diff options
author | Adhemerval Zanella <azanella@linux.vnet.ibm.com> | 2014-12-23 05:59:44 -0600 |
---|---|---|
committer | Adhemerval Zanella <azanella@linux.vnet.ibm.com> | 2015-01-13 11:28:30 -0500 |
commit | 96d6fd6c4060d739abb1822e7ad633af749532b2 (patch) | |
tree | 3958c551f6b4e375b62cf08e3e30fd472a94acf7 /sysdeps/powerpc/powerpc64/multiarch/strcpy.c | |
parent | 0f9e585480edcdf1e30dc3d79e24b84aeee516fa (diff) |
powerpc: Optimized st{r,p}cpy for POWER8/PPC64
This patch adds an optimized POWER8 strcpy using unaligned accesses.
For strings up to 16 bytes the implementation first calculate the
string size, like strlen, and issues a memcpy. For larger strings,
source is first aligned to 16 bytes and then tested over a loop that
reads 16 bytes am combine the cmpb results for speedup. Special case is
added for page cross reads.
It shows 30%-60% improvement over the optimized POWER7 one that uses
only aligned accesses.
Diffstat (limited to 'sysdeps/powerpc/powerpc64/multiarch/strcpy.c')
-rw-r--r-- | sysdeps/powerpc/powerpc64/multiarch/strcpy.c | 7 |
1 files changed, 5 insertions, 2 deletions
diff --git a/sysdeps/powerpc/powerpc64/multiarch/strcpy.c b/sysdeps/powerpc/powerpc64/multiarch/strcpy.c index cd47bf6713..fd0afd408a 100644 --- a/sysdeps/powerpc/powerpc64/multiarch/strcpy.c +++ b/sysdeps/powerpc/powerpc64/multiarch/strcpy.c @@ -23,9 +23,12 @@ extern __typeof (strcpy) __strcpy_ppc attribute_hidden; extern __typeof (strcpy) __strcpy_power7 attribute_hidden; +extern __typeof (strcpy) __strcpy_power8 attribute_hidden; libc_ifunc (strcpy, - (hwcap & PPC_FEATURE_HAS_VSX) - ? __strcpy_power7 + (hwcap2 & PPC_FEATURE2_ARCH_2_07) + ? __strcpy_power8 : + (hwcap & PPC_FEATURE_HAS_VSX) + ? __strcpy_power7 : __strcpy_ppc); #endif |