Noah Goldstein 642933158e x86: Optimize and shrink st{r|p}{n}{cat|cpy}-avx2 functions ... Optimizations are: 1. Use more overlapping stores to avoid branches. 2. Reduce how unrolled the aligning copies are (this is more of a code-size save, its a negative for some sizes in terms of perf). 3. For st{r|p}n{cat|cpy} re-order the branches to minimize the number that are taken. Performance Changes: Times are from N = 10 runs of the benchmark suite and are reported as geometric mean of all ratios of New Implementation / Old Implementation. strcat-avx2 -> 0.998 strcpy-avx2 -> 0.937 stpcpy-avx2 -> 0.971 strncpy-avx2 -> 0.793 stpncpy-avx2 -> 0.775 strncat-avx2 -> 0.962 Code Size Changes: function -> Bytes New / Bytes Old -> Ratio strcat-avx2 -> 685 / 1639 -> 0.418 strcpy-avx2 -> 560 / 903 -> 0.620 stpcpy-avx2 -> 592 / 939 -> 0.630 strncpy-avx2 -> 1176 / 2390 -> 0.492 stpncpy-avx2 -> 1268 / 2438 -> 0.520 strncat-avx2 -> 1042 / 2563 -> 0.407 Notes: 1. Because of the significant difference between the implementations they are split into three files. strcpy-avx2.S -> strcpy, stpcpy, strcat strncpy-avx2.S -> strncpy strncat-avx2.S > strncat I couldn't find a way to merge them without making the ifdefs incredibly difficult to follow. Full check passes on x86-64 and build succeeds for all ISA levels w/ and w/o multiarch.		2022-11-08 19:22:33 -08:00
..
64
fpu	x86: Remove .tfloat usage	2022-10-03 14:03:21 -03:00
multiarch	x86: Optimize and shrink st{r|p}{n}{cat|cpy}-avx2 functions	2022-11-08 19:22:33 -08:00
nptl	elf: Introduce <dl-call_tls_init_tp.h> and call_tls_init_tp (bug 29249)	2022-11-03 17:28:03 +01:00
x32	Fix build with GCC 13 _FloatN, _FloatNx built-in functions	2022-10-31 23:20:08 +00:00
____longjmp_chk.S
__longjmp.S	Introduce <pointer_guard.h>, extracted from <sysdep.h>	2022-10-18 17:03:55 +02:00
_mcount.S
abort-instr.h
add_n.S
addmul_1.S
bsd-_setjmp.S
bsd-setjmp.S
configure
configure.ac
crti.S
crtn.S
dl-hwcaps-subdirs.c
dl-irel.h
dl-machine.h	x86-64: Only define used SSE/AVX/AVX512 run-time resolvers	2022-06-27 14:17:52 -07:00
dl-procinfo.c
dl-runtime.h
dl-tls.c
dl-tls.h
dl-tlsdesc.h
dl-tlsdesc.S
dl-trampoline.h	x86-64: Small improvements to dl-trampoline.S	2022-06-29 19:47:52 -07:00
dl-trampoline.S	x86-64: Small improvements to dl-trampoline.S	2022-06-29 19:47:52 -07:00
ffs.c
ffsll.c
ifuncmain8.c
ifuncmod8.c
Implies
isa-default-impl.h	x86: Remove faulty sanity tests for RTLD build with no multiarch	2022-06-23 11:14:08 -07:00
isa.h
jmpbuf-offsets.h
jmpbuf-unwind.h	Use PTR_MANGLE and PTR_DEMANGLE unconditionally in C sources	2022-10-18 17:04:10 +02:00
l10nflist.c
link-defines.sym
locale-defines.sym
localplt.data	elf: Rework exception handling in the dynamic loader [BZ #25486]	2022-11-03 09:39:31 +01:00
lshift.S
machine-gmon.h
Makefile	x86_64: Remove platform directory library loading test	2022-10-06 07:59:48 -03:00
memchr.S	x86: Add support for compiling {raw|w}memchr with high ISA level	2022-06-22 19:41:35 -07:00
memcmp-isa-default-impl.h	x86: Add support for building {w}memcmp{eq} with explicit ISA level	2022-07-05 16:42:42 -07:00
memcmp.S	x86: Add support for building {w}memcmp{eq} with explicit ISA level	2022-07-05 16:42:42 -07:00
memcmpeq.S	x86: Add support for building {w}memcmp{eq} with explicit ISA level	2022-07-05 16:42:42 -07:00
memcpy_chk.S
memcpy.S
memmove_chk.S
memmove.S	x86: Add support for building {w}memmove{_chk} with explicit ISA level	2022-07-05 16:42:42 -07:00
mempcpy_chk.S
mempcpy.S
memrchr.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
memset_chk.S
memset.S	x86: Add support for building {w}memset{_chk} with explicit ISA level	2022-07-05 16:42:42 -07:00
mp_clz_tab.c
mul_1.S
preconfigure
preconfigure.ac
rawmemchr.S	x86: Add support for compiling {raw|w}memchr with high ISA level	2022-06-22 19:41:35 -07:00
rshift.S
rtld-offsets.sym
setjmp.S	Introduce <pointer_guard.h>, extracted from <sysdep.h>	2022-10-18 17:03:55 +02:00
stackguard-macros.h
stackinfo.h	nptl: x86_64: Use same code for CURRENT_STACK_FRAME and stackinfo_get_sp	2022-08-31 09:04:27 -03:00
start.S
