[libc][math][c23] Add logf16 C23 math function #106072
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
54 tasks
|
@llvm/pr-subscribers-libc Author: OverMighty (overmighty) ChangesPart of #95250. Full diff: https://github.com/llvm/llvm-project/pull/106072.diff 14 Files Affected:
diff --git a/libc/config/gpu/entrypoints.txt b/libc/config/gpu/entrypoints.txt
index b206bb7c936f39..e6a8e381a9212a 100644
--- a/libc/config/gpu/entrypoints.txt
+++ b/libc/config/gpu/entrypoints.txt
@@ -532,6 +532,7 @@ if(LIBC_TYPES_HAS_FLOAT16)
libc.src.math.llrintf16
libc.src.math.llroundf16
libc.src.math.logbf16
+ libc.src.math.logf16
libc.src.math.lrintf16
libc.src.math.lroundf16
libc.src.math.modff16
diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt
index b702a7ff482a5c..69944ce3904562 100644
--- a/libc/config/linux/x86_64/entrypoints.txt
+++ b/libc/config/linux/x86_64/entrypoints.txt
@@ -643,6 +643,7 @@ if(LIBC_TYPES_HAS_FLOAT16)
libc.src.math.llrintf16
libc.src.math.llroundf16
libc.src.math.logbf16
+ libc.src.math.logf16
libc.src.math.lrintf16
libc.src.math.lroundf16
libc.src.math.modff16
diff --git a/libc/docs/math/index.rst b/libc/docs/math/index.rst
index de88f14426799f..e958312b313589 100644
--- a/libc/docs/math/index.rst
+++ b/libc/docs/math/index.rst
@@ -306,7 +306,7 @@ Higher Math Functions
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
| lgamma | | | | | | 7.12.8.3 | F.10.5.3 |
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
-| log | |check| | |check| | | | | 7.12.6.11 | F.10.3.11 |
+| log | |check| | |check| | | |check| | | 7.12.6.11 | F.10.3.11 |
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
| log10 | |check| | |check| | | | | 7.12.6.12 | F.10.3.12 |
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
diff --git a/libc/spec/stdc.td b/libc/spec/stdc.td
index 043762bb227456..c3115b19349ff4 100644
--- a/libc/spec/stdc.td
+++ b/libc/spec/stdc.td
@@ -565,6 +565,7 @@ def StdC : StandardSpec<"stdc"> {
FunctionSpec<"log", RetValSpec<DoubleType>, [ArgSpec<DoubleType>]>,
FunctionSpec<"logf", RetValSpec<FloatType>, [ArgSpec<FloatType>]>,
+ GuardedFunctionSpec<"logf16", RetValSpec<Float16Type>, [ArgSpec<Float16Type>], "LIBC_TYPES_HAS_FLOAT16">,
FunctionSpec<"logb", RetValSpec<DoubleType>, [ArgSpec<DoubleType>]>,
FunctionSpec<"logbf", RetValSpec<FloatType>, [ArgSpec<FloatType>]>,
diff --git a/libc/src/math/CMakeLists.txt b/libc/src/math/CMakeLists.txt
index be05903d4a1c6e..47c4ad188ef6df 100644
--- a/libc/src/math/CMakeLists.txt
+++ b/libc/src/math/CMakeLists.txt
@@ -331,6 +331,7 @@ add_math_entrypoint_object(log2f)
add_math_entrypoint_object(log)
add_math_entrypoint_object(logf)
+add_math_entrypoint_object(logf16)
add_math_entrypoint_object(logb)
add_math_entrypoint_object(logbf)
