floraachy/OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
3,353 Commits 47 Branches 64 Tags
b28f75cd7e61cf5bdcf404ebece07f75553ecde0
Commit Graph

3353 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Arjan van de Ven
b28f75cd7e set GEMM_PREFERED_SIZE for HASWELL 
Haswell likes a GEMM_PREFERED_SIZE of 16 to improve the split that the
threading code does to make it a nice multiple of the SIMD kernel size
 2018-12-16 23:09:27 +00:00
Arjan van de Ven
d321448a63 dgemm: use dgemm_ncopy_8_skylakex.c also for Haswell 
The dgemm_ncopy_8_skylakex.c code is not avx512 specific and gives
a nice performance boost for medium sized matrices
 2018-12-16 23:09:22 +00:00
Arjan van de Ven
c43331ad0a dgemm: Use the skylakex beta function also for haswell 
it's more efficient for certain tall/skinny matrices
 2018-12-16 23:09:17 +00:00
Martin Kroeker
e8ca5a59a9 Merge pull request #1919 from fenrus75/haswelltuning 
(sgemm) Apply some of the SKYLAKEX optimizations also to HASWELL
 2018-12-16 20:11:05 +01:00
Martin Kroeker
c4e23dd016 Update Makefile 2018-12-16 18:14:40 +01:00
Martin Kroeker
cfc4acc221 typo 2018-12-16 16:19:51 +01:00
Martin Kroeker
545c2b1bbb Add -mavx2 on Haswell only if the compiler supports it 2018-12-16 13:09:19 +01:00
Arjan van de Ven
69d206440a Make the skylakex/haswell sgemm code compile and run even with compilers without avx2 support 2018-12-16 00:19:41 +00:00
Martin Kroeker
3843e3e017 use -maxv2 on haswell 2018-12-15 23:30:31 +01:00
Martin Kroeker
fbcb14a74b should be core-avx2 2018-12-15 20:18:59 +01:00
Martin Kroeker
2a3190dc76 fix elseifeq and use older option core2-avx for compatibility 2018-12-15 20:17:44 +01:00
Martin Kroeker
1ebe5c0f49 Add -march=haswell to HASWELL part of DYNAMIC_ARCH build 2018-12-15 19:35:35 +01:00
Arjan van de Ven
0586899a10 Use sgemm_ncopy_4_skylakex.c also for Haswell 
sgemm_ncopy_4_skylakex.c uses SSE transpose operations where the
real perf win happens; this also works great for Haswell.

This gives double digit percentage gains on small and skinny matrices
 2018-12-15 13:49:19 +00:00
Arjan van de Ven
00dc09ad19 Use the skylake sgemm beta code also for haswell 
with a few small changes it's possible to use the skylake sgemm code
also for haswell, this gives a modest gain (10% range) for smallish
matrixes but does wonders for very skinny matrixes
 2018-12-15 13:49:13 +00:00
Martin Kroeker
78d877b54b Merge pull request #1914 from fenrus75/smallmatrix 
Add a "sgemm direct" mode for small matrixes
 2018-12-13 19:08:14 +01:00
Arjan van de Ven
cdc668d82b Add a "sgemm direct" mode for small matrixes 
OpenBLAS has a fancy algorithm for copying the input data while laying
it out in a more CPU friendly memory layout.

This is great for large matrixes; the cost of the copy is easily
ammortized by the gains from the better memory layout.

But for small matrixes (on CPUs that can do efficient unaligned loads) this
copy can be a net loss.

This patch adds (for SKYLAKEX initially) a "sgemm direct" mode, that bypasses
the whole copy machinary for ALPHA=1/BETA=0/... standard arguments,
for small matrixes only.

