summary refs log tree commit diff
path: root/sysdeps/powerpc/powerpc64/multiarch/rawmemchr.c
diff options
context:
space:
mode:
authorAnton Blanchard <anton@ozlabs.org>2020-05-14 21:49:16 +1000
committerPaul E. Murphy <murphyp@linux.vnet.ibm.com>2020-05-18 17:08:54 -0500
commit765de945efc5d5602999b2999fe8abdf04881370 (patch)
tree8c252731dcf3fc61ae700a5dba120c1eafc8f41f /sysdeps/powerpc/powerpc64/multiarch/rawmemchr.c
parent9e38f455a6c602be86b7b5a8d6523cbdcd7ec051 (diff)
downloadglibc-765de945efc5d5602999b2999fe8abdf04881370.tar.gz
glibc-765de945efc5d5602999b2999fe8abdf04881370.tar.xz
glibc-765de945efc5d5602999b2999fe8abdf04881370.zip
powerpc: Optimized rawmemchr for POWER9
This version uses vector instructions and is up to 60% faster on medium
matches and up to 90% faster on long matches, compared to the POWER7
version. A few examples:

                            __rawmemchr_power9  __rawmemchr_power7
Length   32, alignment  0:   2.27566             3.77765
Length   64, alignment  2:   2.46231             3.51064
Length 1024, alignment  0:  17.3059             32.6678
Diffstat (limited to 'sysdeps/powerpc/powerpc64/multiarch/rawmemchr.c')
-rw-r--r--sysdeps/powerpc/powerpc64/multiarch/rawmemchr.c12
1 files changed, 10 insertions, 2 deletions
diff --git a/sysdeps/powerpc/powerpc64/multiarch/rawmemchr.c b/sysdeps/powerpc/powerpc64/multiarch/rawmemchr.c
index 847157e5f0..c49c0c51ff 100644
--- a/sysdeps/powerpc/powerpc64/multiarch/rawmemchr.c
+++ b/sysdeps/powerpc/powerpc64/multiarch/rawmemchr.c
@@ -24,13 +24,21 @@
 
 extern __typeof (__rawmemchr) __rawmemchr_ppc attribute_hidden;
 extern __typeof (__rawmemchr) __rawmemchr_power7 attribute_hidden;
+# ifdef __LITTLE_ENDIAN__
+extern __typeof (__rawmemchr) __rawmemchr_power9 attribute_hidden;
+# endif
+
 # undef __rawmemchr
 
 /* Avoid DWARF definition DIE on ifunc symbol so that GDB can handle
    ifunc symbol properly.  */
 libc_ifunc_redirected (__redirect___rawmemchr, __rawmemchr,
-		       (hwcap & PPC_FEATURE_HAS_VSX)
-		       ? __rawmemchr_power7
+# ifdef __LITTLE_ENDIAN__
+		       (hwcap2 & PPC_FEATURE2_ARCH_3_00)
+		       ? __rawmemchr_power9 :
+# endif
+		         (hwcap & PPC_FEATURE_HAS_VSX)
+		         ? __rawmemchr_power7
 		       : __rawmemchr_ppc);
 
 weak_alias (__rawmemchr, rawmemchr)