diff options
| author | H.J. Lu <hjl.tools@gmail.com> | 2017-06-27 07:55:00 -0700 |
|---|---|---|
| committer | H.J. Lu <hjl.tools@gmail.com> | 2017-06-27 07:55:00 -0700 |
| commit | e94c31035739b693c3699b3c4cad0206631fbee7 (patch) | |
| tree | b0666d325944a10e78d4dd6333c19c108c50d74c /ChangeLog | |
| parent | 6980be7cbf4f108a4936ac64242f58340d56c806 (diff) | |
| download | glibc-e94c31035739b693c3699b3c4cad0206631fbee7.tar glibc-e94c31035739b693c3699b3c4cad0206631fbee7.tar.gz glibc-e94c31035739b693c3699b3c4cad0206631fbee7.tar.bz2 glibc-e94c31035739b693c3699b3c4cad0206631fbee7.zip | |
x86-64: Optimize memcmp-avx2-movbe.S for short difference
Check the first 32 bytes before checking size when size >= 32 bytes
to avoid unnecessary branch if the difference is in the first 32 bytes.
Replace vpmovmskb/subl/jnz with vptest/jnc.
On Haswell, the new version is as fast as the previous one. On Skylake,
the new version is a little bit faster.
* sysdeps/x86_64/multiarch/memcmp-avx2-movbe.S (MEMCMP): Check
the first 32 bytes before checking size when size >= 32 bytes.
Replace vpmovmskb/subl/jnz with vptest/jnc.
Diffstat (limited to 'ChangeLog')
| -rw-r--r-- | ChangeLog | 6 |
1 files changed, 6 insertions, 0 deletions
@@ -1,3 +1,9 @@ +2017-06-27 H.J. Lu <hongjiu.lu@intel.com> + + * sysdeps/x86_64/multiarch/memcmp-avx2-movbe.S (MEMCMP): Check + the first 32 bytes before checking size when size >= 32 bytes. + Replace vpmovmskb/subl/jnz with vptest/jnc. + 2017-06-27 Stefan Liebler <stli@linux.vnet.ibm.com> * sysdeps/s390/s390-32/tls-macros.h (TLS_IE): Use r12 for GOT address. |