What is small? For the non-threaded case this has been measured to be
in the M*N*K = 28 * 512 * 512 range, while in the threaded case it's
less, around M*N*K = 1 * 512 * 512
 2018-12-13 13:47:31 +00:00
Martin Kroeker
87718807f0 Merge pull request #1910 from martin-frbg/issue1909 
Fix for DYNAMIC_ARCH builds made on a AVX512-capable host
 2018-12-12 14:56:25 +01:00
Martin Kroeker
51aec8e96b make sure the added march=skylake-avx512 does not cause problems on Windows 2018-12-11 22:47:32 +01:00
Martin Kroeker
06f7d78d70 Add -march=skylake-avx512 to SkylakeX part of DYNAMIC_ARCH builds 2018-12-11 21:10:38 +01:00
Martin Kroeker
38cc638591 Avoid adding blanket march=skylake-avx512 to dynamic_arch builds 2018-12-11 21:09:26 +01:00
Martin Kroeker
0bf6d74e5f Fix typo in previous commit for arm dynamic arch 2018-12-07 19:37:33 +01:00
Martin Kroeker
133c278ee5 Add DYNAMIC_CORE list for ARM64 
cf #1908
 2018-12-07 17:42:23 +01:00
Martin Kroeker
2b355592e3 Make sure to use the arm version of dynamic.c in ARM64 DYNAMIC_ARCH 
cf. #1908
 2018-12-07 16:25:55 +01:00
Martin Kroeker
ff3eb1d474 Merge pull request #1904 from martin-frbg/issue1870 
Fix cmake parsing of GEMM kernels for ARMV8
 2018-12-06 23:01:23 +01:00
Martin Kroeker
0b09516678 Fix missing parameter in popen call 2018-12-06 18:33:05 +01:00
Martin Kroeker
7639f2e1f0 Rewrite the conditional for OSX to fix cmake parsing on others 
The Makefile variable parser in utils.cmake currently does not handle conditionals. Having the definitions for non-OSX last will at least make cmake builds work again on non-OSX platforms.
 2018-12-06 14:04:27 +01:00
Martin Kroeker
2fc712469d Avoid creating spurious non-suffixed c/zgemm_kernels 
Plain cgemm_kernel and zgemm_kernel are not used anywhere, only cgemm_kernel_b etc.
Needlessly building them (without any define like NN, CN, etc.) just happened to work on most platforms, but not on arm64. See #1870
 2018-12-06 13:56:06 +01:00
Martin Kroeker
6ba30e270d Fix typo that broke CNRM2 on ARMV8 since 0.3.0 
must have happened in my #1449
 2018-12-06 13:42:25 +01:00
Martin Kroeker
bf23518e36 Merge pull request #1903 from rengolin/armv8 
Fix two mistakes on Arm64 builds
 2018-12-05 22:10:53 +01:00
Renato Golin
31a490ea88 Fix two mistakes on Arm64 builds 
* Falkor is an ARMv8.0 with ARMv8.1 features, and chosing armv8.1-a for
   march generates instructions it cannot cope with. Reverting it back
   to armv8-a.
 * ThunderX2's build was left with a #define VULCAN, which made it miss
   the right compiler flags in Makefile.arm64, although it did create
   the right library in the end.
 2018-12-05 18:51:38 +00:00
Martin Kroeker
701ea88347 Use p2align instead of align for OSX compatibility 
fixes #1902
 2018-12-03 13:06:43 +01:00
Martin Kroeker
721c56c224 Merge pull request #1899 from brada4/fbsd12 
Add mutually supported architecture mappings for FreeBSD12 ports
 2018-12-03 12:50:27 +01:00
Martin Kroeker
c5f8aeff2d Merge branch 'develop' into fbsd12 2018-12-03 12:50:14 +01:00
Martin Kroeker
8278cbe7f8 Merge pull request #1894 from pkubaj/patch-2 
Use correct ARCH name on BSD powerpc64
 2018-12-03 12:48:53 +01:00
Martin Kroeker
ea6d1b96bd Update Makefile.system 2018-12-03 08:59:10 +01:00
Martin Kroeker
360374be62 Update with the changes from 0.3.4 2018-12-02 23:44:13 +01:00
Martin Kroeker
f5acaad8f0 Increment version to 0.3.5.dev 2018-12-02 23:43:15 +01:00
Martin Kroeker
93fa6b7b76 Increment version to 0.3.5.dev 2018-12-02 23:42:33 +01:00
Martin Kroeker
b028960aba Merge branch 'release-0.3.0' into develop 2018-12-02 23:38:49 +01:00
Martin Kroeker
3c9e3faedb fixup BSD naming of powerpc arch 2018-12-02 23:24:53 +01:00
Andrew
44c81fd135 oops 2018-12-02 20:27:53 +01:00
Andrew
26b3710485 Add architecture mappings for FreeBSD12 2018-12-02 12:07:41 +01:00
Andrew
84e614d0fd init 2018-12-02 12:05:15 +01:00
Martin Kroeker
dceff5542c Handle Android environments that identify as Linux (#1898) 
* Handle Android environments that identify as Linux

termux terminal emulator does this, causing build failures through missed defines in common.h
 2018-12-01 20:56:11 +01:00
Martin Kroeker
6c7b691083 Really revert xDOT changes from 1832 
neglected to rebase #1892 on merging
 2018-11-30 21:32:01 +01:00
Martin Kroeker
5f4c550c27 Merge pull request #1892 from martin-frbg/mipsdot 
revert MIPS64 xDOT kernel changes from #1832
 2018-11-30 21:28:21 +01:00
pkubaj
731b2722ba Fix build on POWER, remove DragonFly, add NetBSD 
__asm is complete on its own

DBSD developers state they will only support amd64, but NetBSD supports POWER.
 2018-11-30 21:12:05 +01:00
pkubaj
f85ce54d4a Use correct Makefile on powerpc64 
FreeBSD uses powerpc64 name for POWER architecture. Use correct Makefile for this platform.
 2018-11-30 16:05:49 +00:00
Andrew
2601cd58ab remove surplus locking code , only enabled w x86, disabled or never enabled on all others 2018-11-30 11:38:19 +01:00
Martin Kroeker
95a5542e3c Revert DOT kernel changes from #1834 
as the failures seen on Loongson3A appear to be limited to DSDOT/SDSDOT (i.e. my hackish "fix" from #1684)
 2018-11-30 11:16:24 +01:00