floraachy/OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
7,087 Commits 47 Branches 64 Tags
e2779c852f0630b6ff6a2eec9e8cfbb014e300f7
Commit Graph

7087 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Martin Kroeker
e2779c852f Do not build the tests when only the CBLAS interface is selected (#4041) 
* Do not build the tests when only the CBLAS interface is selected
 2023-05-15 20:49:56 +02:00
Martin Kroeker
ccad94162a Merge pull request #4039 from klho/develop 
Bug fix and improvements for [z]imatcopy interface.
 2023-05-14 10:51:24 +02:00
Ken Ho
df1b1f6a91 More detailed error message in [z]imatcopy.c. 2023-05-12 09:41:52 -07:00
Ken Ho
7a86c437b5 Change some "if" statements to "else if" following suggestion by @mmuetzel. 2023-05-10 09:13:04 -07:00
Ken Ho
33ab415f68 Bug fix and improvements for [z]imatcopy interface. 2023-05-08 14:43:56 -07:00
Martin Kroeker
c74ee11376 Add an M1-based OSX crossbuild and a NeoverseN1 build to CIRRUS CI (#3997) 
* Add an M1-based OSX crossbuild and a NeoverseN1 build (plus Windows//LLVM commented out for now)
 2023-05-08 14:24:38 +02:00
Martin Kroeker
65a7941aa5 Merge pull request #4036 from martin-frbg/issue4020 
Mark cblas_xerbla's arguments as const in cblas.h
 2023-05-08 12:54:30 +02:00
Martin Kroeker
c2078b2356 Mark xerbla's arguments as const 2023-05-07 20:15:13 +02:00
Martin Kroeker
d6a42ed574 Merge pull request #4035 from martin-frbg/issue4034 
Fix (redundant) lapack-runtest target in toplevel Makefile
 2023-05-06 15:51:07 +02:00
Martin Kroeker
60226b35e1 Fix (redundant) lapack-runtest target 2023-05-06 12:44:38 +02:00
Martin Kroeker
4e597ae00b Merge pull request #4031 from martin-frbg/issue4026 
Add suggestions to NUM_THREADS/auxiliary buffer message
 2023-05-05 09:32:32 +02:00
Martin Kroeker
e5538a62cb Add suggestions to NUM_THREADS/auxiliary buffer message 2023-05-04 22:56:39 +02:00
Martin Kroeker
6f38a946e8 Merge pull request #4028 from catap/mktemp-fix 
Do not requires GNU mktemp
 2023-05-03 11:25:25 +02:00
Martin Kroeker
29c717050f Merge pull request #4022 from martin-frbg/gemmtm 
fix cblas_?gemmt
 2023-05-03 11:24:54 +02:00
Kirill A. Korinsky
b1781ad338 Do not requires GNU mktemp 
Historically the GNU mktemp was the first one which doesn't requires
`-t` to create a directory.

Here I've introduced a fallback when `-t` is required.

