x86-64: Optimize strcat/strncat, strcpy/strncpy and stpcpy/stpncpy with AVX2
Optimize x86-64 strcat/strncat, strcpy/strncpy and stpcpy/stpncpy with AVX2.
It uses vector comparison as much as possible. In general, the larger the
source string, the greater performance gain observed, reaching speedups of
1.6x compared to SSE2 unaligned routines. Select AVX2 strcat/strncat,
strcpy/strncpy and stpcpy/stpncpy on AVX2 machines where vzeroupper is
preferred and AVX unaligned load is fast.
* sysdeps/x86_64/multiarch/Makefile (sysdep_routines): Add
strcat-avx2, strncat-avx2, strcpy-avx2, strncpy-avx2,
stpcpy-avx2 and stpncpy-avx2.
* sysdeps/x86_64/multiarch/ifunc-impl-list.c:
(__libc_ifunc_impl_list): Add tests for __strcat_avx2,
__strncat_avx2, __strcpy_avx2, __strncpy_avx2, __stpcpy_avx2
and __stpncpy_avx2.
* sysdeps/x86_64/multiarch/{ifunc-unaligned-ssse3.h =>
ifunc-strcpy.h}: rename header for a more generic name.
* sysdeps/x86_64/multiarch/ifunc-strcpy.h:
(IFUNC_SELECTOR): Return OPTIMIZE (avx2) on AVX 2 machines if
AVX unaligned load is fast and vzeroupper is preferred.
* sysdeps/x86_64/multiarch/stpcpy-avx2.S: New file
* sysdeps/x86_64/multiarch/stpncpy-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strcat-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strcpy-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strncat-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strncpy-avx2.S: Likewise
2018-10-08 13:59:50 +00:00
|
|
|
/* strcat with AVX2
|
2020-01-01 00:14:33 +00:00
|
|
|
Copyright (C) 2011-2020 Free Software Foundation, Inc.
|
x86-64: Optimize strcat/strncat, strcpy/strncpy and stpcpy/stpncpy with AVX2
Optimize x86-64 strcat/strncat, strcpy/strncpy and stpcpy/stpncpy with AVX2.
It uses vector comparison as much as possible. In general, the larger the
source string, the greater performance gain observed, reaching speedups of
1.6x compared to SSE2 unaligned routines. Select AVX2 strcat/strncat,
strcpy/strncpy and stpcpy/stpncpy on AVX2 machines where vzeroupper is
preferred and AVX unaligned load is fast.
* sysdeps/x86_64/multiarch/Makefile (sysdep_routines): Add
strcat-avx2, strncat-avx2, strcpy-avx2, strncpy-avx2,
stpcpy-avx2 and stpncpy-avx2.
* sysdeps/x86_64/multiarch/ifunc-impl-list.c:
(__libc_ifunc_impl_list): Add tests for __strcat_avx2,
__strncat_avx2, __strcpy_avx2, __strncpy_avx2, __stpcpy_avx2
and __stpncpy_avx2.
* sysdeps/x86_64/multiarch/{ifunc-unaligned-ssse3.h =>
ifunc-strcpy.h}: rename header for a more generic name.
* sysdeps/x86_64/multiarch/ifunc-strcpy.h:
(IFUNC_SELECTOR): Return OPTIMIZE (avx2) on AVX 2 machines if
AVX unaligned load is fast and vzeroupper is preferred.
* sysdeps/x86_64/multiarch/stpcpy-avx2.S: New file
* sysdeps/x86_64/multiarch/stpncpy-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strcat-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strcpy-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strncat-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strncpy-avx2.S: Likewise
2018-10-08 13:59:50 +00:00
|
|
|
Contributed by Intel Corporation.
|
|
|
|
|
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
|
Prefer https to http for gnu.org and fsf.org URLs
Also, change sources.redhat.com to sourceware.org.
This patch was automatically generated by running the following shell
script, which uses GNU sed, and which avoids modifying files imported
from upstream:
sed -ri '
s,(http|ftp)(://(.*\.)?(gnu|fsf|sourceware)\.org($|[^.]|\.[^a-z])),https\2,g
s,(http|ftp)(://(.*\.)?)sources\.redhat\.com($|[^.]|\.[^a-z]),https\2sourceware.org\4,g
' \
$(find $(git ls-files) -prune -type f \
! -name '*.po' \
! -name 'ChangeLog*' \
! -path COPYING ! -path COPYING.LIB \
! -path manual/fdl-1.3.texi ! -path manual/lgpl-2.1.texi \
! -path manual/texinfo.tex ! -path scripts/config.guess \
! -path scripts/config.sub ! -path scripts/install-sh \
! -path scripts/mkinstalldirs ! -path scripts/move-if-change \
! -path INSTALL ! -path locale/programs/charmap-kw.h \
! -path po/libc.pot ! -path sysdeps/gnu/errlist.c \
! '(' -name configure \
-execdir test -f configure.ac -o -f configure.in ';' ')' \
! '(' -name preconfigure \
-execdir test -f preconfigure.ac ';' ')' \
-print)
and then by running 'make dist-prepare' to regenerate files built
from the altered files, and then executing the following to cleanup:
