[PATCH 09/15] math: Use asinpif from CORE-MATH
Adhemerval Zanella Netto
adhemerval.zanella@linaro.org
Mon Feb 3 12:27:31 GMT 2025
On 03/02/25 03:40, Paul Zimmermann wrote:
> I suggest the following change which should improve performance on x86_64/x86_64-v2:
>
> --- a/sysdeps/ieee754/flt-32/s_asinpif.c
> +++ b/sysdeps/ieee754/flt-32/s_asinpif.c
> @@ -122,7 +122,7 @@ __asinpif (float x)
> else
> {
> double f = sqrt (1 - az);
> - double c0 = fma (az, c[1], c[0]);
> + double c0 = c[0] + az * c[1];
> double c2 = c[2] + az * c[3];
> double c4 = c[4] + az * c[5];
> double c6 = c[6] + az * c[7];
>
> Moreover "fast patch" should be fast path.
Thanks, this allows to drop the last patch that adds a asinpif ifunc variant
for x86_64-v3, since performance is now an improvement on all x86_64 ABIs.
>
> Paul
>
>> From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
>> Cc: DJ Delorie <dj@redhat.com>,
>> Joseph Myers <josmyers@redhat.com>,
>> Paul Zimmermann <Paul.Zimmermann@inria.fr>,
>> Alexei Sibidanov <sibid@uvic.ca>
>> Date: Fri, 31 Jan 2025 16:17:13 -0300
>>
>> The CORE-MATH implementation is correctly rounded (for any rounding mode)
>> and shows better performance to the generic asinpif.
>>
>> The code was adapted to glibc style and to use the definition of
>> math_config.h (to handle errno, overflow, and underflow).
>>
>> Benchtest on x64_64 (Ryzen 9 5900X, gcc 14.2.1), aarch64 (Neoverse-N1,
>> gcc 13.3.1), and powerpc (POWER10, gcc 13.2.1):
>>
>> latency master patched improvement
>> x86_64 46.4996 51.0456 -9.78%
>> x86_64v2 46.7551 52.1317 -11.50%
>> x86_64v3 42.6235 34.8162 18.32%
>> aarch64 (Neoverse) 17.4161 14.3604 17.55%
>> power8 10.7347 9.0193 15.98%
>> power10 10.6420 9.0362 15.09%
>>
>> reciprocal-throughput master patched improvement
>> x86_64 24.7208 29.0812 -17.64%
>> x86_64v2 24.2177 29.7166 -22.71%
>> x86_64v3 20.5617 12.3679 39.85%
>> aarch64 (Neoverse) 13.4827 7.17613 46.78%
>> power8 6.46134 3.56089 44.89%
>> power10 5.79007 3.49544 39.63%
>>
>> x86_64/x86_64-v2 shows slower performance due the use of a fma
>> operation in the fast patch, only x86_64-v3 provides it without a
>> function call.
>> ---
>> SHARED-FILES | 4 +
>> sysdeps/aarch64/libm-test-ulps | 4 -
>> sysdeps/arc/fpu/libm-test-ulps | 4 -
>> sysdeps/arc/nofpu/libm-test-ulps | 1 -
>> sysdeps/arm/libm-test-ulps | 4 -
>> sysdeps/hppa/fpu/libm-test-ulps | 4 -
>> sysdeps/i386/fpu/libm-test-ulps | 4 -
>> .../i386/i686/fpu/multiarch/libm-test-ulps | 4 -
>> sysdeps/ieee754/flt-32/s_asinpif.c | 136 ++++++++++++++++++
>> sysdeps/loongarch/lp64/libm-test-ulps | 4 -
>> sysdeps/mips/mips64/libm-test-ulps | 4 -
>> sysdeps/or1k/fpu/libm-test-ulps | 4 -
>> sysdeps/or1k/nofpu/libm-test-ulps | 1 -
>> sysdeps/powerpc/fpu/libm-test-ulps | 4 -
>> sysdeps/riscv/nofpu/libm-test-ulps | 1 -
>> sysdeps/riscv/rvd/libm-test-ulps | 4 -
>> sysdeps/s390/fpu/libm-test-ulps | 4 -
>> sysdeps/sparc/fpu/libm-test-ulps | 4 -
>> sysdeps/x86_64/fpu/libm-test-ulps | 4 -
>> 19 files changed, 140 insertions(+), 59 deletions(-)
>> create mode 100644 sysdeps/ieee754/flt-32/s_asinpif.c
>>
>> diff --git a/SHARED-FILES b/SHARED-FILES
>> index 3fde72644a..e700f4b155 100644
>> --- a/SHARED-FILES
>> +++ b/SHARED-FILES
>> @@ -338,3 +338,7 @@ sysdeps/ieee754/flt-32/s_acospif.c:
>> (src/binary32/acospi/acospif.c in CORE-MATH)
>> - the code was adapted to use glibc code style and internal
>> functions to handle errno, overflow, and underflow.
