Subversion Repositories QNX 8.QNX8 LLVM/Clang compiler suite

Rev

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

  1. //===- llvm/Support/ScaledNumber.h - Support for scaled numbers -*- C++ -*-===//
  2. //
  3. // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
  4. // See https://llvm.org/LICENSE.txt for license information.
  5. // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
  6. //
  7. //===----------------------------------------------------------------------===//
  8. //
  9. // This file contains functions (and a class) useful for working with scaled
  10. // numbers -- in particular, pairs of integers where one represents digits and
  11. // another represents a scale.  The functions are helpers and live in the
  12. // namespace ScaledNumbers.  The class ScaledNumber is useful for modelling
  13. // certain cost metrics that need simple, integer-like semantics that are easy
  14. // to reason about.
  15. //
  16. // These might remind you of soft-floats.  If you want one of those, you're in
  17. // the wrong place.  Look at include/llvm/ADT/APFloat.h instead.
  18. //
  19. //===----------------------------------------------------------------------===//
  20.  
  21. #ifndef LLVM_SUPPORT_SCALEDNUMBER_H
  22. #define LLVM_SUPPORT_SCALEDNUMBER_H
  23.  
  24. #include "llvm/Support/MathExtras.h"
  25. #include <algorithm>
  26. #include <cstdint>
  27. #include <limits>
  28. #include <string>
  29. #include <tuple>
  30. #include <utility>
  31.  
  32. namespace llvm {
  33. namespace ScaledNumbers {
  34.  
  35. /// Maximum scale; same as APFloat for easy debug printing.
  36. const int32_t MaxScale = 16383;
  37.  
  38. /// Maximum scale; same as APFloat for easy debug printing.
  39. const int32_t MinScale = -16382;
  40.  
  41. /// Get the width of a number.
  42. template <class DigitsT> inline int getWidth() { return sizeof(DigitsT) * 8; }
  43.  
  44. /// Conditionally round up a scaled number.
  45. ///
  46. /// Given \c Digits and \c Scale, round up iff \c ShouldRound is \c true.
  47. /// Always returns \c Scale unless there's an overflow, in which case it
  48. /// returns \c 1+Scale.
  49. ///
  50. /// \pre adding 1 to \c Scale will not overflow INT16_MAX.
  51. template <class DigitsT>
  52. inline std::pair<DigitsT, int16_t> getRounded(DigitsT Digits, int16_t Scale,
  53.                                               bool ShouldRound) {
  54.   static_assert(!std::numeric_limits<DigitsT>::is_signed, "expected unsigned");
  55.  
  56.   if (ShouldRound)
  57.     if (!++Digits)
  58.       // Overflow.
  59.       return std::make_pair(DigitsT(1) << (getWidth<DigitsT>() - 1), Scale + 1);
  60.   return std::make_pair(Digits, Scale);
  61. }
  62.  
  63. /// Convenience helper for 32-bit rounding.
  64. inline std::pair<uint32_t, int16_t> getRounded32(uint32_t Digits, int16_t Scale,
  65.                                                  bool ShouldRound) {
  66.   return getRounded(Digits, Scale, ShouldRound);
  67. }
  68.  
  69. /// Convenience helper for 64-bit rounding.
  70. inline std::pair<uint64_t, int16_t> getRounded64(uint64_t Digits, int16_t Scale,
  71.                                                  bool ShouldRound) {
  72.   return getRounded(Digits, Scale, ShouldRound);
  73. }
  74.  
  75. /// Adjust a 64-bit scaled number down to the appropriate width.
  76. ///
  77. /// \pre Adding 64 to \c Scale will not overflow INT16_MAX.
  78. template <class DigitsT>
  79. inline std::pair<DigitsT, int16_t> getAdjusted(uint64_t Digits,
  80.                                                int16_t Scale = 0) {
  81.   static_assert(!std::numeric_limits<DigitsT>::is_signed, "expected unsigned");
  82.  
  83.   const int Width = getWidth<DigitsT>();
  84.   if (Width == 64 || Digits <= std::numeric_limits<DigitsT>::max())
  85.     return std::make_pair(Digits, Scale);
  86.  
  87.   // Shift right and round.
  88.   int Shift = 64 - Width - countLeadingZeros(Digits);
  89.   return getRounded<DigitsT>(Digits >> Shift, Scale + Shift,
  90.                              Digits & (UINT64_C(1) << (Shift - 1)));
  91. }
  92.  
  93. /// Convenience helper for adjusting to 32 bits.
  94. inline std::pair<uint32_t, int16_t> getAdjusted32(uint64_t Digits,
  95.                                                   int16_t Scale = 0) {
  96.   return getAdjusted<uint32_t>(Digits, Scale);
  97. }
  98.  
  99. /// Convenience helper for adjusting to 64 bits.
  100. inline std::pair<uint64_t, int16_t> getAdjusted64(uint64_t Digits,
  101.                                                   int16_t Scale = 0) {
  102.   return getAdjusted<uint64_t>(Digits, Scale);
  103. }
  104.  
  105. /// Multiply two 64-bit integers to create a 64-bit scaled number.
  106. ///
  107. /// Implemented with four 64-bit integer multiplies.
  108. std::pair<uint64_t, int16_t> multiply64(uint64_t LHS, uint64_t RHS);
  109.  
  110. /// Multiply two 32-bit integers to create a 32-bit scaled number.