diff --git a/libc/src/math/generic/CMakeLists.txt b/libc/src/math/generic/CMakeLists.txt
index 9534d17426486b..6a9d9f02f190f0 100644
--- a/libc/src/math/generic/CMakeLists.txt
+++ b/libc/src/math/generic/CMakeLists.txt
@@ -2210,6 +2210,28 @@ add_entrypoint_object(
-O3
)
+add_entrypoint_object(
+ logf16
+ SRCS
+ logf16.cpp
+ HDRS
+ ../logf16.h
+ DEPENDS
+ .expxf16
+ libc.hdr.errno_macros
+ libc.hdr.fenv_macros
+ libc.src.__support.CPP.array
+ 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.macros.optimization
+ libc.src.__support.macros.properties.cpu_features
+ COMPILE_OPTIONS
+ -O3
+)
+
add_entrypoint_object(
logb
SRCS
diff --git a/libc/src/math/generic/expxf16.h b/libc/src/math/generic/expxf16.h
index 3893305d5bfa18..d64bb432811067 100644
--- a/libc/src/math/generic/expxf16.h
+++ b/libc/src/math/generic/expxf16.h
@@ -287,6 +287,32 @@ template <bool IsSinh> LIBC_INLINE float16 eval_sinh_or_cosh(float16 x) {
lo, half_p_odd * exp2_hi_mid_diff, half_p_even * exp2_hi_mid_sum));
}
+// Generated by Sollya with the following commands:
+// > display = hexadecimal;
+// > for i from 0 to 31 do print(round(log(1 + i * 2^-5), SG, RN));
+constexpr cpp::array<float, 32> LOGF_F = {
+ 0x0p+0, 0x1.f829bp-6, 0x1.f0a30cp-5, 0x1.6f0d28p-4, 0x1.e27076p-4,
+ 0x1.29553p-3, 0x1.5ff308p-3, 0x1.9525aap-3, 0x1.c8ff7cp-3, 0x1.fb9186p-3,
+ 0x1.1675cap-2, 0x1.2e8e2cp-2, 0x1.4618bcp-2, 0x1.5d1bdcp-2, 0x1.739d8p-2,
+ 0x1.89a338p-2, 0x1.9f323ep-2, 0x1.b44f78p-2, 0x1.c8ff7cp-2, 0x1.dd46ap-2,
+ 0x1.f128f6p-2, 0x1.02552ap-1, 0x1.0be72ep-1, 0x1.154c3ep-1, 0x1.1e85f6p-1,
+ 0x1.2795e2p-1, 0x1.307d74p-1, 0x1.393e0ep-1, 0x1.41d8fep-1, 0x1.4a4f86p-1,
+ 0x1.52a2d2p-1, 0x1.5ad404p-1,
+};
+
+// Generated by Sollya with the following commands:
+// > display = hexadecimal;
+// > for i from 0 to 31 do print(round(1 / (1 + i * 2^-5), SG, RN));
+constexpr cpp::array<float, 32> ONE_OVER_F = {
+ 0x1p+0, 0x1.f07c2p-1, 0x1.e1e1e2p-1, 0x1.d41d42p-1, 0x1.c71c72p-1,
+ 0x1.bacf92p-1, 0x1.af286cp-1, 0x1.a41a42p-1, 0x1.99999ap-1, 0x1.8f9c18p-1,
+ 0x1.861862p-1, 0x1.7d05f4p-1, 0x1.745d18p-1, 0x1.6c16c2p-1, 0x1.642c86p-1,
+ 0x1.5c9882p-1, 0x1.555556p-1, 0x1.4e5e0ap-1, 0x1.47ae14p-1, 0x1.414142p-1,
+ 0x1.3b13b2p-1, 0x1.3521dp-1, 0x1.2f684cp-1, 0x1.29e412p-1, 0x1.24924ap-1,
+ 0x1.1f7048p-1, 0x1.1a7b96p-1, 0x1.15b1e6p-1, 0x1.111112p-1, 0x1.0c9714p-1,
+ 0x1.08421p-1, 0x1.041042p-1,
+};
+
} // namespace LIBC_NAMESPACE_DECL
#endif // LLVM_LIBC_SRC_MATH_GENERIC_EXPXF16_H
diff --git a/libc/src/math/generic/logf16.cpp b/libc/src/math/generic/logf16.cpp
new file mode 100644
index 00000000000000..adb24306a0f466
--- /dev/null
+++ b/libc/src/math/generic/logf16.cpp
@@ -0,0 +1,158 @@
+//===-- Half-precision log(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/logf16.h"