>> +sysdeps/ieee754/flt-32/s_asinpif.c:
>> + (src/binary32/asinpi/asinpif.c in CORE-MATH)
>> + - the code was adapted to use glibc code style and internal
>> + functions to handle errno, overflow, and underflow.
>> diff --git a/sysdeps/aarch64/libm-test-ulps b/sysdeps/aarch64/libm-test-ulps
>> index 1a403d95b6..abb0611ee5 100644
>> --- a/sysdeps/aarch64/libm-test-ulps
>> +++ b/sysdeps/aarch64/libm-test-ulps
>> @@ -115,22 +115,18 @@ ldouble: 4
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> ldouble: 1
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> ldouble: 2
>>
>> Function: "atan":
>> diff --git a/sysdeps/arc/fpu/libm-test-ulps b/sysdeps/arc/fpu/libm-test-ulps
>> index c0c5daa589..35aebba38a 100644
>> --- a/sysdeps/arc/fpu/libm-test-ulps
>> +++ b/sysdeps/arc/fpu/libm-test-ulps
>> @@ -63,19 +63,15 @@ double: 3
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>>
>> Function: "atan":
>> double: 1
>> diff --git a/sysdeps/arc/nofpu/libm-test-ulps b/sysdeps/arc/nofpu/libm-test-ulps
>> index 2b34f5a0ab..325546e582 100644
>> --- a/sysdeps/arc/nofpu/libm-test-ulps
>> +++ b/sysdeps/arc/nofpu/libm-test-ulps
>> @@ -18,7 +18,6 @@ double: 2
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>>
>> Function: "atan":
>> double: 1
>> diff --git a/sysdeps/arm/libm-test-ulps b/sysdeps/arm/libm-test-ulps
>> index afb0532a66..0927fdb980 100644
>> --- a/sysdeps/arm/libm-test-ulps
>> +++ b/sysdeps/arm/libm-test-ulps
>> @@ -63,19 +63,15 @@ double: 3
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>>
>> Function: "atan":
>> double: 1
>> diff --git a/sysdeps/hppa/fpu/libm-test-ulps b/sysdeps/hppa/fpu/libm-test-ulps
>> index b9959c8a12..02cc3b5ddc 100644
>> --- a/sysdeps/hppa/fpu/libm-test-ulps
>> +++ b/sysdeps/hppa/fpu/libm-test-ulps
>> @@ -63,19 +63,15 @@ double: 3
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>>
>> Function: "atan":
>> double: 1
>> diff --git a/sysdeps/i386/fpu/libm-test-ulps b/sysdeps/i386/fpu/libm-test-ulps
>> index 85c58f34e9..69d0eb1eec 100644
>> --- a/sysdeps/i386/fpu/libm-test-ulps
>> +++ b/sysdeps/i386/fpu/libm-test-ulps
>> @@ -101,25 +101,21 @@ ldouble: 5
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> float128: 2
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 2
>> -float: 1
>> float128: 2
>> ldouble: 2
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> float128: 1
>> ldouble: 2
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> float128: 2
>> ldouble: 2
>>
>> diff --git a/sysdeps/i386/i686/fpu/multiarch/libm-test-ulps b/sysdeps/i386/i686/fpu/multiarch/libm-test-ulps
>> index bc14e7e115..392d7d252c 100644
>> --- a/sysdeps/i386/i686/fpu/multiarch/libm-test-ulps
>> +++ b/sysdeps/i386/i686/fpu/multiarch/libm-test-ulps
>> @@ -101,25 +101,21 @@ ldouble: 5
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> float128: 2
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 2
>> -float: 1
>> float128: 2
>> ldouble: 2
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> float128: 1
>> ldouble: 2
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> float128: 2
>> ldouble: 2
>>
>> diff --git a/sysdeps/ieee754/flt-32/s_asinpif.c b/sysdeps/ieee754/flt-32/s_asinpif.c
>> new file mode 100644
>> index 0000000000..585dc3f06e
>> --- /dev/null
>> +++ b/sysdeps/ieee754/flt-32/s_asinpif.c
>> @@ -0,0 +1,136 @@
>> +/* Correctly-rounded half-revolution arc-sine function for binary32 value.
