle); |
||
| 10667 | |||
| 10668 | return Result; |
||
| 10669 | |||
| 10670 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10671 | XMVECTOR L1; |
||
| 10672 | XMVECTOR L2; |
||
| 10673 | XMVECTOR Dot; |
||
| 10674 | XMVECTOR CosAngle; |
||
| 10675 | XMVECTOR Result; |
||
| 10676 | |||
| 10677 | L1 = XMVector4ReciprocalLength(V1); |
||
| 10678 | L2 = XMVector4ReciprocalLength(V2); |
||
| 10679 | Dot = XMVector4Dot(V1, V2); |
||
| 10680 | L1 = _mm_mul_ps(L1,L2); |
||
| 10681 | CosAngle = _mm_mul_ps(Dot,L1); |
||
| 10682 | CosAngle = XMVectorClamp(CosAngle, g_XMNegativeOne, g_XMOne); |
||
| 10683 | Result = XMVectorACos(CosAngle); |
||
| 10684 | return Result; |
||
| 10685 | |||
| 10686 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10687 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10688 | } |
||
| 10689 | |||
| 10690 | //------------------------------------------------------------------------------ |
||
| 10691 | |||
| 10692 | XMFINLINE XMVECTOR XMVector4Transform |
||
| 10693 | ( |
||
| 10694 | FXMVECTOR V, |
||
| 10695 | CXMMATRIX M |
||
| 10696 | ) |
||
| 10697 | { |
||
| 10698 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10699 | FLOAT fX = (M.m[0][0]*V.vector4_f32[0])+(M.m[1][0]*V.vector4_f32[1])+(M.m[2][0]*V.vector4_f32[2])+(M.m[3][0]*V.vector4_f32[3]); |
||
| 10700 | FLOAT fY = (M.m[0][1]*V.vector4_f32[0])+(M.m[1][1]*V.vector4_f32[1])+(M.m[2][1]*V.vector4_f32[2])+(M.m[3][1]*V.vector4_f32[3]); |
||
| 10701 | FLOAT fZ = (M.m[0][2]*V.vector4_f32[0])+(M.m[1][2]*V.vector4_f32[1])+(M.m[2][2]*V.vector4_f32[2])+(M.m[3][2]*V.vector4_f32[3]); |
||
| 10702 | FLOAT fW = (M.m[0][3]*V.vector4_f32[0])+(M.m[1][3]*V.vector4_f32[1])+(M.m[2][3]*V.vector4_f32[2])+(M.m[3][3]*V.vector4_f32[3]); |
||
| 10703 | XMVECTOR vResult = { |
||
| 10704 | fX, |
||
| 10705 | fY, |
||
| 10706 | fZ, |
||
| 10707 | fW |
||
| 10708 | }; |
||
| 10709 | return vResult; |
||
| 10710 | |||
| 10711 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10712 | // Splat x,y,z and w |
||
| 10713 | XMVECTOR vTempX = _mm_shuffle_ps(V,V,_MM_SHUFFLE(0,0,0,0)); |
||
| 10714 | XMVECTOR vTempY = _mm_shuffle_ps(V,V,_MM_SHUFFLE(1,1,1,1)); |
||
| 10715 | XMVECTOR vTempZ = _mm_shuffle_ps(V,V,_MM_SHUFFLE(2,2,2,2)); |
||
| 10716 | XMVECTOR vTempW = _mm_shuffle_ps(V,V,_MM_SHUFFLE(3,3,3,3)); |
||
| 10717 | // Mul by the matrix |
||
| 10718 | vTempX = _mm_mul_ps(vTempX,M.r[0]); |
||
| 10719 | vTempY = _mm_mul_ps(vTempY,M.r[1]); |
||
| 10720 | vTempZ = _mm_mul_ps(vTempZ,M.r[2]); |
||
| 10721 | vTempW = _mm_mul_ps(vTempW,M.r[3]); |
||
| 10722 | // Add them all together |
||
| 10723 | vTempX = _mm_add_ps(vTempX,vTempY); |
||
| 10724 | vTempZ = _mm_add_ps(vTempZ,vTempW); |
||
| 10725 | vTempX = _mm_add_ps(vTempX,vTempZ); |
||
| 10726 | return vTempX; |
||
| 10727 | #else // _XM_VMX128_INTRINSICS_ |
||
| 10728 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10729 | } |
||
| 10730 | |||
| 10731 | //------------------------------------------------------------------------------ |
||
| 10732 | |||
| 10733 | XMINLINE XMFLOAT4* XMVector4TransformStream |
||
| 10734 | ( |
||
| 10735 | XMFLOAT4* pOutputStream, |
||
| 10736 | UINT OutputStride, |
||
| 10737 | CONST XMFLOAT4* pInputStream, |
||
| 10738 | UINT InputStride, |
||
| 10739 | UINT VectorCount, |
||
| 10740 | CXMMATRIX M |
||
| 10741 | ) |
||
| 10742 | { |
||
| 10743 | #if defined(_XM_NO_INTRINSICS_) |
||
| 10744 | |||
| 10745 | XMVECTOR V; |
||
| 10746 | XMVECTOR X; |
||
| 10747 | XMVECTOR Y; |
||
| 10748 | XMVECTOR Z; |
||
| 10749 | XMVECTOR W; |
||
| 10750 | XMVECTOR Result; |
||
| 10751 | UINT i; |
||
| 10752 | BYTE* pInputVector = (BYTE*)pInputStream; |
||
| 10753 | BYTE* pOutputVector = (BYTE*)pOutputStream; |
||
| 10754 | |||
| 10755 | XMASSERT(pOutputStream); |
||
| 10756 | XMASSERT(pInputStream); |
||
| 10757 | |||
| 10758 | for (i = 0; i < VectorCount; i++) |
||
| 10759 | { |
||
| 10760 | V = XMLoadFloat4((XMFLOAT4*)pInputVector); |
||
| 10761 | W = XMVectorSplatW(V); |
||
| 10762 | Z = XMVectorSplatZ(V); |
||
| 10763 | Y = XMVectorSplatY(V); |
||
| 10764 | X = XMVectorSplatX(V); |
||
| 10765 | // W = XMVectorReplicate(((XMFLOAT4*)pInputVector)->w); |
||
| 10766 | // Z = XMVectorReplicate(((XMFLOAT4*)pInputVector)->z); |
||
| 10767 | // Y = XMVectorReplicate(((XMFLOAT4*)pInputVector)->y); |
||
| 10768 | // X = XMVectorReplicate(((XMFLOAT4*)pInputVector)->x); |
||
| 10769 | |||
| 10770 | Result = XMVectorMultiply(W, M.r[3]); |
||
| 10771 | Result = XMVectorMultiplyAdd(Z, M.r[2], Result); |
||
| 10772 | Result = XMVectorMultiplyAdd(Y, M.r[1], Result); |
||
| 10773 | Result = XMVectorMultiplyAdd(X, M.r[0], Result); |
||
| 10774 | |||
| 10775 | XMStoreFloat4((XMFLOAT4*)pOutputVector, Result); |
||
| 10776 | |||
| 10777 | pInputVector += InputStride; |
||
| 10778 | pOutputVector += OutputStride; |
||
| 10779 | } |
||
| 10780 | |||
| 10781 | return pOutputStream; |
||
| 10782 | |||
| 10783 | #elif defined(_XM_SSE_INTRINSICS_) |
||
| 10784 | UINT i; |
||
| 10785 | |||
| 10786 | XMASSERT(pOutputStream); |
||
| 10787 | XMASSERT(pInputStream); |
||
| 10788 | |||
| 10789 | const BYTE*pInputVector = reinterpret_cast<const BYTE *>(pInputStream); |
||
| 10790 | BYTE* pOutputVector = reinterpret_cast<BYTE *>(pOutputStream); |
||
| 10791 | for (i = 0; i < VectorCount; i++) |
||
| 10792 | { |
||
| 10793 | // Fetch the row and splat it |
||
| 10794 | XMVECTOR vTempx = _mm_loadu_ps(reinterpret_cast<const float *>(pInputVector)); |
||
| 10795 | XMVECTOR vTempy = _mm_shuffle_ps(vTempx,vTempx,_MM_SHUFFLE(1,1,1,1)); |
||
| 10796 | XMVECTOR vTempz = _mm_shuffle_ps(vTempx,vTempx,_MM_SHUFFLE(2,2,2,2)); |
||
| 10797 | XMVECTOR vTempw = _mm_shuffle_ps(vTempx,vTempx,_MM_SHUFFLE(3,3,3,3)); |
||
| 10798 | vTempx = _mm_shuffle_ps(vTempx,vTempx,_MM_SHUFFLE(0,0,0,0)); |
||
| 10799 | vTempx = _mm_mul_ps(vTempx,M.r[0]); |
||
| 10800 | vTempy = _mm_mul_ps(vTempy,M.r[1]); |
||
| 10801 | vTempz = _mm_mul_ps(vTempz,M.r[2]); |
||
| 10802 | vTempw = _mm_mul_ps(vTempw,M.r[3]); |
||
| 10803 | vTempx = _mm_add_ps(vTempx,vTempy); |
||
| 10804 | vTempw = _mm_add_ps(vTempw,vTempz); |
||
| 10805 | vTempw = _mm_add_ps(vTempw,vTempx); |
||
| 10806 | // Store the transformed vector |
||
| 10807 | _mm_storeu_ps(reinterpret_cast<float *>(pOutputVector),vTempw); |
||
| 10808 | |||
| 10809 | pInputVector += InputStride; |
||
| 10810 | pOutputVector += OutputStride; |
||
| 10811 | } |
||
| 10812 | return pOutputStream; |
||
| 10813 | #elif defined(XM_NO_MISALIGNED_VECTOR_ACCESS) |
||
| 10814 | #endif // _XM_VMX128_INTRINSICS_ |
||
| 10815 | } |
||
| 10816 | |||
| 10817 | #ifdef __cplusplus |
||
| 10818 | |||
| 10819 | /**************************************************************************** |
||
| 10820 | * |
||
| 10821 | * XMVECTOR operators |
||
| 10822 | * |
||
| 10823 | ****************************************************************************/ |
||
| 10824 | |||
| 10825 | #ifndef XM_NO_OPERATOR_OVERLOADS |
||
| 10826 | |||
| 10827 | //------------------------------------------------------------------------------ |
||
| 10828 | |||
| 10829 | XMFINLINE XMVECTOR operator+ (FXMVECTOR V) |
||
| 10830 | { |
||
| 10831 | return V; |
||
| 10832 | } |
||
| 10833 | |||
| 10834 | //------------------------------------------------------------------------------ |
||
| 10835 | |||
| 10836 | XMFINLINE XMVECTOR operator- (FXMVECTOR V) |
||
| 10837 | { |
||
| 10838 | return XMVectorNegate(V); |
||
| 10839 | } |
||
| 10840 | |||
| 10841 | //------------------------------------------------------------------------------ |
||
| 10842 | |||
| 10843 | XMFINLINE XMVECTOR& operator+= |
||
| 10844 | ( |
||
| 10845 | XMVECTOR& V1, |
||
| 10846 | FXMVECTOR V2 |
||
| 10847 | ) |
||
| 10848 | { |
||
| 10849 | V1 = XMVectorAdd(V1, V2); |
||
| 10850 | return V1; |
||
| 10851 | } |
||
| 10852 | |||
| 10853 | //------------------------------------------------------------------------------ |
||
| 10854 | |||
| 10855 | XMFINLINE XMVECTOR& operator-= |
||
| 10856 | ( |
||
| 10857 | XMVECTOR& V1, |
||
| 10858 | FXMVECTOR V2 |
||
| 10859 | ) |
||
| 10860 | { |
||
| 10861 | V1 = XMVectorSubtract(V1, V2); |
||
| 10862 | return V1; |
||
| 10863 | } |
||
| 10864 | |||
| 10865 | //------------------------------------------------------------------------------ |
||
| 10866 | |||
| 10867 | XMFINLINE XMVECTOR& operator*= |
||
| 10868 | ( |
||
| 10869 | XMVECTOR& V1, |
||
| 10870 | FXMVECTOR V2 |
||
| 10871 | ) |
||
| 10872 | { |
||
| 10873 | V1 = XMVectorMultiply(V1, V2); |
||
| 10874 | return V1; |
||
| 10875 | } |
||
| 10876 | |||
| 10877 | //------------------------------------------------------------------------------ |
||
| 10878 | |||
| 10879 | XMFINLINE XMVECTOR& operator/= |
||
| 10880 | ( |
||
| 10881 | XMVECTOR& V1, |
||
| 10882 | FXMVECTOR V2 |
||
| 10883 | ) |
||
| 10884 | { |
||
| 10885 | XMVECTOR InvV = XMVectorReciprocal(V2); |
||
| 10886 | V1 = XMVectorMultiply(V1, InvV); |
||
| 10887 | return V1; |
||
| 10888 | } |
||
| 10889 | |||
| 10890 | //------------------------------------------------------------------------------ |
||
| 10891 | |||
| 10892 | XMFINLINE XMVECTOR& operator*= |
||
| 10893 | ( |
||
| 10894 | XMVECTOR& V, |
||
| 10895 | CONST FLOAT S |
||
| 10896 | ) |
||
| 10897 | { |
||
| 10898 | V = XMVectorScale(V, S); |
||
| 10899 | return V; |
||
| 10900 | } |
||
| 10901 | |||
| 10902 | //------------------------------------------------------------------------------ |
||
| 10903 | |||
| 10904 | XMFINLINE XMVECTOR& operator/= |
||
| 10905 | ( |
||
| 10906 | XMVECTOR& V, |
||
| 10907 | CONST FLOAT S |
||
| 10908 | ) |
||
| 10909 | { |
||
| 10910 | V = XMVectorScale(V, 1.0f / S); |
||
| 10911 | return V; |
||
| 10912 | } |
||
| 10913 | |||
| 10914 | //------------------------------------------------------------------------------ |
||
| 10915 | |||
| 10916 | XMFINLINE XMVECTOR operator+ |
||
| 10917 | ( |
||
| 10918 | FXMVECTOR V1, |
||
| 10919 | FXMVECTOR V2 |
||
| 10920 | ) |
||
| 10921 | { |
||
| 10922 | return XMVectorAdd(V1, V2); |
||
| 10923 | } |
||
| 10924 | |||
| 10925 | //------------------------------------------------------------------------------ |
||
| 10926 | |||
| 10927 | XMFINLINE XMVECTOR operator- |
||
| 10928 | ( |
||
| 10929 | FXMVECTOR V1, |
||
| 10930 | FXMVECTOR V2 |
||
| 10931 | ) |
||
| 10932 | { |
||
| 10933 | return XMVectorSubtract(V1, V2); |
||
| 10934 | } |
||
| 10935 | |||
| 10936 | //------------------------------------------------------------------------------ |
||
| 10937 | |||
| 10938 | XMFINLINE XMVECTOR operator* |
||
| 10939 | ( |
||
| 10940 | FXMVECTOR V1, |
||
| 10941 | FXMVECTOR V2 |
