36bd3eeddf
Vectorize BF16 GEMV (VSX & MMA). Use GEMM_GEMV_FORWARD_BF16 (for Power).
2024-10-13 13:46:11 -05:00
722e4ae07a
MAINT: Explicitly replace instead of unknown
2024-10-05 16:23:02 +00:00
a6b7751881
BUG: Allow tests to be run multiple times
...
Without failures due to existing files
2024-10-05 16:22:56 +00:00
89702e1f4a
Fix zero element GEMV test.
2024-08-16 11:37:39 -05:00
77f85c7c00
GEMV tests don't like zero elements.
2024-08-16 11:15:32 -05:00
868aa857bc
Change malloc zero to return one byte and update the SBGEMM test to again use sizes of zero.
2024-08-16 10:28:10 -05:00
b1802f4dc8
Fix unit test to start at 1 instead of 0 - since malloc zero bytes fails on some systems.
2024-08-16 09:51:37 -05:00
c23897f585
Add GEMV testing to SBGEMx vs SGEMx testing.
2024-08-14 15:55:23 -05:00
6452f7b46d
Merge pull request #4873 from ChipKerchner/fixSBGEMMDefaults
...
[POWER] Problem with multi-threaded SBGEMM
2024-08-14 19:22:03 +02:00
31226740d6
Cleanup of SBGEMM unit test.
2024-08-14 08:10:25 -05:00
20bdb65882
Fix recursive variable expansion in Makefiles for LOONGSON3A
2024-08-14 15:08:32 +08:00
b1737698db
Fix DEFAULTS in SBGEMM for POWER10. Also comparisons for SBGEMM unit test can be exactly due to epilison differences.
2024-08-13 07:01:21 -05:00
76db713e79
fix invocation of GEMM3M tests
2024-08-07 21:37:20 +02:00
56e1782ffb
Add another missing parenthesis
2024-07-15 15:15:23 -07:00
f708944fea
Add all 4 variations of the SBGEMM to compare_sgemm_sbgemm
2024-07-10 13:07:48 -05:00
edacf9b397
Work around spurious BLAS3 test errors on LOONGSON3R3/4 ( #4667 )
...
Force compilation with gfortran to use O0 on older Loongson hardware to avoid spurious test failures
2024-04-30 08:50:47 +02:00
28f151808e
Avoid overriding the global USE_GEMM3M
2024-02-26 21:01:46 +01:00
ba201c1939
Enable GEMM3M tests on supported platforms
2024-02-25 23:39:24 +01:00
4adfe4d531
Avoid linking both libgomp and libomp in mixed clang/gfortran builds
2024-02-22 22:16:01 +01:00
e9f480111e
fix sbgemm bfloat16 conversion errors introduced in PR 4488
2024-02-07 19:57:18 +01:00
fb99fc2e6e
fix type conversion warnings
2024-02-07 13:42:08 +01:00
61c8e19f95
Fix Makefile to support OpenMP on AIX for xlc (clang) with xlf.
2024-01-31 15:27:50 -06:00
6b2651ece3
Fix building test_sbgemm
2023-11-19 02:57:13 -06:00
d46eba06a7
Pack structure only on AIX.
2023-10-12 09:41:33 -05:00
e98e3c4783
Fix float32_bits union so that it always the sizeof float.
2023-10-11 18:05:55 -05:00
97a61d0577
Fix bfloat16_bits union so that it always the sizeof unsigned short.
2023-10-11 17:36:43 -05:00
2a9981a244
Add -lgomp when IBM xlf is combined with gcc in OPENMP builds
2023-09-24 10:19:11 +02:00
44e6e5479b
Use the C compiler for the C SBGEMM test source
2023-09-21 23:01:21 +02:00
b209915121
Fix build with clang
...
There are two instances when building the tests where OpenBLAS fails to
build with OpenMP and clang due to library paths getting reset as flags
are set rather than appended. This seems to only affect certain
clang/libomp installations, but if it's already grabbing the correct
library paths we might as well use them.
2023-07-28 12:59:44 -07:00
3d338b57de
remove spurious loops
2022-08-08 17:09:45 +02:00
d9dc015cfc
Use blasint for INTERFACE64 compatibility
2022-08-08 14:52:10 +02:00
1d97405c02
POWER: Enable bfloat16 kernels by default
...
This patch enables bfloat16 kernels by default for POWER processors.
Tested on Linux POWER8, POWER9, POWER10 and AIX POWER10 systems.
2022-07-28 07:43:53 -05:00
74dba0c90a
Fix missing braces from previous commit (PR3600)
2022-04-09 20:03:36 +02:00
bef3871833
Disable flang (over-)optimizations in BLAS tests ( #3600 )
...
* limit flang optimizations to -O2
2022-04-09 17:14:24 +02:00
145ccf8943
Prevent powershell from adding a BOM to test input ( #3595 )
...
* Prevent addition of a BOM to test input (which would distort the names of output files)
2022-04-07 14:24:19 +02:00
2408315d10
Skip tests if Windows powershell added a BOM
2022-03-22 21:37:55 +01:00
5dc6aa74f0
Disable gfortran tree vectorizer to avoid gcc11+ miscompilation at O3
2021-07-30 14:46:19 +02:00
b4f4ed378b
Disable gfortran tree vectorizer to avoid gcc11+ miscompilation at O3
2021-07-30 14:21:08 +02:00
3cfdb1770c
Remove code that disabled EXTRALIB on RISCV C910V
2021-07-06 20:21:07 +02:00
30d835168a
Merge pull request #3088 from xoviat/msvc
...
add misc fixes.
2021-03-14 17:14:28 +01:00
6ae7af78a3
Support compilation with nagfor
2021-03-11 11:53:51 +01:00
3165c915b6
fix test helpers
2021-01-27 15:24:49 -06:00
457ccc42c9
Merge branch 'develop' into msvc
2021-01-27 14:15:59 -06:00
2b114c3f30
Restore proper Makefile
2020-11-22 17:16:22 +01:00
65eb7afaf4
Change ifndef CROSS to ifneq
2020-11-22 16:25:36 +01:00
ef8e7d0279
Add the support for RISC-V Vector.
...
Change-Id: Iae7800a32f5af3903c330882cdf6f292d885f266
2020-10-15 16:09:02 +08:00
c1f4f5d4e7
Replace Makefile with simplified version again
2020-10-14 01:08:50 +02:00
bc5c7f9578
Cleanup
2020-10-13 19:56:09 +02:00
b5d30b390d
Fix build issues with bfloat16
...
This patch fixes compilation errors due to recent renaming from SH to SB
with BUILD_BFLOAT16.
2020-10-13 11:00:22 -05:00
5800758b43
Rename "HALF" and "sh" to "BFLOAT16" and "sb"
2020-10-11 23:44:38 +02:00
Chip Kerchner
Rohit Goswami
Martin Kroeker
Henry Chen
Vladimir Nikolić
Isuru Fernando
Aiden Grossman
Rajalakshmi Srinivasaraghavan
xoviat
xoviat
damonyu