>> +
>> +Copyright (c) 2022-2025 Alexei Sibidanov.
>> +
>> +The original version of this file was copied from the CORE-MATH
>> +project (file src/binary32/asinpi/asinpif.c, revision 49e223e).
>> +
>> +Permission is hereby granted, free of charge, to any person obtaining a copy
>> +of this software and associated documentation files (the "Software"), to deal
>> +in the Software without restriction, including without limitation the rights
>> +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
>> +copies of the Software, and to permit persons to whom the Software is
>> +furnished to do so, subject to the following conditions:
>> +
>> +The above copyright notice and this permission notice shall be included in all
>> +copies or substantial portions of the Software.
>> +
>> +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
>> +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
>> +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
>> +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
>> +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
>> +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
>> +SOFTWARE.
>> +
>> +*/
>> +
>> +#include <errno.h>
>> +#include <math.h>
>> +#include <stdint.h>
>> +#include <libm-alias-float.h>
>> +#include "math_config.h"
>> +
>> +float
>> +__asinpif (float x)
>> +{
>> + float ax = fabsf (x);
>> + double az = ax;
>> + double z = x;
>> + uint32_t t = asuint (x);
>> + int32_t e = (t >> 23) & 0xff;
>> + if (__glibc_unlikely (e >= 127))
>> + {
>> + if (ax == 1.0f)
>> + return copysignf (0.5f, x);
>> + if (e == 0xff && (t << 9))
>> + return x + x; /* nan */
>> + return __math_edomf ((x - x) / (x - x)); /* nan */
>> + }
>> + int32_t s = 146 - e;
>> + int32_t i = 0;
>> + if (__glibc_likely (s < 32))
>> + i = ((t & (~0u >> 9)) | 1 << 23) >> s;
>> + static const double ch[][8] =
>> + {
>> + { 0x1.45f306dc9c882p-2, 0x1.b2995e7b7dc2fp-5, 0x1.8723a1cf50c7ep-6,
>> + 0x1.d1a4591d16a29p-7, 0x1.3ce3aa68ddaeep-7, 0x1.d3182ab0cc1bfp-8,
>> + 0x1.62b379a8b88e3p-8, 0x1.6811411fcfec2p-8 },
>> + { 0x1.ffffffffd3cdap-2, -0x1.17cc1b3355fdcp-4, 0x1.d067a1e8d5a99p-6,
>> + -0x1.08e16fb09314ap-6, 0x1.5eed43d42dcb2p-7, -0x1.f58baca7acc71p-8,
>> + 0x1.5dab64e2dcf15p-8, -0x1.59270e30797acp-9 },
>> + { 0x1.fffffff7c4617p-2, -0x1.17cc149ded3a2p-4, 0x1.d0654d4cb2c1ap-6,
>> + -0x1.08c3ba713d33ap-6, 0x1.5d2053481079cp-7, -0x1.e485ebc545e7ep-8,
>> + 0x1.303baca167dddp-8, -0x1.dee8d16d06b38p-10 },
>> + { 0x1.ffffffa749848p-2, -0x1.17cbe7155935p-4, 0x1.d05a312269adfp-6,