+#include "expxf16.h"
+#include "hdr/errno_macros.h"
+#include "hdr/fenv_macros.h"
+#include "src/__support/CPP/array.h"
+#include "src/__support/FPUtil/FEnvImpl.h"
+#include "src/__support/FPUtil/FPBits.h"
+#include "src/__support/FPUtil/PolyEval.h"
+#include "src/__support/FPUtil/except_value_utils.h"
+#include "src/__support/FPUtil/multiply_add.h"
+#include "src/__support/common.h"
+#include "src/__support/macros/config.h"
+#include "src/__support/macros/optimization.h"
+#include "src/__support/macros/properties/cpu_features.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+#ifdef LIBC_TARGET_CPU_HAS_FMA
+static constexpr size_t N_LOGF16_EXCEPTS = 5;
+#else
+static constexpr size_t N_LOGF16_EXCEPTS = 11;
+#endif
+
+static constexpr fputil::ExceptValues<float16, N_LOGF16_EXCEPTS>
+ LOGF16_EXCEPTS = {{
+// (input, RZ output, RU offset, RD offset, RN offset)
+#ifndef LIBC_TARGET_CPU_HAS_FMA
+ // x = 0x1.61cp-13, logf16(x) = -0x1.16p+3 (RZ)
+ {0x0987U, 0xc858U, 0U, 1U, 0U},
+ // x = 0x1.f2p-12, logf16(x) = -0x1.e98p+2 (RZ)
+ {0x0fc8U, 0xc7a6U, 0U, 1U, 1U},
+// x = 0x1.4d4p-9, logf16(x) = -0x1.7e4p+2 (RZ)
+#endif
+ // x = 0x1.4d4p-9, logf16(x) = -0x1.7e4p+2 (RZ)
+ {0x1935U, 0xc5f9U, 0U, 1U, 0U},
+ // x = 0x1.5ep-8, logf16(x) = -0x1.4ecp+2 (RZ)
+ {0x1d78U, 0xc53bU, 0U, 1U, 0U},
+#ifndef LIBC_TARGET_CPU_HAS_FMA
+ // x = 0x1.fdp-1, logf16(x) = -0x1.81p-8 (RZ)
+ {0x3bf4U, 0x9e04U, 0U, 1U, 1U},
+ // x = 0x1.fep-1, logf16(x) = -0x1.008p-8 (RZ)
+ {0x3bf8U, 0x9c02U, 0U, 1U, 0U},
+#endif
+ // x = 0x1.ffp-1, logf16(x) = -0x1.004p-9 (RZ)
+ {0x3bfcU, 0x9801U, 0U, 1U, 0U},
+ // x = 0x1.ff8p-1, logf16(x) = -0x1p-10 (RZ)
+ {0x3bfeU, 0x9400U, 0U, 1U, 1U},
+#ifdef LIBC_TARGET_CPU_HAS_FMA
+ // x = 0x1.4c4p+1, logf16(x) = 0x1.e84p-1 (RZ)
+ {0x4131U, 0x3ba1U, 1U, 0U, 1U},
+#else
+ // x = 0x1.75p+2, logf16(x) = 0x1.c34p+0 (RZ)
+ {0x45d4U, 0x3f0dU, 1U, 0U, 0U},
+ // x = 0x1.75p+2, logf16(x) = 0x1.c34p+0 (RZ)
+ {0x45d4U, 0x3f0dU, 1U, 0U, 0U},
+ // x = 0x1.d5p+9, logf16(x) = 0x1.b5cp+2 (RZ)
+ {0x6354U, 0x46d7U, 1U, 0U, 1U},
+#endif
+ }};
+
+LLVM_LIBC_FUNCTION(float16, logf16, (float16 x)) {
+ using FPBits = fputil::FPBits<float16>;
+ FPBits x_bits(x);
+
+ uint16_t x_u = x_bits.uintval();
+
+ // If x <= 0, or x is 1, or x is +inf, or x is NaN.
+ if (LIBC_UNLIKELY(x_u == 0U || x_u == 0x3c00U || x_u >= 0x7c00U)) {
+ // log(NaN) = NaN
+ if (x_bits.is_nan()) {
+ if (x_bits.is_signaling_nan()) {
+ fputil::raise_except_if_required(FE_INVALID);
+ return FPBits::quiet_nan().get_val();
+ }
+
+ return x;
+ }
+
+ // log(+/-0) = −inf
+ if ((x_u & 0x7fffU) == 0U) {
+ fputil::raise_except_if_required(FE_DIVBYZERO);
+ return FPBits::inf(Sign::NEG).get_val();
+ }
+
+ if (x_u == 0x3c00U)
+ return FPBits::zero().get_val();
+
+ // When x < 0.
