Mateusz Guzik
20ca271fdd
amd64: depessimize bcmp for small buffers ...
Adapt assembly generated by clang for memcmp and use it for <= 64 sized
compares (which are the vast majority).
Sample result of doing stats on Broadwell (% of samples):
before: 4.0 kernel bcmp cache_lookup
after : 0.7 kernel bcmp cache_lookup
The routine is most definitely still not optimal. Anyone interested in
spending time improving it is welcome to take over.
Reviewed by: kib
2018-05-09 15:16:25 +00:00
2018-05-09 15:16:25 +00:00
2018-05-06 14:37:11 +00:00
2018-05-07 07:31:25 +00:00
2018-05-01 21:42:27 +00:00
2018-05-04 20:54:27 +00:00
2018-05-09 08:50:42 +00:00
2018-05-08 21:14:29 +00:00
2018-05-09 02:07:09 +00:00
2018-03-29 04:02:50 +00:00
2018-04-17 15:44:05 +00:00
2018-05-09 04:09:49 +00:00
2018-05-04 14:38:48 +00:00
2018-05-06 21:29:29 +00:00
2018-03-06 14:28:37 +00:00
2018-05-06 00:22:38 +00:00
2018-04-20 19:37:08 +00:00
2018-05-08 17:00:34 +00:00
2018-03-11 20:13:15 +00:00
2018-05-09 14:39:24 +00:00
2018-02-26 18:23:36 +00:00
2018-05-01 17:32:43 +00:00
2018-05-09 11:59:24 +00:00
2018-05-09 10:50:51 +00:00
2018-03-30 18:50:13 +00:00
2018-02-23 18:15:50 +00:00
2018-05-09 14:11:35 +00:00
2018-05-08 11:39:01 +00:00
2018-04-26 12:23:31 +00:00
2018-05-09 11:59:24 +00:00
2018-04-11 17:28:24 +00:00
2018-03-27 20:51:49 +00:00
2018-04-19 15:24:21 +00:00
2018-05-08 13:23:39 +00:00
2018-04-17 18:04:28 +00:00
2018-03-30 18:50:13 +00:00
2018-05-02 15:59:15 +00:00
2018-05-09 14:11:35 +00:00
2018-04-08 19:23:50 +00:00
2018-03-24 21:30:24 +00:00
2018-04-05 00:32:01 +00:00
2018-05-04 17:17:30 +00:00
2018-05-07 21:09:21 +00:00
2018-05-02 10:19:17 +00:00
20ca271fdd