chmod a+x sysdeps/unix/sysv/linux/riscv/configure
# Omit irrelevant whitespace and comment-only changes,
# perhaps from a slightly-different Autoconf version.
git checkout -f \
sysdeps/csky/configure \
sysdeps/hppa/configure \
sysdeps/riscv/configure \
sysdeps/unix/sysv/linux/csky/configure
# Omit changes that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/powerpc/powerpc64/ppc-mcount.S: trailing lines
git checkout -f \
sysdeps/powerpc/powerpc64/ppc-mcount.S \
sysdeps/unix/sysv/linux/s390/s390-64/syscall.S
# Omit change that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S: last line does not end in newline
git checkout -f sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S
2019-09-07 05:40:42 +00:00
|
|
|
<https://www.gnu.org/licenses/>. */
|
x86-64: Optimize strcat/strncat, strcpy/strncpy and stpcpy/stpncpy with AVX2
Optimize x86-64 strcat/strncat, strcpy/strncpy and stpcpy/stpncpy with AVX2.
It uses vector comparison as much as possible. In general, the larger the
source string, the greater performance gain observed, reaching speedups of
1.6x compared to SSE2 unaligned routines. Select AVX2 strcat/strncat,
strcpy/strncpy and stpcpy/stpncpy on AVX2 machines where vzeroupper is
preferred and AVX unaligned load is fast.
* sysdeps/x86_64/multiarch/Makefile (sysdep_routines): Add
strcat-avx2, strncat-avx2, strcpy-avx2, strncpy-avx2,
stpcpy-avx2 and stpncpy-avx2.
* sysdeps/x86_64/multiarch/ifunc-impl-list.c:
(__libc_ifunc_impl_list): Add tests for __strcat_avx2,
__strncat_avx2, __strcpy_avx2, __strncpy_avx2, __stpcpy_avx2
and __stpncpy_avx2.
* sysdeps/x86_64/multiarch/{ifunc-unaligned-ssse3.h =>
ifunc-strcpy.h}: rename header for a more generic name.
* sysdeps/x86_64/multiarch/ifunc-strcpy.h:
(IFUNC_SELECTOR): Return OPTIMIZE (avx2) on AVX 2 machines if
AVX unaligned load is fast and vzeroupper is preferred.
* sysdeps/x86_64/multiarch/stpcpy-avx2.S: New file
* sysdeps/x86_64/multiarch/stpncpy-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strcat-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strcpy-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strncat-avx2.S: Likewise
* sysdeps/x86_64/multiarch/strncpy-avx2.S: Likewise
2018-10-08 13:59:50 +00:00
|
|
|
|
|
|
|
|
#if IS_IN (libc)
|
|
|
|
|
|
|
|
|
|
# include <sysdep.h>
|
|
|
|
|
|
|
|
|
|
# ifndef STRCAT
|
|
|
|
|
# define STRCAT __strcat_avx2
|
|
|
|
|
# endif
|
|
|
|
|
|
|
|
|
|
# define USE_AS_STRCAT
|
|
|
|
|
|
|
|
|
|
/* Number of bytes in a vector register */
|
|
|
|
|
# define VEC_SIZE 32
|
|
|
|
|
|
|
|
|
|
.section .text.avx,"ax",@progbits
|
|
|
|
|
ENTRY (STRCAT)
|
|
|
|
|
mov %rdi, %r9
|
|
|
|
|
# ifdef USE_AS_STRNCAT
|
|
|
|
|
mov %rdx, %r8
|
|
|
|
|
# endif
|
|
|
|
|
|
|
|
|
|
xor %eax, %eax
|
|
|
|
|
mov %edi, %ecx
|
|
|
|
|
and $((VEC_SIZE * 4) - 1), %ecx
|
|
|
|
|
vpxor %xmm6, %xmm6, %xmm6
|
|
|
|
|
cmp $(VEC_SIZE * 3), %ecx
|
|
|
|
|
ja L(fourth_vector_boundary)
|
|
|
|
|
vpcmpeqb (%rdi), %ymm6, %ymm0
|
|
|
|
|
vpmovmskb %ymm0, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_first_vector)
|
|
|
|
|
mov %rdi, %rax
|
|
|
|
|
and $-VEC_SIZE, %rax
|
|
|
|
|
jmp L(align_vec_size_start)
|
|
|
|
|
L(fourth_vector_boundary):
|
|
|
|
|
mov %rdi, %rax
|
|
|
|
|
and $-VEC_SIZE, %rax
|
|
|
|
|
vpcmpeqb (%rax), %ymm6, %ymm0
|
|
|
|
|
mov $-1, %r10d
|
|
|
|
|
sub %rax, %rcx
|
|
|
|
|