>> + -0x1.0862b3ee617d7p-6, 0x1.5920708db2a73p-7, -0x1.cb0463b3862c3p-8,
>> + 0x1.02b82478f95d7p-8, -0x1.52a7b8579e729p-10 },
>> + { 0x1.fffffe1f92bb5p-2, -0x1.17cb3e74c64e3p-4, 0x1.d03af67311cbfp-6,
>> + -0x1.079441cbfc7ap-6, 0x1.52b4287805a61p-7, -0x1.ac3286d604a98p-8,
>> + 0x1.b2f1210d9701bp-9, -0x1.e740ddc25afd6p-11 },
>> + { 0x1.fffff92beb6e2p-2, -0x1.17c986fe9518bp-4, 0x1.cff98167c9a5ep-6,
>> + -0x1.0638b591eae52p-6, 0x1.4a0803828959ep-7, -0x1.8adeca229f11dp-8,
>> + 0x1.6b9a7ba05dfcep-9, -0x1.640521a43b2dp-11 },
>> + { 0x1.ffffeccee5bfcp-2, -0x1.17c5f1753f5eap-4, 0x1.cf874e4fe258fp-6,
>> + -0x1.043e6cf77b256p-6, 0x1.3f7db42227d92p-7, -0x1.691a6fa2a2882p-8,
>> + 0x1.2f6543162bc61p-9, -0x1.07d5da05822b6p-11 },
>> + { 0x1.ffffd2f64431dp-2, -0x1.17bf8208c10c1p-4, 0x1.ced7487cdb124p-6,
>> + -0x1.01a0d30932905p-6, 0x1.3388f99b254dap-7, -0x1.4844e245c65bdp-8,
>> + 0x1.fa777150197c6p-10, -0x1.8c1ecf16a05c8p-12 },
>> + { 0x1.ffffa36d1712ep-2, -0x1.17b523971bd4ep-4, 0x1.cddee26de2deep-6,
>> + -0x1.fccb00abaaabcp-7, 0x1.269afc3622342p-7, -0x1.2933152686752p-8,
>> + 0x1.a76d4956cc9a3p-10, -0x1.2ce7d6dc651cep-12 },
>> + { 0x1.ffff5402ab3a1p-2, -0x1.17a5ba85da77ap-4, 0x1.cc96894e05c02p-6,
>> + -0x1.f532143cb832ep-7, 0x1.19180b660ff09p-7, -0x1.0c57417a78b3cp-8,
>> + 0x1.62e26cbd7bb1ep-10, -0x1.ce28d33fe1df3p-13 },
>> + { 0x1.fffed8d639751p-2, -0x1.1790349f3ae76p-4, 0x1.caf9a4fd1b398p-6,
>> + -0x1.ec986b111342ep-7, 0x1.0b53c3ad4baa4p-7, -0x1.e3c2282eeace4p-9,
>> + 0x1.2a55369f55bbep-10, -0x1.667fe48c396e8p-13 },
>> + { 0x1.fffe24b714161p-2, -0x1.177394fbcb719p-4, 0x1.c90652d920ebdp-6,
>> + -0x1.e3239197bddf1p-7, 0x1.fb2188525b025p-8, -0x1.b3aadd451afc7p-9,
>> + 0x1.f74020f31fdabp-11, -0x1.18b0cb246768dp-13 },
>> + { 0x1.fffd298bec9e2p-2, -0x1.174efbfd34648p-4, 0x1.c6bcfe48ea92bp-6,
>> + -0x1.d8f9f2a16157cp-7, 0x1.e0044f56c8864p-8, -0x1.883e2347fe76cp-9,
>> + 0x1.a9f0e3c1b7af5p-11, -0x1.bb5acc0e60825p-14 },
>> + { 0x1.fffbd8b784c4dp-2, -0x1.1721abdd3722ep-4, 0x1.c41fee756d4bp-6,
>> + -0x1.ce40bccf8065fp-7, 0x1.c59b684b70ef9p-8, -0x1.6133d027996b3p-9,
>> + 0x1.69cad01106397p-11, -0x1.60f8e45494156p-14 },
>> + { 0x1.fffa23749cf88p-2, -0x1.16eb0a8285c06p-4, 0x1.c132d762e1b0dp-6,
>> + -0x1.c31a959398f4ep-7, 0x1.ac1c5b46bc8ap-8, -0x1.3e34f1abe51dcp-9,
>> + 0x1.346738737c0b9p-11, -0x1.1b227a3f5c75p-14 },
>> + { 0x1.fff7fb25bb407p-2, -0x1.16aaa14d7564p-4, 0x1.bdfa75fca5ff2p-6,
>> + -0x1.b7a6e260d079cp-7, 0x1.93ab06911033cp-8, -0x1.1ee5560967fd5p-9,
>> + 0x1.07d31060838bfp-11, -0x1.c96f33a283115p-15
>> + }
>> + };
>> + const double *c = ch[i];
