diff --git a/libc/config/linux/aarch64/entrypoints.txt b/libc/config/linux/aarch64/entrypoints.txt index c1e688ea7e86b..1558209055de6 100644 --- a/libc/config/linux/aarch64/entrypoints.txt +++ b/libc/config/linux/aarch64/entrypoints.txt @@ -644,6 +644,7 @@ endif() if(LIBC_TYPES_HAS_FLOAT16) list(APPEND TARGET_LIBM_ENTRYPOINTS # math.h C23 _Float16 entrypoints + libc.src.math.acoshf16 libc.src.math.canonicalizef16 libc.src.math.ceilf16 libc.src.math.copysignf16 diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt index 3cb9ee82752b3..cfb42344e7953 100644 --- a/libc/config/linux/x86_64/entrypoints.txt +++ b/libc/config/linux/x86_64/entrypoints.txt @@ -653,6 +653,7 @@ if(LIBC_TYPES_HAS_FLOAT16) # math.h C23 _Float16 entrypoints libc.src.math.asinf16 libc.src.math.acosf16 + libc.src.math.acoshf16 libc.src.math.canonicalizef16 libc.src.math.ceilf16 libc.src.math.copysignf16 diff --git a/libc/docs/headers/math/index.rst b/libc/docs/headers/math/index.rst index 5b855ce4881c3..8dc7c8c1a0966 100644 --- a/libc/docs/headers/math/index.rst +++ b/libc/docs/headers/math/index.rst @@ -251,7 +251,7 @@ Higher Math Functions +===========+==================+=================+========================+======================+========================+========================+============================+ | acos | |check| | | | |check| | | 7.12.4.1 | F.10.1.1 | +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+ -| acosh | |check| | | | | | 7.12.5.1 | F.10.2.1 | +| acosh | |check| | | | |check| | | 7.12.5.1 | F.10.2.1 | +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+ | acospi | | | | | | 7.12.4.8 | F.10.1.8 | +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+ diff --git a/libc/include/math.yaml b/libc/include/math.yaml index a66f981030864..7fd73445cc74e 100644 --- a/libc/include/math.yaml +++ b/libc/include/math.yaml @@ -27,6 +27,13 @@ functions: return_type: float arguments: - type: float + name: acoshf16 + standards: + - stdc + return_type: _Float16 + arguments: + - type: _Float16 + guard: LIBC_TYPES_HAS_FLOAT16 - name: asin standards: - stdc diff --git a/libc/src/math/CMakeLists.txt b/libc/src/math/CMakeLists.txt index f18a73d46f9aa..bdf8b923416d8 100644 --- a/libc/src/math/CMakeLists.txt +++ b/libc/src/math/CMakeLists.txt @@ -46,6 +46,7 @@ add_math_entrypoint_object(acosf16) add_math_entrypoint_object(acosh) add_math_entrypoint_object(acoshf) +add_math_entrypoint_object(acoshf16) add_math_entrypoint_object(asin) add_math_entrypoint_object(asinf) diff --git a/libc/src/math/acoshf16.h b/libc/src/math/acoshf16.h new file mode 100644 index 0000000000000..f471ecfeb6154 --- /dev/null +++ b/libc/src/math/acoshf16.h @@ -0,0 +1,21 @@ +//===-- Implementation header for acoshf16 ----------------------*- C++ -*-===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_LIBC_SRC_MATH_ACOSHF16_H +#define LLVM_LIBC_SRC_MATH_ACOSHF16_H + +#include "src/__support/macros/config.h" +#include "src/__support/macros/properties/types.h" + +namespace LIBC_NAMESPACE_DECL { + +float16 acoshf16(float16 x); + +} // namespace LIBC_NAMESPACE_DECL + +#endif // LLVM_LIBC_SRC_MATH_ACOSHF16_H diff --git a/libc/src/math/generic/CMakeLists.txt b/libc/src/math/generic/CMakeLists.txt index 3114289bad486..d9650f9fd6b35 100644 --- a/libc/src/math/generic/CMakeLists.txt +++ b/libc/src/math/generic/CMakeLists.txt @@ -3944,6 +3944,27 @@ add_entrypoint_object( libc.src.