brintos

brintos / llvm-project-archived public Read only

0
0
Text · 3.7 KiB · f75d145 Raw
120 lines · c
1//===-- Implementation header for atanf16 -----------------------*- 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-exception6//7//===----------------------------------------------------------------------===//8 9#ifndef LLVM_LIBC_SRC___SUPPORT_MATH_ATANF16_H10#define LLVM_LIBC_SRC___SUPPORT_MATH_ATANF16_H11 12#include "include/llvm-libc-macros/float16-macros.h"13 14#ifdef LIBC_TYPES_HAS_FLOAT1615 16#include "src/__support/FPUtil/FEnvImpl.h"17#include "src/__support/FPUtil/FPBits.h"18#include "src/__support/FPUtil/PolyEval.h"19#include "src/__support/FPUtil/cast.h"20#include "src/__support/FPUtil/except_value_utils.h"21#include "src/__support/FPUtil/multiply_add.h"22#include "src/__support/FPUtil/sqrt.h"23#include "src/__support/macros/optimization.h"24 25namespace LIBC_NAMESPACE_DECL {26 27namespace math {28 29LIBC_INLINE static constexpr float16 atanf16(float16 x) {30  // Generated by Solly using the following command:31  // > round(pi/2, SG, RN);32  constexpr float PI_2 = 0x1.921fb6p0;33 34#ifndef LIBC_MATH_HAS_SKIP_ACCURATE_PASS35  constexpr size_t N_EXCEPTS = 6;36 37  constexpr fputil::ExceptValues<float16, N_EXCEPTS> ATANF16_EXCEPTS{{38      // (input, RZ output, RU offset, RD offset, RN offset)39      {0x2745, 0x2744, 1, 0, 1},40      {0x3099, 0x3090, 1, 0, 1},41      {0x3c6c, 0x3aae, 1, 0, 1},42      {0x466e, 0x3daa, 1, 0, 1},43      {0x48ae, 0x3ddb, 1, 0, 0},44      {0x5619, 0x3e3d, 1, 0, 1},45  }};46#endif // !LIBC_MATH_HAS_SKIP_ACCURATE_PASS47 48  using FPBits = fputil::FPBits<float16>;49  FPBits xbits(x);50 51  uint16_t x_u = xbits.uintval();52  uint16_t x_abs = x_u & 0x7fff;53  bool x_sign = x_u >> 15;54  float sign = (x_sign ? -1.0 : 1.0);55 56  // |x| >= +/-inf57  if (LIBC_UNLIKELY(x_abs >= 0x7c00)) {58    if (xbits.is_nan()) {59      if (xbits.is_signaling_nan()) {60        fputil::raise_except_if_required(FE_INVALID);61        return FPBits::quiet_nan().get_val();62      }63      return x;64    }65 66    // atanf16(+/-inf) = +/-pi/267    return fputil::cast<float16>(sign * PI_2);68  }69 70  float xf = x;71  float xsq = xf * xf;72#ifndef LIBC_MATH_HAS_SKIP_ACCURATE_PASS73  // Handle exceptional values74  if (auto r = ATANF16_EXCEPTS.lookup_odd(x_abs, x_sign);75      LIBC_UNLIKELY(r.has_value()))76    return r.value();77#endif78 79  // |x| <= 0x1p0, |x| <= 180  if (x_abs <= 0x3c00) {81    // atanf16(+/-0) = +/-082    if (LIBC_UNLIKELY(x_abs == 0))83      return x;84 85    // Degree-14 minimax odd polynomial of atan(x) generated by Sollya with:86    // > P = fpminimax(atan(x)/x, [|0, 2, 4, 6, 8, 10, 12, 14|], [|SG...|],87    // [0, 1]);88    float result = fputil::polyeval(89        xsq, 0x1.fffffcp-1f, -0x1.55519ep-2f, 0x1.98f6a8p-3f, -0x1.1f0a92p-3f,90        0x1.95b654p-4f, -0x1.e65492p-5f, 0x1.8c0c36p-6f, -0x1.32316ep-8f);91    return fputil::cast<float16>(xf * result);92  }93 94  // If |x| > 195  // y = atan(x) = sign(x) * atan(|x|)96  // atan(|x|) = pi/2 - atan(1/|x|)97  // Recall, 1/|x| < 198  float x_inv_sq = 1.0f / xsq;99  float x_inv = fputil::sqrt<float>(x_inv_sq);100 101  // Degree-14 minimax odd polynomial of atan(x) generated by Sollya with:102  // > P = fpminimax(atan(x)/x, [|0, 2, 4, 6, 8, 10, 12, 14|], [|SG...|],103  // [0, 1]);104  float interm =105      fputil::polyeval(x_inv_sq, 0x1.fffffcp-1f, -0x1.55519ep-2f,106                       0x1.98f6a8p-3f, -0x1.1f0a92p-3f, 0x1.95b654p-4f,107                       -0x1.e65492p-5f, 0x1.8c0c36p-6f, -0x1.32316ep-8f);108 109  return fputil::cast<float16>(sign *110                               fputil::multiply_add(x_inv, -interm, PI_2));111}112 113} // namespace math114 115} // namespace LIBC_NAMESPACE_DECL116 117#endif // LIBC_TYPES_HAS_FLOAT16118 119#endif // LLVM_LIBC_SRC___SUPPORT_MATH_ATANF16_H120