  111. ///
  112. /// Implemented with one 64-bit integer multiply.
  113. template <class DigitsT>
  114. inline std::pair<DigitsT, int16_t> getProduct(DigitsT LHS, DigitsT RHS) {
  115.   static_assert(!std::numeric_limits<DigitsT>::is_signed, "expected unsigned");
  116.  
  117.   if (getWidth<DigitsT>() <= 32 || (LHS <= UINT32_MAX && RHS <= UINT32_MAX))
  118.     return getAdjusted<DigitsT>(uint64_t(LHS) * RHS);
  119.  
  120.   return multiply64(LHS, RHS);
  121. }
  122.  
  123. /// Convenience helper for 32-bit product.
  124. inline std::pair<uint32_t, int16_t> getProduct32(uint32_t LHS, uint32_t RHS) {
  125.   return getProduct(LHS, RHS);
  126. }
  127.  
  128. /// Convenience helper for 64-bit product.
  129. inline std::pair<uint64_t, int16_t> getProduct64(uint64_t LHS, uint64_t RHS) {
  130.   return getProduct(LHS, RHS);
  131. }
  132.  
  133. /// Divide two 64-bit integers to create a 64-bit scaled number.
  134. ///
  135. /// Implemented with long division.
  136. ///
  137. /// \pre \c Dividend and \c Divisor are non-zero.
  138. std::pair<uint64_t, int16_t> divide64(uint64_t Dividend, uint64_t Divisor);
  139.  
  140. /// Divide two 32-bit integers to create a 32-bit scaled number.
  141. ///
  142. /// Implemented with one 64-bit integer divide/remainder pair.
  143. ///
  144. /// \pre \c Dividend and \c Divisor are non-zero.
  145. std::pair<uint32_t, int16_t> divide32(uint32_t Dividend, uint32_t Divisor);
  146.  
  147. /// Divide two 32-bit numbers to create a 32-bit scaled number.
  148. ///
  149. /// Implemented with one 64-bit integer divide/remainder pair.
  150. ///
  151. /// Returns \c (DigitsT_MAX, MaxScale) for divide-by-zero (0 for 0/0).
  152. template <class DigitsT>
  153. std::pair<DigitsT, int16_t> getQuotient(DigitsT Dividend, DigitsT Divisor) {
  154.   static_assert(!std::numeric_limits<DigitsT>::is_signed, "expected unsigned");
  155.   static_assert(sizeof(DigitsT) == 4 || sizeof(DigitsT) == 8,
  156.                 "expected 32-bit or 64-bit digits");
  157.  
  158.   // Check for zero.
  159.   if (!Dividend)
  160.     return std::make_pair(0, 0);
  161.   if (!Divisor)
  162.     return std::make_pair(std::numeric_limits<DigitsT>::max(), MaxScale);
  163.  
  164.   if (getWidth<DigitsT>() == 64)
  165.     return divide64(Dividend, Divisor);
  166.   return divide32(Dividend, Divisor);
  167. }
  168.  
  169. /// Convenience helper for 32-bit quotient.
  170. inline std::pair<uint32_t, int16_t> getQuotient32(uint32_t Dividend,
  171.                                                   uint32_t Divisor) {
  172.   return getQuotient(Dividend, Divisor);
  173. }
  174.  
  175. /// Convenience helper for 64-bit quotient.
  176. inline std::pair<uint64_t, int16_t> getQuotient64(uint64_t Dividend,
  177.                                                   uint64_t Divisor) {
  178.   return getQuotient(Dividend, Divisor);
  179. }
  180.  
  181. /// Implementation of getLg() and friends.
  182. ///
  183. /// Returns the rounded lg of \c Digits*2^Scale and an int specifying whether
  184. /// this was rounded up (1), down (-1), or exact (0).
  185. ///
  186. /// Returns \c INT32_MIN when \c Digits is zero.
  187. template <class DigitsT>
  188. inline std::pair<int32_t, int> getLgImpl(DigitsT Digits, int16_t Scale) {
  189.   static_assert(!std::numeric_limits<DigitsT>::is_signed, "expected unsigned");
  190.  
  191.   if (!Digits)
  192.     return std::make_pair(INT32_MIN, 0);
  193.  
  194.   // Get the floor of the lg of Digits.
  195.   int32_t LocalFloor = sizeof(Digits) * 8 - countLeadingZeros(Digits) - 1;
  196.  
  197.   // Get the actual floor.
  198.   int32_t Floor = Scale + LocalFloor;
  199.   if (Digits == UINT64_C(1) << LocalFloor)
  200.     return std::make_pair(Floor, 0);
  201.  
  202.   // Round based on the next digit.
  203.   assert(LocalFloor >= 1);
  204.   bool Round = Digits & UINT64_C(1) << (LocalFloor - 1);
  205.   return std::make_pair(Floor + Round, Round ? 1 : -1);
  206. }
  207.  
  208. /// Get the lg (rounded) of a scaled number.
  209. ///
  210. /// Get the lg of \c Digits*2^Scale.
  211. ///
  212. /// Returns \c INT32_MIN when \c Digits is zero.
  213. template <class DigitsT> int32_t getLg(DigitsT Digits, int16_t Scale) {
  214.   return getLgImpl(Digits, Scale).first;
  215. }
  216.  
  217. /// Get the lg floor of a scaled number.