||
| 10942 | ) |
||
| 10943 | { |
||
| 10944 | return XMVectorMultiply(V1, V2); |
||
| 10945 | } |
||
| 10946 | |||
| 10947 | //------------------------------------------------------------------------------ |
||
| 10948 | |||
| 10949 | XMFINLINE XMVECTOR operator/ |
||
| 10950 | ( |
||
| 10951 | FXMVECTOR V1, |
||
| 10952 | FXMVECTOR V2 |
||
| 10953 | ) |
||
| 10954 | { |
||
| 10955 | XMVECTOR InvV = XMVectorReciprocal(V2); |
||
| 10956 | return XMVectorMultiply(V1, InvV); |
||
| 10957 | } |
||
| 10958 | |||
| 10959 | //------------------------------------------------------------------------------ |
||
| 10960 | |||
| 10961 | XMFINLINE XMVECTOR operator* |
||
| 10962 | ( |
||
| 10963 | FXMVECTOR V, |
||
| 10964 | CONST FLOAT S |
||
| 10965 | ) |
||
| 10966 | { |
||
| 10967 | return XMVectorScale(V, S); |
||
| 10968 | } |
||
| 10969 | |||
| 10970 | //------------------------------------------------------------------------------ |
||
| 10971 | |||
| 10972 | XMFINLINE XMVECTOR operator/ |
||
| 10973 | ( |
||
| 10974 | FXMVECTOR V, |
||
| 10975 | CONST FLOAT S |
||
| 10976 | ) |
||
| 10977 | { |
||
| 10978 | return XMVectorScale(V, 1.0f / S); |
||
| 10979 | } |
||
| 10980 | |||
| 10981 | //------------------------------------------------------------------------------ |
||
| 10982 | |||
| 10983 | XMFINLINE XMVECTOR operator* |
||
| 10984 | ( |
||
| 10985 | FLOAT S, |
||
| 10986 | FXMVECTOR V |
||
| 10987 | ) |
||
| 10988 | { |
||
| 10989 | return XMVectorScale(V, S); |
||
| 10990 | } |
||
| 10991 | |||
| 10992 | #endif // !XM_NO_OPERATOR_OVERLOADS |
||
| 10993 | |||
| 10994 | /**************************************************************************** |
||
| 10995 | * |
||
| 10996 | * XMFLOAT2 operators |
||
| 10997 | * |
||
| 10998 | ****************************************************************************/ |
||
| 10999 | |||
| 11000 | //------------------------------------------------------------------------------ |
||
| 11001 | |||
| 11002 | XMFINLINE _XMFLOAT2::_XMFLOAT2 |
||
| 11003 | ( |
||
| 11004 | CONST FLOAT* pArray |
||
| 11005 | ) |
||
| 11006 | { |
||
| 11007 | x = pArray[0]; |
||
| 11008 | y = pArray[1]; |
||
| 11009 | } |
||
| 11010 | |||
| 11011 | //------------------------------------------------------------------------------ |
||
| 11012 | |||
| 11013 | XMFINLINE _XMFLOAT2& _XMFLOAT2::operator= |
||
| 11014 | ( |
||
| 11015 | CONST _XMFLOAT2& Float2 |
||
| 11016 | ) |
||
| 11017 | { |
||
| 11018 | x = Float2.x; |
||
| 11019 | y = Float2.y; |
||
| 11020 | return *this; |
||
| 11021 | } |
||
| 11022 | |||
| 11023 | /**************************************************************************** |
||
| 11024 | * |
||
| 11025 | * XMHALF2 operators |
||
| 11026 | * |
||
| 11027 | ****************************************************************************/ |
||
| 11028 | |||
| 11029 | //------------------------------------------------------------------------------ |
||
| 11030 | |||
| 11031 | XMFINLINE _XMHALF2::_XMHALF2 |
||
| 11032 | ( |
||
| 11033 | CONST HALF* pArray |
||
| 11034 | ) |
||
| 11035 | { |
||
| 11036 | x = pArray[0]; |
||
| 11037 | y = pArray[1]; |
||
| 11038 | } |
||
| 11039 | |||
| 11040 | //------------------------------------------------------------------------------ |
||
| 11041 | |||
| 11042 | XMFINLINE _XMHALF2::_XMHALF2 |
||
| 11043 | ( |
||
| 11044 | FLOAT _x, |
||
| 11045 | FLOAT _y |
||
| 11046 | ) |
||
| 11047 | { |
||
| 11048 | x = XMConvertFloatToHalf(_x); |
||
| 11049 | y = XMConvertFloatToHalf(_y); |
||
| 11050 | } |
||
| 11051 | |||
| 11052 | //------------------------------------------------------------------------------ |
||
| 11053 | |||
| 11054 | XMFINLINE _XMHALF2::_XMHALF2 |
||
| 11055 | ( |
||
| 11056 | CONST FLOAT* pArray |
||
| 11057 | ) |
||
| 11058 | { |
||
| 11059 | x = XMConvertFloatToHalf(pArray[0]); |
||
| 11060 | y = XMConvertFloatToHalf(pArray[1]); |
||
| 11061 | } |
||
| 11062 | |||
| 11063 | //------------------------------------------------------------------------------ |
||
| 11064 | |||
| 11065 | XMFINLINE _XMHALF2& _XMHALF2::operator= |
||
| 11066 | ( |
||
| 11067 | CONST _XMHALF2& Half2 |
||
| 11068 | ) |
||
| 11069 | { |
||
| 11070 | x = Half2.x; |
||
| 11071 | y = Half2.y; |
||
| 11072 | return *this; |
||
| 11073 | } |
||
| 11074 | |||
| 11075 | /**************************************************************************** |
||
| 11076 | * |
||
| 11077 | * XMSHORTN2 operators |
||
| 11078 | * |
||
| 11079 | ****************************************************************************/ |
||
| 11080 | |||
| 11081 | //------------------------------------------------------------------------------ |
||
| 11082 | |||
| 11083 | XMFINLINE _XMSHORTN2::_XMSHORTN2 |
||
| 11084 | ( |
||
| 11085 | CONST SHORT* pArray |
||
| 11086 | ) |
||
| 11087 | { |
||
| 11088 | x = pArray[0]; |
||
| 11089 | y = pArray[1]; |
||
| 11090 | } |
||
| 11091 | |||
| 11092 | //------------------------------------------------------------------------------ |
||
| 11093 | |||
| 11094 | XMFINLINE _XMSHORTN2::_XMSHORTN2 |
||
| 11095 | ( |
||
| 11096 | FLOAT _x, |
||
| 11097 | FLOAT _y |
||
| 11098 | ) |
||
| 11099 | { |
||
| 11100 | XMStoreShortN2(this, XMVectorSet(_x, _y, 0.0f, 0.0f)); |
||
| 11101 | } |
||
| 11102 | |||
| 11103 | //------------------------------------------------------------------------------ |
||
| 11104 | |||
| 11105 | XMFINLINE _XMSHORTN2::_XMSHORTN2 |
||
| 11106 | ( |
||
| 11107 | CONST FLOAT* pArray |
||
| 11108 | ) |
||
| 11109 | { |
||
| 11110 | XMStoreShortN2(this, XMLoadFloat2((XMFLOAT2*)pArray)); |
||
| 11111 | } |
||
| 11112 | |||
| 11113 | //------------------------------------------------------------------------------ |
||
| 11114 | |||
| 11115 | XMFINLINE _XMSHORTN2& _XMSHORTN2::operator= |
||
| 11116 | ( |
||
| 11117 | CONST _XMSHORTN2& ShortN2 |
||
| 11118 | ) |
||
| 11119 | { |
||
| 11120 | x = ShortN2.x; |
||
| 11121 | y = ShortN2.y; |
||
| 11122 | return *this; |
||
| 11123 | } |
||
| 11124 | |||
| 11125 | /**************************************************************************** |
||
| 11126 | * |
||
| 11127 | * XMSHORT2 operators |
||
| 11128 | * |
||
| 11129 | ****************************************************************************/ |
||
| 11130 | |||
| 11131 | //------------------------------------------------------------------------------ |
||
| 11132 | |||
| 11133 | XMFINLINE _XMSHORT2::_XMSHORT2 |
||
| 11134 | ( |
||
| 11135 | CONST SHORT* pArray |
||
| 11136 | ) |
||
| 11137 | { |
||
| 11138 | x = pArray[0]; |
||
| 11139 | y = pArray[1]; |
||
| 11140 | } |
||
| 11141 | |||
| 11142 | //------------------------------------------------------------------------------ |
||
| 11143 | |||
| 11144 | XMFINLINE _XMSHORT2::_XMSHORT2 |
||
| 11145 | ( |
||
| 11146 | FLOAT _x, |
||
| 11147 | FLOAT _y |
||
| 11148 | ) |
||
| 11149 | { |
||
| 11150 | XMStoreShort2(this, XMVectorSet(_x, _y, 0.0f, 0.0f)); |
||
| 11151 | } |
||
| 11152 | |||
| 11153 | //------------------------------------------------------------------------------ |
||
| 11154 | |||
| 11155 | XMFINLINE _XMSHORT2::_XMSHORT2 |
||
| 11156 | ( |
||
| 11157 | CONST FLOAT* pArray |
||
| 11158 | ) |
||
| 11159 | { |
||
| 11160 | XMStoreShort2(this, XMLoadFloat2((XMFLOAT2*)pArray)); |
||
| 11161 | } |
||
| 11162 | |||
| 11163 | //------------------------------------------------------------------------------ |
||
| 11164 | |||
| 11165 | XMFINLINE _XMSHORT2& _XMSHORT2::operator= |
||
| 11166 | ( |
||
| 11167 | CONST _XMSHORT2& Short2 |
||
| 11168 | ) |
||
| 11169 | { |
||
| 11170 | x = Short2.x; |
||
| 11171 | y = Short2.y; |
||
| 11172 | return *this; |
||
| 11173 | } |
||
| 11174 | |||
| 11175 | /**************************************************************************** |
||
| 11176 | * |
||
| 11177 | * XMUSHORTN2 operators |
||
| 11178 | * |
||
| 11179 | ****************************************************************************/ |
||
| 11180 | |||
| 11181 | //------------------------------------------------------------------------------ |
||
| 11182 | |||
| 11183 | XMFINLINE _XMUSHORTN2::_XMUSHORTN2 |
||
| 11184 | ( |
||
| 11185 | CONST USHORT* pArray |
||
| 11186 | ) |
||
| 11187 | { |
||
| 11188 | x = pArray[0]; |
||
| 11189 | y = pArray[1]; |
||
| 11190 | } |
||
| 11191 | |||
| 11192 | //------------------------------------------------------------------------------ |
||
| 11193 | |||
| 11194 | XMFINLINE _XMUSHORTN2::_XMUSHORTN2 |
||
| 11195 | ( |
||
| 11196 | FLOAT _x, |
||
| 11197 | FLOAT _y |
||
| 11198 | ) |
||
| 11199 | { |
||
| 11200 | XMStoreUShortN2(this, XMVectorSet(_x, _y, 0.0f, 0.0f)); |
||
| 11201 | } |
||
| 11202 | |||
| 11203 | //------------------------------------------------------------------------------ |
||
| 11204 | |||
| 11205 | XMFINLINE _XMUSHORTN2::_XMUSHORTN2 |
||
| 11206 | ( |
||
| 11207 | CONST FLOAT* pArray |
||
| 11208 | ) |
||
| 11209 | { |
||
| 11210 | XMStoreUShortN2(this, XMLoadFloat2((XMFLOAT2*)pArray)); |
||
| 11211 | } |
||
| 11212 | |||
| 11213 | //------------------------------------------------------------------------------ |
||
| 11214 | |||
| 11215 | XMFINLINE _XMUSHORTN2& _XMUSHORTN2::operator= |
||
| 11216 | ( |
||
| 11217 | CONST _XMUSHORTN2& UShortN2 |
||
| 11218 | ) |
||
| 11219 | { |
||
| 11220 | x = UShortN2.x; |
||
| 11221 | y = UShortN2.y; |
||
| 11222 | return *this; |
||
| 11223 | } |
||
| 11224 | |||
| 11225 | /**************************************************************************** |
||
| 11226 | * |
||
| 11227 | * XMUSHORT2 operators |
||
| 11228 | * |
||
| 11229 | ****************************************************************************/ |
||
| 11230 | |||
| 11231 | //------------------------------------------------------------------------------ |
||
| 11232 | |||
| 11233 | XMFINLINE _XMUSHORT2::_XMUSHORT2 |
||
| 11234 | ( |
||
| 11235 | CONST USHORT* pArray |
||
| 11236 | ) |
||
| 11237 | { |
||
| 11238 | x = pArray[0]; |
||
| 11239 | y = pArray[1]; |
||
| 11240 | } |
||
| 11241 | |||
| 11242 | //------------------------------------------------------------------------------ |
||
| 11243 | |||
| 11244 | XMFINLINE _XMUSHORT2::_XMUSHORT2 |
||
| 11245 | ( |
||
| 11246 | FLOAT _x, |
||
| 11247 | FLOAT _y |
||
| 11248 | ) |
||
| 11249 | { |
||
| 11250 | XMStoreUShort2(this, XMVectorSet(_x, _y, 0.0f, 0.0f)); |
||
| 11251 | } |
||
| 11252 | |||
| 11253 | //------------------------------------------------------------------------------ |
||
| 11254 | |||
| 11255 | XMFINLINE _XMUSHORT2::_XMUSHORT2 |
||
| 11256 | ( |
||
| 11257 | CONST FLOAT* pArray |
||
| 11258 | ) |
||
| 11259 | { |
||
