[libc-commits] [libc] Add sinf16 function (PR #116674)
via libc-commits
libc-commits at lists.llvm.org
Wed Nov 27 14:48:20 PST 2024
https://github.com/wldfngrs updated https://github.com/llvm/llvm-project/pull/116674
>From 49b868b06308000df1c8bf50ccf41ea0d4c4dcfb Mon Sep 17 00:00:00 2001
From: wldfngrs <wldfngrs at gmail.com>
Date: Mon, 18 Nov 2024 18:57:45 +0100
Subject: [PATCH 1/8] Add sinf16 function
---
libc/config/linux/x86_64/entrypoints.txt | 1 +
libc/newhdrgen/yaml/math.yaml | 7 +++
libc/src/math/CMakeLists.txt | 1 +
libc/src/math/generic/CMakeLists.txt | 19 ++++++
libc/src/math/generic/sincosf16_utils.h | 35 +++++++++--
libc/src/math/generic/sinf16.cpp | 77 ++++++++++++++++++++++++
libc/src/math/sinf16.h | 21 +++++++
libc/test/src/math/CMakeLists.txt | 11 ++++
libc/test/src/math/sinf16_test.cpp | 38 ++++++++++++
libc/test/src/math/smoke/CMakeLists.txt | 11 ++++
libc/test/src/math/smoke/sinf16_test.cpp | 34 +++++++++++
11 files changed, 251 insertions(+), 4 deletions(-)
create mode 100644 libc/src/math/generic/sinf16.cpp
create mode 100644 libc/src/math/sinf16.h
create mode 100644 libc/test/src/math/sinf16_test.cpp
create mode 100644 libc/test/src/math/smoke/sinf16_test.cpp
diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt
index 957e28bd66cc4c..99f4e34262a8df 100644
--- a/libc/config/linux/x86_64/entrypoints.txt
+++ b/libc/config/linux/x86_64/entrypoints.txt
@@ -699,6 +699,7 @@ if(LIBC_TYPES_HAS_FLOAT16)
libc.src.math.scalbnf16
libc.src.math.setpayloadf16
libc.src.math.setpayloadsigf16
+ libc.src.math.sinf16
libc.src.math.sinhf16
libc.src.math.sinpif16
libc.src.math.sqrtf16
diff --git a/libc/newhdrgen/yaml/math.yaml b/libc/newhdrgen/yaml/math.yaml
index e09f0929e45f80..00efc34789667e 100644
--- a/libc/newhdrgen/yaml/math.yaml
+++ b/libc/newhdrgen/yaml/math.yaml
@@ -2339,6 +2339,13 @@ functions:
return_type: float
arguments:
- type: float
+ - name: sinf16
+ standards:
+ - stdc
+ return_type: _Float16
+ arguments:
+ - type: _Float16
+ guard: LIBC_TYPES_HAS_FLOAT16
- name: sinhf
standards:
- stdc
diff --git a/libc/src/math/CMakeLists.txt b/libc/src/math/CMakeLists.txt
index 76a5e491effa05..390a59d07a28bc 100644
--- a/libc/src/math/CMakeLists.txt
+++ b/libc/src/math/CMakeLists.txt
@@ -484,6 +484,7 @@ add_math_entrypoint_object(sincosf)
add_math_entrypoint_object(sin)
add_math_entrypoint_object(sinf)
+add_math_entrypoint_object(sinf16)
add_math_entrypoint_object(sinpif)
add_math_entrypoint_object(sinpif16)
diff --git a/libc/src/math/generic/CMakeLists.txt b/libc/src/math/generic/CMakeLists.txt
index 34b0f1424e8fd3..51396865b99d45 100644
--- a/libc/src/math/generic/CMakeLists.txt
+++ b/libc/src/math/generic/CMakeLists.txt
@@ -496,6 +496,25 @@ add_entrypoint_object(
-O3
)
+add_entrypoint_object(
+ sinf16
+ SRCS
+ sinf16.cpp
+ HDRS
+ ../sinf16.h
+ DEPENDS
+ .sincosf16_utils
+ libc.hdr.errno_macros
+ libc.hdr.fenv_macros
+ libc.src.__support.FPUtil.cast
+ libc.src.__support.FPUtil.fenv_impl
+ libc.src.__support.FPUtil.fp_bits
+ libc.src.__support.FPUtil.multiply_add
+ libc.src.__support.macros.optimization
+ COMPILE_OPTIONS
+ -O3
+)
+
add_entrypoint_object(
sincos
SRCS
diff --git a/libc/src/math/generic/sincosf16_utils.h b/libc/src/math/generic/sincosf16_utils.h
index 83511755a56c42..fe6e58c1f919c4 100644
--- a/libc/src/math/generic/sincosf16_utils.h
+++ b/libc/src/math/generic/sincosf16_utils.h
@@ -39,6 +39,11 @@ constexpr float SIN_K_PI_OVER_32[64] = {
-0x1.6a09e6p-1, -0x1.44cf32p-1, -0x1.1c73b4p-1, -0x1.e2b5d4p-2,
-0x1.87de2ap-2, -0x1.294062p-2, -0x1.8f8b84p-3, -0x1.917a6cp-4};
+static constexpr float THIRTYTWO_OVER_PI[4] = {
+//0x1.45fp3, 0x1.837p-12, -0x1.b1b8p-28, -0x1.f568p-43
+ 0x1.46p3, -0x1.9f4p-10, 0x1.b94p-22, -0x1.bcp-36
+};
+
LIBC_INLINE int32_t range_reduction_sincospif16(float x, float &y) {