stpcpy.S	x86: Add support to build st{p|r}{n}{cpy|cat} with explicit ISA level	2022-07-16 03:07:59 -07:00
stpncpy.S	x86: Add support to build st{p|r}{n}{cpy|cat} with explicit ISA level	2022-07-16 03:07:59 -07:00
strcasecmp_l-nonascii.c
strcasecmp_l.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strcasecmp.S
strcat.S	x86: Add support to build st{p|r}{n}{cpy|cat} with explicit ISA level	2022-07-16 03:07:59 -07:00
strchr-isa-default-impl.h	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strchr.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strchrnul.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strcmp.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strcpy.S	x86: Add support to build st{p|r}{n}{cpy|cat} with explicit ISA level	2022-07-16 03:07:59 -07:00
strcspn-generic.c	x86: Add support for building str{c|p}{brk|spn} with explicit ISA level	2022-07-05 16:42:42 -07:00
strcspn.c	x86: Add support for building str{c|p}{brk|spn} with explicit ISA level	2022-07-05 16:42:42 -07:00
strlen.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strncase_l-nonascii.c
strncase_l.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strncase.S
strncat.S	x86: Add support to build st{p|r}{n}{cpy|cat} with explicit ISA level	2022-07-16 03:07:59 -07:00
strncmp.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strncpy.S	x86: Add support to build st{p|r}{n}{cpy|cat} with explicit ISA level	2022-07-16 03:07:59 -07:00
strnlen.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strpbrk-generic.c	x86: Add support for building str{c|p}{brk|spn} with explicit ISA level	2022-07-05 16:42:42 -07:00
strpbrk.c	x86: Add support for building str{c|p}{brk|spn} with explicit ISA level	2022-07-05 16:42:42 -07:00
strrchr.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
strspn-generic.c	x86: Add support for building str{c|p}{brk|spn} with explicit ISA level	2022-07-05 16:42:42 -07:00
strspn.c	x86: Add support for building str{c|p}{brk|spn} with explicit ISA level	2022-07-05 16:42:42 -07:00
sub_n.S
submul_1.S
sysdep.h	x86-64: Move LP_SIZE definition to its own header	2022-10-18 17:02:08 +02:00
tls_get_addr.S
tlsdesc.c
tlsdesc.sym
tst-audit3.c
tst-audit4-aux.c
tst-audit4.c
tst-audit5.c
tst-audit6.c
tst-audit7.c
tst-audit10-aux.c
tst-audit10.c
tst-audit.h
tst-auditmod3a.c
tst-auditmod3b.c
tst-auditmod4a.c
tst-auditmod4b.c
tst-auditmod5a.c
tst-auditmod5b.c
tst-auditmod6a.c
tst-auditmod6b.c
tst-auditmod6c.c
tst-auditmod7a.c
tst-auditmod7b.c
tst-auditmod10a.c
tst-auditmod10b.c
tst-avx512-aux.c
tst-avx512.c
tst-avx512mod.c
tst-avx-aux.c
tst-avx.c
tst-avxmod.c
tst-glibc-hwcaps.c
tst-platform-1.c
tst-platformmod-1.c
tst-platformmod-2.c
tst-quad1.c
tst-quad1pie.c
tst-quad2.c
tst-quad2pie.c
tst-quadmod1.S
tst-quadmod1pie.S
tst-quadmod2.S
tst-quadmod2pie.S
tst-rsi-strlen.c
tst-rsi-wcslen.c
tst-split-dynreloc.c
tst-split-dynreloc.lds
tst-sse.c
tst-ssemod.c
tst-x86-64-tls-1.c
varshift.c	x86: Add support for building str{c|p}{brk|spn} with explicit ISA level	2022-07-05 16:42:42 -07:00
Versions
wcschr.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
wcscmp.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
wcscpy-generic.c	x86: Add support to build wcscpy with explicit ISA level	2022-07-16 03:07:59 -07:00
wcscpy.S	x86: Add support to build wcscpy with explicit ISA level	2022-07-16 03:07:59 -07:00
wcslen.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
wcsncmp-generic.c	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
wcsncmp.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
wcsnlen-generic.c	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
wcsnlen.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
wcsrchr.S	x86: Add support to build strcmp/strlen/strchr with explicit ISA level	2022-07-16 03:07:59 -07:00
wmemchr.S	x86: Add support for compiling {raw|w}memchr with high ISA level	2022-06-22 19:41:35 -07:00
wmemcmp.S	x86: Add support for building {w}memcmp{eq} with explicit ISA level	2022-07-05 16:42:42 -07:00
wmemset_chk.S
wmemset.S
wordcopy.c
x86-lp_size.h	x86-64: Move LP_SIZE definition to its own header	2022-10-18 17:02:08 +02:00