| 11260 | XMStoreUShort2(this, XMLoadFloat2((XMFLOAT2*)pArray)); |
||
| 11261 | } |
||
| 11262 | |||
| 11263 | //------------------------------------------------------------------------------ |
||
| 11264 | |||
| 11265 | XMFINLINE _XMUSHORT2& _XMUSHORT2::operator= |
||
| 11266 | ( |
||
| 11267 | CONST _XMUSHORT2& UShort2 |
||
| 11268 | ) |
||
| 11269 | { |
||
| 11270 | x = UShort2.x; |
||
| 11271 | y = UShort2.y; |
||
| 11272 | return *this; |
||
| 11273 | } |
||
| 11274 | |||
| 11275 | /**************************************************************************** |
||
| 11276 | * |
||
| 11277 | * XMFLOAT3 operators |
||
| 11278 | * |
||
| 11279 | ****************************************************************************/ |
||
| 11280 | |||
| 11281 | //------------------------------------------------------------------------------ |
||
| 11282 | |||
| 11283 | XMFINLINE _XMFLOAT3::_XMFLOAT3 |
||
| 11284 | ( |
||
| 11285 | CONST FLOAT* pArray |
||
| 11286 | ) |
||
| 11287 | { |
||
| 11288 | x = pArray[0]; |
||
| 11289 | y = pArray[1]; |
||
| 11290 | z = pArray[2]; |
||
| 11291 | } |
||
| 11292 | |||
| 11293 | //------------------------------------------------------------------------------ |
||
| 11294 | |||
| 11295 | XMFINLINE _XMFLOAT3& _XMFLOAT3::operator= |
||
| 11296 | ( |
||
| 11297 | CONST _XMFLOAT3& Float3 |
||
| 11298 | ) |
||
| 11299 | { |
||
| 11300 | x = Float3.x; |
||
| 11301 | y = Float3.y; |
||
| 11302 | z = Float3.z; |
||
| 11303 | return *this; |
||
| 11304 | } |
||
| 11305 | |||
| 11306 | /**************************************************************************** |
||
| 11307 | * |
||
| 11308 | * XMHENDN3 operators |
||
| 11309 | * |
||
| 11310 | ****************************************************************************/ |
||
| 11311 | |||
| 11312 | //------------------------------------------------------------------------------ |
||
| 11313 | |||
| 11314 | XMFINLINE _XMHENDN3::_XMHENDN3 |
||
| 11315 | ( |
||
| 11316 | FLOAT _x, |
||
| 11317 | FLOAT _y, |
||
| 11318 | FLOAT _z |
||
| 11319 | ) |
||
| 11320 | { |
||
| 11321 | XMStoreHenDN3(this, XMVectorSet(_x, _y, _z, 0.0f)); |
||
| 11322 | } |
||
| 11323 | |||
| 11324 | //------------------------------------------------------------------------------ |
||
| 11325 | |||
| 11326 | XMFINLINE _XMHENDN3::_XMHENDN3 |
||
| 11327 | ( |
||
| 11328 | CONST FLOAT* pArray |
||
| 11329 | ) |
||
| 11330 | { |
||
| 11331 | XMStoreHenDN3(this, XMLoadFloat3((XMFLOAT3*)pArray)); |
||
| 11332 | } |
||
| 11333 | |||
| 11334 | //------------------------------------------------------------------------------ |
||
| 11335 | |||
| 11336 | XMFINLINE _XMHENDN3& _XMHENDN3::operator= |
||
| 11337 | ( |
||
| 11338 | CONST _XMHENDN3& HenDN3 |
||
| 11339 | ) |
||
| 11340 | { |
||
| 11341 | v = HenDN3.v; |
||
| 11342 | return *this; |
||
| 11343 | } |
||
| 11344 | |||
| 11345 | //------------------------------------------------------------------------------ |
||
| 11346 | |||
| 11347 | XMFINLINE _XMHENDN3& _XMHENDN3::operator= |
||
| 11348 | ( |
||
| 11349 | CONST UINT Packed |
||
| 11350 | ) |
||
| 11351 | { |
||
| 11352 | v = Packed; |
||
| 11353 | return *this; |
||
| 11354 | } |
||
| 11355 | |||
| 11356 | /**************************************************************************** |
||
| 11357 | * |
||
| 11358 | * XMHEND3 operators |
||
| 11359 | * |
||
| 11360 | ****************************************************************************/ |
||
| 11361 | |||
| 11362 | //------------------------------------------------------------------------------ |
||
| 11363 | |||
| 11364 | XMFINLINE _XMHEND3::_XMHEND3 |
||
| 11365 | ( |
||
| 11366 | FLOAT _x, |
||
| 11367 | FLOAT _y, |
||
| 11368 | FLOAT _z |
||
| 11369 | ) |
||
| 11370 | { |
||
| 11371 | XMStoreHenD3(this, XMVectorSet(_x, _y, _z, 0.0f)); |
||
| 11372 | } |
||
| 11373 | |||
| 11374 | //------------------------------------------------------------------------------ |
||
| 11375 | |||
| 11376 | XMFINLINE _XMHEND3::_XMHEND3 |
||
| 11377 | ( |
||
| 11378 | CONST FLOAT* pArray |
||
| 11379 | ) |
||
| 11380 | { |
||
| 11381 | XMStoreHenD3(this, XMLoadFloat3((XMFLOAT3*)pArray)); |
||
| 11382 | } |
||
| 11383 | |||
| 11384 | //------------------------------------------------------------------------------ |
||
| 11385 | |||
| 11386 | XMFINLINE _XMHEND3& _XMHEND3::operator= |
||
| 11387 | ( |
||
| 11388 | CONST _XMHEND3& HenD3 |
||
| 11389 | ) |
||
| 11390 | { |
||
| 11391 | v = HenD3.v; |
||
| 11392 | return *this; |
||
| 11393 | } |
||
| 11394 | |||
| 11395 | //------------------------------------------------------------------------------ |
||
| 11396 | |||
| 11397 | XMFINLINE _XMHEND3& _XMHEND3::operator= |
||
| 11398 | ( |
||
| 11399 | CONST UINT Packed |
||
| 11400 | ) |
||
| 11401 | { |
||
| 11402 | v = Packed; |
||
| 11403 | return *this; |
||
| 11404 | } |
||
| 11405 | |||
| 11406 | /**************************************************************************** |
||
| 11407 | * |
||
| 11408 | * XMUHENDN3 operators |
||
| 11409 | * |
||
| 11410 | ****************************************************************************/ |
||
| 11411 | |||
| 11412 | //------------------------------------------------------------------------------ |
||
| 11413 | |||
| 11414 | XMFINLINE _XMUHENDN3::_XMUHENDN3 |
||
| 11415 | ( |
||
| 11416 | FLOAT _x, |
||
| 11417 | FLOAT _y, |
||
| 11418 | FLOAT _z |
||
| 11419 | ) |
||
| 11420 | { |
||
| 11421 | XMStoreUHenDN3(this, XMVectorSet(_x, _y, _z, 0.0f)); |
||
| 11422 | } |
||
| 11423 | |||
| 11424 | //------------------------------------------------------------------------------ |
||
| 11425 | |||
| 11426 | XMFINLINE _XMUHENDN3::_XMUHENDN3 |
||
| 11427 | ( |
||
| 11428 | CONST FLOAT* pArray |
||
| 11429 | ) |
||
| 11430 | { |
||
| 11431 | XMStoreUHenDN3(this, XMLoadFloat3((XMFLOAT3*)pArray)); |
||
| 11432 | } |
||
| 11433 | |||
| 11434 | //------------------------------------------------------------------------------ |
||
| 11435 | |||
| 11436 | XMFINLINE _XMUHENDN3& _XMUHENDN3::operator= |
||
| 11437 | ( |
||
| 11438 | CONST _XMUHENDN3& UHenDN3 |
||
| 11439 | ) |
||
| 11440 | { |
||
| 11441 | v = UHenDN3.v; |
||
| 11442 | return *this; |
||
| 11443 | } |
||
| 11444 | |||
| 11445 | //------------------------------------------------------------------------------ |
||
| 11446 | |||
| 11447 | XMFINLINE _XMUHENDN3& _XMUHENDN3::operator= |
||
| 11448 | ( |
||
| 11449 | CONST UINT Packed |
||
| 11450 | ) |
||
| 11451 | { |
||
| 11452 | v = Packed; |
||
| 11453 | return *this; |
||
| 11454 | } |
||
| 11455 | |||
| 11456 | /**************************************************************************** |
||
| 11457 | * |
||
| 11458 | * XMUHEND3 operators |
||
| 11459 | * |
||
| 11460 | ****************************************************************************/ |
||
| 11461 | |||
| 11462 | //------------------------------------------------------------------------------ |
||
| 11463 | |||
| 11464 | XMFINLINE _XMUHEND3::_XMUHEND3 |
||
| 11465 | ( |
||
| 11466 | FLOAT _x, |
||
| 11467 | FLOAT _y, |
||
| 11468 | FLOAT _z |
||
| 11469 | ) |
||
| 11470 | { |
||
| 11471 | XMStoreUHenD3(this, XMVectorSet(_x, _y, _z, 0.0f)); |
||
| 11472 | } |
||
| 11473 | |||
| 11474 | //------------------------------------------------------------------------------ |
||
| 11475 | |||
| 11476 | XMFINLINE _XMUHEND3::_XMUHEND3 |
||
| 11477 | ( |
||
| 11478 | CONST FLOAT* pArray |
||
| 11479 | ) |
||
| 11480 | { |
||
| 11481 | XMStoreUHenD3(this, XMLoadFloat3((XMFLOAT3*)pArray)); |
||
| 11482 | } |
||
| 11483 | |||
| 11484 | //------------------------------------------------------------------------------ |
||
| 11485 | |||
| 11486 | XMFINLINE _XMUHEND3& _XMUHEND3::operator= |
||
| 11487 | ( |
||
| 11488 | CONST _XMUHEND3& UHenD3 |
||
| 11489 | ) |
||
| 11490 | { |
||
| 11491 | v = UHenD3.v; |
||
| 11492 | return *this; |
||
| 11493 | } |
||
| 11494 | |||
| 11495 | //------------------------------------------------------------------------------ |
||
| 11496 | |||
| 11497 | XMFINLINE _XMUHEND3& _XMUHEND3::operator= |
||
| 11498 | ( |
||
| 11499 | CONST UINT Packed |
||
| 11500 | ) |
||
| 11501 | { |
||
| 11502 | v = Packed; |
||
| 11503 | return *this; |
||
| 11504 | } |
||
| 11505 | |||
| 11506 | /**************************************************************************** |
||
| 11507 | * |
||
| 11508 | * XMDHENN3 operators |
||
| 11509 | * |
||
| 11510 | ****************************************************************************/ |
||
| 11511 | |||
| 11512 | //------------------------------------------------------------------------------ |
||
| 11513 | |||
| 11514 | XMFINLINE _XMDHENN3::_XMDHENN3 |
||
| 11515 | ( |
||
| 11516 | FLOAT _x, |
||
| 11517 | FLOAT _y, |
||
| 11518 | FLOAT _z |
||
| 11519 | ) |
||
| 11520 | { |
||
| 11521 | XMStoreDHenN3(this, XMVectorSet(_x, _y, _z, 0.0f)); |
||
| 11522 | } |
||
| 11523 | |||
| 11524 | //------------------------------------------------------------------------------ |
||
| 11525 | |||
| 11526 | XMFINLINE _XMDHENN3::_XMDHENN3 |
||
| 11527 | ( |
||
| 11528 | CONST FLOAT* pArray |
||
| 11529 | ) |
||
| 11530 | { |
||
| 11531 | XMStoreDHenN3(this, XMLoadFloat3((XMFLOAT3*)pArray)); |
||
| 11532 | } |
||
| 11533 | |||
| 11534 | //------------------------------------------------------------------------------ |
||
| 11535 | |||
| 11536 | XMFINLINE _XMDHENN3& _XMDHENN3::operator= |
||
| 11537 | ( |
||
| 11538 | CONST _XMDHENN3& DHenN3 |
||
| 11539 | ) |
||
| 11540 | { |
||
| 11541 | v = DHenN3.v; |
||
| 11542 | return *this; |
||
| 11543 | } |
||
| 11544 | |||
| 11545 | //------------------------------------------------------------------------------ |
||
| 11546 | |||
| 11547 | XMFINLINE _XMDHENN3& _XMDHENN3::operator= |
||
| 11548 | ( |
||
| 11549 | CONST UINT Packed |
||
| 11550 | ) |
||
| 11551 | { |
||
| 11552 | v = Packed; |
||
| 11553 | return *this; |
||
| 11554 | } |
||
| 11555 | |||
| 11556 | /**************************************************************************** |
||
| 11557 | * |
||
| 11558 | * XMDHEN3 operators |
||
| 11559 | * |
||
| 11560 | ****************************************************************************/ |
||
| 11561 | |||
| 11562 | //------------------------------------------------------------------------------ |
||
| 11563 | |||
| 11564 | XMFINLINE _XMDHEN3::_XMDHEN3 |
||
| 11565 | ( |
||
| 11566 | FLOAT _x, |
||
| 11567 | FLOAT _y, |
||
| 11568 | FLOAT _z |
||
| 11569 | ) |
||
| 11570 | { |
||
| 11571 | XMStoreDHen3(this, XMVectorSet(_x, _y, _z, 0.0f)); |
||
| 11572 | } |
||
| 11573 | |||
| 11574 | //------------------------------------------------------------------------------ |