float kf = fputil::nearest_integer(x * 32);
y = fputil::multiply_add<float>(x, 32.0, -kf);
@@ -46,11 +51,18 @@ LIBC_INLINE int32_t range_reduction_sincospif16(float x, float &y) {
return static_cast<int32_t>(kf);
}
-LIBC_INLINE void sincospif16_eval(float xf, float &sin_k, float &cos_k,
- float &sin_y, float &cosm1_y) {
- float y;
- int32_t k = range_reduction_sincospif16(xf, y);
+LIBC_INLINE int32_t range_reduction_sincosf16(float x, float &y) {
+ double xd = x;
+ double prod = xd * 0x1.45f306dc9c883p3;
+ double kf = fputil::nearest_integer(prod);
+ double yd = prod - kf;
+ y = static_cast<float>(yd);
+ return static_cast<int32_t>(kf);
+}
+
+static LIBC_INLINE void sincosf16_poly_eval(int32_t k, float y, float &sin_k, float &cos_k, float &sin_y, float &cosm1_y) {
+
sin_k = SIN_K_PI_OVER_32[k & 63];
cos_k = SIN_K_PI_OVER_32[(k + 16) & 63];
@@ -72,6 +84,21 @@ LIBC_INLINE void sincospif16_eval(float xf, float &sin_k, float &cos_k,
0x1.a6f7a2p-29f);
}
+LIBC_INLINE void sincosf16_eval(float xf, float &sin_k, float &cos_k, float &sin_y, float &cosm1_y) {
+ float y;
+ int32_t k = range_reduction_sincosf16(xf, y);
+
+ sincosf16_poly_eval(k, y, sin_k, cos_k, sin_y, cosm1_y);
+}
+
+LIBC_INLINE void sincospif16_eval(float xf, float &sin_k, float &cos_k,
+ float &sin_y, float &cosm1_y) {
+ float y;
+ int32_t k = range_reduction_sincospif16(xf, y);
+
+ sincosf16_poly_eval(k, y, sin_k, cos_k, sin_y, cosm1_y);
+}
+
} // namespace LIBC_NAMESPACE_DECL
#endif // LLVM_LIBC_SRC_MATH_GENERIC_SINCOSF16_UTILS_H
diff --git a/libc/src/math/generic/sinf16.cpp b/libc/src/math/generic/sinf16.cpp
new file mode 100644
index 00000000000000..7f1c8c7cf65217
--- /dev/null
+++ b/libc/src/math/generic/sinf16.cpp
@@ -0,0 +1,77 @@
+//===-- Half-precision sin 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/sinf16.h"
+#include "hdr/errno_macros.h"
+#include "hdr/fenv_macros.h"
+#include "sincosf16_utils.h"
+#include "src/__support/FPUtil/FEnvImpl.h"
+#include "src/__support/FPUtil/FPBits.h"
+#include "src/__support/FPUtil/cast.h"
+#include "src/__support/FPUtil/multiply_add.h"
+#include "src/__support/macros/optimization.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
+ using FPBits = typename fputil::FPBits<float16>;
+ FPBits xbits(x);
+
+ uint16_t x_u = xbits.uintval();
+ uint16_t x_abs = x_u & 0x7fff;
+ float xf = x;
+
+ // Range reduction:
+ // For !x| > pi/32, we perform range reduction as follows:
+ // Find k and y such that:
+ // x = (k + y) * pi/32
+ // k is an integer, |y| < 0.5
+ //
+ // This is done by performing:
+ // k = round(x * 32/pi)
+ // y = x * 32/pi - k
+ //
+ // Once k and y are computed, we then deduce the answer by the sine of sum
+ // formula:
+ // sin(x) = sin((k + y) * pi/32)
+ // = sin(k * pi/32) * cos(y * pi/32) +
+ // sin(y * pi/32) * cos(k * pi/32)
+
+ if (LIBC_UNLIKELY(x_abs <= 0x13d0)) {
+ int rounding = fputil::quick_get_round();
+
+ // For signed zeros
+ if ((LIBC_UNLIKELY(x_abs == 0U)) ||
+ (rounding == FE_UPWARD && xbits.is_pos()) ||
+ (rounding == FE_DOWNWARD && xbits.is_neg()))
+ return x;
+ }
+
+ if (LIBC_UNLIKELY(x_abs >= 0x7c00)) {
+ // If value is equal to infinity
+ if (x_abs == 0x7c00) {
+ fputil::set_errno_if_required(EDOM);
+ fputil::raise_except_if_required(FE_INVALID);
+ }
+
+ return x + FPBits::quiet_nan().get_val();
+ }
+
+
+ float sin_k, cos_k, sin_y, cosm1_y;
+ sincosf16_eval(xf, sin_k, cos_k, sin_y, cosm1_y);
+
+ if (LIBC_UNLIKELY(sin_y == 0 && sin_k == 0))
+ return FPBits::zero(xbits.sign()).get_val();
+
+ // Since, cosm1_y = cos_y - 1, therfore:
+ // sin(x) = cos_k * sin_y + sin_k + (cosm1_y * sin_k)
+ return fputil::cast<float16>(fputil::multiply_add(sin_y, cos_k, fputil::multiply_add(cosm1_y, sin_k, sin_k)));
+}
+
+} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/src/math/sinf16.h b/libc/src/math/sinf16.h