+ if (x_u > 0x8000) {
+ fputil::set_errno_if_required(EDOM);
+ fputil::raise_except_if_required(FE_INVALID);
+ return FPBits::quiet_nan().get_val();
+ }
+
+ // log(+inf) = +inf
+ return FPBits::inf().get_val();
+ }
+
+ if (auto r = LOGF16_EXCEPTS.lookup(x_u); LIBC_UNLIKELY(r.has_value()))
+ return r.value();
+
+ // To compute log(x), we perform the following range reduction:
+ // x = 2^m * 1.mant,
+ // log(x) = m * log(2) + log(1.mant).
+ // To compute log(1.mant), let f be the highest 6 bits including the hidden
+ // bit, and d be the difference (1.mant - f), i.e., the remaining 5 bits of
+ // the mantissa, then:
+ // log(1.mant) = log(f) + log(1.mant / f)
+ // = log(f) + log(1 + d/f)
+ // since d/f is sufficiently small.
+ // We store log(f) and 1/f in the lookup tables LOGF_F and ONE_OVER_F
+ // respectively.
+
+ int m = -FPBits::EXP_BIAS;
+
+ // When x is subnormal.
+ if ((x_u & FPBits::EXP_MASK) == 0U) {
+ // Normalize x.
+ x_bits = FPBits(x_bits.get_val() *
+ static_cast<float16>((1U << FPBits::FRACTION_LEN)));
+ x_u = x_bits.uintval();
+ m -= FPBits::FRACTION_LEN;
+ }
+
+ uint16_t mant = x_bits.get_mantissa();
+ // Leading 10 - 5 = 5 bits of the mantissa.
+ int f = mant >> 5;
+ // Unbiased exponent.
+ m += x_u >> FPBits::FRACTION_LEN;
+
+ // Set bits to 1.mant instead of 2^m * 1.mant.
+ x_bits.set_biased_exponent(FPBits::EXP_BIAS);
+ float mant_f = x_bits.get_val();
+ // v = 1.mant * 1/f - 1 = d/f
+ float v = fputil::multiply_add(mant_f, ONE_OVER_F[f], -1.0f);
+
+ // Degree-3 minimax polynomial generated by Sollya with the following
+ // commands:
+ // > display = hexadecimal;
+ // > P = fpminimax(log(1 + x)/x, 2, [|SG...|], [-2^-5, 2^-5]);
+ // > x * P;
+ float log1p_d_over_f =
+ v * fputil::polyeval(v, 0x1p+0f, -0x1.001804p-1f, 0x1.557ef6p-2f);
+ // log(1.mant) = log(f) + log(1 + d/f)
+ float log_1_mant = LOGF_F[f] + log1p_d_over_f;
+ return static_cast<float16>(
+ fputil::multiply_add(static_cast<float>(m), LOGF_2, log_1_mant));
+}
+
+} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/src/math/logf16.h b/libc/src/math/logf16.h
new file mode 100644
index 00000000000000..e2d296b1d90881
--- /dev/null
+++ b/libc/src/math/logf16.h
@@ -0,0 +1,21 @@
+//===-- Implementation header for logf16 ------------------------*- 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_LOGF16_H
+#define LLVM_LIBC_SRC_MATH_LOGF16_H
+
+#include "src/__support/macros/config.h"
+#include "src/__support/macros/properties/types.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+float16 logf16(float16 x);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_MATH_LOGF16_H
diff --git a/libc/test/UnitTest/FPMatcher.h b/libc/test/UnitTest/FPMatcher.h
index 43752a4942ad56..9a695885101013 100644
--- a/libc/test/UnitTest/FPMatcher.h
+++ b/libc/test/UnitTest/FPMatcher.h
@@ -266,4 +266,17 @@ template <typename T> struct FPTest : public Test {
EXPECT_FP_EQ_WITH_EXCEPTION_ROUNDING_MODE( \