>> + double z2 = z * z;
>> + double z4 = z2 * z2;
>> + if (__glibc_unlikely (i == 0))
>> + {
>> + double c0 = c[0] + z2 * c[1];
>> + double c2 = c[2] + z2 * c[3];
>> + double c4 = c[4] + z2 * c[5];
>> + double c6 = c[6] + z2 * c[7];
>> + c0 += c2 * z4;
>> + c4 += c6 * z4;
>> + c0 += c4 * (z4 * z4);
>> + if (__glibc_unlikely (ax != 0.0 && ax <= 0x1.921fb4p-126f))
>> + __set_errno (ERANGE);
>> + return z * c0;
>> + }
>> + else
>> + {
>> + double f = sqrt (1 - az);
>> + double c0 = fma (az, c[1], c[0]);
>> + double c2 = c[2] + az * c[3];
>> + double c4 = c[4] + az * c[5];
>> + double c6 = c[6] + az * c[7];
>> + c0 += c2 * z2;
>> + c4 += c6 * z2;
>> + c0 += c4 * z4;
>> + double r = fma (-c0, copysign (f, x), copysign (0.5, x));
>> + return r;
>> + }
>> +}
>> +libm_alias_float (__asinpi, asinpi)
>> diff --git a/sysdeps/loongarch/lp64/libm-test-ulps b/sysdeps/loongarch/lp64/libm-test-ulps
>> index ce84ddf1e6..33dd6718ba 100644
>> --- a/sysdeps/loongarch/lp64/libm-test-ulps
>> +++ b/sysdeps/loongarch/lp64/libm-test-ulps
>> @@ -83,22 +83,18 @@ ldouble: 4
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> ldouble: 1
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> ldouble: 2
>>
>> Function: "atan":
>> diff --git a/sysdeps/mips/mips64/libm-test-ulps b/sysdeps/mips/mips64/libm-test-ulps
>> index 67c37dfd5e..869ceff928 100644
>> --- a/sysdeps/mips/mips64/libm-test-ulps
>> +++ b/sysdeps/mips/mips64/libm-test-ulps
>> @@ -83,22 +83,18 @@ ldouble: 4
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> ldouble: 1
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> ldouble: 2
>>
>> Function: "atan":
>> diff --git a/sysdeps/or1k/fpu/libm-test-ulps b/sysdeps/or1k/fpu/libm-test-ulps
>> index d3b1036d29..75db236e09 100644
>> --- a/sysdeps/or1k/fpu/libm-test-ulps
>> +++ b/sysdeps/or1k/fpu/libm-test-ulps
>> @@ -63,19 +63,15 @@ double: 3
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>>
>> Function: "atan":
>> double: 1
>> diff --git a/sysdeps/or1k/nofpu/libm-test-ulps b/sysdeps/or1k/nofpu/libm-test-ulps
>> index 14b7e0f3f9..a1f7c80097 100644
>> --- a/sysdeps/or1k/nofpu/libm-test-ulps
>> +++ b/sysdeps/or1k/nofpu/libm-test-ulps
>> @@ -54,7 +54,6 @@ double: 3
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>>
>> Function: "atan":
>> double: 1
>> diff --git a/sysdeps/powerpc/fpu/libm-test-ulps b/sysdeps/powerpc/fpu/libm-test-ulps
>> index c9c86de147..fa3cf2e844 100644
>> --- a/sysdeps/powerpc/fpu/libm-test-ulps
>> +++ b/sysdeps/powerpc/fpu/libm-test-ulps
>> @@ -107,25 +107,21 @@ ldouble: 7