new file mode 100644
index 00000000000000..23f1aa99b62337
--- /dev/null
+++ b/libc/src/math/sinf16.h
@@ -0,0 +1,21 @@
+//===-- Implementation header for sinf16 ------------------------*- 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_SINF16_H
+#define LLVM_LIBC_SRC_MATH_SINF16_H
+
+#include "src/__support/macros/config.h"
+#include "src/__support/macros/properties/types.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+float16 sinf16(float16 x);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_MATH_SINF16_H
diff --git a/libc/test/src/math/CMakeLists.txt b/libc/test/src/math/CMakeLists.txt
index 610f4d9fc1a3b9..ea75720df4f430 100644
--- a/libc/test/src/math/CMakeLists.txt
+++ b/libc/test/src/math/CMakeLists.txt
@@ -85,6 +85,17 @@ add_fp_unittest(
libc.src.__support.FPUtil.fp_bits
)
+add_fp_unittest(
+ sinf16_test
+ NEED_MPFR
+ SUITE
+ libc-math-unittests
+ SRCS
+ sinf16_test.cpp
+ DEPENDS
+ libc.src.math.sinf16
+)
+
add_fp_unittest(
sinpif_test
NEED_MPFR
diff --git a/libc/test/src/math/sinf16_test.cpp b/libc/test/src/math/sinf16_test.cpp
new file mode 100644
index 00000000000000..cdb69d388f7f70
--- /dev/null
+++ b/libc/test/src/math/sinf16_test.cpp
@@ -0,0 +1,38 @@
+//===-- Exhaustive test for sinf16 ----------------------------------------===//
+//
+// 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/sinf16.h"
+#include "test/UnitTest/FPMatcher.h"
+#include "test/UnitTest/Test.h"
+#include "utils/MPFRWrapper/MPFRUtils.h"
+
+using LlvmLibcSinf16Test = 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(LlvmLibcSinf16Test, PositiveRange) {
+ for (uint16_t v = POS_START; v <= POS_STOP; ++v) {
+ float16 x = FPBits(v).get_val();
+ EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Sin, x, LIBC_NAMESPACE::sinf16(x), 0.5);
+ }
+}
+
+TEST_F(LlvmLibcSinf16Test, NegativeRange) {
+ for (uint16_t v = NEG_START; v <= NEG_STOP; ++v) {
+ float16 x = FPBits(v).get_val();
+ EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Sin, x, LIBC_NAMESPACE::sinf16(x), 0.5);
+ }
+}
diff --git a/libc/test/src/math/smoke/CMakeLists.txt b/libc/test/src/math/smoke/CMakeLists.txt
index e9c785f7d93300..2c1c4dba738465 100644
--- a/libc/test/src/math/smoke/CMakeLists.txt
+++ b/libc/test/src/math/smoke/CMakeLists.txt
@@ -49,6 +49,17 @@ add_fp_unittest(
libc.src.__support.FPUtil.fp_bits
)
+add_fp_unittest(
+ sinf16_test
+ SUITE
+ libc-math-smoke-tests
+ SRCS
+ sinf16_test.cpp
+ DEPENDS
+ libc.src.errno.errno
+ libc.src.math.sinf16
+)
+
add_fp_unittest(
sinpif_test
SUITE
diff --git a/libc/test/src/math/smoke/sinf16_test.cpp b/libc/test/src/math/smoke/sinf16_test.cpp
new file mode 100644
index 00000000000000..301e91a683a3ec
--- /dev/null
+++ b/libc/test/src/math/smoke/sinf16_test.cpp
@@ -0,0 +1,34 @@
+//===-- Unittests for sinf16 ----------------------------------------------===//
+//
+// 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/sinf16.h"
+#include "test/UnitTest/FPMatcher.h"
+#include "test/UnitTest/Test.h"
+
+using LlvmLibcSinf16Test = LIBC_NAMESPACE::testing::FPTest<float16>;
+
+TEST_F(LlvmLibcSinf16Test, SpecialNumbers) {
+ LIBC_NAMESPACE::libc_errno = 0;
+
+ EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::sinf16(aNaN));
+ EXPECT_MATH_ERRNO(0);
+
+ EXPECT_FP_EQ(zero, LIBC_NAMESPACE::sinf16(zero));
+ EXPECT_MATH_ERRNO(0);
+
+ EXPECT_FP_EQ(neg_zero, LIBC_NAMESPACE::sinf16(neg_zero));
+ EXPECT_MATH_ERRNO(0);
+
+ EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::sinf16(inf));
+ EXPECT_MATH_ERRNO(EDOM);
+
+ EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::sinf16(neg_inf));
+ EXPECT_MATH_ERRNO(EDOM);
+}
>From 1d5f809cf8f22a33d01038c407f9fa800bb9b05c Mon Sep 17 00:00:00 2001
From: wldfngrs <wldfngrs at gmail.com>
Date: Fri, 22 Nov 2024 20:19:13 +0100
