diff options
author | Adhemerval Zanella <azanella@linux.vnet.ibm.com> | 2011-12-29 19:43:33 -0500 |
---|---|---|
committer | Adhemerval Zanella <azanella@linux.vnet.ibm.com> | 2011-12-29 19:43:33 -0500 |
commit | b20eec6a2b7056d4a1737a92574cab6d15225796 (patch) | |
tree | fb4b8e1422fa11219e79370fc5718534fb5b313b /ChangeLog | |
parent | a960be3fad233bd721f64dfe62d3f0ee798d26e3 (diff) | |
download | glibc-b20eec6a2b7056d4a1737a92574cab6d15225796.tar.gz |
PowerPC - Optimization for str[n]casecmp functions
This patch provides throughput boost for the strcasecmp function
(25% on ppc32 and 40% on ppc64) and strncasecmp (15% on both ppc32
and ppc64) for POWER7. The optimization is done by manually
(strcasecmp) or automatically (strncasecmp) unrolling the test loop
to avoid CPU stalls caused by a test followed by a load.
Diffstat (limited to 'ChangeLog')
-rw-r--r-- | ChangeLog | 18 |
1 files changed, 18 insertions, 0 deletions
@@ -1,3 +1,21 @@ +2011-11-16 Adhemerval Zanella <azanella@linux.vnet.ibm.com> + + * sysdeps/powerpc/Makefile: Added locale-defines.sym generation. + * sysdeps/powerpc/locale-defines.sym: Locale definitions for strcasecmp + optimized code. + * sysdeps/powerpc/powerpc32/power7/Makefile: New file: added unroll-loop + option for strncasecmp/strncasecmp_l compilation. + * sysdeps/powerpc/powerpc32/power7/strcasecmp.S: New file: strcasecmp + optimization for PPC32. + * sysdeps/powerpc/powerpc32/power7/strcasecmp_l.S: New file: strcasecmp_l + optimization for PPC32. + * sysdeps/powerpc/powerpc64/power7/Makefile: Added unroll-loop option for + strncasecmp/strncasecmp_l compilation. + * sysdeps/powerpc/powerpc64/power7/strcasecmp.S: New file: strcasecmp + optimization for PPC64. + * sysdeps/powerpc/powerpc64/power7/strcasecmp_l.S: New file: strcasecmp_l + optimization for PPC64. + 2011-11-18 Adhemerval Zanella <azanella@linux.vnet.ibm.com> * math/libm-test.inc: Added more nerabyint tests. |