||
| 11575 | |||
| 11576 | XMFINLINE _XMDHEN3::_XMDHEN3 |
||
| 11577 | ( |
||
| 11578 | CONST FLOAT* pArray |
||
| 11579 | ) |
||
| 11580 | { |
||
| 11581 | XMStoreDHen3(this, XMLoadFloat3((XMFLOAT3*)pArray)); |
||
| 11582 | } |
||
| 11583 | |||
| 11584 | //------------------------------------------------------------------------------ |
||
| 11585 | |||
| 11586 | XMFINLINE _XMDHEN3& _XMDHEN3::operator= |
||
| 11587 | ( |
||
| 11588 | CONST _XMDHEN3& DHen3 |
||
| 11589 | ) |
||
| 11590 | { |
||
| 11591 | v = DHen3.v; |
||
| 11592 | return *this; |
||
| 11593 | } |
||
| 11594 | |||
| 11595 | //------------------------------------------------------------------------------ |
||
| 11596 | |||
| 11597 | XMFINLINE _XMDHEN3& _XMDHEN3::operator= |
||
| 11598 | ( |
||
| 11599 | CONST UINT Packed |
||
| 11600 | ) |
||
| 11601 | { |
||
| 11602 | v = Packed; |
||
| 11603 | return *this; |
||
| 11604 | } |
||
| 11605 | |||
| 11606 | /**************************************************************************** |
||
| 11607 | * |
||
| 11608 | * XMUDHENN3 operators |
||
| 11609 | * |
||
| 11610 | ****************************************************************************/ |
||
| 11611 | |||
| 11612 | //------------------------------------------------------------------------------ |
||
| 11613 | |||
| 11614 | XMFINLINE _XMUDHENN3::_XMUDHENN3 |
||
| 11615 | ( |
||
| 11616 | FLOAT _x, |
||
| 11617 | FLOAT _y, |
||
| 11618 | FLOAT _z |
||
| 11619 | ) |
||
| 11620 | { |
||
| 11621 | XMStoreUDHenN3(this, XMVectorSet(_x, _y, _z, 0.0f)); |
||
| 11622 | } |
||
| 11623 | |||
| 11624 | //------------------------------------------------------------------------------ |
||
| 11625 | |||
| 11626 | XMFINLINE _XMUDHENN3::_XMUDHENN3 |
||
| 11627 | ( |
||
| 11628 | CONST FLOAT* pArray |
||
| 11629 | ) |
||
| 11630 | { |
||
| 11631 | XMStoreUDHenN3(this, XMLoadFloat3((XMFLOAT3*)pArray)); |
||
| 11632 | } |
||
| 11633 | |||
| 11634 | //------------------------------------------------------------------------------ |
||
| 11635 | |||
| 11636 | XMFINLINE _XMUDHENN3& _XMUDHENN3::operator= |
||
| 11637 | ( |
||
| 11638 | CONST _XMUDHENN3& UDHenN3 |
||
| 11639 | ) |
||
| 11640 | { |
||
| 11641 | v = UDHenN3.v; |
||
| 11642 | return *this; |
||
| 11643 | } |
||
| 11644 | |||
| 11645 | //------------------------------------------------------------------------------ |
||
| 11646 | |||
| 11647 | XMFINLINE _XMUDHENN3& _XMUDHENN3::operator= |
||
| 11648 | ( |
||
| 11649 | CONST UINT Packed |
||
| 11650 | ) |
||
| 11651 | { |
||
| 11652 | v = Packed; |
||
| 11653 | return *this; |
||
| 11654 | } |
||
| 11655 | |||
| 11656 | /**************************************************************************** |
||
| 11657 | * |
||
| 11658 | * XMUDHEN3 operators |
||
| 11659 | * |
||
| 11660 | ****************************************************************************/ |
||
| 11661 | |||
| 11662 | //------------------------------------------------------------------------------ |
||
| 11663 | |||
| 11664 | XMFINLINE _XMUDHEN3::_XMUDHEN3 |
||
| 11665 | ( |
||
| 11666 | FLOAT _x, |
||
| 11667 | FLOAT _y, |
||
| 11668 | FLOAT _z |
||
| 11669 | ) |
||
| 11670 | { |
||
| 11671 | XMStoreUDHen3(this, XMVectorSet(_x, _y, _z, 0.0f)); |
||
| 11672 | } |
||
| 11673 | |||
| 11674 | //------------------------------------------------------------------------------ |
||
| 11675 | |||
| 11676 | XMFINLINE _XMUDHEN3::_XMUDHEN3 |
||
| 11677 | ( |
||
| 11678 | CONST FLOAT* pArray |
||
| 11679 | ) |
||
| 11680 | { |
||
| 11681 | XMStoreUDHen3(this, XMLoadFloat3((XMFLOAT3*)pArray)); |
||
| 11682 | } |
||
| 11683 | |||
| 11684 | //------------------------------------------------------------------------------ |
||
| 11685 | |||
| 11686 | XMFINLINE _XMUDHEN3& _XMUDHEN3::operator= |
||
| 11687 | ( |
||
| 11688 | CONST _XMUDHEN3& UDHen3 |
||
| 11689 | ) |
||
| 11690 | { |
||
| 11691 | v = UDHen3.v; |
||
| 11692 | return *this; |
||
| 11693 | } |
||
| 11694 | |||
| 11695 | //------------------------------------------------------------------------------ |
||
| 11696 | |||
| 11697 | XMFINLINE _XMUDHEN3& _XMUDHEN3::operator= |
||
| 11698 | ( |
||
| 11699 | CONST UINT Packed |
||
| 11700 | ) |
||
| 11701 | { |
||
| 11702 | v = Packed; |
||
| 11703 | return *this; |
||
| 11704 | } |
||
| 11705 | |||
| 11706 | /**************************************************************************** |
||
| 11707 | * |
||
| 11708 | * XMU565 operators |
||
| 11709 | * |
||
| 11710 | ****************************************************************************/ |
||
| 11711 | |||
| 11712 | XMFINLINE _XMU565::_XMU565 |
||
| 11713 | ( |
||
| 11714 | CONST CHAR *pArray |
||
| 11715 | ) |
||
| 11716 | { |
||
| 11717 | x = pArray[0]; |
||
| 11718 | y = pArray[1]; |
||
| 11719 | z = pArray[2]; |
||
| 11720 | } |
||
| 11721 | |||
| 11722 | XMFINLINE _XMU565::_XMU565 |
||
| 11723 | ( |
||
| 11724 | FLOAT _x, |
||
| 11725 | FLOAT _y, |
||
| 11726 | FLOAT _z |
||
| 11727 | ) |
||
| 11728 | { |
||
| 11729 | XMStoreU565(this, XMVectorSet( _x, _y, _z, 0.0f )); |
||
| 11730 | } |
||
| 11731 | |||
| 11732 | XMFINLINE _XMU565::_XMU565 |
||
| 11733 | ( |
||
| 11734 | CONST FLOAT *pArray |
||
| 11735 | ) |
||
| 11736 | { |
||
| 11737 | XMStoreU565(this, XMLoadFloat3((XMFLOAT3*)pArray )); |
||
| 11738 | } |
||
| 11739 | |||
| 11740 | XMFINLINE _XMU565& _XMU565::operator= |
||
| 11741 | ( |
||
| 11742 | CONST _XMU565& U565 |
||
| 11743 | ) |
||
| 11744 | { |
||
| 11745 | v = U565.v; |
||
| 11746 | return *this; |
||
| 11747 | } |
||
| 11748 | |||
| 11749 | XMFINLINE _XMU565& _XMU565::operator= |
||
| 11750 | ( |
||
| 11751 | CONST USHORT Packed |
||
| 11752 | ) |
||
| 11753 | { |
||
| 11754 | v = Packed; |
||
| 11755 | return *this; |
||
| 11756 | } |
||
| 11757 | |||
| 11758 | /**************************************************************************** |
||
| 11759 | * |
||
| 11760 | * XMFLOAT3PK operators |
||
| 11761 | * |
||
| 11762 | ****************************************************************************/ |
||
| 11763 | |||
| 11764 | XMFINLINE _XMFLOAT3PK::_XMFLOAT3PK |
||
| 11765 | ( |
||
| 11766 | FLOAT _x, |
||
| 11767 | FLOAT _y, |
||
| 11768 | FLOAT _z |
||
| 11769 | ) |
||
| 11770 | { |
||
| 11771 | XMStoreFloat3PK(this, XMVectorSet( _x, _y, _z, 0.0f )); |
||
| 11772 | } |
||
| 11773 | |||
| 11774 | XMFINLINE _XMFLOAT3PK::_XMFLOAT3PK |
||
| 11775 | ( |
||
| 11776 | CONST FLOAT *pArray |
||
| 11777 | ) |
||
| 11778 | { |
||
| 11779 | XMStoreFloat3PK(this, XMLoadFloat3((XMFLOAT3*)pArray )); |
||
| 11780 | } |
||
| 11781 | |||
| 11782 | XMFINLINE _XMFLOAT3PK& _XMFLOAT3PK::operator= |
||
| 11783 | ( |
||
| 11784 | CONST _XMFLOAT3PK& float3pk |
||
| 11785 | ) |
||
| 11786 | { |
||
| 11787 | v = float3pk.v; |
||
| 11788 | return *this; |
||
| 11789 | } |
||
| 11790 | |||
| 11791 | XMFINLINE _XMFLOAT3PK& _XMFLOAT3PK::operator= |
||
| 11792 | ( |
||
| 11793 | CONST UINT Packed |
||
| 11794 | ) |
||
| 11795 | { |
||
| 11796 | v = Packed; |
||
| 11797 | return *this; |
||
| 11798 | } |
||
| 11799 | |||
| 11800 | /**************************************************************************** |
||
| 11801 | * |
||
| 11802 | * XMFLOAT3SE operators |
||
| 11803 | * |
||
| 11804 | ****************************************************************************/ |
||
| 11805 | |||
| 11806 | XMFINLINE _XMFLOAT3SE::_XMFLOAT3SE |
||
| 11807 | ( |
||
| 11808 | FLOAT _x, |
||
| 11809 | FLOAT _y, |
||
| 11810 | FLOAT _z |
||
| 11811 | ) |
||
| 11812 | { |
||
| 11813 | XMStoreFloat3SE(this, XMVectorSet( _x, _y, _z, 0.0f )); |
||
| 11814 | } |
||
| 11815 | |||
| 11816 | XMFINLINE _XMFLOAT3SE::_XMFLOAT3SE |
||
| 11817 | ( |
||
| 11818 | CONST FLOAT *pArray |
||
| 11819 | ) |
||
| 11820 | { |
||
| 11821 | XMStoreFloat3SE(this, XMLoadFloat3((XMFLOAT3*)pArray )); |
||
| 11822 | } |
||
| 11823 | |||
| 11824 | XMFINLINE _XMFLOAT3SE& _XMFLOAT3SE::operator= |
||
| 11825 | ( |
||
| 11826 | CONST _XMFLOAT3SE& float3se |
||
| 11827 | ) |
||
| 11828 | { |
||
| 11829 | v = float3se.v; |
||
| 11830 | return *this; |
||
| 11831 | } |
||
| 11832 | |||
| 11833 | XMFINLINE _XMFLOAT3SE& _XMFLOAT3SE::operator= |
||
| 11834 | ( |
||
| 11835 | CONST UINT Packed |
||
| 11836 | ) |
||
| 11837 | { |
||
| 11838 | v = Packed; |
||
| 11839 | return *this; |
||
| 11840 | } |
||
| 11841 | |||
| 11842 | /**************************************************************************** |
||
| 11843 | * |
||
| 11844 | * XMFLOAT4 operators |
||
| 11845 | * |
||
| 11846 | ****************************************************************************/ |
||
| 11847 | |||
| 11848 | //------------------------------------------------------------------------------ |
||
| 11849 | |||
| 11850 | XMFINLINE _XMFLOAT4::_XMFLOAT4 |
||
| 11851 | ( |
||
| 11852 | CONST FLOAT* pArray |
||
| 11853 | ) |
||
| 11854 | { |
||
| 11855 | x = pArray[0]; |
||
| 11856 | y = pArray[1]; |
||
| 11857 | z = pArray[2]; |
||
| 11858 | w = pArray[3]; |
||
| 11859 | } |
||
| 11860 | |||
| 11861 | //------------------------------------------------------------------------------ |
||
| 11862 | |||
| 11863 | XMFINLINE _XMFLOAT4& _XMFLOAT4::operator= |
||
| 11864 | ( |
||
| 11865 | CONST _XMFLOAT4& Float4 |
||
| 11866 | ) |
||
| 11867 | { |
||
| 11868 | x = Float4.x; |
||
| 11869 | y = Float4.y; |
||
| 11870 | z = Float4.z; |
||
| 11871 | w = Float4.w; |
||
| 11872 | return *this; |
||
| 11873 | } |
||
| 11874 | |||
| 11875 | /**************************************************************************** |
||
| 11876 | * |
||
| 11877 | * XMHALF4 operators |
||
| 11878 | * |
||
| 11879 | ****************************************************************************/ |
||
| 11880 | |||
| 11881 | //------------------------------------------------------------------------------ |
||
| 11882 | |||
| 11883 | XMFINLINE _XMHALF4::_XMHALF4 |
||
| 11884 | ( |
||
| 11885 | CONST HALF* pArray |
||
| 11886 | ) |
||
| 11887 | { |
||
| 11888 | x = pArray[0]; |
||
| 11889 | y = pArray[1]; |
||
| 11890 | z = pArray[2]; |
||
| 11891 | w = pArray[3]; |
||
| 11892 | } |
||
| 11893 | |||
| 11894 | //------------------------------------------------------------------------------ |
||
| 11895 | |||
| 11896 | XMFINLINE _XMHALF4::_XMHALF4 |
||
| 11897 | ( |
||
| 11898 | FLOAT _x, |
||
| 11899 | FLOAT _y, |
||
| 11900 | FLOAT _z, |
||
| 11901 | FLOAT _w |
||
| 11902 | ) |
||
| 11903 | { |
||
| 11904 | x = XMConvertFloatToHalf(_x); |
||
| 11905 | y = XMConvertFloatToHalf(_y); |
||
| 11906 | z = XMConvertFloatToHalf(_z); |
||
| 11907 | w = XMConvertFloatToHalf(_w); |
||
| 11908 | } |
||
| 11909 | |||