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> float128: 1
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>> float128: 2
>> ldouble: 4
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> float128: 1
>> ldouble: 4
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> float128: 2
>> ldouble: 4
>>
>> diff --git a/sysdeps/riscv/nofpu/libm-test-ulps b/sysdeps/riscv/nofpu/libm-test-ulps
>> index 6206a9531a..a5184ecad9 100644
>> --- a/sysdeps/riscv/nofpu/libm-test-ulps
>> +++ b/sysdeps/riscv/nofpu/libm-test-ulps
>> @@ -71,7 +71,6 @@ ldouble: 4
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "atan":
>> diff --git a/sysdeps/riscv/rvd/libm-test-ulps b/sysdeps/riscv/rvd/libm-test-ulps
>> index 124ca4b719..3bfc9668d5 100644
>> --- a/sysdeps/riscv/rvd/libm-test-ulps
>> +++ b/sysdeps/riscv/rvd/libm-test-ulps
>> @@ -83,22 +83,18 @@ ldouble: 4
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> ldouble: 1
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> ldouble: 2
>>
>> Function: "atan":
>> diff --git a/sysdeps/s390/fpu/libm-test-ulps b/sysdeps/s390/fpu/libm-test-ulps
>> index 364ccf3326..7d61bf1cef 100644
>> --- a/sysdeps/s390/fpu/libm-test-ulps
>> +++ b/sysdeps/s390/fpu/libm-test-ulps
>> @@ -83,22 +83,18 @@ ldouble: 4
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> ldouble: 1
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> ldouble: 2
>>
>> Function: "atan":
>> diff --git a/sysdeps/sparc/fpu/libm-test-ulps b/sysdeps/sparc/fpu/libm-test-ulps
>> index 1174972002..426f45893e 100644
>> --- a/sysdeps/sparc/fpu/libm-test-ulps
>> +++ b/sysdeps/sparc/fpu/libm-test-ulps
>> @@ -83,22 +83,18 @@ ldouble: 4
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>> ldouble: 2
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> ldouble: 1
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> ldouble: 2
>>
>> Function: "atan":
>> diff --git a/sysdeps/x86_64/fpu/libm-test-ulps b/sysdeps/x86_64/fpu/libm-test-ulps
>> index 5ed5112b49..d4c4bfa42b 100644
>> --- a/sysdeps/x86_64/fpu/libm-test-ulps
>> +++ b/sysdeps/x86_64/fpu/libm-test-ulps
>> @@ -180,25 +180,21 @@ float: 1
>>
>> Function: "asinpi":
>> double: 1
>> -float: 1
>> float128: 2
>> ldouble: 2
>>
>> Function: "asinpi_downward":
>> double: 1
>> -float: 1
>> float128: 2
>> ldouble: 2
>>
>> Function: "asinpi_towardzero":
>> double: 1
>> -float: 2
>> float128: 1
>> ldouble: 2
>>
>> Function: "asinpi_upward":
>> double: 2
>> -float: 2
>> float128: 2
>> ldouble: 2
>>
>> --
>> 2.43.0
>>
>>
More information about the Libc-alpha
mailing list