Subject: [PATCH 2/8] Fixed rounding errors
---
libc/src/math/generic/CMakeLists.txt | 1 +
libc/src/math/generic/sincosf16_utils.h | 23 +++++------
libc/src/math/generic/sinf16.cpp | 49 +++++++++++++++++++-----
libc/src/math/generic/tanpif16.cpp | 4 +-
libc/test/src/math/sinf16_test.cpp | 6 ++-
libc/test/src/math/smoke/sinf16_test.cpp | 3 +-
6 files changed, 57 insertions(+), 29 deletions(-)
diff --git a/libc/src/math/generic/CMakeLists.txt b/libc/src/math/generic/CMakeLists.txt
index 51396865b99d45..97a6784ef3c5f1 100644
--- a/libc/src/math/generic/CMakeLists.txt
+++ b/libc/src/math/generic/CMakeLists.txt
@@ -509,6 +509,7 @@ add_entrypoint_object(
libc.src.__support.FPUtil.cast
libc.src.__support.FPUtil.fenv_impl
libc.src.__support.FPUtil.fp_bits
+ libc.src.__support.FPUtil.except_value_utils
libc.src.__support.FPUtil.multiply_add
libc.src.__support.macros.optimization
COMPILE_OPTIONS
diff --git a/libc/src/math/generic/sincosf16_utils.h b/libc/src/math/generic/sincosf16_utils.h
index fe6e58c1f919c4..13d27fb9828a3d 100644
--- a/libc/src/math/generic/sincosf16_utils.h
+++ b/libc/src/math/generic/sincosf16_utils.h
@@ -11,6 +11,7 @@
#include "src/__support/FPUtil/FPBits.h"
#include "src/__support/FPUtil/PolyEval.h"
+#include "src/__support/FPUtil/cast.h"
#include "src/__support/FPUtil/nearest_integer.h"
#include "src/__support/common.h"
#include "src/__support/macros/config.h"
@@ -39,11 +40,6 @@ constexpr float SIN_K_PI_OVER_32[64] = {
-0x1.6a09e6p-1, -0x1.44cf32p-1, -0x1.1c73b4p-1, -0x1.e2b5d4p-2,
-0x1.87de2ap-2, -0x1.294062p-2, -0x1.8f8b84p-3, -0x1.917a6cp-4};
-static constexpr float THIRTYTWO_OVER_PI[4] = {
-//0x1.45fp3, 0x1.837p-12, -0x1.b1b8p-28, -0x1.f568p-43
- 0x1.46p3, -0x1.9f4p-10, 0x1.b94p-22, -0x1.bcp-36
-};
-
LIBC_INLINE int32_t range_reduction_sincospif16(float x, float &y) {
float kf = fputil::nearest_integer(x * 32);
y = fputil::multiply_add<float>(x, 32.0, -kf);
@@ -52,17 +48,17 @@ LIBC_INLINE int32_t range_reduction_sincospif16(float x, float &y) {
}
LIBC_INLINE int32_t range_reduction_sincosf16(float x, float &y) {
- double xd = x;
- double prod = xd * 0x1.45f306dc9c883p3;
+ double prod = x * 0x1.45f306dc9c883p3;
double kf = fputil::nearest_integer(prod);
- double yd = prod - kf;
+ y = static_cast<float>(prod - kf);
- y = static_cast<float>(yd);
return static_cast<int32_t>(kf);
}
-static LIBC_INLINE void sincosf16_poly_eval(int32_t k, float y, float &sin_k, float &cos_k, float &sin_y, float &cosm1_y) {
-
+static LIBC_INLINE void sincosf16_poly_eval(int32_t k, float y, float &sin_k,
+ float &cos_k, float &sin_y,
+ float &cosm1_y) {
+
sin_k = SIN_K_PI_OVER_32[k & 63];
cos_k = SIN_K_PI_OVER_32[(k + 16) & 63];
@@ -84,7 +80,8 @@ static LIBC_INLINE void sincosf16_poly_eval(int32_t k, float y, float &sin_k, fl
0x1.a6f7a2p-29f);
}
-LIBC_INLINE void sincosf16_eval(float xf, float &sin_k, float &cos_k, float &sin_y, float &cosm1_y) {
+LIBC_INLINE void sincosf16_eval(float xf, float &sin_k, float &cos_k,
+ float &sin_y, float &cosm1_y) {
float y;
int32_t k = range_reduction_sincosf16(xf, y);
@@ -95,7 +92,7 @@ LIBC_INLINE void sincospif16_eval(float xf, float &sin_k, float &cos_k,
float &sin_y, float &cosm1_y) {
float y;
int32_t k = range_reduction_sincospif16(xf, y);
-
+
sincosf16_poly_eval(k, y, sin_k, cos_k, sin_y, cosm1_y);
}
diff --git a/libc/src/math/generic/sinf16.cpp b/libc/src/math/generic/sinf16.cpp
index 7f1c8c7cf65217..f307b3ea330c5e 100644
--- a/libc/src/math/generic/sinf16.cpp
+++ b/libc/src/math/generic/sinf16.cpp
@@ -13,11 +13,20 @@
#include "src/__support/FPUtil/FEnvImpl.h"
#include "src/__support/FPUtil/FPBits.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/macros/optimization.h"
namespace LIBC_NAMESPACE_DECL {
+constexpr size_t N_EXCEPTS = 2;
+
+constexpr fputil::ExceptValues<float16, N_EXCEPTS> SINF16_EXCEPTS{{
