Commit Graph

20 Commits

Author SHA1 Message Date
Danila Kutenin
08e3f77a9c Optimize most important parts with NEON SIMD
First part, updated most UTF8, hashing, memory and codecs. Except
utf8lower and upper, maybe a little later.

That includes huge amount of research with movemask dealing. Exact
details and blog post TBD.
2022-06-15 13:19:29 +00:00
Alexander Gololobov
4350935377 Re-enable SSE2 for PowerPC 2022-05-20 08:51:48 +02:00
Robert Schulze
e06db56c67
Move ppc cpu features out of the x86 block 2022-05-15 09:48:48 +02:00
Alexander Gololobov
1dc3d04837 Removed __SSE2__ flag for PPC64 2022-05-12 16:20:19 +02:00
Alexey Milovidov
380f1a0052 Limit PowerPC code generation to Power8 2022-04-22 06:26:40 +02:00
hanqf-git
71ce4b46d8 Add x86 feature avx512 support for memcmpSmall. Co-authored-by: @vesslanjin jun.i.jin@intel.com 2022-01-17 16:13:49 +08:00
hanqf-git
d2bc77ac47 Add x86 feature avx512 support for memcmpSmall 2022-01-17 12:03:12 +08:00
Alexey Milovidov
541f3ae1f9 Fix build on AArch64 2021-11-21 14:30:36 +03:00
Alexey Milovidov
6246cec409 Fix build on AArch64 2021-11-21 14:30:04 +03:00
jasperzhu
0217ed09ac revise compile flag 2021-10-27 05:19:15 +05:30
jasperzhu
3d0e595579 revise compile flag 2021-10-13 22:49:21 +05:30
jasperzhu
4b080489b5 revise flags definitions 2021-10-13 07:34:21 +05:30
jasperzhu
80760c0a0c revise compile definitions 2021-10-13 05:59:20 +05:30
jasperzhu
f08dd0e8d8 modify flags definitions 2021-10-13 05:13:34 +05:30
jasperzhu
002c8d3a80 add disable vectorize flag due to llvm bug 2021-10-13 03:38:18 +05:30
jasperzhu
be28e94471 add compile flag -force-vector-width=16 due to llvm autovectorization bug with avx512 2021-10-12 06:30:56 +05:30
jasperzhu
0ef2624408 add specific flags for compiling filter operation source files 2021-10-12 04:04:27 +05:30
jasperzhu
a3d629a5b5 add x86 feature avx2/avx512 support for filter implementation 2021-10-11 22:51:13 +05:30
Alexey Milovidov
eda489dde8 Untangle CPU features specification in CMake 2021-08-10 02:58:51 +03:00
Ivan Lezhankin
ac47427b76 cmake/test_cpu → cmake/cpu_features 2019-09-13 17:54:22 +03:00