| 11910 | //------------------------------------------------------------------------------ |
||
| 11911 | |||
| 11912 | XMFINLINE _XMHALF4::_XMHALF4 |
||
| 11913 | ( |
||
| 11914 | CONST FLOAT* pArray |
||
| 11915 | ) |
||
| 11916 | { |
||
| 11917 | XMConvertFloatToHalfStream(&x, sizeof(HALF), pArray, sizeof(FLOAT), 4); |
||
| 11918 | } |
||
| 11919 | |||
| 11920 | //------------------------------------------------------------------------------ |
||
| 11921 | |||
| 11922 | XMFINLINE _XMHALF4& _XMHALF4::operator= |
||
| 11923 | ( |
||
| 11924 | CONST _XMHALF4& Half4 |
||
| 11925 | ) |
||
| 11926 | { |
||
| 11927 | x = Half4.x; |
||
| 11928 | y = Half4.y; |
||
| 11929 | z = Half4.z; |
||
| 11930 | w = Half4.w; |
||
| 11931 | return *this; |
||
| 11932 | } |
||
| 11933 | |||
| 11934 | /**************************************************************************** |
||
| 11935 | * |
||
| 11936 | * XMSHORTN4 operators |
||
| 11937 | * |
||
| 11938 | ****************************************************************************/ |
||
| 11939 | |||
| 11940 | //------------------------------------------------------------------------------ |
||
| 11941 | |||
| 11942 | XMFINLINE _XMSHORTN4::_XMSHORTN4 |
||
| 11943 | ( |
||
| 11944 | CONST SHORT* pArray |
||
| 11945 | ) |
||
| 11946 | { |
||
| 11947 | x = pArray[0]; |
||
| 11948 | y = pArray[1]; |
||
| 11949 | z = pArray[2]; |
||
| 11950 | w = pArray[3]; |
||
| 11951 | } |
||
| 11952 | |||
| 11953 | //------------------------------------------------------------------------------ |
||
| 11954 | |||
| 11955 | XMFINLINE _XMSHORTN4::_XMSHORTN4 |
||
| 11956 | ( |
||
| 11957 | FLOAT _x, |
||
| 11958 | FLOAT _y, |
||
| 11959 | FLOAT _z, |
||
| 11960 | FLOAT _w |
||
| 11961 | ) |
||
| 11962 | { |
||
| 11963 | XMStoreShortN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 11964 | } |
||
| 11965 | |||
| 11966 | //------------------------------------------------------------------------------ |
||
| 11967 | |||
| 11968 | XMFINLINE _XMSHORTN4::_XMSHORTN4 |
||
| 11969 | ( |
||
| 11970 | CONST FLOAT* pArray |
||
| 11971 | ) |
||
| 11972 | { |
||
| 11973 | XMStoreShortN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 11974 | } |
||
| 11975 | |||
| 11976 | //------------------------------------------------------------------------------ |
||
| 11977 | |||
| 11978 | XMFINLINE _XMSHORTN4& _XMSHORTN4::operator= |
||
| 11979 | ( |
||
| 11980 | CONST _XMSHORTN4& ShortN4 |
||
| 11981 | ) |
||
| 11982 | { |
||
| 11983 | x = ShortN4.x; |
||
| 11984 | y = ShortN4.y; |
||
| 11985 | z = ShortN4.z; |
||
| 11986 | w = ShortN4.w; |
||
| 11987 | return *this; |
||
| 11988 | } |
||
| 11989 | |||
| 11990 | /**************************************************************************** |
||
| 11991 | * |
||
| 11992 | * XMSHORT4 operators |
||
| 11993 | * |
||
| 11994 | ****************************************************************************/ |
||
| 11995 | |||
| 11996 | //------------------------------------------------------------------------------ |
||
| 11997 | |||
| 11998 | XMFINLINE _XMSHORT4::_XMSHORT4 |
||
| 11999 | ( |
||
| 12000 | CONST SHORT* pArray |
||
| 12001 | ) |
||
| 12002 | { |
||
| 12003 | x = pArray[0]; |
||
| 12004 | y = pArray[1]; |
||
| 12005 | z = pArray[2]; |
||
| 12006 | w = pArray[3]; |
||
| 12007 | } |
||
| 12008 | |||
| 12009 | //------------------------------------------------------------------------------ |
||
| 12010 | |||
| 12011 | XMFINLINE _XMSHORT4::_XMSHORT4 |
||
| 12012 | ( |
||
| 12013 | FLOAT _x, |
||
| 12014 | FLOAT _y, |
||
| 12015 | FLOAT _z, |
||
| 12016 | FLOAT _w |
||
| 12017 | ) |
||
| 12018 | { |
||
| 12019 | XMStoreShort4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12020 | } |
||
| 12021 | |||
| 12022 | //------------------------------------------------------------------------------ |
||
| 12023 | |||
| 12024 | XMFINLINE _XMSHORT4::_XMSHORT4 |
||
| 12025 | ( |
||
| 12026 | CONST FLOAT* pArray |
||
| 12027 | ) |
||
| 12028 | { |
||
| 12029 | XMStoreShort4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12030 | } |
||
| 12031 | |||
| 12032 | //------------------------------------------------------------------------------ |
||
| 12033 | |||
| 12034 | XMFINLINE _XMSHORT4& _XMSHORT4::operator= |
||
| 12035 | ( |
||
| 12036 | CONST _XMSHORT4& Short4 |
||
| 12037 | ) |
||
| 12038 | { |
||
| 12039 | x = Short4.x; |
||
| 12040 | y = Short4.y; |
||
| 12041 | z = Short4.z; |
||
| 12042 | w = Short4.w; |
||
| 12043 | return *this; |
||
| 12044 | } |
||
| 12045 | |||
| 12046 | /**************************************************************************** |
||
| 12047 | * |
||
| 12048 | * XMUSHORTN4 operators |
||
| 12049 | * |
||
| 12050 | ****************************************************************************/ |
||
| 12051 | |||
| 12052 | //------------------------------------------------------------------------------ |
||
| 12053 | |||
| 12054 | XMFINLINE _XMUSHORTN4::_XMUSHORTN4 |
||
| 12055 | ( |
||
| 12056 | CONST USHORT* pArray |
||
| 12057 | ) |
||
| 12058 | { |
||
| 12059 | x = pArray[0]; |
||
| 12060 | y = pArray[1]; |
||
| 12061 | z = pArray[2]; |
||
| 12062 | w = pArray[3]; |
||
| 12063 | } |
||
| 12064 | |||
| 12065 | //------------------------------------------------------------------------------ |
||
| 12066 | |||
| 12067 | XMFINLINE _XMUSHORTN4::_XMUSHORTN4 |
||
| 12068 | ( |
||
| 12069 | FLOAT _x, |
||
| 12070 | FLOAT _y, |
||
| 12071 | FLOAT _z, |
||
| 12072 | FLOAT _w |
||
| 12073 | ) |
||
| 12074 | { |
||
| 12075 | XMStoreUShortN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12076 | } |
||
| 12077 | |||
| 12078 | //------------------------------------------------------------------------------ |
||
| 12079 | |||
| 12080 | XMFINLINE _XMUSHORTN4::_XMUSHORTN4 |
||
| 12081 | ( |
||
| 12082 | CONST FLOAT* pArray |
||
| 12083 | ) |
||
| 12084 | { |
||
| 12085 | XMStoreUShortN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12086 | } |
||
| 12087 | |||
| 12088 | //------------------------------------------------------------------------------ |
||
| 12089 | |||
| 12090 | XMFINLINE _XMUSHORTN4& _XMUSHORTN4::operator= |
||
| 12091 | ( |
||
| 12092 | CONST _XMUSHORTN4& UShortN4 |
||
| 12093 | ) |
||
| 12094 | { |
||
| 12095 | x = UShortN4.x; |
||
| 12096 | y = UShortN4.y; |
||
| 12097 | z = UShortN4.z; |
||
| 12098 | w = UShortN4.w; |
||
| 12099 | return *this; |
||
| 12100 | } |
||
| 12101 | |||
| 12102 | /**************************************************************************** |
||
| 12103 | * |
||
| 12104 | * XMUSHORT4 operators |
||
| 12105 | * |
||
| 12106 | ****************************************************************************/ |
||
| 12107 | |||
| 12108 | //------------------------------------------------------------------------------ |
||
| 12109 | |||
| 12110 | XMFINLINE _XMUSHORT4::_XMUSHORT4 |
||
| 12111 | ( |
||
| 12112 | CONST USHORT* pArray |
||
| 12113 | ) |
||
| 12114 | { |
||
| 12115 | x = pArray[0]; |
||
| 12116 | y = pArray[1]; |
||
| 12117 | z = pArray[2]; |
||
| 12118 | w = pArray[3]; |
||
| 12119 | } |
||
| 12120 | |||
| 12121 | //------------------------------------------------------------------------------ |
||
| 12122 | |||
| 12123 | XMFINLINE _XMUSHORT4::_XMUSHORT4 |
||
| 12124 | ( |
||
| 12125 | FLOAT _x, |
||
| 12126 | FLOAT _y, |
||
| 12127 | FLOAT _z, |
||
| 12128 | FLOAT _w |
||
| 12129 | ) |
||
| 12130 | { |
||
| 12131 | XMStoreUShort4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12132 | } |
||
| 12133 | |||
| 12134 | //------------------------------------------------------------------------------ |
||
| 12135 | |||
| 12136 | XMFINLINE _XMUSHORT4::_XMUSHORT4 |
||
| 12137 | ( |
||
| 12138 | CONST FLOAT* pArray |
||
| 12139 | ) |
||
| 12140 | { |
||
| 12141 | XMStoreUShort4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12142 | } |
||
| 12143 | |||
| 12144 | //------------------------------------------------------------------------------ |
||
| 12145 | |||
| 12146 | XMFINLINE _XMUSHORT4& _XMUSHORT4::operator= |
||
| 12147 | ( |
||
| 12148 | CONST _XMUSHORT4& UShort4 |
||
| 12149 | ) |
||
| 12150 | { |
||
| 12151 | x = UShort4.x; |
||
| 12152 | y = UShort4.y; |
||
| 12153 | z = UShort4.z; |
||
| 12154 | w = UShort4.w; |
||
| 12155 | return *this; |
||
| 12156 | } |
||
| 12157 | |||
| 12158 | /**************************************************************************** |
||
| 12159 | * |
||
| 12160 | * XMXDECN4 operators |
||
| 12161 | * |
||
| 12162 | ****************************************************************************/ |
||
| 12163 | |||
| 12164 | //------------------------------------------------------------------------------ |
||
| 12165 | |||
| 12166 | XMFINLINE _XMXDECN4::_XMXDECN4 |
||
| 12167 | ( |
||
| 12168 | FLOAT _x, |
||
| 12169 | FLOAT _y, |
||
| 12170 | FLOAT _z, |
||
| 12171 | FLOAT _w |
||
| 12172 | ) |
||
| 12173 | { |
||
| 12174 | XMStoreXDecN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12175 | } |
||
| 12176 | |||
| 12177 | //------------------------------------------------------------------------------ |
||
| 12178 | |||
| 12179 | XMFINLINE _XMXDECN4::_XMXDECN4 |
||
| 12180 | ( |
||
| 12181 | CONST FLOAT* pArray |
||
| 12182 | ) |
||
| 12183 | { |
||
| 12184 | XMStoreXDecN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12185 | } |
||
| 12186 | |||
| 12187 | //------------------------------------------------------------------------------ |
||
| 12188 | |||
| 12189 | XMFINLINE _XMXDECN4& _XMXDECN4::operator= |
||
| 12190 | ( |
||
| 12191 | CONST _XMXDECN4& XDecN4 |
||
| 12192 | ) |
||
| 12193 | { |
||
| 12194 | v = XDecN4.v; |
||
| 12195 | return *this; |
||
| 12196 | } |
||
| 12197 | |||
| 12198 | //------------------------------------------------------------------------------ |
||
| 12199 | |||
| 12200 | XMFINLINE _XMXDECN4& _XMXDECN4::operator= |
||
| 12201 | ( |
||
| 12202 | CONST UINT Packed |
||
| 12203 | ) |
||
| 12204 | { |
||
| 12205 | v = Packed; |
||
| 12206 | return *this; |
||
| 12207 | } |
||
| 12208 | |||
| 12209 | /**************************************************************************** |
||
| 12210 | * |
||
| 12211 | * XMXDEC4 operators |
||
| 12212 | * |
||
| 12213 | ****************************************************************************/ |
||
| 12214 | |||
| 12215 | //------------------------------------------------------------------------------ |
||
| 12216 | |||
| 12217 | XMFINLINE _XMXDEC4::_XMXDEC4 |
||
| 12218 | ( |
||
| 12219 | FLOAT _x, |
||
| 12220 | FLOAT _y, |
||
| 12221 | FLOAT _z, |
||
| 12222 | FLOAT _w |
||
| 12223 | ) |
||
| 12224 | { |
||