  218. ///
  219. /// Get the floor of the lg of \c Digits*2^Scale.
  220. ///
  221. /// Returns \c INT32_MIN when \c Digits is zero.
  222. template <class DigitsT> int32_t getLgFloor(DigitsT Digits, int16_t Scale) {
  223.   auto Lg = getLgImpl(Digits, Scale);
  224.   return Lg.first - (Lg.second > 0);
  225. }
  226.  
  227. /// Get the lg ceiling of a scaled number.
  228. ///
  229. /// Get the ceiling of the lg of \c Digits*2^Scale.
  230. ///
  231. /// Returns \c INT32_MIN when \c Digits is zero.
  232. template <class DigitsT> int32_t getLgCeiling(DigitsT Digits, int16_t Scale) {
  233.   auto Lg = getLgImpl(Digits, Scale);
  234.   return Lg.first + (Lg.second < 0);
  235. }
  236.  
  237. /// Implementation for comparing scaled numbers.
  238. ///
  239. /// Compare two 64-bit numbers with different scales.  Given that the scale of
  240. /// \c L is higher than that of \c R by \c ScaleDiff, compare them.  Return -1,
  241. /// 1, and 0 for less than, greater than, and equal, respectively.
  242. ///
  243. /// \pre 0 <= ScaleDiff < 64.
  244. int compareImpl(uint64_t L, uint64_t R, int ScaleDiff);
  245.  
  246. /// Compare two scaled numbers.
  247. ///
  248. /// Compare two scaled numbers.  Returns 0 for equal, -1 for less than, and 1
  249. /// for greater than.
  250. template <class DigitsT>
  251. int compare(DigitsT LDigits, int16_t LScale, DigitsT RDigits, int16_t RScale) {
  252.   static_assert(!std::numeric_limits<DigitsT>::is_signed, "expected unsigned");
  253.  
  254.   // Check for zero.
  255.   if (!LDigits)
  256.     return RDigits ? -1 : 0;
  257.   if (!RDigits)
  258.     return 1;
  259.  
  260.   // Check for the scale.  Use getLgFloor to be sure that the scale difference
  261.   // is always lower than 64.
  262.   int32_t lgL = getLgFloor(LDigits, LScale), lgR = getLgFloor(RDigits, RScale);
  263.   if (lgL != lgR)
  264.     return lgL < lgR ? -1 : 1;
  265.  
  266.   // Compare digits.
  267.   if (LScale < RScale)
  268.     return compareImpl(LDigits, RDigits, RScale - LScale);
  269.  
  270.   return -compareImpl(RDigits, LDigits, LScale - RScale);
  271. }
  272.  
  273. /// Match scales of two numbers.
  274. ///
  275. /// Given two scaled numbers, match up their scales.  Change the digits and
  276. /// scales in place.  Shift the digits as necessary to form equivalent numbers,
  277. /// losing precision only when necessary.
  278. ///
  279. /// If the output value of \c LDigits (\c RDigits) is \c 0, the output value of
  280. /// \c LScale (\c RScale) is unspecified.
  281. ///
  282. /// As a convenience, returns the matching scale.  If the output value of one
  283. /// number is zero, returns the scale of the other.  If both are zero, which
  284. /// scale is returned is unspecified.
  285. template <class DigitsT>
  286. int16_t matchScales(DigitsT &LDigits, int16_t &LScale, DigitsT &RDigits,
  287.                     int16_t &RScale) {
  288.   static_assert(!std::numeric_limits<DigitsT>::is_signed, "expected unsigned");
  289.  
  290.   if (LScale < RScale)
  291.     // Swap arguments.
  292.     return matchScales(RDigits, RScale, LDigits, LScale);
  293.   if (!LDigits)
  294.     return RScale;
  295.   if (!RDigits || LScale == RScale)
  296.     return LScale;
  297.  
  298.   // Now LScale > RScale.  Get the difference.
  299.   int32_t ScaleDiff = int32_t(LScale) - RScale;
  300.   if (ScaleDiff >= 2 * getWidth<DigitsT>()) {
  301.     // Don't bother shifting.  RDigits will get zero-ed out anyway.
  302.     RDigits = 0;
  303.     return LScale;
  304.   }
  305.  
  306.   // Shift LDigits left as much as possible, then shift RDigits right.
  307.   int32_t ShiftL = std::min<int32_t>(countLeadingZeros(LDigits), ScaleDiff);
  308.   assert(ShiftL < getWidth<DigitsT>() && "can't shift more than width");
  309.  
  310.   int32_t ShiftR = ScaleDiff - ShiftL;
  311.   if (ShiftR >= getWidth<DigitsT>()) {
  312.     // Don't bother shifting.  RDigits will get zero-ed out anyway.
  313.     RDigits = 0;
  314.     return LScale;
  315.   }
  316.  
  317.   LDigits <<= ShiftL;
  318.   RDigits >>= ShiftR;
  319.  
  320.   LScale -= ShiftL;
  321.   RScale += ShiftR;
  322.   assert(LScale == RScale && "scales should match");
  323.   return LScale;
  324. }
  325.  
  326. /// Get the sum of two scaled numbers.
  327. ///
  328. /// Get the sum of two scaled numbers with as much precision as possible.
  329. ///
  330. /// \pre Adding 1 to \c LScale (or \c RScale) will not overflow INT16_MAX.
