mirror of git://sourceware.org/git/glibc.git
i386: Replace assembly versions of e_exp2f with generic e_exp2f.c
This patch replaces i386 assembly versions of e_exp2f with generic e_exp2f.c. For workload-spec2017.wrf, on Nehalem, it improves performance by: Before After Improvement reciprocal-throughput 112.996 40.0454 182% latency 126.581 54.4479 132% On Skylake, it improves performance by: Before After Improvement reciprocal-throughput 113.14 39.447 186% latency 136.068 55.684 144% On IvyBridge with --disable-multi-arch, it improves performance by: Before After Improvement reciprocal-throughput 132.521 40.3759 228% latency 145.791 58.4587 149% * sysdeps/i386/fpu/e_exp2f.S: Removed. * sysdeps/i386/fpu/w_exp2f.c: Likewise. * sysdeps/i386/fpu/libm-test-ulps: Updated for generic e_exp2f.c. * sysdeps/i386/i686/fpu/multiarch/libm-test-ulps: Likewise. * sysdeps/i386/i686/fpu/multiarch/Makefile (libm-sysdep_routines): Add e_exp2f-sse2. (CFLAGS-e_exp2f-sse2.c): New. * sysdeps/i386/i686/fpu/multiarch/e_exp2f-sse2.c: New file. * sysdeps/i386/i686/fpu/multiarch/e_exp2f.c: Likewise.
This commit is contained in:
parent
5d15c96975
commit
7eda65f69e
12
ChangeLog
12
ChangeLog
|
@ -1,3 +1,15 @@
|
|||
2017-10-22 H.J. Lu <hongjiu.lu@intel.com>
|
||||
|
||||
* sysdeps/i386/fpu/e_exp2f.S: Removed.
|
||||
* sysdeps/i386/fpu/w_exp2f.c: Likewise.
|
||||
* sysdeps/i386/fpu/libm-test-ulps: Updated for generic e_exp2f.c.
|
||||
* sysdeps/i386/i686/fpu/multiarch/libm-test-ulps: Likewise.
|
||||
* sysdeps/i386/i686/fpu/multiarch/Makefile (libm-sysdep_routines):
|
||||
Add e_exp2f-sse2.
|
||||
(CFLAGS-e_exp2f-sse2.c): New.
|
||||
* sysdeps/i386/i686/fpu/multiarch/e_exp2f-sse2.c: New file.
|
||||
* sysdeps/i386/i686/fpu/multiarch/e_exp2f.c: Likewise.
|
||||
|
||||
2017-10-22 H.J. Lu <hongjiu.lu@intel.com>
|
||||
|
||||
* sysdeps/x86_64/fpu/multiarch/Makefile (libm-sysdep_routines):
|
||||
|
|
|
@ -1,52 +0,0 @@
|
|||
/*
|
||||
* Written by J.T. Conklin <jtc@netbsd.org>.
|
||||
* Adapted for exp2 by Ulrich Drepper <drepper@cygnus.com>.
|
||||
* Public domain.
|
||||
*/
|
||||
|
||||
#include <machine/asm.h>
|
||||
#include <i386-math-asm.h>
|
||||
|
||||
DEFINE_FLT_MIN
|
||||
|
||||
#ifdef PIC
|
||||
# define MO(op) op##@GOTOFF(%ecx)
|
||||
#else
|
||||
# define MO(op) op
|
||||
#endif
|
||||
|
||||
.text
|
||||
ENTRY(__ieee754_exp2f)
|
||||
#ifdef PIC
|
||||
LOAD_PIC_REG (cx)
|
||||
#endif
|
||||
flds 4(%esp)
|
||||
/* I added the following ugly construct because exp(+-Inf) resulted
|
||||
in NaN. The ugliness results from the bright minds at Intel.
|
||||
For the i686 the code can be written better.