| 12225 | XMStoreXDec4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12226 | } |
||
| 12227 | |||
| 12228 | //------------------------------------------------------------------------------ |
||
| 12229 | |||
| 12230 | XMFINLINE _XMXDEC4::_XMXDEC4 |
||
| 12231 | ( |
||
| 12232 | CONST FLOAT* pArray |
||
| 12233 | ) |
||
| 12234 | { |
||
| 12235 | XMStoreXDec4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12236 | } |
||
| 12237 | |||
| 12238 | //------------------------------------------------------------------------------ |
||
| 12239 | |||
| 12240 | XMFINLINE _XMXDEC4& _XMXDEC4::operator= |
||
| 12241 | ( |
||
| 12242 | CONST _XMXDEC4& XDec4 |
||
| 12243 | ) |
||
| 12244 | { |
||
| 12245 | v = XDec4.v; |
||
| 12246 | return *this; |
||
| 12247 | } |
||
| 12248 | |||
| 12249 | //------------------------------------------------------------------------------ |
||
| 12250 | |||
| 12251 | XMFINLINE _XMXDEC4& _XMXDEC4::operator= |
||
| 12252 | ( |
||
| 12253 | CONST UINT Packed |
||
| 12254 | ) |
||
| 12255 | { |
||
| 12256 | v = Packed; |
||
| 12257 | return *this; |
||
| 12258 | } |
||
| 12259 | |||
| 12260 | /**************************************************************************** |
||
| 12261 | * |
||
| 12262 | * XMDECN4 operators |
||
| 12263 | * |
||
| 12264 | ****************************************************************************/ |
||
| 12265 | |||
| 12266 | //------------------------------------------------------------------------------ |
||
| 12267 | |||
| 12268 | XMFINLINE _XMDECN4::_XMDECN4 |
||
| 12269 | ( |
||
| 12270 | FLOAT _x, |
||
| 12271 | FLOAT _y, |
||
| 12272 | FLOAT _z, |
||
| 12273 | FLOAT _w |
||
| 12274 | ) |
||
| 12275 | { |
||
| 12276 | XMStoreDecN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12277 | } |
||
| 12278 | |||
| 12279 | //------------------------------------------------------------------------------ |
||
| 12280 | |||
| 12281 | XMFINLINE _XMDECN4::_XMDECN4 |
||
| 12282 | ( |
||
| 12283 | CONST FLOAT* pArray |
||
| 12284 | ) |
||
| 12285 | { |
||
| 12286 | XMStoreDecN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12287 | } |
||
| 12288 | |||
| 12289 | //------------------------------------------------------------------------------ |
||
| 12290 | |||
| 12291 | XMFINLINE _XMDECN4& _XMDECN4::operator= |
||
| 12292 | ( |
||
| 12293 | CONST _XMDECN4& DecN4 |
||
| 12294 | ) |
||
| 12295 | { |
||
| 12296 | v = DecN4.v; |
||
| 12297 | return *this; |
||
| 12298 | } |
||
| 12299 | |||
| 12300 | //------------------------------------------------------------------------------ |
||
| 12301 | |||
| 12302 | XMFINLINE _XMDECN4& _XMDECN4::operator= |
||
| 12303 | ( |
||
| 12304 | CONST UINT Packed |
||
| 12305 | ) |
||
| 12306 | { |
||
| 12307 | v = Packed; |
||
| 12308 | return *this; |
||
| 12309 | } |
||
| 12310 | |||
| 12311 | /**************************************************************************** |
||
| 12312 | * |
||
| 12313 | * XMDEC4 operators |
||
| 12314 | * |
||
| 12315 | ****************************************************************************/ |
||
| 12316 | |||
| 12317 | //------------------------------------------------------------------------------ |
||
| 12318 | |||
| 12319 | XMFINLINE _XMDEC4::_XMDEC4 |
||
| 12320 | ( |
||
| 12321 | FLOAT _x, |
||
| 12322 | FLOAT _y, |
||
| 12323 | FLOAT _z, |
||
| 12324 | FLOAT _w |
||
| 12325 | ) |
||
| 12326 | { |
||
| 12327 | XMStoreDec4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12328 | } |
||
| 12329 | |||
| 12330 | //------------------------------------------------------------------------------ |
||
| 12331 | |||
| 12332 | XMFINLINE _XMDEC4::_XMDEC4 |
||
| 12333 | ( |
||
| 12334 | CONST FLOAT* pArray |
||
| 12335 | ) |
||
| 12336 | { |
||
| 12337 | XMStoreDec4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12338 | } |
||
| 12339 | |||
| 12340 | //------------------------------------------------------------------------------ |
||
| 12341 | |||
| 12342 | XMFINLINE _XMDEC4& _XMDEC4::operator= |
||
| 12343 | ( |
||
| 12344 | CONST _XMDEC4& Dec4 |
||
| 12345 | ) |
||
| 12346 | { |
||
| 12347 | v = Dec4.v; |
||
| 12348 | return *this; |
||
| 12349 | } |
||
| 12350 | |||
| 12351 | //------------------------------------------------------------------------------ |
||
| 12352 | |||
| 12353 | XMFINLINE _XMDEC4& _XMDEC4::operator= |
||
| 12354 | ( |
||
| 12355 | CONST UINT Packed |
||
| 12356 | ) |
||
| 12357 | { |
||
| 12358 | v = Packed; |
||
| 12359 | return *this; |
||
| 12360 | } |
||
| 12361 | |||
| 12362 | /**************************************************************************** |
||
| 12363 | * |
||
| 12364 | * XMUDECN4 operators |
||
| 12365 | * |
||
| 12366 | ****************************************************************************/ |
||
| 12367 | |||
| 12368 | //------------------------------------------------------------------------------ |
||
| 12369 | |||
| 12370 | XMFINLINE _XMUDECN4::_XMUDECN4 |
||
| 12371 | ( |
||
| 12372 | FLOAT _x, |
||
| 12373 | FLOAT _y, |
||
| 12374 | FLOAT _z, |
||
| 12375 | FLOAT _w |
||
| 12376 | ) |
||
| 12377 | { |
||
| 12378 | XMStoreUDecN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12379 | } |
||
| 12380 | |||
| 12381 | //------------------------------------------------------------------------------ |
||
| 12382 | |||
| 12383 | XMFINLINE _XMUDECN4::_XMUDECN4 |
||
| 12384 | ( |
||
| 12385 | CONST FLOAT* pArray |
||
| 12386 | ) |
||
| 12387 | { |
||
| 12388 | XMStoreUDecN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12389 | } |
||
| 12390 | |||
| 12391 | //------------------------------------------------------------------------------ |
||
| 12392 | |||
| 12393 | XMFINLINE _XMUDECN4& _XMUDECN4::operator= |
||
| 12394 | ( |
||
| 12395 | CONST _XMUDECN4& UDecN4 |
||
| 12396 | ) |
||
| 12397 | { |
||
| 12398 | v = UDecN4.v; |
||
| 12399 | return *this; |
||
| 12400 | } |
||
| 12401 | |||
| 12402 | //------------------------------------------------------------------------------ |
||
| 12403 | |||
| 12404 | XMFINLINE _XMUDECN4& _XMUDECN4::operator= |
||
| 12405 | ( |
||
| 12406 | CONST UINT Packed |
||
| 12407 | ) |
||
| 12408 | { |
||
| 12409 | v = Packed; |
||
| 12410 | return *this; |
||
| 12411 | } |
||
| 12412 | |||
| 12413 | /**************************************************************************** |
||
| 12414 | * |
||
| 12415 | * XMUDEC4 operators |
||
| 12416 | * |
||
| 12417 | ****************************************************************************/ |
||
| 12418 | |||
| 12419 | //------------------------------------------------------------------------------ |
||
| 12420 | |||
| 12421 | XMFINLINE _XMUDEC4::_XMUDEC4 |
||
| 12422 | ( |
||
| 12423 | FLOAT _x, |
||
| 12424 | FLOAT _y, |
||
| 12425 | FLOAT _z, |
||
| 12426 | FLOAT _w |
||
| 12427 | ) |
||
| 12428 | { |
||
| 12429 | XMStoreUDec4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12430 | } |
||
| 12431 | |||
| 12432 | //------------------------------------------------------------------------------ |
||
| 12433 | |||
| 12434 | XMFINLINE _XMUDEC4::_XMUDEC4 |
||
| 12435 | ( |
||
| 12436 | CONST FLOAT* pArray |
||
| 12437 | ) |
||
| 12438 | { |
||
| 12439 | XMStoreUDec4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12440 | } |
||
| 12441 | |||
| 12442 | //------------------------------------------------------------------------------ |
||
| 12443 | |||
| 12444 | XMFINLINE _XMUDEC4& _XMUDEC4::operator= |
||
| 12445 | ( |
||
| 12446 | CONST _XMUDEC4& UDec4 |
||
| 12447 | ) |
||
| 12448 | { |
||
| 12449 | v = UDec4.v; |
||
| 12450 | return *this; |
||
| 12451 | } |
||
| 12452 | |||
| 12453 | //------------------------------------------------------------------------------ |
||
| 12454 | |||
| 12455 | XMFINLINE _XMUDEC4& _XMUDEC4::operator= |
||
| 12456 | ( |
||
| 12457 | CONST UINT Packed |
||
| 12458 | ) |
||
| 12459 | { |
||
| 12460 | v = Packed; |
||
| 12461 | return *this; |
||
| 12462 | } |
||
| 12463 | |||
| 12464 | /**************************************************************************** |
||
| 12465 | * |
||
| 12466 | * XMXICON4 operators |
||
| 12467 | * |
||
| 12468 | ****************************************************************************/ |
||
| 12469 | |||
| 12470 | //------------------------------------------------------------------------------ |
||
| 12471 | |||
| 12472 | XMFINLINE _XMXICON4::_XMXICON4 |
||
| 12473 | ( |
||
| 12474 | FLOAT _x, |
||
| 12475 | FLOAT _y, |
||
| 12476 | FLOAT _z, |
||
| 12477 | FLOAT _w |
||
| 12478 | ) |
||
| 12479 | { |
||
| 12480 | XMStoreXIcoN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12481 | } |
||
| 12482 | |||
| 12483 | //------------------------------------------------------------------------------ |
||
| 12484 | |||
| 12485 | XMFINLINE _XMXICON4::_XMXICON4 |
||
| 12486 | ( |
||
| 12487 | CONST FLOAT* pArray |
||
| 12488 | ) |
||
| 12489 | { |
||
| 12490 | XMStoreXIcoN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12491 | } |
||
| 12492 | |||
| 12493 | //------------------------------------------------------------------------------ |
||
| 12494 | |||
| 12495 | XMFINLINE _XMXICON4& _XMXICON4::operator= |
||
| 12496 | ( |
||
| 12497 | CONST _XMXICON4& XIcoN4 |
||
| 12498 | ) |
||
| 12499 | { |
||
| 12500 | v = XIcoN4.v; |
||
| 12501 | return *this; |
||
| 12502 | } |
||
| 12503 | |||
| 12504 | //------------------------------------------------------------------------------ |
||
| 12505 | |||
| 12506 | XMFINLINE _XMXICON4& _XMXICON4::operator= |
||
| 12507 | ( |
||
| 12508 | CONST UINT64 Packed |
||
| 12509 | ) |
||
| 12510 | { |
||
| 12511 | v = Packed; |
||
| 12512 | return *this; |
||
| 12513 | } |
||
| 12514 | |||
| 12515 | /**************************************************************************** |
||
| 12516 | * |
||
| 12517 | * XMXICO4 operators |
||
| 12518 | * |
||
| 12519 | ****************************************************************************/ |
||
| 12520 | |||
| 12521 | //------------------------------------------------------------------------------ |
||
| 12522 | |||
| 12523 | XMFINLINE _XMXICO4::_XMXICO4 |
||
| 12524 | ( |
||
| 12525 | FLOAT _x, |
||
| 12526 | FLOAT _y, |
||
| 12527 | FLOAT _z, |
||
| 12528 | FLOAT _w |
||
| 12529 | ) |
||
| 12530 | { |
||
| 12531 | XMStoreXIco4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12532 | } |
||
| 12533 | |||
| 12534 | //------------------------------------------------------------------------------ |
||
| 12535 | |||
| 12536 | XMFINLINE _XMXICO4::_XMXICO4 |
||
| 12537 | ( |
||
| 12538 | CONST FLOAT* pArray |
||
| 12539 | ) |
||
| 12540 | { |
||
| 12541 | XMStoreXIco4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12542 | } |