  331. template <class DigitsT>
  332. std::pair<DigitsT, int16_t> getSum(DigitsT LDigits, int16_t LScale,
  333.                                    DigitsT RDigits, int16_t RScale) {
  334.   static_assert(!std::numeric_limits<DigitsT>::is_signed, "expected unsigned");
  335.  
  336.   // Check inputs up front.  This is only relevant if addition overflows, but
  337.   // testing here should catch more bugs.
  338.   assert(LScale < INT16_MAX && "scale too large");
  339.   assert(RScale < INT16_MAX && "scale too large");
  340.  
  341.   // Normalize digits to match scales.
  342.   int16_t Scale = matchScales(LDigits, LScale, RDigits, RScale);
  343.  
  344.   // Compute sum.
  345.   DigitsT Sum = LDigits + RDigits;
  346.   if (Sum >= RDigits)
  347.     return std::make_pair(Sum, Scale);
  348.  
  349.   // Adjust sum after arithmetic overflow.
  350.   DigitsT HighBit = DigitsT(1) << (getWidth<DigitsT>() - 1);
  351.   return std::make_pair(HighBit | Sum >> 1, Scale + 1);
  352. }
  353.  
  354. /// Convenience helper for 32-bit sum.
  355. inline std::pair<uint32_t, int16_t> getSum32(uint32_t LDigits, int16_t LScale,
  356.                                              uint32_t RDigits, int16_t RScale) {
  357.   return getSum(LDigits, LScale, RDigits, RScale);
  358. }
  359.  
  360. /// Convenience helper for 64-bit sum.
  361. inline std::pair<uint64_t, int16_t> getSum64(uint64_t LDigits, int16_t LScale,
  362.                                              uint64_t RDigits, int16_t RScale) {
  363.   return getSum(LDigits, LScale, RDigits, RScale);
  364. }
  365.  
  366. /// Get the difference of two scaled numbers.
  367. ///
  368. /// Get LHS minus RHS with as much precision as possible.
  369. ///
  370. /// Returns \c (0, 0) if the RHS is larger than the LHS.
  371. template <class DigitsT>
  372. std::pair<DigitsT, int16_t> getDifference(DigitsT LDigits, int16_t LScale,
  373.                                           DigitsT RDigits, int16_t RScale) {
  374.   static_assert(!std::numeric_limits<DigitsT>::is_signed, "expected unsigned");
  375.  
  376.   // Normalize digits to match scales.
  377.   const DigitsT SavedRDigits = RDigits;
  378.   const int16_t SavedRScale = RScale;
  379.   matchScales(LDigits, LScale, RDigits, RScale);
  380.  
  381.   // Compute difference.
  382.   if (LDigits <= RDigits)
  383.     return std::make_pair(0, 0);
  384.   if (RDigits || !SavedRDigits)
  385.     return std::make_pair(LDigits - RDigits, LScale);
  386.  
  387.   // Check if RDigits just barely lost its last bit.  E.g., for 32-bit:
  388.   //
  389.   //   1*2^32 - 1*2^0 == 0xffffffff != 1*2^32
  390.   const auto RLgFloor = getLgFloor(SavedRDigits, SavedRScale);
  391.   if (!compare(LDigits, LScale, DigitsT(1), RLgFloor + getWidth<DigitsT>()))
  392.     return std::make_pair(std::numeric_limits<DigitsT>::max(), RLgFloor);
  393.  
  394.   return std::make_pair(LDigits, LScale);
  395. }
  396.  
  397. /// Convenience helper for 32-bit difference.
  398. inline std::pair<uint32_t, int16_t> getDifference32(uint32_t LDigits,
  399.                                                     int16_t LScale,
  400.                                                     uint32_t RDigits,
  401.                                                     int16_t RScale) {
  402.   return getDifference(LDigits, LScale, RDigits, RScale);
  403. }
  404.  
  405. /// Convenience helper for 64-bit difference.
  406. inline std::pair<uint64_t, int16_t> getDifference64(uint64_t LDigits,
  407.                                                     int16_t LScale,
  408.                                                     uint64_t RDigits,
  409.                                                     int16_t RScale) {
  410.   return getDifference(LDigits, LScale, RDigits, RScale);
  411. }
  412.  
  413. } // end namespace ScaledNumbers
  414. } // end namespace llvm
  415.  
  416. namespace llvm {
  417.  
  418. class raw_ostream;
  419. class ScaledNumberBase {
  420. public:
  421.   static constexpr int DefaultPrecision = 10;
  422.  
  423.   static void dump(uint64_t D, int16_t E, int Width);
  424.   static raw_ostream &print(raw_ostream &OS, uint64_t D, int16_t E, int Width,
  425.                             unsigned Precision);
  426.   static std::string toString(uint64_t D, int16_t E, int Width,
  427.                               unsigned Precision);
  428.   static int countLeadingZeros32(uint32_t N) { return countLeadingZeros(N); }
  429.   static int countLeadingZeros64(uint64_t N) { return countLeadingZeros(N); }
  430.   static uint64_t getHalf(uint64_t N) { return (N >> 1) + (N & 1); }
  431.  