__support.macros.optimization ) +add_entrypoint_object( + acoshf16 + SRCS + acoshf16.cpp + HDRS + ../acoshf16.h + DEPENDS + .explogxf + libc.hdr.errno_macros + libc.hdr.fenv_macros + libc.src.__support.FPUtil.cast + libc.src.__support.FPUtil.except_value_utils + libc.src.__support.FPUtil.fenv_impl + libc.src.__support.FPUtil.fp_bits + libc.src.__support.FPUtil.multiply_add + libc.src.__support.FPUtil.polyeval + libc.src.__support.FPUtil.sqrt + libc.src.__support.macros.optimization + libc.src.__support.macros.properties.types +) + add_entrypoint_object( asinhf SRCS diff --git a/libc/src/math/generic/acoshf16.cpp b/libc/src/math/generic/acoshf16.cpp new file mode 100644 index 0000000000000..44783a8749ac2 --- /dev/null +++ b/libc/src/math/generic/acoshf16.cpp @@ -0,0 +1,110 @@ +//===-- Half-precision acosh(x) function ----------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#include "src/math/acoshf16.h" +#include "explogxf.h" +#include "hdr/errno_macros.h" +#include "hdr/fenv_macros.h" +#include "src/__support/FPUtil/FEnvImpl.h" +#include "src/__support/FPUtil/FPBits.h" +#include "src/__support/FPUtil/PolyEval.h" +#include "src/__support/FPUtil/cast.h" +#include "src/__support/FPUtil/except_value_utils.h" +#include "src/__support/FPUtil/multiply_add.h" +#include "src/__support/FPUtil/sqrt.h" +#include "src/__support/common.h" +#include "src/__support/macros/config.h" +#include "src/__support/macros/optimization.h" + +namespace LIBC_NAMESPACE_DECL { + +static constexpr size_t N_EXCEPTS = 2; +static constexpr fputil::ExceptValues ACOSHF16_EXCEPTS{{ + // (input, RZ output, RU offset, RD offset, RN offset) + // x = 0x1.6dcp+1, acoshf16(x) = 0x1.b6p+0 (RZ) + {0x41B7, 0x3ED8, 1, 0, 0}, + // x = 0x1.39p+0, acoshf16(x) = 0x1.4f8p-1 (RZ) + {0x3CE4, 0x393E, 1, 0, 1}, +}}; + +LLVM_LIBC_FUNCTION(float16, acoshf16, (float16 x)) { + using FPBits = fputil::FPBits; + FPBits xbits(x); + uint16_t x_u = xbits.uintval(); + + // Check for NaN input first. + if (LIBC_UNLIKELY(xbits.is_inf_or_nan())) { + if (xbits.is_signaling_nan()) { + fputil::raise_except_if_required(FE_INVALID); + return FPBits::quiet_nan().get_val(); + } + if (xbits.is_neg()) { + fputil::set_errno_if_required(EDOM); + fputil::raise_except_if_required(FE_INVALID); + return FPBits::quiet_nan().get_val(); + } + return x; + } + + // Domain error for inputs less than 1.0. + if (LIBC_UNLIKELY(x <= 1.0f)) { + if (x == 1.0f) + return FPBits::zero().get_val(); + fputil::set_errno_if_required(EDOM); + fputil::raise_except_if_required(FE_INVALID); + return