||
| 12543 | |||
| 12544 | //------------------------------------------------------------------------------ |
||
| 12545 | |||
| 12546 | XMFINLINE _XMXICO4& _XMXICO4::operator= |
||
| 12547 | ( |
||
| 12548 | CONST _XMXICO4& XIco4 |
||
| 12549 | ) |
||
| 12550 | { |
||
| 12551 | v = XIco4.v; |
||
| 12552 | return *this; |
||
| 12553 | } |
||
| 12554 | |||
| 12555 | //------------------------------------------------------------------------------ |
||
| 12556 | |||
| 12557 | XMFINLINE _XMXICO4& _XMXICO4::operator= |
||
| 12558 | ( |
||
| 12559 | CONST UINT64 Packed |
||
| 12560 | ) |
||
| 12561 | { |
||
| 12562 | v = Packed; |
||
| 12563 | return *this; |
||
| 12564 | } |
||
| 12565 | |||
| 12566 | /**************************************************************************** |
||
| 12567 | * |
||
| 12568 | * XMICON4 operators |
||
| 12569 | * |
||
| 12570 | ****************************************************************************/ |
||
| 12571 | |||
| 12572 | //------------------------------------------------------------------------------ |
||
| 12573 | |||
| 12574 | XMFINLINE _XMICON4::_XMICON4 |
||
| 12575 | ( |
||
| 12576 | FLOAT _x, |
||
| 12577 | FLOAT _y, |
||
| 12578 | FLOAT _z, |
||
| 12579 | FLOAT _w |
||
| 12580 | ) |
||
| 12581 | { |
||
| 12582 | XMStoreIcoN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12583 | } |
||
| 12584 | |||
| 12585 | //------------------------------------------------------------------------------ |
||
| 12586 | |||
| 12587 | XMFINLINE _XMICON4::_XMICON4 |
||
| 12588 | ( |
||
| 12589 | CONST FLOAT* pArray |
||
| 12590 | ) |
||
| 12591 | { |
||
| 12592 | XMStoreIcoN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12593 | } |
||
| 12594 | |||
| 12595 | //------------------------------------------------------------------------------ |
||
| 12596 | |||
| 12597 | XMFINLINE _XMICON4& _XMICON4::operator= |
||
| 12598 | ( |
||
| 12599 | CONST _XMICON4& IcoN4 |
||
| 12600 | ) |
||
| 12601 | { |
||
| 12602 | v = IcoN4.v; |
||
| 12603 | return *this; |
||
| 12604 | } |
||
| 12605 | |||
| 12606 | //------------------------------------------------------------------------------ |
||
| 12607 | |||
| 12608 | XMFINLINE _XMICON4& _XMICON4::operator= |
||
| 12609 | ( |
||
| 12610 | CONST UINT64 Packed |
||
| 12611 | ) |
||
| 12612 | { |
||
| 12613 | v = Packed; |
||
| 12614 | return *this; |
||
| 12615 | } |
||
| 12616 | |||
| 12617 | /**************************************************************************** |
||
| 12618 | * |
||
| 12619 | * XMICO4 operators |
||
| 12620 | * |
||
| 12621 | ****************************************************************************/ |
||
| 12622 | |||
| 12623 | //------------------------------------------------------------------------------ |
||
| 12624 | |||
| 12625 | XMFINLINE _XMICO4::_XMICO4 |
||
| 12626 | ( |
||
| 12627 | FLOAT _x, |
||
| 12628 | FLOAT _y, |
||
| 12629 | FLOAT _z, |
||
| 12630 | FLOAT _w |
||
| 12631 | ) |
||
| 12632 | { |
||
| 12633 | XMStoreIco4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12634 | } |
||
| 12635 | |||
| 12636 | //------------------------------------------------------------------------------ |
||
| 12637 | |||
| 12638 | XMFINLINE _XMICO4::_XMICO4 |
||
| 12639 | ( |
||
| 12640 | CONST FLOAT* pArray |
||
| 12641 | ) |
||
| 12642 | { |
||
| 12643 | XMStoreIco4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12644 | } |
||
| 12645 | |||
| 12646 | //------------------------------------------------------------------------------ |
||
| 12647 | |||
| 12648 | XMFINLINE _XMICO4& _XMICO4::operator= |
||
| 12649 | ( |
||
| 12650 | CONST _XMICO4& Ico4 |
||
| 12651 | ) |
||
| 12652 | { |
||
| 12653 | v = Ico4.v; |
||
| 12654 | return *this; |
||
| 12655 | } |
||
| 12656 | |||
| 12657 | //------------------------------------------------------------------------------ |
||
| 12658 | |||
| 12659 | XMFINLINE _XMICO4& _XMICO4::operator= |
||
| 12660 | ( |
||
| 12661 | CONST UINT64 Packed |
||
| 12662 | ) |
||
| 12663 | { |
||
| 12664 | v = Packed; |
||
| 12665 | return *this; |
||
| 12666 | } |
||
| 12667 | |||
| 12668 | /**************************************************************************** |
||
| 12669 | * |
||
| 12670 | * XMUICON4 operators |
||
| 12671 | * |
||
| 12672 | ****************************************************************************/ |
||
| 12673 | |||
| 12674 | //------------------------------------------------------------------------------ |
||
| 12675 | |||
| 12676 | XMFINLINE _XMUICON4::_XMUICON4 |
||
| 12677 | ( |
||
| 12678 | FLOAT _x, |
||
| 12679 | FLOAT _y, |
||
| 12680 | FLOAT _z, |
||
| 12681 | FLOAT _w |
||
| 12682 | ) |
||
| 12683 | { |
||
| 12684 | XMStoreUIcoN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12685 | } |
||
| 12686 | |||
| 12687 | //------------------------------------------------------------------------------ |
||
| 12688 | |||
| 12689 | XMFINLINE _XMUICON4::_XMUICON4 |
||
| 12690 | ( |
||
| 12691 | CONST FLOAT* pArray |
||
| 12692 | ) |
||
| 12693 | { |
||
| 12694 | XMStoreUIcoN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12695 | } |
||
| 12696 | |||
| 12697 | //------------------------------------------------------------------------------ |
||
| 12698 | |||
| 12699 | XMFINLINE _XMUICON4& _XMUICON4::operator= |
||
| 12700 | ( |
||
| 12701 | CONST _XMUICON4& UIcoN4 |
||
| 12702 | ) |
||
| 12703 | { |
||
| 12704 | v = UIcoN4.v; |
||
| 12705 | return *this; |
||
| 12706 | } |
||
| 12707 | |||
| 12708 | //------------------------------------------------------------------------------ |
||
| 12709 | |||
| 12710 | XMFINLINE _XMUICON4& _XMUICON4::operator= |
||
| 12711 | ( |
||
| 12712 | CONST UINT64 Packed |
||
| 12713 | ) |
||
| 12714 | { |
||
| 12715 | v = Packed; |
||
| 12716 | return *this; |
||
| 12717 | } |
||
| 12718 | |||
| 12719 | /**************************************************************************** |
||
| 12720 | * |
||
| 12721 | * XMUICO4 operators |
||
| 12722 | * |
||
| 12723 | ****************************************************************************/ |
||
| 12724 | |||
| 12725 | //------------------------------------------------------------------------------ |
||
| 12726 | |||
| 12727 | XMFINLINE _XMUICO4::_XMUICO4 |
||
| 12728 | ( |
||
| 12729 | FLOAT _x, |
||
| 12730 | FLOAT _y, |
||
| 12731 | FLOAT _z, |
||
| 12732 | FLOAT _w |
||
| 12733 | ) |
||
| 12734 | { |
||
| 12735 | XMStoreUIco4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12736 | } |
||
| 12737 | |||
| 12738 | //------------------------------------------------------------------------------ |
||
| 12739 | |||
| 12740 | XMFINLINE _XMUICO4::_XMUICO4 |
||
| 12741 | ( |
||
| 12742 | CONST FLOAT* pArray |
||
| 12743 | ) |
||
| 12744 | { |
||
| 12745 | XMStoreUIco4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12746 | } |
||
| 12747 | |||
| 12748 | //------------------------------------------------------------------------------ |
||
| 12749 | |||
| 12750 | XMFINLINE _XMUICO4& _XMUICO4::operator= |
||
| 12751 | ( |
||
| 12752 | CONST _XMUICO4& UIco4 |
||
| 12753 | ) |
||
| 12754 | { |
||
| 12755 | v = UIco4.v; |
||
| 12756 | return *this; |
||
| 12757 | } |
||
| 12758 | |||
| 12759 | //------------------------------------------------------------------------------ |
||
| 12760 | |||
| 12761 | XMFINLINE _XMUICO4& _XMUICO4::operator= |
||
| 12762 | ( |
||
| 12763 | CONST UINT64 Packed |
||
| 12764 | ) |
||
| 12765 | { |
||
| 12766 | v = Packed; |
||
| 12767 | return *this; |
||
| 12768 | } |
||
| 12769 | |||
| 12770 | /**************************************************************************** |
||
| 12771 | * |
||
| 12772 | * XMCOLOR4 operators |
||
| 12773 | * |
||
| 12774 | ****************************************************************************/ |
||
| 12775 | |||
| 12776 | //------------------------------------------------------------------------------ |
||
| 12777 | |||
| 12778 | XMFINLINE _XMCOLOR::_XMCOLOR |
||
| 12779 | ( |
||
| 12780 | FLOAT _x, |
||
| 12781 | FLOAT _y, |
||
| 12782 | FLOAT _z, |
||
| 12783 | FLOAT _w |
||
| 12784 | ) |
||
| 12785 | { |
||
| 12786 | XMStoreColor(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12787 | } |
||
| 12788 | |||
| 12789 | //------------------------------------------------------------------------------ |
||
| 12790 | |||
| 12791 | XMFINLINE _XMCOLOR::_XMCOLOR |
||
| 12792 | ( |
||
| 12793 | CONST FLOAT* pArray |
||
| 12794 | ) |
||
| 12795 | { |
||
| 12796 | XMStoreColor(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12797 | } |
||
| 12798 | |||
| 12799 | //------------------------------------------------------------------------------ |
||
| 12800 | |||
| 12801 | XMFINLINE _XMCOLOR& _XMCOLOR::operator= |
||
| 12802 | ( |
||
| 12803 | CONST _XMCOLOR& Color |
||
| 12804 | ) |
||
| 12805 | { |
||
| 12806 | c = Color.c; |
||
| 12807 | return *this; |
||
| 12808 | } |
||
| 12809 | |||
| 12810 | //------------------------------------------------------------------------------ |
||
| 12811 | |||
| 12812 | XMFINLINE _XMCOLOR& _XMCOLOR::operator= |
||
| 12813 | ( |
||
| 12814 | CONST UINT Color |
||
| 12815 | ) |
||
| 12816 | { |
||
| 12817 | c = Color; |
||
| 12818 | return *this; |
||
| 12819 | } |
||
| 12820 | |||
| 12821 | /**************************************************************************** |
||
| 12822 | * |
||
| 12823 | * XMBYTEN4 operators |
||
| 12824 | * |
||
| 12825 | ****************************************************************************/ |
||
| 12826 | |||
| 12827 | //------------------------------------------------------------------------------ |
||
| 12828 | |||
| 12829 | XMFINLINE _XMBYTEN4::_XMBYTEN4 |
||
| 12830 | ( |
||
| 12831 | CONST CHAR* pArray |
||
| 12832 | ) |
||
| 12833 | { |
||
| 12834 | x = pArray[0]; |
||
| 12835 | y = pArray[1]; |
||
| 12836 | z = pArray[2]; |
||
| 12837 | w = pArray[3]; |
||
| 12838 | } |
||
| 12839 | |||
| 12840 | //------------------------------------------------------------------------------ |
||
| 12841 | |||
| 12842 | XMFINLINE _XMBYTEN4::_XMBYTEN4 |
||
| 12843 | ( |
||
| 12844 | FLOAT _x, |
||
| 12845 | FLOAT _y, |
||
| 12846 | FLOAT _z, |
||
| 12847 | FLOAT _w |
||
| 12848 | ) |
||
| 12849 | { |
||
| 12850 | XMStoreByteN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12851 | } |
||
| 12852 | |||
| 12853 | //------------------------------------------------------------------------------ |
||
| 12854 | |||
| 12855 | XMFINLINE _XMBYTEN4::_XMBYTEN4 |
||
| 12856 | ( |
||
| 12857 | CONST FLOAT* pArray |
||
| 12858 | ) |
||
| 12859 | { |
||
| 12860 | XMStoreByteN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12861 | } |