(expected), (actual), (expected_except), RoundingMode::TowardZero)
+#define EXPECT_FP_EQ_WITH_EXCEPTION_ALL_ROUNDING(expected, actual, \
+ expected_except) \
+ do { \
+ EXPECT_FP_EQ_WITH_EXCEPTION_ROUNDING_NEAREST((expected), (actual), \
+ (expected_except)); \
+ EXPECT_FP_EQ_WITH_EXCEPTION_ROUNDING_UPWARD((expected), (actual), \
+ (expected_except)); \
+ EXPECT_FP_EQ_WITH_EXCEPTION_ROUNDING_DOWNWARD((expected), (actual), \
+ (expected_except)); \
+ EXPECT_FP_EQ_WITH_EXCEPTION_ROUNDING_TOWARD_ZERO((expected), (actual), \
+ (expected_except)); \
+ } while (0)
+
#endif // LLVM_LIBC_TEST_UNITTEST_FPMATCHER_H
diff --git a/libc/test/src/math/CMakeLists.txt b/libc/test/src/math/CMakeLists.txt
index d2a75bd7c8e1f9..fabc1b51f5b0c8 100644
--- a/libc/test/src/math/CMakeLists.txt
+++ b/libc/test/src/math/CMakeLists.txt
@@ -1761,6 +1761,17 @@ add_fp_unittest(
libc.src.__support.FPUtil.fp_bits
)
+add_fp_unittest(
+ logf16_test
+ NEED_MPFR
+ SUITE
+ libc-math-unittests
+ SRCS
+ logf16_test.cpp
+ DEPENDS
+ libc.src.math.logf16
+)
+
add_fp_unittest(
log2_test
NEED_MPFR
diff --git a/libc/test/src/math/logf16_test.cpp b/libc/test/src/math/logf16_test.cpp
new file mode 100644
index 00000000000000..67e5e172e53b8c
--- /dev/null
+++ b/libc/test/src/math/logf16_test.cpp
@@ -0,0 +1,40 @@
+//===-- Exhaustive test for logf16 ---------------------------------------===//
+//
+// 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/logf16.h"
+#include "test/UnitTest/FPMatcher.h"
+#include "test/UnitTest/Test.h"
+#include "utils/MPFRWrapper/MPFRUtils.h"
+
+using LlvmLibcLogf16Test = LIBC_NAMESPACE::testing::FPTest<float16>;
+
+namespace mpfr = LIBC_NAMESPACE::testing::mpfr;
+
+// Range: [0, Inf];
+static constexpr uint16_t POS_START = 0x0000U;
+static constexpr uint16_t POS_STOP = 0x7c00U;
+
+// Range: [-Inf, 0];
+static constexpr uint16_t NEG_START = 0x8000U;
+static constexpr uint16_t NEG_STOP = 0xfc00U;
+
+TEST_F(LlvmLibcLogf16Test, PositiveRange) {
+ for (uint16_t v = POS_START; v <= POS_STOP; ++v) {
+ float16 x = FPBits(v).get_val();
+ EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Log, x,
+ LIBC_NAMESPACE::logf16(x), 0.5);
+ }
+}
+
+TEST_F(LlvmLibcLogf16Test, NegativeRange) {
+ for (uint16_t v = NEG_START; v <= NEG_STOP; ++v) {
+ float16 x = FPBits(v).get_val();
+ EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Log, x,
+ LIBC_NAMESPACE::logf16(x), 0.5);
+ }
+}
diff --git a/libc/test/src/math/smoke/CMakeLists.txt b/libc/test/src/math/smoke/CMakeLists.txt
index db21c90acc1f3f..f3aa81506ac4ac 100644
--- a/libc/test/src/math/smoke/CMakeLists.txt
+++ b/libc/test/src/math/smoke/CMakeLists.txt
@@ -3408,6 +3408,18 @@ add_fp_unittest(
libc.src.__support.FPUtil.fp_bits
)
+add_fp_unittest(
+ logf16_test
+ SUITE
+ libc-math-smoke-tests
+ SRCS
+ logf16_test.cpp
+ DEPENDS
+ libc.hdr.fenv_macros
+ libc.src.errno.errno
+ libc.src.math.logf16
+)
+