+ // (input, RZ output, RU offset, RD offset, RN offset)
+ {0x585c, 0x3ba3, 1, 0, 1},
+ {0x5cb0, 0xbbff, 0, 1, 0},
+}};
+
LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
using FPBits = typename fputil::FPBits<float16>;
FPBits xbits(x);
@@ -25,7 +34,7 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
uint16_t x_u = xbits.uintval();
uint16_t x_abs = x_u & 0x7fff;
float xf = x;
-
+
// Range reduction:
// For !x| > pi/32, we perform range reduction as follows:
// Find k and y such that:
@@ -41,15 +50,35 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
// sin(x) = sin((k + y) * pi/32)
// = sin(k * pi/32) * cos(y * pi/32) +
// sin(y * pi/32) * cos(k * pi/32)
-
+
+ // Handle exceptional values
+ if (LIBC_UNLIKELY(x_abs == 0x585C || x_abs == 0x5cb0)) {
+ bool x_sign = x_u >> 15;
+ if (auto r = SINF16_EXCEPTS.lookup_odd(x_abs, x_sign);
+ LIBC_UNLIKELY(r.has_value()))
+ return r.value();
+ }
+
+ int rounding = fputil::quick_get_round();
if (LIBC_UNLIKELY(x_abs <= 0x13d0)) {
- int rounding = fputil::quick_get_round();
-
- // For signed zeros
- if ((LIBC_UNLIKELY(x_abs == 0U)) ||
- (rounding == FE_UPWARD && xbits.is_pos()) ||
- (rounding == FE_DOWNWARD && xbits.is_neg()))
+ if (LIBC_UNLIKELY(x_abs == 0U))
return x;
+
+ if ((rounding == FE_UPWARD && xbits.is_pos()) ||
+ (rounding == FE_DOWNWARD && xbits.is_neg()))
+ return x;
+
+ if (rounding == FE_UPWARD && xbits.is_neg()) {
+ x_u--;
+ return FPBits(x_u).get_val();
+ }
+ }
+
+ if (LIBC_UNLIKELY(x_abs == 0x2b45)) {
+ if (rounding == FE_DOWNWARD && xbits.is_neg()) {
+ x_u--;
+ return FPBits(x_u).get_val();
+ }
}
if (LIBC_UNLIKELY(x_abs >= 0x7c00)) {
@@ -61,7 +90,6 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
return x + FPBits::quiet_nan().get_val();
}
-
float sin_k, cos_k, sin_y, cosm1_y;
sincosf16_eval(xf, sin_k, cos_k, sin_y, cosm1_y);
@@ -71,7 +99,8 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
// Since, cosm1_y = cos_y - 1, therfore:
// sin(x) = cos_k * sin_y + sin_k + (cosm1_y * sin_k)
- return fputil::cast<float16>(fputil::multiply_add(sin_y, cos_k, fputil::multiply_add(cosm1_y, sin_k, sin_k)));
+ return fputil::cast<float16>(fputil::multiply_add(
+ sin_y, cos_k, fputil::multiply_add(cosm1_y, sin_k, sin_k)));
}
} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/src/math/generic/tanpif16.cpp b/libc/src/math/generic/tanpif16.cpp
index ab3c9cb2122ba9..67635536ee3193 100644
--- a/libc/src/math/generic/tanpif16.cpp
+++ b/libc/src/math/generic/tanpif16.cpp
@@ -21,7 +21,7 @@ namespace LIBC_NAMESPACE_DECL {
constexpr size_t N_EXCEPTS = 21;
-constexpr fputil::ExceptValues<float16, N_EXCEPTS> TANF16_EXCEPTS{{
+constexpr fputil::ExceptValues<float16, N_EXCEPTS> TANPIF16_EXCEPTS{{
// (input, RZ output, RU offset, RD offset, RN offset)
{0x07f2, 0x0e3d, 1, 0, 0}, {0x086a, 0x0eee, 1, 0, 1},
{0x08db, 0x0fa0, 1, 0, 0}, {0x094c, 0x1029, 1, 0, 0},
@@ -49,7 +49,7 @@ LLVM_LIBC_FUNCTION(float16, tanpif16, (float16 x)) {
return x;
bool x_sign = x_u >> 15;
- if (auto r = TANF16_EXCEPTS.lookup_odd(x_abs, x_sign);
+ if (auto r = TANPIF16_EXCEPTS.lookup_odd(x_abs, x_sign);
LIBC_UNLIKELY(r.has_value()))
return r.value();
}
diff --git a/libc/test/src/math/sinf16_test.cpp b/libc/test/src/math/sinf16_test.cpp
index cdb69d388f7f70..b469e1dcb8d33f 100644
--- a/libc/test/src/math/sinf16_test.cpp
+++ b/libc/test/src/math/sinf16_test.cpp
@@ -26,13 +26,15 @@ static constexpr uint16_t NEG_STOP = 0xfc00U;
TEST_F(LlvmLibcSinf16Test, PositiveRange) {
for (uint16_t v = POS_START; v <= POS_STOP; ++v) {
float16 x = FPBits(v).get_val();
- EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Sin, x, LIBC_NAMESPACE::sinf16(x), 0.5);
+ EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Sin, x,
+ LIBC_NAMESPACE::sinf16(x), 0.5);
}
}
TEST_F(LlvmLibcSinf16Test, NegativeRange) {