FPBits::quiet_nan().get_val(); + } + + if (auto r = ACOSHF16_EXCEPTS.lookup(xbits.uintval()); + LIBC_UNLIKELY(r.has_value())) + return r.value(); + + float xf = x; + // High-precision polynomial approximation for inputs close to 1.0 + // ([1, 1.25)). + // + // Brief derivation: + // 1. Expand acosh(1 + delta) using Taylor series around delta=0: + // acosh(1 + delta) ≈ sqrt(2 * delta) * [1 - delta/12 + 3*delta^2/160 + // - 5*delta^3/896 + 35*delta^4/18432 + ...] + // 2. Truncate the series to fit accurately for delta in [0, 0.25]. + // 3. Polynomial coefficients (from sollya) used here are: + // P(delta) ≈ 1 - 0x1.555556p-4 * delta + 0x1.333334p-6 * delta^2 + // - 0x1.6db6dcp-8 * delta^3 + 0x1.f1c71cp-10 * delta^4 + // 4. The Sollya commands used to generate these coefficients were: + // > display = hexadecimal; + // > round(1/12, SG, RN); + // > round(3/160, SG, RN); + // > round(5/896, SG, RN); + // > round(35/18432, SG, RN); + // With hexadecimal display mode enabled, the outputs were: + // 0x1.555556p-4 + // 0x1.333334p-6 + // 0x1.6db6dcp-8 + // 0x1.f1c71cp-10 + // 5. The maximum absolute error, estimated using: + // dirtyinfnorm(acosh(1 + x) - sqrt(2*x) * P(x), [0, 0.25]) + // is: + // 0x1.d84281p-22 + if (LIBC_UNLIKELY(x_u < 0x3D00U)) { + float delta = xf - 1.0f; + float sqrt_2_delta = fputil::sqrt(2.0 * delta); + float pe = fputil::polyeval(delta, 0x1p+0f, -0x1.555556p-4f, 0x1.333334p-6f, + -0x1.6db6dcp-8f, 0x1.f1c71cp-10f); + float approx = sqrt_2_delta * pe; + return fputil::cast(approx); + } + + // acosh(x) = log(x + sqrt(x^2 - 1)) + float sqrt_term = fputil::sqrt(fputil::multiply_add(xf, xf, -1.0f)); + float result = static_cast(log_eval(xf + sqrt_term)); + + return fputil::cast(result); +} + +} // namespace LIBC_NAMESPACE_DECL diff --git a/libc/test/src/math/CMakeLists.txt b/libc/test/src/math/CMakeLists.txt index 032050bb06ec3..5dd53de0234a7 100644 --- a/libc/test/src/math/CMakeLists.txt +++ b/libc/test/src/math/CMakeLists.txt @@ -2173,6 +2173,17 @@ add_fp_unittest( libc.src.__support.FPUtil.fp_bits ) +add_fp_unittest( + acoshf16_test + NEED_MPFR + SUITE + libc-math-unittests + SRCS + acoshf16_test.cpp + DEPENDS + libc.src.math.acoshf16 +) + add_fp_unittest( asinf_test NEED_MPFR diff --git a/libc/test/src/math/acoshf16_test.cpp b/libc/test/src/math/acoshf16_test.cpp new file mode 100644 index 0000000000000..7348018396bd7 --- /dev/null +++ b/libc/test/src/math/acoshf16_test.cpp @@ -0,0 +1,28 @@ +//===-- Unittests for acoshf16 --------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#include "src/errno/libc_errno.h" +#include "src/math/acoshf16.h" +#include "test/UnitTest/FPMatcher.h" +#include "test/UnitTest/Test.h" +#include "utils/MPFRWrapper/MPFRUtils.h" +#include + +using LlvmLibcAcoshf16Test = LIBC_NAMESPACE::testing::FPTest; +namespace mpfr = LIBC_NAMESPACE::testing::mpfr; + +static constexpr uint16_t START = 0x3c00U; +static constexpr uint16_t STOP = 