add_fp_unittest(
log2_test
SUITE
diff --git a/libc/test/src/math/smoke/logf16_test.cpp b/libc/test/src/math/smoke/logf16_test.cpp
new file mode 100644
index 00000000000000..e2f21003e8d81b
--- /dev/null
+++ b/libc/test/src/math/smoke/logf16_test.cpp
@@ -0,0 +1,47 @@
+//===-- Unittests for logf16 ----------------------------------------------===//
+//
+// 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 "hdr/fenv_macros.h"
+#include "src/errno/libc_errno.h"
+#include "src/math/logf16.h"
+#include "test/UnitTest/FPMatcher.h"
+#include "test/UnitTest/Test.h"
+
+using LlvmLibcLogf16Test = LIBC_NAMESPACE::testing::FPTest<float16>;
+
+TEST_F(LlvmLibcLogf16Test, SpecialNumbers) {
+ LIBC_NAMESPACE::libc_errno = 0;
+
+ EXPECT_FP_EQ_ALL_ROUNDING(aNaN, LIBC_NAMESPACE::logf16(aNaN));
+ EXPECT_MATH_ERRNO(0);
+
+ EXPECT_FP_EQ_WITH_EXCEPTION(aNaN, LIBC_NAMESPACE::logf16(sNaN), FE_INVALID);
+ EXPECT_MATH_ERRNO(0);
+
+ EXPECT_FP_EQ_ALL_ROUNDING(inf, LIBC_NAMESPACE::logf16(inf));
+ EXPECT_MATH_ERRNO(0);
+
+ EXPECT_FP_EQ_ALL_ROUNDING(aNaN, LIBC_NAMESPACE::logf16(neg_inf));
+ EXPECT_MATH_ERRNO(EDOM);
+
+ EXPECT_FP_EQ_WITH_EXCEPTION_ALL_ROUNDING(
+ neg_inf, LIBC_NAMESPACE::logf16(zero), FE_DIVBYZERO);
+ EXPECT_MATH_ERRNO(0);
+
+ EXPECT_FP_EQ_WITH_EXCEPTION_ALL_ROUNDING(
+ neg_inf, LIBC_NAMESPACE::logf16(neg_zero), FE_DIVBYZERO);
+ EXPECT_MATH_ERRNO(0);
+
+ EXPECT_FP_EQ_ALL_ROUNDING(zero,
+ LIBC_NAMESPACE::logf16(static_cast<float16>(1.0)));
+ EXPECT_MATH_ERRNO(0);
+
+ EXPECT_FP_EQ_ALL_ROUNDING(aNaN,
+ LIBC_NAMESPACE::logf16(static_cast<float16>(-1.0)));
+ EXPECT_MATH_ERRNO(EDOM);
+}
|
overmighty
commented
Aug 26, 2024
libc/src/math/generic/expxf16.h
Outdated
| // Generated by Sollya with the following commands: | ||
| // > display = hexadecimal; | ||
| // > for i from 0 to 31 do print(round(1 / (1 + i * 2^-5), SG, RN)); | ||
| constexpr cpp::array<float, 32> ONE_OVER_F = { |
There was a problem hiding this comment.
Maybe I shouldn't have used the same ONE_OVER_F name that common_constants.h uses for the double 1/f lookup table.
There was a problem hiding this comment.
yes,I think we should at least add type suffix to these tables.
There was a problem hiding this comment.
Couldn't think of a better name than ONE_OVER_F_F since I usually add _F as suffix for float constants.
overmighty
force-pushed
the
users/overmighty/libc-math-tanhf16
branch
from
eadda81 to
c343f44
Compare
overmighty
force-pushed
the
users/overmighty/libc-math-logf16
branch
from
37db333 to
cbc95e7
Compare
overmighty
force-pushed
the
users/overmighty/libc-math-tanhf16
branch
from
c343f44 to
4758caf
Compare
overmighty
force-pushed
the
users/overmighty/libc-math-logf16
branch
from
cbc95e7 to
0541aa4
Compare
overmighty
force-pushed
the
users/overmighty/libc-math-logf16
branch
from
0541aa4 to
9450748
Compare
lntue
approved these changes
Oct 18, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Part of #95250.