for (uint16_t v = NEG_START; v <= NEG_STOP; ++v) {
float16 x = FPBits(v).get_val();
- EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Sin, x, LIBC_NAMESPACE::sinf16(x), 0.5);
+ EXPECT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Sin, x,
+ LIBC_NAMESPACE::sinf16(x), 0.5);
}
}
diff --git a/libc/test/src/math/smoke/sinf16_test.cpp b/libc/test/src/math/smoke/sinf16_test.cpp
index 301e91a683a3ec..2966c3c952fd21 100644
--- a/libc/test/src/math/smoke/sinf16_test.cpp
+++ b/libc/test/src/math/smoke/sinf16_test.cpp
@@ -6,7 +6,6 @@
//
//===----------------------------------------------------------------------===//
-
#include "src/errno/libc_errno.h"
#include "src/math/sinf16.h"
#include "test/UnitTest/FPMatcher.h"
@@ -16,7 +15,7 @@ using LlvmLibcSinf16Test = LIBC_NAMESPACE::testing::FPTest<float16>;
TEST_F(LlvmLibcSinf16Test, SpecialNumbers) {
LIBC_NAMESPACE::libc_errno = 0;
-
+
EXPECT_FP_EQ(aNaN, LIBC_NAMESPACE::sinf16(aNaN));
EXPECT_MATH_ERRNO(0);
>From 09b79304b02c29a1cf5756a3765c7e97836ffdea Mon Sep 17 00:00:00 2001
From: wldfngrs <wldfngrs at gmail.com>
Date: Fri, 22 Nov 2024 21:29:20 +0100
Subject: [PATCH 3/8] small fix
---
libc/src/math/generic/sinf16.cpp | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/libc/src/math/generic/sinf16.cpp b/libc/src/math/generic/sinf16.cpp
index f307b3ea330c5e..75d96659da0542 100644
--- a/libc/src/math/generic/sinf16.cpp
+++ b/libc/src/math/generic/sinf16.cpp
@@ -19,12 +19,13 @@
namespace LIBC_NAMESPACE_DECL {
-constexpr size_t N_EXCEPTS = 2;
+constexpr size_t N_EXCEPTS = 3;
constexpr fputil::ExceptValues<float16, N_EXCEPTS> SINF16_EXCEPTS{{
// (input, RZ output, RU offset, RD offset, RN offset)
{0x585c, 0x3ba3, 1, 0, 1},
{0x5cb0, 0xbbff, 0, 1, 0},
+ {0x51f5, 0xb80f, 0, 1, 0},
}};
LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
@@ -52,7 +53,7 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
// sin(y * pi/32) * cos(k * pi/32)
// Handle exceptional values
- if (LIBC_UNLIKELY(x_abs == 0x585C || x_abs == 0x5cb0)) {
+ if (LIBC_UNLIKELY(x_abs == 0x585C || x_abs == 0x5cb0 || x_abs == 0x51f5)) {
bool x_sign = x_u >> 15;
if (auto r = SINF16_EXCEPTS.lookup_odd(x_abs, x_sign);
LIBC_UNLIKELY(r.has_value()))
>From 0eefcc29e067087ddce6d24323d559346e67e923 Mon Sep 17 00:00:00 2001
From: wldfngrs <wldfngrs at gmail.com>
Date: Fri, 22 Nov 2024 21:46:04 +0100
Subject: [PATCH 4/8] Update Docs
---
libc/docs/math/index.rst | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libc/docs/math/index.rst b/libc/docs/math/index.rst
index 2b86f49a3619e6..4934e93ccb1645 100644
--- a/libc/docs/math/index.rst
+++ b/libc/docs/math/index.rst
@@ -336,7 +336,7 @@ Higher Math Functions
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
| rsqrt | | | | | | 7.12.7.9 | F.10.4.9 |
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
-| sin | |check| | |check| | | | | 7.12.4.6 | F.10.1.6 |
+| sin | |check| | |check| | | |check| | | 7.12.4.6 | F.10.1.6 |
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
| sincos | |check| | |check| | | | | | |
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
>From fd28fcd44ffbf3bcb0ab1247860816ea7fbeb775 Mon Sep 17 00:00:00 2001
From: wldfngrs <wldfngrs at gmail.com>
Date: Sun, 24 Nov 2024 00:55:08 +0100
Subject: [PATCH 5/8] nit formatting
---
libc/src/math/generic/sinf16.cpp | 2 +-
libc/test/src/math/sinf16_test.cpp | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/libc/src/math/generic/sinf16.cpp b/libc/src/math/generic/sinf16.cpp
index 75d96659da0542..dcee9be7487703 100644
--- a/libc/src/math/generic/sinf16.cpp
+++ b/libc/src/math/generic/sinf16.cpp
@@ -1,4 +1,4 @@
-//===-- Half-precision sin function ------------------------------------===//
+//===-- Half-precision sin(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.