  432.   static std::pair<uint64_t, bool> splitSigned(int64_t N) {
  433.     if (N >= 0)
  434.       return std::make_pair(N, false);
  435.     uint64_t Unsigned = N == INT64_MIN ? UINT64_C(1) << 63 : uint64_t(-N);
  436.     return std::make_pair(Unsigned, true);
  437.   }
  438.   static int64_t joinSigned(uint64_t U, bool IsNeg) {
  439.     if (U > uint64_t(INT64_MAX))
  440.       return IsNeg ? INT64_MIN : INT64_MAX;
  441.     return IsNeg ? -int64_t(U) : int64_t(U);
  442.   }
  443. };
  444.  
  445. /// Simple representation of a scaled number.
  446. ///
  447. /// ScaledNumber is a number represented by digits and a scale.  It uses simple
  448. /// saturation arithmetic and every operation is well-defined for every value.
  449. /// It's somewhat similar in behaviour to a soft-float, but is *not* a
  450. /// replacement for one.  If you're doing numerics, look at \a APFloat instead.
  451. /// Nevertheless, we've found these semantics useful for modelling certain cost
  452. /// metrics.
  453. ///
  454. /// The number is split into a signed scale and unsigned digits.  The number
  455. /// represented is \c getDigits()*2^getScale().  In this way, the digits are
  456. /// much like the mantissa in the x87 long double, but there is no canonical
  457. /// form so the same number can be represented by many bit representations.
  458. ///
  459. /// ScaledNumber is templated on the underlying integer type for digits, which
  460. /// is expected to be unsigned.
  461. ///
  462. /// Unlike APFloat, ScaledNumber does not model architecture floating point
  463. /// behaviour -- while this might make it a little faster and easier to reason
  464. /// about, it certainly makes it more dangerous for general numerics.
  465. ///
  466. /// ScaledNumber is totally ordered.  However, there is no canonical form, so
  467. /// there are multiple representations of most scalars.  E.g.:
  468. ///
  469. ///     ScaledNumber(8u, 0) == ScaledNumber(4u, 1)
  470. ///     ScaledNumber(4u, 1) == ScaledNumber(2u, 2)
  471. ///     ScaledNumber(2u, 2) == ScaledNumber(1u, 3)
  472. ///
  473. /// ScaledNumber implements most arithmetic operations.  Precision is kept
  474. /// where possible.  Uses simple saturation arithmetic, so that operations
  475. /// saturate to 0.0 or getLargest() rather than under or overflowing.  It has
  476. /// some extra arithmetic for unit inversion.  0.0/0.0 is defined to be 0.0.
  477. /// Any other division by 0.0 is defined to be getLargest().
  478. ///
  479. /// As a convenience for modifying the exponent, left and right shifting are
  480. /// both implemented, and both interpret negative shifts as positive shifts in
  481. /// the opposite direction.
  482. ///
  483. /// Scales are limited to the range accepted by x87 long double.  This makes
  484. /// it trivial to add functionality to convert to APFloat (this is already
  485. /// relied on for the implementation of printing).
  486. ///
  487. /// Possible (and conflicting) future directions:
  488. ///
  489. ///  1. Turn this into a wrapper around \a APFloat.
  490. ///  2. Share the algorithm implementations with \a APFloat.
  491. ///  3. Allow \a ScaledNumber to represent a signed number.
  492. template <class DigitsT> class ScaledNumber : ScaledNumberBase {
  493. public:
  494.   static_assert(!std::numeric_limits<DigitsT>::is_signed,
  495.                 "only unsigned floats supported");
  496.  
  497.   typedef DigitsT DigitsType;
  498.  
  499. private:
  500.   typedef std::numeric_limits<DigitsType> DigitsLimits;
  501.  
  502.   static constexpr int Width = sizeof(DigitsType) * 8;
  503.   static_assert(Width <= 64, "invalid integer width for digits");
  504.  
  505. private:
  506.   DigitsType Digits = 0;
  507.   int16_t Scale = 0;
  508.  
  509. public:
  510.   ScaledNumber() = default;
  511.  
  512.   constexpr ScaledNumber(DigitsType Digits, int16_t Scale)
  513.       : Digits(Digits), Scale(Scale) {}
  514.  
  515. private:
  516.   ScaledNumber(const std::pair<DigitsT, int16_t> &X)
  517.       : Digits(X.first), Scale(X.second) {}
  518.  
  519. public:
  520.   static ScaledNumber getZero() { return ScaledNumber(0, 0); }
  521.   static ScaledNumber getOne() { return ScaledNumber(1, 0); }
  522.   static ScaledNumber getLargest() {
  523.     return ScaledNumber(DigitsLimits::max(), ScaledNumbers::MaxScale);
  524.   }
  525.   static ScaledNumber get(uint64_t N) { return adjustToWidth(N, 0); }
  526.   static ScaledNumber getInverse(uint64_t N) {
  527.     return get(N).invert();
  528.   }
  529.   static ScaledNumber getFraction(DigitsType N, DigitsType D) {
  530.     return getQuotient(N, D);
  531.   }
  532.  
  533.   int16_t getScale() const { return Scale; }
  534.   DigitsType getDigits() const { return Digits; }
  535.  
  536.   /// Convert to the given integer type.
  537.   ///
  538.   /// Convert to \c IntT using simple saturating arithmetic, truncating if
  539.   /// necessary.
  540.   template <class IntT> IntT toInt() const;
  541.  
  542.   bool isZero() const { return !Digits; }
  543.   bool isLargest() const { return *this == getLargest(); }
  544.   bool isOne() const {
  545.     if (Scale > 0 || Scale <= -Width)
  546.       return false;
  547.     return Digits == DigitsType(1) << -Scale;
  548.   }
  549.  