0x7c00; + +TEST_F(LlvmLibcAcoshf16Test, PositiveRange) { + for (uint16_t v = START; v <= STOP; ++v) { + float16 x = FPBits(v).get_val(); + EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Acosh, x, + LIBC_NAMESPACE::acoshf16(x), 0.5); + } +} diff --git a/libc/test/src/math/smoke/CMakeLists.txt b/libc/test/src/math/smoke/CMakeLists.txt index aef7c83ba0215..13cdbd6c3f8a8 100644 --- a/libc/test/src/math/smoke/CMakeLists.txt +++ b/libc/test/src/math/smoke/CMakeLists.txt @@ -3945,6 +3945,18 @@ add_fp_unittest( libc.src.__support.FPUtil.fp_bits ) +add_fp_unittest( + acoshf16_test + SUITE + libc-math-smoke-tests + SRCS + acoshf16_test.cpp + DEPENDS + libc.src.errno.errno + libc.src.math.acoshf16 + libc.src.__support.FPUtil.cast +) + add_fp_unittest( asinf_test SUITE diff --git a/libc/test/src/math/smoke/acoshf16_test.cpp b/libc/test/src/math/smoke/acoshf16_test.cpp new file mode 100644 index 0000000000000..7681c2a4e7fbc --- /dev/null +++ b/libc/test/src/math/smoke/acoshf16_test.cpp @@ -0,0 +1,70 @@ +//===-- Unittests for acoshf16 --------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#include "src/__support/FPUtil/cast.h" +#include "src/errno/libc_errno.h" +#include "src/math/acoshf16.h" +#include "test/UnitTest/FPMatcher.h" +#include "test/UnitTest/Test.h" + +using LlvmLibcAcoshf16Test = LIBC_NAMESPACE::testing::FPTest; + +TEST_F(LlvmLibcAcoshf16Test, SpecialNumbers) { + LIBC_NAMESPACE::libc_errno = 0; + EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::acoshf16(aNaN)); + EXPECT_MATH_ERRNO(0); + + EXPECT_FP_EQ_WITH_EXCEPTION(aNaN, LIBC_NAMESPACE::acoshf16(sNaN), FE_INVALID); + EXPECT_MATH_ERRNO(0); + + EXPECT_FP_EQ_WITH_EXCEPTION(aNaN, LIBC_NAMESPACE::acoshf16(zero), FE_INVALID); + EXPECT_MATH_ERRNO(EDOM); + + EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::acoshf16(neg_zero)); + EXPECT_MATH_ERRNO(EDOM); + + EXPECT_FP_EQ_WITH_EXCEPTION(aNaN, LIBC_NAMESPACE::acoshf16(neg_zero), + FE_INVALID); + EXPECT_MATH_ERRNO(EDOM); + + EXPECT_FP_EQ(inf, LIBC_NAMESPACE::acoshf16(inf)); + EXPECT_MATH_ERRNO(0); + + EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::acoshf16(neg_inf)); + EXPECT_MATH_ERRNO(EDOM); + + EXPECT_FP_EQ_WITH_EXCEPTION(aNaN, LIBC_NAMESPACE::acoshf16(neg_inf), + FE_INVALID); + EXPECT_MATH_ERRNO(EDOM); + + EXPECT_FP_EQ(zero, LIBC_NAMESPACE::acoshf16( + LIBC_NAMESPACE::fputil::cast(1.0))); + EXPECT_MATH_ERRNO(0); + + EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::acoshf16( + LIBC_NAMESPACE::fputil::cast(0.5))); + EXPECT_MATH_ERRNO(EDOM); + + EXPECT_FP_EQ_WITH_EXCEPTION( + aNaN, + LIBC_NAMESPACE::acoshf16(LIBC_NAMESPACE::fputil::cast(-1.0)), + FE_INVALID); + EXPECT_MATH_ERRNO(EDOM); + + EXPECT_FP_EQ_WITH_EXCEPTION( + aNaN, + LIBC_NAMESPACE::acoshf16(LIBC_NAMESPACE::fputil::cast(-2.0)), + FE_INVALID); + EXPECT_MATH_ERRNO(EDOM); + + EXPECT_FP_EQ_WITH_EXCEPTION( + aNaN, + LIBC_NAMESPACE::acoshf16(LIBC_NAMESPACE::fputil::cast(-3.0)), + FE_INVALID); + EXPECT_MATH_ERRNO(EDOM); +}