AuroraMiddleware/glibc
1
0
Fork 0
mirror of https://sourceware.org/git/glibc.git synced 2024-11-21 20:40:05 +00:00
Code Issues Packages Projects Releases Wiki Activity
5c22fd25c1
glibc/sysdeps/i386
History
Adhemerval Zanella 5c22fd25c1 math: Use exp2m1f from CORE-MATH 
The CORE-MATH implementation is correctly rounded (for any rounding mode)
and shows better performance compared to the generic exp2m1f.

The code was adapted to glibc style and to use the definition of
math_config.h (to handle errno, overflow, and underflow).  The
only change is to handle FLT_MAX_EXP for FE_DOWNWARD or FE_TOWARDZERO.

The benchmark inputs are based on exp2f ones.

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                      40.6042        48.7104       -19.96%
x86_64v2                    40.7506        35.9032        11.90%
x86_64v3                    35.2301        31.7956        9.75%
i686                        102.094        94.6657        7.28%
aarch64                     18.2704        15.1387        17.14%
power10                     11.9444         8.2402        31.01%

reciprocal-throughput        master        patched   improvement
x86_64                      20.8683        16.1428        22.64%
x86_64v2                    19.5076        10.4474        46.44%
x86_64v3                    19.2106        10.4014        45.86%
i686                        56.4054        59.3004        -5.13%
aarch64                     12.0781         7.3953        38.77%
power10                      6.5306         5.9388         9.06%

The generic implementation calls __ieee754_exp2f and x86_64 provides
an optimized ifunc version (built with -mfma -mavx2, not correctly
rounded).  This explains the performance difference for x86_64.

Same for i686, where the ABI provides an optimized __ieee754_exp2f
version built with '-msse2 -mfpmath=sse'.  When built wth same
flags, the new algorithm shows a better performance:

                            master        patched    improvement
latency                    102.094        91.2823         10.59%
reciprocal-throughput      56.4054        52.7984          6.39%

Signed-off-by: Alexei Sibidanov <sibid@uvic.ca>
Signed-off-by: Paul Zimmermann <Paul.Zimmermann@inria.fr>
Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Reviewed-by: DJ Delorie <dj@redhat.com>
2024-11-01 11:27:35 -03:00
..
fpu math: Use exp2m1f from CORE-MATH 2024-11-01 11:27:35 -03:00
htl Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
i586 i586: Fix multiple definitions of __memcpy_chk and __mempcpy_chk 2024-05-02 11:50:21 +01:00
i686 math: Use exp2m1f from CORE-MATH 2024-11-01 11:27:35 -03:00
i786 Also applying directories. 1999-01-24 10:39:22 +00:00
nptl Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
sys Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
____longjmp_chk.S Add sigstack handling to Linux ____longjmp_chk on i386. 2009-07-30 21:50:14 -07:00
__longjmp.S i386: Remove CET support 2024-01-09 13:55:51 -03:00
abort-instr.h update from main archive 961220 1996-12-21 04:13:58 +00:00
add_n.S i386: Remove CET support 2024-01-09 13:55:51 -03:00
addmul_1.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
asm-syntax.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
backtrace.c Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
bsd-_setjmp.S i386: Remove CET support 2024-01-09 13:55:51 -03:00
bsd-setjmp.S i386: Remove CET support 2024-01-09 13:55:51 -03:00
configure Convert to autoconf 2.72 (vanilla release, no distribution patches) 2024-06-17 21:15:28 +02:00
configure.ac i386: Fail if configured with --enable-cet 2024-01-09 13:55:51 -03:00
crti.S i386: Remove CET support 2024-01-09 13:55:51 -03:00
crtn.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
dl-fixup-attribute.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
dl-irel.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
dl-machine-rel.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
dl-machine.h x86: Update _dl_tlsdesc_dynamic to preserve caller-saved registers 2024-02-28 09:02:56 -08:00
dl-procinfo.c Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
dl-tls.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
dl-tlsdesc-dynamic.h x86: Update _dl_tlsdesc_dynamic to preserve caller-saved registers 2024-02-28 09:02:56 -08:00
dl-tlsdesc.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
dl-tlsdesc.S x86: Update _dl_tlsdesc_dynamic to preserve caller-saved registers 2024-02-28 09:02:56 -08:00
dl-trampoline.S i386: Remove CET support bits 2024-01-10 05:20:20 -08:00
gccframe.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
i386-mcount.S i386: Remove CET support 2024-01-09 13:55:51 -03:00
Implies Add float128 support for x86_64, x86. 2017-06-26 22:02:24 +00:00
isa.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
jmpbuf-offsets.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
jmpbuf-unwind.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
link-defines.sym elf: Remove Intel MPX support (lazy PLT, ld.so profile, and LD_AUDIT) 2021-10-11 11:14:02 -07:00
lshift.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
machine-gmon.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
Makefile mcheck: Align struct hdr to MALLOC_ALIGNMENT bytes [BZ #28068] 2021-07-12 18:13:32 -07:00
malloc-alignment.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
math-use-builtins-ffs.h string: Use builtins for ffs and ffsll 2024-02-01 09:31:33 -03:00
memchr.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
memcmp.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
memcopy.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
memcpy_chk.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
memcpy.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
memmove_chk.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
memmove.S Add i386 memset and memcpy assembly functions 2015-08-27 09:04:54 -07:00
mempcpy_chk.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
mempcpy.S Add i386 memset and memcpy assembly functions 2015-08-27 09:04:54 -07:00
memset_chk.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
memset.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
mp_clz_tab.c Update. 2002-03-14 20:48:50 +00:00
mul_1.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
preconfigure Move base_machine and machine settings from configure.ac to sysdeps preconfigure fragments. 2014-06-25 17:52:56 +00:00
pthread_spin_trylock.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
rawmemchr.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
rshift.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
setfpucw.c Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
setjmp.S i386: Remove CET support 2024-01-09 13:55:51 -03:00
stackguard-macros.h BZ #15754: CVE-2013-4788 2013-09-23 00:52:09 -04:00
stackinfo.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
start.S i386: Remove CET support 2024-01-09 13:55:51 -03:00
stpcpy.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
stpncpy.S i386: Don't define stpncpy alias when used in IFUNC [BZ #31768] 2024-05-20 19:35:00 -07:00
strcat.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
strchr.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
strchrnul.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
strcspn.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
string-inlines.c Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
string-opthr.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
strlen.c Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
strlen.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
strpbrk.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
strrchr.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
strspn.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
sub_n.S i386: Remove CET support 2024-01-09 13:55:51 -03:00
submul_1.S Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
symbol-hacks.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
sysdep.h i386: Remove CET support 2024-01-09 13:55:51 -03:00
tlsdesc.c Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
tlsdesc.sym Introduce TLS descriptors for i386 and x86_64. 2008-05-13 05:41:30 +00:00
tst-audit3.c Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
tst-audit3.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
tst-audit.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
tst-auditmod3a.c Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
tst-auditmod3b.c Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
tst-ld-sse-use.sh Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
unwind-arch.h Update copyright dates with scripts/update-copyrights 2024-01-01 10:53:40 -08:00
Versions Move __fentry__ version definition to sysdeps/{i386,x86_64} 2018-08-10 09:07:44 +02:00