|
||||
-- drepper@cygnus.com. */
|
||||
fxam /* Is NaN or +-Inf? */
|
||||
fstsw %ax
|
||||
movb $0x45, %dh
|
||||
andb %ah, %dh
|
||||
cmpb $0x05, %dh
|
||||
je 1f /* Is +-Inf, jump. */
|
||||
fld %st
|
||||
frndint /* int(x) */
|
||||
fsubr %st,%st(1) /* fract(x) */
|
||||
fxch
|
||||
f2xm1 /* 2^(fract(x)) - 1 */
|
||||
fld1
|
||||
faddp /* 2^(fract(x)) */
|
||||
fscale /* e^x */
|
||||
fstp %st(1)
|
||||
FLT_NARROW_EVAL_UFLOW_NONNEG_NAN
|
||||
ret
|
||||
|
||||
1: testl $0x200, %eax /* Test sign. */
|
||||
jz 2f /* If positive, jump. */
|
||||
fstp %st
|
||||
fldz /* Set result to 0. */
|
||||
2: ret
|
||||
END (__ieee754_exp2f)
|
||||
strong_alias (__ieee754_exp2f, __exp2f_finite)
|
|
@ -1921,7 +1921,9 @@ ildouble: 1
|
|||
ldouble: 1
|
||||
|
||||
Function: "exp2_upward":
|
||||
float: 1
|
||||
float128: 2
|
||||
ifloat: 1
|
||||
ifloat128: 2
|
||||
ildouble: 1
|
||||
ldouble: 1
|
||||
|
|
|
@ -1 +0,0 @@
|
|||
#include <sysdeps/../math/w_exp2f.c>
|
|
@ -1,6 +1,7 @@
|
|||
ifeq ($(subdir),math)
|
||||
libm-sysdep_routines += e_expf-sse2 s_sinf-sse2 s_cosf-sse2 \
|
||||
libm-sysdep_routines += e_exp2f-sse2 e_expf-sse2 s_sinf-sse2 s_cosf-sse2 \
|
||||
s_sincosf-sse2
|
||||
|
||||
CFLAGS-e_exp2f-sse2.c = -msse2 -mfpmath=sse
|
||||
CFLAGS-e_expf-sse2.c = -msse2 -mfpmath=sse
|
||||
endif
|
||||
|
|
|
@ -0,0 +1,3 @@
|
|||
#define __exp2f __exp2f_sse2
|
||||
|
||||
#include <sysdeps/ieee754/flt-32/e_exp2f.c>
|
|
@ -0,0 +1,37 @@
|
|||
/* Multiple versions of exp2f.
|
||||
Copyright (C) 2017 Free Software Foundation, Inc.
|
||||
This file is part of the GNU C Library.
|
||||
|
||||
The GNU C Library is free software; you can redistribute it and/or
|
||||
modify it under the terms of the GNU Lesser General Public
|
||||
License as published by the Free Software Foundation; either
|
||||
version 2.1 of the License, or (at your option) any later version.
|
||||
|
||||
The GNU C Library is distributed in the hope that it will be useful,
|
||||
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||||
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
||||
Lesser General Public License for more details.
|
||||
|
||||
You should have received a copy of the GNU Lesser General Public
|
||||
License along with the GNU C Library; if not, see
|
||||
<http://www.gnu.org/licenses/>. */
|
||||
|
||||
extern float __redirect_exp2f (float);
|
||||
|
||||
#define SYMBOL_NAME exp2f
|
||||
#include "ifunc-sse2.h"
|
||||
|
||||
libc_ifunc_redirected (__redirect_exp2f, __exp2f, IFUNC_SELECTOR ());
|
||||
|
||||
#ifdef SHARED
|
||||
# include <shlib-compat.h>
|
||||
versioned_symbol (libm, __exp2f, exp2f, GLIBC_2_27);
|
||||
#else
|
||||
weak_alias (__exp2f, exp2f)
|
||||
#endif
|
||||
|
||||
strong_alias (__exp2f, __ieee754_exp2f)
|
||||
strong_alias (__exp2f, __exp2f_finite)
|
||||
|
||||
#define __exp2f __exp2f_ia32
|
||||
#include <sysdeps/ieee754/flt-32/e_exp2f.c>
|
|
@ -1921,7 +1921,9 @@ ildouble: 1
|
|||
ldouble: 1
|
||||
|
||||
Function: "exp2_upward":
|
||||
float: 1
|
||||
float128: 2
|
||||
ifloat: 1
|
||||
ifloat128: 2
|
||||
ildouble: 1
|
||||
ldouble: 1
|
||||
|
|
Loading…
Reference in New Issue