60005eb47b
Don't overwrite blas_thread_buffer if already set
...
After a fork it is possible that blas_thread_buffer has already
allocated memory buffers: goto_set_num_threads does allocate those
already and it may be called by num_cpu_avail in case the OpenBLAS
NUM_THREADS differ from the OMP num threads.
This leads to a memory leak which can cause subsequent execution of BLAS
kernels to fail.
Fixes #2993
2020-11-19 14:51:51 +01:00
7e9cb39a25
Merge pull request #2981 from Qiyu8/fix-sum
...
Fix sum optimize issues
2020-11-16 08:40:46 +01:00
be075d53cf
Merge pull request #2983 from Qiyu8/optimize-srot
...
Optimize the performance of rot by using universal intrinsics
2020-11-16 08:38:37 +01:00
b00a0de132
remove the -mfma flag in when the host has AVX.
2020-11-16 09:14:56 +08:00
d341a0fea0
Merge pull request #2989 from martin-frbg/cmake-fma
...
Fix missing -mfma compiler flag in cmake builds without DYNAMIC_ARCH
2020-11-13 12:35:09 +01:00
ec4d77c47c
Add -mfma for HAVE_FMA3 in the non-DYNAMIC_ARCH case as well
2020-11-13 09:16:34 +01:00
02699226d0
Merge pull request #111 from xianyi/develop
...
rebase
2020-11-13 09:14:23 +01:00
ae0b1dea19
modify system.cmake to enable fma flag
2020-11-13 10:20:24 +08:00
e0dac6b53b
fix the CI failure of target specific option mismatch
2020-11-12 20:31:03 +08:00
e5c2ceb675
fix the CI failure of lack the head
2020-11-12 17:35:17 +08:00
a87e537b8c
modify macro
2020-11-11 15:53:48 +08:00
5bc0a7583f
only FMA3 and vector larger than 128 have positive effects.
2020-11-11 15:18:01 +08:00
8c0b206d4c
Optimize the performance of rot by using universal intrinsics
2020-11-11 14:33:12 +08:00
c4c591ac5a
fix sum optimize issues
2020-11-10 16:16:38 +08:00
ff16329cb7
Merge pull request #2972 from xiegengxin/rot-intrinsic
...
Improve the performance of rot by using AVX512 and AVX2 intrinsic
2020-11-08 22:43:00 +01:00
433637ccd8
Merge pull request #2980 from martin-frbg/fixgetarch
...
Fix missing AVX2 and FMA3 capabilities in FORCE_target mode
2020-11-08 17:39:05 +01:00
ec088bf33a
Fix missing AVX2 and FMA3 capabilities in FORCE_target mode
2020-11-08 13:15:40 +01:00
110c7a6de0
Merge pull request #2979 from RajalakshmiSR/dot_power10
...
Optimize sdot/ddot for POWER10
2020-11-08 10:19:34 +01:00
d2faa1be4e
Merge pull request #2978 from martin-frbg/fixdynfeatures
...
Fix handling of cpu capability flags in DYNAMIC_ARCH builds
2020-11-08 10:19:17 +01:00
1c4cfdc139
Stay compatible with old gmake that did not support undefine
2020-11-08 00:12:55 +01:00
f6a57d8f63
Update Makefile.system
2020-11-08 00:01:36 +01:00
f4b7ba12b7
Update Makefile.system
2020-11-07 23:37:21 +01:00
6e364981a8
Optimize sdot/ddot for POWER10
...
This patch makes use of new POWER10 vector pair instructions for
loads and stores.