For example MacPorts contains similar patch: bbe8abfe26/math/OpenBLAS/files/patch-MacOSX-mktemp.diff
 2023-04-29 11:13:26 +02:00
Martin Kroeker
1f6f7328eb remove redundant declaration 2023-04-27 09:14:12 +02:00
Martin Kroeker
7152d6b06d fix cblas_gemmt 2023-04-27 08:36:20 +02:00
Martin Kroeker
e9a8d5b45f Merge pull request #4015 from martin-frbg/issue4013-2 
[WIP] Disable gcc's tree-vectorizer for x86_64 CGEMV
 2023-04-23 18:51:12 +02:00
Martin Kroeker
72caceb324 Merge pull request #4009 from Mousius/sve-gemm 
Use SVE kernel for SGEMM/DGEMM on Arm(R) Neoverse(TM) V1
 2023-04-22 13:56:45 +02:00
Martin Kroeker
d1b631899b Merge pull request #4018 from mmuetzel/ci 
Adapt CI rules for MSYS2 for updated ccache
 2023-04-21 23:52:13 +02:00
Markus Mützel
e27e9a50b1 CI (MSYS2): Save ccache before running tests. 2023-04-21 14:10:40 +02:00
Markus Mützel
67d33e5b98 CI (MSYS2): Update location of compiler cache. 2023-04-21 13:02:23 +02:00
Martin Kroeker
84bcf6639f Disable gcc's tree-vectorizer pass on all operating systems 2023-04-20 23:24:52 +02:00
Martin Kroeker
30a0ccbd14 Merge pull request #4014 from martin-frbg/issue4013 
Generally disable gcc's tree-vectorizer in x86_64 SGEMV,SSYMV,ZGEMV,C/ZDOT
 2023-04-20 10:45:15 +02:00
Martin Kroeker
c9174ae8d7 Disable gcc's tree-vectorizer pass on all operating systems 2023-04-19 23:45:44 +02:00
Martin Kroeker
c2fe9cb91f Disable gcc's tree-vectorizer pass on all operating systems 2023-04-19 23:45:14 +02:00
Martin Kroeker
66b39b835c Disable gcc's tree-vectorizer pass on all operating systems 2023-04-19 23:44:45 +02:00
Martin Kroeker
bb6d6735bf Disable gcc's tree-vectorizer pass on all operating systems 2023-04-19 23:44:15 +02:00
Martin Kroeker
d18efaed20 Disable gcc's tree-vectorizer pass on all operating systems 2023-04-19 23:43:43 +02:00
Martin Kroeker
99f6d31ed5 Disable gcc's tree-vectorizer pass on all operating systems 2023-04-19 23:42:55 +02:00
Martin Kroeker
7de9335c56 Disable gcc's tree-vectorizer pass on all operating systems 2023-04-19 23:42:09 +02:00
Martin Kroeker
437c0bf2b4 Merge pull request #3843 from Mousius/switch-ratio 
Propagate SWITCH_RATIO to DYNAMIC_ARCH builds
 2023-04-19 11:51:54 +02:00
Martin Kroeker
c628030669 Merge pull request #3855 from Mousius/more-switch-ratio-tuning 
SWITCH_RATIO for Arm(R) Neoverse(TM) architecture
 2023-04-18 22:45:51 +02:00
Martin Kroeker
efcf71255a Merge pull request #4003 from martin-frbg/issue3995 
Fix instabilities in CGEMM/CTRMM/DNRM2 on Apple M1/M2 under OSX
 2023-04-18 14:55:23 +02:00
Martin Kroeker
51dd1339e7 Merge pull request #4010 from martin-frbg/issue3989-2 
Remove any stray trailing dash from CROSS_SUFFIX
 2023-04-18 14:55:02 +02:00
Martin Kroeker
479509bb37 Remove any stray trailing dash from CROSS_SUFFIX (as would result from clang -arch) 2023-04-17 21:57:25 +02:00
Chris Sidebottom
ec334e69dc Use SVE kernel for SGEMM/DGEMM on Arm(R) Neoverse(TM) V1 
This re-spins #3869 with some additional copy unrolling which helps maintain SYRK performance.

After #3868, the SVE kernels represent a pretty good boost.

This re-uses ARMV8SVE as a base and I'm going to incrementally move everything to use ARMV8SVE in additional patches (as well as fix up anything that's not already in ARMV8SVE).
 2023-04-17 17:38:42 +01:00
Chris Sidebottom
5b165420b5 SWITCH_RATIO for Arm(R) Neoverse(TM) architecture 
This seems like a good balance of values for reasonably sized matrices. With `SWITCH_RATIO=16` the DGEMM scales better to bigger sizes but the better solution would be some kind of
thread throttling so I've gone with `SWITCH_RATIO=8`.
 2023-04-17 15:42:55 +01:00
Chris Sidebottom
32f2fafde7 Propagate SWITCH_RATIO to DYNAMIC_ARCH builds 
Previously dynamic builds were either using the default SWITCH_RATIO
or one from the higher level architecture; this patch ensures the
dynamic builds can use this parameter as well.
 2023-04-17 15:34:12 +01:00
Martin Kroeker
a5e1fdd525 Merge pull request #4007 from Mousius/update-contributors 
Add Chris Sidebottom to CONTRIBUTORS.md
 2023-04-17 15:45:39 +02:00
Martin Kroeker
44164e3a3d revert "move alpha out of register 18" (out of PR scope, no SVE on Apple hw) 2023-04-17 14:23:13 +02:00
Chris Sidebottom
bfc20c2e97 Add Chris Sidebottom to CONTRIBUTORS.md 2023-04-17 11:53:31 +01:00
Martin Kroeker
a44422f0d5 Merge pull request #3983 from thrasibule/makeflags 
parallel build fixes
 2023-04-16 13:49:05 +02:00
Martin Kroeker
73e6fcb925 Merge pull request #4006 from martin-frbg/issue4005 
Fix ?GEMMT implementation
 2023-04-16 13:30:17 +02:00
Martin Kroeker
38d7a7b562 Fix ?GEMMT 2023-04-16 00:07:58 +02:00
Martin Kroeker
8be68fa7f4 move declaration of sca to really keep the compiler from throwing it out (for now) 2023-04-15 12:02:39 +02:00
Martin Kroeker
4eac244c9a Merge pull request #4004 from martin-frbg/ccheckif 
fix missing blank in c_check
 2023-04-14 22:57:18 +02:00
Martin Kroeker
970e611e00 fix missing blank in test 2023-04-14 19:42:34 +02:00
Martin Kroeker
f096a339e4 Use long value fields for cpu ident on OSX 2023-04-13 18:16:09 +02:00
Martin Kroeker
3727672a74 Improve workaround and keep compilers from optimizing it out 2023-04-13 18:07:52 +02:00