  550.   /// The log base 2, rounded.
  551.   ///
  552.   /// Get the lg of the scalar.  lg 0 is defined to be INT32_MIN.
  553.   int32_t lg() const { return ScaledNumbers::getLg(Digits, Scale); }
  554.  
  555.   /// The log base 2, rounded towards INT32_MIN.
  556.   ///
  557.   /// Get the lg floor.  lg 0 is defined to be INT32_MIN.
  558.   int32_t lgFloor() const { return ScaledNumbers::getLgFloor(Digits, Scale); }
  559.  
  560.   /// The log base 2, rounded towards INT32_MAX.
  561.   ///
  562.   /// Get the lg ceiling.  lg 0 is defined to be INT32_MIN.
  563.   int32_t lgCeiling() const {
  564.     return ScaledNumbers::getLgCeiling(Digits, Scale);
  565.   }
  566.  
  567.   bool operator==(const ScaledNumber &X) const { return compare(X) == 0; }
  568.   bool operator<(const ScaledNumber &X) const { return compare(X) < 0; }
  569.   bool operator!=(const ScaledNumber &X) const { return compare(X) != 0; }
  570.   bool operator>(const ScaledNumber &X) const { return compare(X) > 0; }
  571.   bool operator<=(const ScaledNumber &X) const { return compare(X) <= 0; }
  572.   bool operator>=(const ScaledNumber &X) const { return compare(X) >= 0; }
  573.  
  574.   bool operator!() const { return isZero(); }
  575.  
  576.   /// Convert to a decimal representation in a string.
  577.   ///
  578.   /// Convert to a string.  Uses scientific notation for very large/small
  579.   /// numbers.  Scientific notation is used roughly for numbers outside of the
  580.   /// range 2^-64 through 2^64.
  581.   ///
  582.   /// \c Precision indicates the number of decimal digits of precision to use;
  583.   /// 0 requests the maximum available.
  584.   ///
  585.   /// As a special case to make debugging easier, if the number is small enough
  586.   /// to convert without scientific notation and has more than \c Precision
  587.   /// digits before the decimal place, it's printed accurately to the first
  588.   /// digit past zero.  E.g., assuming 10 digits of precision:
  589.   ///
  590.   ///     98765432198.7654... => 98765432198.8
  591.   ///      8765432198.7654... =>  8765432198.8
  592.   ///       765432198.7654... =>   765432198.8
  593.   ///        65432198.7654... =>    65432198.77
  594.   ///         5432198.7654... =>     5432198.765
  595.   std::string toString(unsigned Precision = DefaultPrecision) {
  596.     return ScaledNumberBase::toString(Digits, Scale, Width, Precision);
  597.   }
  598.  
  599.   /// Print a decimal representation.
  600.   ///
  601.   /// Print a string.  See toString for documentation.
  602.   raw_ostream &print(raw_ostream &OS,
  603.                      unsigned Precision = DefaultPrecision) const {
  604.     return ScaledNumberBase::print(OS, Digits, Scale, Width, Precision);
  605.   }
  606.   void dump() const { return ScaledNumberBase::dump(Digits, Scale, Width); }
  607.  
  608.   ScaledNumber &operator+=(const ScaledNumber &X) {
  609.     std::tie(Digits, Scale) =
  610.         ScaledNumbers::getSum(Digits, Scale, X.Digits, X.Scale);
  611.     // Check for exponent past MaxScale.
  612.     if (Scale > ScaledNumbers::MaxScale)
  613.       *this = getLargest();
  614.     return *this;
  615.   }
  616.   ScaledNumber &operator-=(const ScaledNumber &X) {
  617.     std::tie(Digits, Scale) =
  618.         ScaledNumbers::getDifference(Digits, Scale, X.Digits, X.Scale);
  619.     return *this;
  620.   }
  621.   ScaledNumber &operator*=(const ScaledNumber &X);
  622.   ScaledNumber &operator/=(const ScaledNumber &X);
  623.   ScaledNumber &operator<<=(int16_t Shift) {
  624.     shiftLeft(Shift);
  625.     return *this;
  626.   }
  627.   ScaledNumber &operator>>=(int16_t Shift) {
  628.     shiftRight(Shift);
  629.     return *this;
  630.   }
  631.  
  632. private:
  633.   void shiftLeft(int32_t Shift);
  634.   void shiftRight(int32_t Shift);
  635.  
  636.   /// Adjust two floats to have matching exponents.
  637.   ///
  638.   /// Adjust \c this and \c X to have matching exponents.  Returns the new \c X
  639.   /// by value.  Does nothing if \a isZero() for either.
  640.   ///
  641.   /// The value that compares smaller will lose precision, and possibly become
  642.   /// \a isZero().
  643.   ScaledNumber matchScales(ScaledNumber X) {
  644.     ScaledNumbers::matchScales(Digits, Scale, X.Digits, X.Scale);
  645.     return X;
  646.   }
  647.  
  648. public:
  649.   /// Scale a large number accurately.
  650.   ///
  651.   /// Scale N (multiply it by this).  Uses full precision multiplication, even
  652.   /// if Width is smaller than 64, so information is not lost.