diff --git a/libc/test/src/math/sinf16_test.cpp b/libc/test/src/math/sinf16_test.cpp
index b469e1dcb8d33f..b05501cb0f1454 100644
--- a/libc/test/src/math/sinf16_test.cpp
+++ b/libc/test/src/math/sinf16_test.cpp
@@ -4,7 +4,7 @@
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
-//===---------------------------------------------------------------------===//
+//===----------------------------------------------------------------------===//
#include "src/math/sinf16.h"
#include "test/UnitTest/FPMatcher.h"
>From e234abd17bf426f228caf46bfafc5369bc821cf7 Mon Sep 17 00:00:00 2001
From: wldfngrs <wldfngrs at gmail.com>
Date: Sun, 24 Nov 2024 00:57:02 +0100
Subject: [PATCH 6/8] formatting
---
libc/src/math/generic/sinf16.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libc/src/math/generic/sinf16.cpp b/libc/src/math/generic/sinf16.cpp
index dcee9be7487703..bb36b20c696586 100644
--- a/libc/src/math/generic/sinf16.cpp
+++ b/libc/src/math/generic/sinf16.cpp
@@ -1,4 +1,4 @@
-//===-- Half-precision sin(x) function -------------------------------------===//
+//===-- Half-precision sin(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.
>From c6bc8571fffb30a5f5f3d31d09f1ffe0f612beb4 Mon Sep 17 00:00:00 2001
From: wldfngrs <wldfngrs at gmail.com>
Date: Wed, 27 Nov 2024 23:45:02 +0100
Subject: [PATCH 7/8] Added comments, improved code readability
---
libc/src/math/generic/sincosf16_utils.h | 14 +++++++++++++
libc/src/math/generic/sinf16.cpp | 27 ++++++++++++-------------
2 files changed, 27 insertions(+), 14 deletions(-)
diff --git a/libc/src/math/generic/sincosf16_utils.h b/libc/src/math/generic/sincosf16_utils.h
index 13d27fb9828a3d..b3d262519e1ad0 100644
--- a/libc/src/math/generic/sincosf16_utils.h
+++ b/libc/src/math/generic/sincosf16_utils.h
@@ -47,6 +47,20 @@ LIBC_INLINE int32_t range_reduction_sincospif16(float x, float &y) {
return static_cast<int32_t>(kf);
}
+// Recall, range reduction:
+// k = round(x * 32/pi)
+// y = x * 32/pi - k
+//
+// The constant 0x1.45f306dc9c883p3 is 32/pi rounded to double-precision.
+// 32/pi is generated by Sollya with the following commands:
+// > display = hexadecimal;
+// > round(32/pi, D, RN);
+//
+// The precision choice of 'double' is to minimize rounding errors
+// in this initial scaling step, preserving enough bits so errors accumulated
+// while computing the subtraction: y = x * 32/pi - round(x * 32/pi)
+// are beyond the least-significant bit of single-precision used during
+// intermediate computation.
LIBC_INLINE int32_t range_reduction_sincosf16(float x, float &y) {
double prod = x * 0x1.45f306dc9c883p3;
double kf = fputil::nearest_integer(prod);
diff --git a/libc/src/math/generic/sinf16.cpp b/libc/src/math/generic/sinf16.cpp
index bb36b20c696586..7bc7daed2eb731 100644
--- a/libc/src/math/generic/sinf16.cpp
+++ b/libc/src/math/generic/sinf16.cpp
@@ -19,17 +19,18 @@
namespace LIBC_NAMESPACE_DECL {
-constexpr size_t N_EXCEPTS = 3;
+constexpr size_t N_EXCEPTS = 4;
constexpr fputil::ExceptValues<float16, N_EXCEPTS> SINF16_EXCEPTS{{
// (input, RZ output, RU offset, RD offset, RN offset)
+ {0x2b45, 0x2b43, 1, 0, 1},
{0x585c, 0x3ba3, 1, 0, 1},
{0x5cb0, 0xbbff, 0, 1, 0},
{0x51f5, 0xb80f, 0, 1, 0},
}};
LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
- using FPBits = typename fputil::FPBits<float16>;
+ using FPBits = fputil::FPBits<float16>;
FPBits xbits(x);
uint16_t x_u = xbits.uintval();
@@ -37,7 +38,7 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
float xf = x;
// Range reduction:
- // For !x| > pi/32, we perform range reduction as follows:
+ // For |x| > pi/32, we perform range reduction as follows:
// Find k and y such that:
// x = (k + y) * pi/32
// k is an integer, |y| < 0.5
@@ -53,13 +54,19 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
// sin(y * pi/32) * cos(k * pi/32)
// Handle exceptional values
- if (LIBC_UNLIKELY(x_abs == 0x585C || x_abs == 0x5cb0 || x_abs == 0x51f5)) {
+ if (LIBC_UNLIKELY(x_abs == 0x585c || x_abs == 0x5cb0 || x_abs == 0x51f5 ||
+ x_abs == 0x2b45)) {
bool x_sign = x_u >> 15;
if (auto r = SINF16_EXCEPTS.lookup_odd(x_abs, x_sign);
LIBC_UNLIKELY(r.has_value()))
return r.value();
}
+ // Exhaustive tests show that for |x| <= 0x13d0, 1ULP rounding errors occur.