||
| 12862 | |||
| 12863 | //------------------------------------------------------------------------------ |
||
| 12864 | |||
| 12865 | XMFINLINE _XMBYTEN4& _XMBYTEN4::operator= |
||
| 12866 | ( |
||
| 12867 | CONST _XMBYTEN4& ByteN4 |
||
| 12868 | ) |
||
| 12869 | { |
||
| 12870 | x = ByteN4.x; |
||
| 12871 | y = ByteN4.y; |
||
| 12872 | z = ByteN4.z; |
||
| 12873 | w = ByteN4.w; |
||
| 12874 | return *this; |
||
| 12875 | } |
||
| 12876 | |||
| 12877 | /**************************************************************************** |
||
| 12878 | * |
||
| 12879 | * XMBYTE4 operators |
||
| 12880 | * |
||
| 12881 | ****************************************************************************/ |
||
| 12882 | |||
| 12883 | //------------------------------------------------------------------------------ |
||
| 12884 | |||
| 12885 | XMFINLINE _XMBYTE4::_XMBYTE4 |
||
| 12886 | ( |
||
| 12887 | CONST CHAR* pArray |
||
| 12888 | ) |
||
| 12889 | { |
||
| 12890 | x = pArray[0]; |
||
| 12891 | y = pArray[1]; |
||
| 12892 | z = pArray[2]; |
||
| 12893 | w = pArray[3]; |
||
| 12894 | } |
||
| 12895 | |||
| 12896 | //------------------------------------------------------------------------------ |
||
| 12897 | |||
| 12898 | XMFINLINE _XMBYTE4::_XMBYTE4 |
||
| 12899 | ( |
||
| 12900 | FLOAT _x, |
||
| 12901 | FLOAT _y, |
||
| 12902 | FLOAT _z, |
||
| 12903 | FLOAT _w |
||
| 12904 | ) |
||
| 12905 | { |
||
| 12906 | XMStoreByte4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12907 | } |
||
| 12908 | |||
| 12909 | //------------------------------------------------------------------------------ |
||
| 12910 | |||
| 12911 | XMFINLINE _XMBYTE4::_XMBYTE4 |
||
| 12912 | ( |
||
| 12913 | CONST FLOAT* pArray |
||
| 12914 | ) |
||
| 12915 | { |
||
| 12916 | XMStoreByte4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12917 | } |
||
| 12918 | |||
| 12919 | //------------------------------------------------------------------------------ |
||
| 12920 | |||
| 12921 | XMFINLINE _XMBYTE4& _XMBYTE4::operator= |
||
| 12922 | ( |
||
| 12923 | CONST _XMBYTE4& Byte4 |
||
| 12924 | ) |
||
| 12925 | { |
||
| 12926 | x = Byte4.x; |
||
| 12927 | y = Byte4.y; |
||
| 12928 | z = Byte4.z; |
||
| 12929 | w = Byte4.w; |
||
| 12930 | return *this; |
||
| 12931 | } |
||
| 12932 | |||
| 12933 | /**************************************************************************** |
||
| 12934 | * |
||
| 12935 | * XMUBYTEN4 operators |
||
| 12936 | * |
||
| 12937 | ****************************************************************************/ |
||
| 12938 | |||
| 12939 | //------------------------------------------------------------------------------ |
||
| 12940 | |||
| 12941 | XMFINLINE _XMUBYTEN4::_XMUBYTEN4 |
||
| 12942 | ( |
||
| 12943 | CONST BYTE* pArray |
||
| 12944 | ) |
||
| 12945 | { |
||
| 12946 | x = pArray[0]; |
||
| 12947 | y = pArray[1]; |
||
| 12948 | z = pArray[2]; |
||
| 12949 | w = pArray[3]; |
||
| 12950 | } |
||
| 12951 | |||
| 12952 | //------------------------------------------------------------------------------ |
||
| 12953 | |||
| 12954 | XMFINLINE _XMUBYTEN4::_XMUBYTEN4 |
||
| 12955 | ( |
||
| 12956 | FLOAT _x, |
||
| 12957 | FLOAT _y, |
||
| 12958 | FLOAT _z, |
||
| 12959 | FLOAT _w |
||
| 12960 | ) |
||
| 12961 | { |
||
| 12962 | XMStoreUByteN4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 12963 | } |
||
| 12964 | |||
| 12965 | //------------------------------------------------------------------------------ |
||
| 12966 | |||
| 12967 | XMFINLINE _XMUBYTEN4::_XMUBYTEN4 |
||
| 12968 | ( |
||
| 12969 | CONST FLOAT* pArray |
||
| 12970 | ) |
||
| 12971 | { |
||
| 12972 | XMStoreUByteN4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 12973 | } |
||
| 12974 | |||
| 12975 | //------------------------------------------------------------------------------ |
||
| 12976 | |||
| 12977 | XMFINLINE _XMUBYTEN4& _XMUBYTEN4::operator= |
||
| 12978 | ( |
||
| 12979 | CONST _XMUBYTEN4& UByteN4 |
||
| 12980 | ) |
||
| 12981 | { |
||
| 12982 | x = UByteN4.x; |
||
| 12983 | y = UByteN4.y; |
||
| 12984 | z = UByteN4.z; |
||
| 12985 | w = UByteN4.w; |
||
| 12986 | return *this; |
||
| 12987 | } |
||
| 12988 | |||
| 12989 | /**************************************************************************** |
||
| 12990 | * |
||
| 12991 | * XMUBYTE4 operators |
||
| 12992 | * |
||
| 12993 | ****************************************************************************/ |
||
| 12994 | |||
| 12995 | //------------------------------------------------------------------------------ |
||
| 12996 | |||
| 12997 | XMFINLINE _XMUBYTE4::_XMUBYTE4 |
||
| 12998 | ( |
||
| 12999 | CONST BYTE* pArray |
||
| 13000 | ) |
||
| 13001 | { |
||
| 13002 | x = pArray[0]; |
||
| 13003 | y = pArray[1]; |
||
| 13004 | z = pArray[2]; |
||
| 13005 | w = pArray[3]; |
||
| 13006 | } |
||
| 13007 | |||
| 13008 | //------------------------------------------------------------------------------ |
||
| 13009 | |||
| 13010 | XMFINLINE _XMUBYTE4::_XMUBYTE4 |
||
| 13011 | ( |
||
| 13012 | FLOAT _x, |
||
| 13013 | FLOAT _y, |
||
| 13014 | FLOAT _z, |
||
| 13015 | FLOAT _w |
||
| 13016 | ) |
||
| 13017 | { |
||
| 13018 | XMStoreUByte4(this, XMVectorSet(_x, _y, _z, _w)); |
||
| 13019 | } |
||
| 13020 | |||
| 13021 | //------------------------------------------------------------------------------ |
||
| 13022 | |||
| 13023 | XMFINLINE _XMUBYTE4::_XMUBYTE4 |
||
| 13024 | ( |
||
| 13025 | CONST FLOAT* pArray |
||
| 13026 | ) |
||
| 13027 | { |
||
| 13028 | XMStoreUByte4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 13029 | } |
||
| 13030 | |||
| 13031 | //------------------------------------------------------------------------------ |
||
| 13032 | |||
| 13033 | XMFINLINE _XMUBYTE4& _XMUBYTE4::operator= |
||
| 13034 | ( |
||
| 13035 | CONST _XMUBYTE4& UByte4 |
||
| 13036 | ) |
||
| 13037 | { |
||
| 13038 | x = UByte4.x; |
||
| 13039 | y = UByte4.y; |
||
| 13040 | z = UByte4.z; |
||
| 13041 | w = UByte4.w; |
||
| 13042 | return *this; |
||
| 13043 | } |
||
| 13044 | |||
| 13045 | /**************************************************************************** |
||
| 13046 | * |
||
| 13047 | * XMUNIBBLE4 operators |
||
| 13048 | * |
||
| 13049 | ****************************************************************************/ |
||
| 13050 | |||
| 13051 | //------------------------------------------------------------------------------ |
||
| 13052 | |||
| 13053 | XMFINLINE _XMUNIBBLE4::_XMUNIBBLE4 |
||
| 13054 | ( |
||
| 13055 | CONST CHAR *pArray |
||
| 13056 | ) |
||
| 13057 | { |
||
| 13058 | x = pArray[0]; |
||
| 13059 | y = pArray[1]; |
||
| 13060 | z = pArray[2]; |
||
| 13061 | w = pArray[3]; |
||
| 13062 | } |
||
| 13063 | |||
| 13064 | //------------------------------------------------------------------------------ |
||
| 13065 | |||
| 13066 | XMFINLINE _XMUNIBBLE4::_XMUNIBBLE4 |
||
| 13067 | ( |
||
| 13068 | FLOAT _x, |
||
| 13069 | FLOAT _y, |
||
| 13070 | FLOAT _z, |
||
| 13071 | FLOAT _w |
||
| 13072 | ) |
||
| 13073 | { |
||
| 13074 | XMStoreUNibble4(this, XMVectorSet( _x, _y, _z, _w )); |
||
| 13075 | } |
||
| 13076 | |||
| 13077 | //------------------------------------------------------------------------------ |
||
| 13078 | |||
| 13079 | XMFINLINE _XMUNIBBLE4::_XMUNIBBLE4 |
||
| 13080 | ( |
||
| 13081 | CONST FLOAT *pArray |
||
| 13082 | ) |
||
| 13083 | { |
||
| 13084 | XMStoreUNibble4(this, XMLoadFloat4((XMFLOAT4*)pArray)); |
||
| 13085 | } |
||
| 13086 | |||
| 13087 | //------------------------------------------------------------------------------ |
||
| 13088 | |||
| 13089 | XMFINLINE _XMUNIBBLE4& _XMUNIBBLE4::operator= |
||
| 13090 | ( |
||
| 13091 | CONST _XMUNIBBLE4& UNibble4 |
||
| 13092 | ) |
||
| 13093 | { |
||
| 13094 | v = UNibble4.v; |
||
| 13095 | return *this; |
||
| 13096 | } |
||
| 13097 | |||
| 13098 | //------------------------------------------------------------------------------ |
||
| 13099 | |||
| 13100 | XMFINLINE _XMUNIBBLE4& _XMUNIBBLE4::operator= |
||
| 13101 | ( |
||
| 13102 | CONST USHORT Packed |
||
| 13103 | ) |
||
| 13104 | { |
||
| 13105 | v = Packed; |
||
| 13106 | return *this; |
||
| 13107 | } |
||
| 13108 | |||
| 13109 | /**************************************************************************** |
||
| 13110 | * |
||
| 13111 | * XMU555 operators |
||
| 13112 | * |
||
| 13113 | ****************************************************************************/ |
||
| 13114 | |||
| 13115 | //------------------------------------------------------------------------------ |
||
| 13116 | |||
| 13117 | XMFINLINE _XMU555::_XMU555 |
||
| 13118 | ( |
||
| 13119 | CONST CHAR *pArray, |
||
| 13120 | BOOL _w |
||
| 13121 | ) |
||
| 13122 | { |
||
| 13123 | x = pArray[0]; |
||
| 13124 | y = pArray[1]; |
||
| 13125 | z = pArray[2]; |
||
| 13126 | w = _w; |
||
| 13127 | } |
||
| 13128 | |||
| 13129 | //------------------------------------------------------------------------------ |
||
| 13130 | |||
| 13131 | XMFINLINE _XMU555::_XMU555 |
||
| 13132 | ( |
||
| 13133 | FLOAT _x, |
||
| 13134 | FLOAT _y, |
||
| 13135 | FLOAT _z, |
||
| 13136 | BOOL _w |
||
| 13137 | ) |
||
| 13138 | { |
||
| 13139 | XMStoreU555(this, XMVectorSet(_x, _y, _z, ((_w) ? 1.0f : 0.0f) )); |
||
| 13140 | } |
||
| 13141 | |||
| 13142 | //------------------------------------------------------------------------------ |
||
| 13143 | |||
| 13144 | XMFINLINE _XMU555::_XMU555 |
||
| 13145 | ( |
||
| 13146 | CONST FLOAT *pArray, |
||
| 13147 | BOOL _w |
||
| 13148 | ) |
||
| 13149 | { |
||
| 13150 | XMVECTOR V = XMLoadFloat3((XMFLOAT3*)pArray); |
||
| 13151 | XMStoreU555(this, XMVectorSetW(V, ((_w) ? 1.0f : 0.0f) )); |
||
| 13152 | } |
||
| 13153 | |||
| 13154 | //------------------------------------------------------------------------------ |
||
| 13155 | |||
| 13156 | XMFINLINE _XMU555& _XMU555::operator= |
||
| 13157 | ( |
||
| 13158 | CONST _XMU555& U555 |
||
| 13159 | ) |
||
| 13160 | { |
||
| 13161 | v = U555.v; |
||
| 13162 | return *this; |
||
| 13163 | } |
||
| 13164 | |||
| 13165 | //------------------------------------------------------------------------------ |
||
| 13166 | |||
| 13167 | XMFINLINE _XMU555& _XMU555::operator= |
||
| 13168 | ( |
||
| 13169 | CONST USHORT Packed |
||
| 13170 | ) |
||
| 13171 | { |
||
| 13172 | v = Packed; |
||
| 13173 | return *this; |
||
| 13174 | } |
||
| 13175 | |||
| 13176 | #endif // __cplusplus |
||
| 13177 | |||
| 13178 | #if defined(_XM_NO_INTRINSICS_) |
||
| 13179 | #undef XMISNAN |
||
| 13180 | #undef XMISINF |
||
| 13181 | #endif |
||
| 13182 | |||
| 13183 | #endif // __XNAMATHVECTOR_INL__ |
||
| 13184 |