  653.   uint64_t scale(uint64_t N) const;
  654.   uint64_t scaleByInverse(uint64_t N) const {
  655.     // TODO: implement directly, rather than relying on inverse.  Inverse is
  656.     // expensive.
  657.     return inverse().scale(N);
  658.   }
  659.   int64_t scale(int64_t N) const {
  660.     std::pair<uint64_t, bool> Unsigned = splitSigned(N);
  661.     return joinSigned(scale(Unsigned.first), Unsigned.second);
  662.   }
  663.   int64_t scaleByInverse(int64_t N) const {
  664.     std::pair<uint64_t, bool> Unsigned = splitSigned(N);
  665.     return joinSigned(scaleByInverse(Unsigned.first), Unsigned.second);
  666.   }
  667.  
  668.   int compare(const ScaledNumber &X) const {
  669.     return ScaledNumbers::compare(Digits, Scale, X.Digits, X.Scale);
  670.   }
  671.   int compareTo(uint64_t N) const {
  672.     return ScaledNumbers::compare<uint64_t>(Digits, Scale, N, 0);
  673.   }
  674.   int compareTo(int64_t N) const { return N < 0 ? 1 : compareTo(uint64_t(N)); }
  675.  
  676.   ScaledNumber &invert() { return *this = ScaledNumber::get(1) / *this; }
  677.   ScaledNumber inverse() const { return ScaledNumber(*this).invert(); }
  678.  
  679. private:
  680.   static ScaledNumber getProduct(DigitsType LHS, DigitsType RHS) {
  681.     return ScaledNumbers::getProduct(LHS, RHS);
  682.   }
  683.   static ScaledNumber getQuotient(DigitsType Dividend, DigitsType Divisor) {
  684.     return ScaledNumbers::getQuotient(Dividend, Divisor);
  685.   }
  686.  
  687.   static int countLeadingZerosWidth(DigitsType Digits) {
  688.     if (Width == 64)
  689.       return countLeadingZeros64(Digits);
  690.     if (Width == 32)
  691.       return countLeadingZeros32(Digits);
  692.     return countLeadingZeros32(Digits) + Width - 32;
  693.   }
  694.  
  695.   /// Adjust a number to width, rounding up if necessary.
  696.   ///
  697.   /// Should only be called for \c Shift close to zero.
  698.   ///
  699.   /// \pre Shift >= MinScale && Shift + 64 <= MaxScale.
  700.   static ScaledNumber adjustToWidth(uint64_t N, int32_t Shift) {
  701.     assert(Shift >= ScaledNumbers::MinScale && "Shift should be close to 0");
  702.     assert(Shift <= ScaledNumbers::MaxScale - 64 &&
  703.            "Shift should be close to 0");
  704.     auto Adjusted = ScaledNumbers::getAdjusted<DigitsT>(N, Shift);
  705.     return Adjusted;
  706.   }
  707.  
  708.   static ScaledNumber getRounded(ScaledNumber P, bool Round) {
  709.     // Saturate.
  710.     if (P.isLargest())
  711.       return P;
  712.  
  713.     return ScaledNumbers::getRounded(P.Digits, P.Scale, Round);
  714.   }
  715. };
  716.  
  717. #define SCALED_NUMBER_BOP(op, base)                                            \
  718.   template <class DigitsT>                                                     \
  719.   ScaledNumber<DigitsT> operator op(const ScaledNumber<DigitsT> &L,            \
  720.                                     const ScaledNumber<DigitsT> &R) {          \
  721.     return ScaledNumber<DigitsT>(L) base R;                                    \
  722.   }
  723. SCALED_NUMBER_BOP(+, += )
  724. SCALED_NUMBER_BOP(-, -= )
  725. SCALED_NUMBER_BOP(*, *= )
  726. SCALED_NUMBER_BOP(/, /= )
  727. #undef SCALED_NUMBER_BOP
  728.  
  729. template <class DigitsT>
  730. ScaledNumber<DigitsT> operator<<(const ScaledNumber<DigitsT> &L,
  731.                                  int16_t Shift) {
  732.   return ScaledNumber<DigitsT>(L) <<= Shift;
  733. }
  734.  
  735. template <class DigitsT>
  736. ScaledNumber<DigitsT> operator>>(const ScaledNumber<DigitsT> &L,
  737.                                  int16_t Shift) {
  738.   return ScaledNumber<DigitsT>(L) >>= Shift;
  739. }
  740.  
  741. template <class DigitsT>
  742. raw_ostream &operator<<(raw_ostream &OS, const ScaledNumber<DigitsT> &X) {
  743.   return X.print(OS, 10);
  744. }
  745.  