+ // To fix this, the following apply:
+ // - When x >= 0, and rounding upward, sin(x) == x.
+ // - When x < 0, and rounding downward, sin(x) == x.
+ // - When x < 0, and rounding upward, sin(x) == (x - 1ULP)
int rounding = fputil::quick_get_round();
if (LIBC_UNLIKELY(x_abs <= 0x13d0)) {
if (LIBC_UNLIKELY(x_abs == 0U))
@@ -75,16 +82,8 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
}
}
- if (LIBC_UNLIKELY(x_abs == 0x2b45)) {
- if (rounding == FE_DOWNWARD && xbits.is_neg()) {
- x_u--;
- return FPBits(x_u).get_val();
- }
- }
-
- if (LIBC_UNLIKELY(x_abs >= 0x7c00)) {
- // If value is equal to infinity
- if (x_abs == 0x7c00) {
+ if (xbits.is_inf_or_nan()) {
+ if (xbits.is_inf()){
fputil::set_errno_if_required(EDOM);
fputil::raise_except_if_required(FE_INVALID);
}
>From 92bdef59839511be287ffe1e3a236eac25fa0cf6 Mon Sep 17 00:00:00 2001
From: wldfngrs <wldfngrs at gmail.com>
Date: Wed, 27 Nov 2024 23:47:43 +0100
Subject: [PATCH 8/8] clang-format
---
libc/src/math/generic/sincosf16_utils.h | 8 ++++----
libc/src/math/generic/sinf16.cpp | 6 +++---
2 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/libc/src/math/generic/sincosf16_utils.h b/libc/src/math/generic/sincosf16_utils.h
index b3d262519e1ad0..5e5edd4a8c85bd 100644
--- a/libc/src/math/generic/sincosf16_utils.h
+++ b/libc/src/math/generic/sincosf16_utils.h
@@ -50,17 +50,17 @@ LIBC_INLINE int32_t range_reduction_sincospif16(float x, float &y) {
// Recall, range reduction:
// k = round(x * 32/pi)
// y = x * 32/pi - k
-//
+//
// The constant 0x1.45f306dc9c883p3 is 32/pi rounded to double-precision.
// 32/pi is generated by Sollya with the following commands:
// > display = hexadecimal;
// > round(32/pi, D, RN);
-//
+//
// The precision choice of 'double' is to minimize rounding errors
// in this initial scaling step, preserving enough bits so errors accumulated
// while computing the subtraction: y = x * 32/pi - round(x * 32/pi)
-// are beyond the least-significant bit of single-precision used during
-// intermediate computation.
+// are beyond the least-significant bit of single-precision used during
+// further intermediate computation.
LIBC_INLINE int32_t range_reduction_sincosf16(float x, float &y) {
double prod = x * 0x1.45f306dc9c883p3;
double kf = fputil::nearest_integer(prod);
diff --git a/libc/src/math/generic/sinf16.cpp b/libc/src/math/generic/sinf16.cpp
index 7bc7daed2eb731..f8c90264f17b69 100644
--- a/libc/src/math/generic/sinf16.cpp
+++ b/libc/src/math/generic/sinf16.cpp
@@ -55,7 +55,7 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
// Handle exceptional values
if (LIBC_UNLIKELY(x_abs == 0x585c || x_abs == 0x5cb0 || x_abs == 0x51f5 ||
- x_abs == 0x2b45)) {
+ x_abs == 0x2b45)) {
bool x_sign = x_u >> 15;
if (auto r = SINF16_EXCEPTS.lookup_odd(x_abs, x_sign);
LIBC_UNLIKELY(r.has_value()))
@@ -63,7 +63,7 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
}
// Exhaustive tests show that for |x| <= 0x13d0, 1ULP rounding errors occur.
- // To fix this, the following apply:
+ // To fix this, the following apply:
// - When x >= 0, and rounding upward, sin(x) == x.
// - When x < 0, and rounding downward, sin(x) == x.
// - When x < 0, and rounding upward, sin(x) == (x - 1ULP)
@@ -83,7 +83,7 @@ LLVM_LIBC_FUNCTION(float16, sinf16, (float16 x)) {
}
if (xbits.is_inf_or_nan()) {
- if (xbits.is_inf()){
+ if (xbits.is_inf()) {
fputil::set_errno_if_required(EDOM);
fputil::raise_except_if_required(FE_INVALID);
}
More information about the libc-commits
mailing list