| 1 | //===-- Implementation for atanbf16(x) function ---------------------------===// |
| 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 | #ifndef LLVM_LIBC_SRC___SUPPORT_MATH_ATANBF16_H |
| 10 | #define LLVM_LIBC_SRC___SUPPORT_MATH_ATANBF16_H |
| 11 | |
| 12 | #include "src/__support/FPUtil/FEnvImpl.h" |
| 13 | #include "src/__support/FPUtil/FPBits.h" |
| 14 | #include "src/__support/FPUtil/PolyEval.h" |
| 15 | #include "src/__support/FPUtil/bfloat16.h" |
| 16 | #include "src/__support/FPUtil/cast.h" |
| 17 | #include "src/__support/FPUtil/multiply_add.h" |
| 18 | #include "src/__support/FPUtil/sqrt.h" |
| 19 | #include "src/__support/macros/optimization.h" |
| 20 | |
| 21 | namespace LIBC_NAMESPACE_DECL { |
| 22 | |
| 23 | namespace math { |
| 24 | |
| 25 | LIBC_INLINE bfloat16 atanbf16(bfloat16 x) { |
| 26 | // Generated by Sollya using the following command: |
| 27 | // > display = hexadecimal; |
| 28 | // > round(pi/2, SG, RN); |
| 29 | constexpr float PI_2 = 0x1.921fb6p0f; |
| 30 | |
| 31 | using FPBits = fputil::FPBits<bfloat16>; |
| 32 | FPBits xbits(x); |
| 33 | |
| 34 | uint16_t x_u = xbits.uintval(); |
| 35 | uint16_t x_abs = x_u & 0x7fff; |
| 36 | bool x_sign = x_u >> 15; |
| 37 | float sign = (x_sign ? -1.0f : 1.0f); |
| 38 | |
| 39 | // Taylor series -> [x - x^3/3 + x^5/5 - x^7/7 ...] |
| 40 | // x * [1 - x^2/3 + x^4/5 - x^6/7...] -> x * P(x) |
| 41 | // atan(x) = x * poly(x^2) |
| 42 | // atan(x)/x = poly(x^2) |
| 43 | |
| 44 | // Degree 14 polynomial of atan(x) generated using Sollya with command : |
| 45 | // > display = hexadecimal ; |
| 46 | // > P = fpminimax(atan(x)/x, [|0, 2, 4, 6, 8, 10, 12, 14|], [|1, SG, |
| 47 | // SG..SG|], [0, 1]); |
| 48 | // |
| 49 | // relative error for the polynomial given by: |
| 50 | // > dirtyinfnorm(atan(x)/x - P(x), [0, 1]); |
| 51 | // gives error ~ 0x1.ee44001p-24 |
| 52 | // worst case error for it being ~ 0x1.dcf750p-23 |
| 53 | // satisfying -> error < worst_case |
| 54 | auto atan_eval = [](float x0) { |
| 55 | return fputil::polyeval(x: x0, a0: -0x1.5552c4p-2f, a: 0x1.990f2p-3f, a: -0x1.1f7dccp-3f, |
| 56 | a: 0x1.97e49p-4f, a: -0x1.ebff34p-5f, a: 0x1.938e46p-6f, |
| 57 | a: -0x1.3a28bcp-8f); |
| 58 | }; |
| 59 | |
| 60 | float xf = x; |
| 61 | float x_sq = xf * xf; |
| 62 | |
| 63 | // |x| <= 1 |
| 64 | if (x_abs <= 0x3f80) { |
| 65 | // atanbf16(+/-0) = +/-0 |
| 66 | if (LIBC_UNLIKELY(x_abs == 0)) |
| 67 | return x; |
| 68 | |
| 69 | // For smaller x |
| 70 | if (LIBC_UNLIKELY(x_abs <= 0x3db8)) |
| 71 | return fputil::cast<bfloat16>(x: fputil::multiply_add(x: xf, y: -0x1p-9f, z: xf)); |
| 72 | |
| 73 | float result = atan_eval(x_sq); |
| 74 | return fputil::cast<bfloat16>(x: fputil::multiply_add(x: xf * x_sq, y: result, z: xf)); |
| 75 | } |
| 76 | |
| 77 | // |x| is +/-inf or NaN |
| 78 | if (LIBC_UNLIKELY(x_abs >= 0x7F80)) { |
| 79 | // NaN |
| 80 | if (xbits.is_nan()) { |
| 81 | if (xbits.is_signaling_nan()) { |
| 82 | fputil::raise_except_if_required(FE_INVALID); |
| 83 | return FPBits::quiet_nan().get_val(); |
| 84 | } |
| 85 | return x; |
| 86 | } |
| 87 | // atanbf16(+/-inf) = +/-pi/2 |
| 88 | return fputil::cast<bfloat16>(x: sign * PI_2); |
| 89 | } |
| 90 | |
| 91 | // If |x| > 1: |
| 92 | // atan(x) = sign(x) * (pi/2 - atan(1/|x|)) |
| 93 | // Since 1/|x| < 1, we can use the same polynomial. |
| 94 | float x_inv_sq = 1.0f / x_sq; |
| 95 | float x_inv = fputil::sqrt<float>(x: x_inv_sq); |
| 96 | |
| 97 | float result = atan_eval(x_inv_sq); |
| 98 | float atan_inv = fputil::multiply_add(x: x_inv * x_inv_sq, y: result, z: x_inv); |
| 99 | return fputil::cast<bfloat16>(x: sign * (PI_2 - atan_inv)); |
| 100 | } |
| 101 | |
| 102 | } // namespace math |
| 103 | } // namespace LIBC_NAMESPACE_DECL |
| 104 | |
| 105 | #endif // LLVM_LIBC_SRC___SUPPORT_MATH_ATANBF16_H |
| 106 | |