  746. #define SCALED_NUMBER_COMPARE_TO_TYPE(op, T1, T2)                              \
  747.   template <class DigitsT>                                                     \
  748.   bool operator op(const ScaledNumber<DigitsT> &L, T1 R) {                     \
  749.     return L.compareTo(T2(R)) op 0;                                            \
  750.   }                                                                            \
  751.   template <class DigitsT>                                                     \
  752.   bool operator op(T1 L, const ScaledNumber<DigitsT> &R) {                     \
  753.     return 0 op R.compareTo(T2(L));                                            \
  754.   }
  755. #define SCALED_NUMBER_COMPARE_TO(op)                                           \
  756.   SCALED_NUMBER_COMPARE_TO_TYPE(op, uint64_t, uint64_t)                        \
  757.   SCALED_NUMBER_COMPARE_TO_TYPE(op, uint32_t, uint64_t)                        \
  758.   SCALED_NUMBER_COMPARE_TO_TYPE(op, int64_t, int64_t)                          \
  759.   SCALED_NUMBER_COMPARE_TO_TYPE(op, int32_t, int64_t)
  760. SCALED_NUMBER_COMPARE_TO(< )
  761. SCALED_NUMBER_COMPARE_TO(> )
  762. SCALED_NUMBER_COMPARE_TO(== )
  763. SCALED_NUMBER_COMPARE_TO(!= )
  764. SCALED_NUMBER_COMPARE_TO(<= )
  765. SCALED_NUMBER_COMPARE_TO(>= )
  766. #undef SCALED_NUMBER_COMPARE_TO
  767. #undef SCALED_NUMBER_COMPARE_TO_TYPE
  768.  
  769. template <class DigitsT>
  770. uint64_t ScaledNumber<DigitsT>::scale(uint64_t N) const {
  771.   if (Width == 64 || N <= DigitsLimits::max())
  772.     return (get(N) * *this).template toInt<uint64_t>();
  773.  
  774.   // Defer to the 64-bit version.
  775.   return ScaledNumber<uint64_t>(Digits, Scale).scale(N);
  776. }
  777.  
  778. template <class DigitsT>
  779. template <class IntT>
  780. IntT ScaledNumber<DigitsT>::toInt() const {
  781.   typedef std::numeric_limits<IntT> Limits;
  782.   if (*this < 1)
  783.     return 0;
  784.   if (*this >= Limits::max())
  785.     return Limits::max();
  786.  
  787.   IntT N = Digits;
  788.   if (Scale > 0) {
  789.     assert(size_t(Scale) < sizeof(IntT) * 8);
  790.     return N << Scale;
  791.   }
  792.   if (Scale < 0) {
  793.     assert(size_t(-Scale) < sizeof(IntT) * 8);
  794.     return N >> -Scale;
  795.   }
  796.   return N;
  797. }
  798.  
  799. template <class DigitsT>
  800. ScaledNumber<DigitsT> &ScaledNumber<DigitsT>::
  801. operator*=(const ScaledNumber &X) {
  802.   if (isZero())
  803.     return *this;
  804.   if (X.isZero())
  805.     return *this = X;
  806.  
  807.   // Save the exponents.
  808.   int32_t Scales = int32_t(Scale) + int32_t(X.Scale);
  809.  
  810.   // Get the raw product.
  811.   *this = getProduct(Digits, X.Digits);
  812.  
  813.   // Combine with exponents.
  814.   return *this <<= Scales;
  815. }
  816. template <class DigitsT>
  817. ScaledNumber<DigitsT> &ScaledNumber<DigitsT>::
  818. operator/=(const ScaledNumber &X) {
  819.   if (isZero())
  820.     return *this;
  821.   if (X.isZero())
  822.     return *this = getLargest();
  823.  
  824.   // Save the exponents.
  825.   int32_t Scales = int32_t(Scale) - int32_t(X.Scale);
  826.  
  827.   // Get the raw quotient.
  828.   *this = getQuotient(Digits, X.Digits);
  829.  
  830.   // Combine with exponents.
  831.   return *this <<= Scales;
  832. }
  833. template <class DigitsT> void ScaledNumber<DigitsT>::shiftLeft(int32_t Shift) {
  834.   if (!Shift || isZero())
  835.     return;
  836.   assert(Shift != INT32_MIN);
  837.   if (Shift < 0) {
  838.     shiftRight(-Shift);
  839.     return;
  840.   }
  841.  
  842.   // Shift as much as we can in the exponent.
  843.   int32_t ScaleShift = std::min(Shift, ScaledNumbers::MaxScale - Scale);
  844.   Scale += ScaleShift;
  845.   if (ScaleShift == Shift)
  846.     return;
  847.  
  848.   // Check this late, since it's rare.
  849.   if (isLargest())
  850.     return;
  851.  
  852.   // Shift the digits themselves.
  853.   Shift -= ScaleShift;
  854.   if (Shift > countLeadingZerosWidth(Digits)) {
  855.     // Saturate.
  856.     *this = getLargest();
  857.     return;
  858.   }
  859.  
  860.   Digits <<= Shift;
  861. }
  862.  
  863. template <class DigitsT> void ScaledNumber<DigitsT>::shiftRight(int32_t Shift) {
  864.   if (!Shift || isZero())
  865.     return;
  866.   assert(Shift != INT32_MIN);
  867.   if (Shift < 0) {
  868.     shiftLeft(-Shift);
  869.     return;
  870.   }
  871.  
  872.   // Shift as much as we can in the exponent.
  873.   int32_t ScaleShift = std::min(Shift, Scale - ScaledNumbers::MinScale);
  874.   Scale -= ScaleShift;
  875.   if (ScaleShift == Shift)
  876.     return;
  877.  
  878.   // Shift the digits themselves.
  879.   Shift -= ScaleShift;
  880.   if (Shift >= Width) {
  881.     // Saturate.
  882.     *this = getZero();
  883.     return;
  884.   }
  885.  
  886.   Digits >>= Shift;
  887. }
  888.  
  889.  
  890. } // end namespace llvm
  891.  
  892. #endif // LLVM_SUPPORT_SCALEDNUMBER_H
  893.