shl %cl, %r10d
|
|
|
|
|
vpmovmskb %ymm0, %edx
|
|
|
|
|
and %r10d, %edx
|
|
|
|
|
jnz L(exit)
|
|
|
|
|
|
|
|
|
|
L(align_vec_size_start):
|
|
|
|
|
vpcmpeqb VEC_SIZE(%rax), %ymm6, %ymm0
|
|
|
|
|
vpmovmskb %ymm0, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_second_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 2)(%rax), %ymm6, %ymm1
|
|
|
|
|
vpmovmskb %ymm1, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_third_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 3)(%rax), %ymm6, %ymm2
|
|
|
|
|
vpmovmskb %ymm2, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_fourth_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 4)(%rax), %ymm6, %ymm3
|
|
|
|
|
vpmovmskb %ymm3, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_fifth_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 5)(%rax), %ymm6, %ymm0
|
|
|
|
|
add $(VEC_SIZE * 4), %rax
|
|
|
|
|
vpmovmskb %ymm0, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_second_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 2)(%rax), %ymm6, %ymm1
|
|
|
|
|
vpmovmskb %ymm1, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_third_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 3)(%rax), %ymm6, %ymm2
|
|
|
|
|
vpmovmskb %ymm2, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_fourth_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 4)(%rax), %ymm6, %ymm3
|
|
|
|
|
vpmovmskb %ymm3, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_fifth_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 5)(%rax), %ymm6, %ymm0
|
|
|
|
|
add $(VEC_SIZE * 4), %rax
|
|
|
|
|
vpmovmskb %ymm0, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_second_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 2)(%rax), %ymm6, %ymm1
|
|
|
|
|
vpmovmskb %ymm1, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_third_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 3)(%rax), %ymm6, %ymm2
|
|
|
|
|
vpmovmskb %ymm2, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_fourth_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 4)(%rax), %ymm6, %ymm3
|
|
|
|
|
vpmovmskb %ymm3, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_fifth_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 5)(%rax), %ymm6, %ymm0
|
|
|
|
|
add $(VEC_SIZE * 4), %rax
|
|
|
|
|
vpmovmskb %ymm0, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_second_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 2)(%rax), %ymm6, %ymm1
|
|
|
|
|
vpmovmskb %ymm1, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_third_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 3)(%rax), %ymm6, %ymm2
|
|
|
|
|
vpmovmskb %ymm2, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_fourth_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 4)(%rax), %ymm6, %ymm3
|
|
|
|
|
vpmovmskb %ymm3, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_fifth_vector)
|
|
|
|
|
|
|
|
|
|
test $((VEC_SIZE * 4) - 1), %rax
|
|
|
|
|
jz L(align_four_vec_loop)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 5)(%rax), %ymm6, %ymm0
|
|
|
|
|
add $(VEC_SIZE * 5), %rax
|
|
|
|
|
vpmovmskb %ymm0, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit)
|
|
|
|
|
|
|
|
|
|
test $((VEC_SIZE * 4) - 1), %rax
|
|
|
|
|
jz L(align_four_vec_loop)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb VEC_SIZE(%rax), %ymm6, %ymm1
|
|
|
|
|
add $VEC_SIZE, %rax
|
|
|