2020-11-07 15:21:58 -06:00
b976a0bf40
Remove previous workaround for compiler flags related to cpu capabilities in x86_64 DYNAMIC_ARCH builds
2020-11-07 20:39:56 +01:00
a04f532edf
Reset cpu property flags between build cycles in DYNAMIC_ARCH mode
2020-11-07 20:37:03 +01:00
ccb9731c7b
Fix propagation of cpu properties to compiler options
2020-11-07 20:30:15 +01:00
a29338aaa6
Remove extraneous quotes that caused a cmake policy warning
2020-11-07 20:27:42 +01:00
438a8e5624
Fix placement of getarch call and spurious cpu property accumulation in DYNAMIC_ARCH builds
2020-11-07 20:26:12 +01:00
e5967810b7
Merge pull request #110 from xianyi/develop
...
rebase
2020-11-07 20:22:41 +01:00
ff74319ea5
Merge pull request #2977 from martin-frbg/issue2976
...
Fix macro name used in ifdef for POWERPC/PGI
2020-11-07 14:41:34 +01:00
28d2dfe2b3
Fix macro name used in ifdef
2020-11-07 12:17:49 +01:00
725ffbf041
fix typo
2020-11-05 16:25:17 +08:00
d9ba49165a
Improve the performance of rot by using AVX512 and AVX2 intrinsic
2020-11-05 15:12:36 +08:00
60ab9c783f
Merge pull request #2966 from martin-frbg/issue2964
...
Ensure that EXPRECISION is disabled for DYNAMIC_ARCH with TARGET=GENERIC and fix CMAKE DYNAMIC_ARCH builds
2020-11-04 16:02:46 +01:00
8cc73fee98
Export NO_EXPRECISION after overriding for DYNAMIC_ARCH with GENERIC target
2020-11-03 23:47:04 +01:00
0155cd53a3
Add -msse3 where needed for DYNAMIC_ARCH builds
2020-11-03 23:45:49 +01:00
a9f9354296
Fix target test
2020-11-02 23:17:46 +01:00
b9bc76aec4
Add files via upload
2020-11-02 22:43:50 +01:00
f071245939
Merge pull request #2967 from RajalakshmiSR/dgemm88
...
POWER10: Change dgemm unroll factors
2020-11-02 18:54:36 +01:00
e5f8c2bf8a
typo fix
2020-11-01 22:25:43 +01:00
6baf8af658
Disable EXPRECISION for the combination of DYNAMIC_CORE and GENERIC target
2020-11-01 22:11:48 +01:00
40a93c232b
Disable EXPRECISION for DYNAMIC_ARCH in combination with TARGET=GENERIC
...
NO_EXPRECISION is disabled for the GENERIC_TARGET already, so prevent mixing with code parts that use a different float size by default
2020-11-01 21:58:26 +01:00
fab952bee4
Merge pull request #2962 from brada4/develop
...
add openbsd 68+ gfortran name
2020-11-01 14:24:40 +01:00
1cf04a6f0e
Merge pull request #2963 from martin-frbg/issue2959
...
Reunify default BUFFER_SIZE on ARM64 to avoid crashes in DYNAMIC_ARCH mode
2020-11-01 09:14:54 +01:00
dd7a9cc5bf
POWER10: Change dgemm unroll factors
...
Changing the unroll factors for dgemm to 8 shows improved performance with
POWER10 MMA feature. Also made some minor changes in sgemm for edge cases.
2020-10-31 18:28:57 -05:00
7f26be4802
Reunify BUFFERSIZE across arm64 platforms to avoid segfaults in DYNAMIC_ARCH
2020-11-01 00:00:43 +01:00
9fab65e90a
add openbsd gfortran
2020-11-01 00:38:08 +02:00
9efc3f0815
Merge pull request #109 from xianyi/develop
...
rebase
2020-10-31 22:33:52 +01:00
aa21cb5217
Merge pull request #2960 from thrasibule/avx2_detection
...
fix avx2 detection
2020-10-31 20:24:21 +01:00
1f564d729b
fix avx2 detection
...
reword commits to make it clearer
2020-10-31 10:00:48 -04:00
Alexander Grund
Martin Kroeker
Qiyu8
Rajalakshmi Srinivasaraghavan
Gengxin Xie
User User-User
Guillaume Horel