|
|
vpmovmskb %ymm1, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit)
|
|
|
|
|
|
|
|
|
|
test $((VEC_SIZE * 4) - 1), %rax
|
|
|
|
|
jz L(align_four_vec_loop)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb VEC_SIZE(%rax), %ymm6, %ymm2
|
|
|
|
|
add $VEC_SIZE, %rax
|
|
|
|
|
vpmovmskb %ymm2, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit)
|
|
|
|
|
|
|
|
|
|
test $((VEC_SIZE * 4) - 1), %rax
|
|
|
|
|
jz L(align_four_vec_loop)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb VEC_SIZE(%rax), %ymm6, %ymm3
|
|
|
|
|
add $VEC_SIZE, %rax
|
|
|
|
|
vpmovmskb %ymm3, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit)
|
|
|
|
|
|
|
|
|
|
add $VEC_SIZE, %rax
|
|
|
|
|
|
|
|
|
|
.p2align 4
|
|
|
|
|
L(align_four_vec_loop):
|
|
|
|
|
vmovaps (%rax), %ymm4
|
|
|
|
|
vpminub VEC_SIZE(%rax), %ymm4, %ymm4
|
|
|
|
|
vmovaps (VEC_SIZE * 2)(%rax), %ymm5
|
|
|
|
|
vpminub (VEC_SIZE * 3)(%rax), %ymm5, %ymm5
|
|
|
|
|
add $(VEC_SIZE * 4), %rax
|
|
|
|
|
vpminub %ymm4, %ymm5, %ymm5
|
|
|
|
|
vpcmpeqb %ymm5, %ymm6, %ymm5
|
|
|
|
|
vpmovmskb %ymm5, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jz L(align_four_vec_loop)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb -(VEC_SIZE * 4)(%rax), %ymm6, %ymm0
|
|
|
|
|
sub $(VEC_SIZE * 5), %rax
|
|
|
|
|
vpmovmskb %ymm0, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_second_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 2)(%rax), %ymm6, %ymm1
|
|
|
|
|
vpmovmskb %ymm1, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_third_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 3)(%rax), %ymm6, %ymm2
|
|
|
|
|
vpmovmskb %ymm2, %edx
|
|
|
|
|
test %edx, %edx
|
|
|
|
|
jnz L(exit_null_on_fourth_vector)
|
|
|
|
|
|
|
|
|
|
vpcmpeqb (VEC_SIZE * 4)(%rax), %ymm6, %ymm3
|
|
|
|
|
vpmovmskb %ymm3, %edx
|
|
|
|
|
sub %rdi, %rax
|
|
|
|
|
bsf %rdx, %rdx
|
|
|
|
|
add %rdx, %rax
|
|
|
|
|
add $(VEC_SIZE * 4), %rax
|
|
|
|
|
jmp L(StartStrcpyPart)
|
|
|
|
|
|
|
|
|
|
.p2align 4
|
|
|
|
|
L(exit):
|
|
|
|
|
sub %rdi, %rax
|
|
|
|
|
L(exit_null_on_first_vector):
|
|
|
|
|
bsf %rdx, %rdx
|
|
|
|
|
add %rdx, %rax
|
|
|
|
|
jmp L(StartStrcpyPart)
|
|
|
|
|
|
|
|
|
|
.p2align 4
|
|
|
|
|
L(exit_null_on_second_vector):
|
|
|
|
|
sub %rdi, %rax
|
|
|
|
|
bsf %rdx, %rdx
|
|
|
|
|
add %rdx, %rax
|
|
|
|
|
add $VEC_SIZE, %rax
|
|
|
|
|
jmp L(StartStrcpyPart)
|
|
|
|
|
|
|
|
|
|
.p2align 4
|
|
|
|
|
L(exit_null_on_third_vector):
|
|
|
|
|
sub %rdi, %rax
|
|
|
|
|
bsf %rdx, %rdx
|
|
|
|
|
add %rdx, %rax
|
|
|
|
|
add $(VEC_SIZE * 2), %rax
|
|
|
|
|
jmp L(StartStrcpyPart)
|
|
|
|
|
|
|
|
|
|
.p2align 4
|
|
|
|
|
L(exit_null_on_fourth_vector):
|
|
|
|
|
sub %rdi, %rax
|
|
|
|
|
bsf %rdx, %rdx
|
|
|
|
|
add %rdx, %rax
|
|
|
|
|
add $(VEC_SIZE * 3), %rax
|
|
|
|
|
jmp L(StartStrcpyPart)
|
|
|
|
|
|
|
|
|
|
.p2align 4
|
|
|
|
|
L(exit_null_on_fifth_vector):
|
|
|
|
|
sub %rdi, %rax
|
|
|
|
|
bsf %rdx, %rdx
|
|
|
|
|
add %rdx, %rax
|
|
|
|
|
add $(VEC_SIZE * 4), %rax
|
|
|
|
|
|
|
|
|
|
.p2align 4
|
|
|
|
|
L(StartStrcpyPart):
|
|
|
|
|
lea (%r9, %rax), %rdi
|
|
|
|
|
mov %rsi, %rcx
|
|
|
|
|
mov %r9, %rax /* save result */
|
|
|
|
|
|
|
|
|
|
# ifdef USE_AS_STRNCAT
|
|
|
|
|
test %r8, %r8
|
|
|
|
|
jz L(ExitZero)
|
|
|
|
|
# define USE_AS_STRNCPY
|
|
|
|
|
# endif
|
|
|
|
|
|
|
|
|
|
# include "strcpy-avx2.S"
|
|
|
|
|
#endif
|
