OpenMathLib/OpenBLAS
1
0
Fork 0
mirror of https://github.com/OpenMathLib/OpenBLAS synced 2026-06-15 07:51:43 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity
9,104 Commits 57 Branches 69 Tags
edaf51dd99bb979f15fa4f2774ba068cfec0c09e
Commit Graph

9104 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Annop Wongwathanarat
edaf51dd99 Add sbgemv_t_bfdot kernel for ARM64 
This improves performance for sbgemv_t by up to 100x on NEOVERSEV1.
The geometric mean speedup is ~61x for M=N=[2,512].
 2025-02-28 12:31:50 +00:00
Martin Kroeker
ef9e3f7159 Merge pull request #5149 from martin-frbg/fixup5077-5088 
Make the Neoverse GEMM/GEMV throttling code conditional on SMP
 2025-02-25 14:01:13 +01:00
Martin Kroeker
09ba099461 make throttling code conditional on SMP 2025-02-25 12:10:48 +01:00
Martin Kroeker
1533fe49be Merge pull request #5144 from taoye9/dispatch_neoversve2_to_neoversven2 
dispatch NEOVERSEV2 to NEOVERSEN2 under dynamic setting
 2025-02-24 16:07:06 +01:00
Martin Kroeker
c03a81b927 Merge pull request #5141 from michalowski-arm/fork-throttle 
Add throttling profile for SGEMM and SGEMV on `NEOVERSEV2`
 2025-02-23 12:16:09 +01:00
Martin Kroeker
643966d9c7 Merge pull request #5146 from martin-frbg/issue5123 
Fix "dummy2" flag reading in PPC970 S/DSCAL
 2025-02-22 21:57:09 +01:00
Martin Kroeker
77fba0f400 Fix "dummy2" flag handling 2025-02-22 20:09:21 +01:00
Ye Tao
f0bea79a6e dispatch NEOVERSEV2 to NEOVERSEN2 under dynamic setting 2025-02-21 10:30:11 +00:00
Martin Kroeker
20d1118865 Merge pull request #5143 from martin-frbg/issue5111 
Fix GEMMT transforming the input array B in some complex cases
 2025-02-21 09:20:39 +01:00
Martin Kroeker
75b958a018 Transform the B array back if necessary before returning 2025-02-20 23:54:12 +01:00
Marek Michalowski
650a062e19 Add thread throttling profile for SGEMV on NEOVERSEV2 2025-02-20 10:28:31 +00:00
Marek Michalowski
b723c1b7b7 Add thread throttling profile for SGEMM on NEOVERSEV2 2025-02-20 10:28:21 +00:00
Martin Kroeker
ceb8f1e34b Merge pull request #5140 from martin-frbg/issue5139 
Add ARM64 options for NVIDIA HPC
 2025-02-19 18:17:15 +01:00
Martin Kroeker
f1fa370579 fix missing endif 2025-02-19 15:22:26 +01:00
Martin Kroeker
6d1444be3a Add ARM64 options for NVIDIA HPC 2025-02-19 14:26:43 +01:00
Martin Kroeker
eb84aac7ad Merge pull request #5084 from quic/topic/sgemm_direct_sme1 
Support for SGEMM_DIRECT Kernel based on SME1
 2025-02-19 10:56:49 +01:00
Martin Kroeker
abbd78aa59 Merge pull request #5138 from martin-frbg/issue5131 
Ensure that gmake builds with flang-new link the flang runtime into the shared library
 2025-02-18 09:53:31 +01:00
Martin Kroeker
ebcab90976 Handle flang-new runtime library linking on Linux like classic-flang 2025-02-17 23:12:58 +01:00
Martin Kroeker
ed1584666c Merge pull request #5137 from martin-frbg/issue5136 
Fix the CMake build to define USE_TRMM for RISCV64 targets as well
 2025-02-17 07:37:07 +01:00
Martin Kroeker
b9ae246f20 define USE_TRMM for RISCV64 targets as well 2025-02-16 23:18:04 +01:00
Martin Kroeker
86cf9d8a2e Merge pull request #5133 from OpenMathLib/revert-4920-issue4917 
Revert "Fix potential inaccuracy in multithreaded level3 related to SWITCH_RATIO"
 2025-02-16 19:16:43 +01:00
Martin Kroeker
0b3c56968d Merge pull request #5135 from martin-frbg/ghwf-n2 
CI: remove the express NeoverseN2 target from the Cobalt100 job in the gh workflow
 2025-02-16 19:16:10 +01:00
Martin Kroeker
c1bb90a823 remove the express NeoverseN2 target from the Cobalt100 job 2025-02-16 14:23:07 +01:00
Martin Kroeker
77c638db67 Revert "Fix potential inaccuracy in multithreaded level3 related to SWITCH_RATIO" 2025-02-15 20:37:48 +01:00
Vaisakh K V
f66ca05b31 Merge branch 'develop' into topic/sgemm_direct_sme1 2025-02-13 14:54:37 +05:30
Vaisakh K V
d23eb3b93e Support for SME1 based sgemm_direct kernel for cblas_sgemm level 3 API 
* Added ARMV9SME target
* Added SGEMM_DIRECT kernel based on SME1
 2025-02-13 14:51:21 +05:30
Martin Kroeker
a64b75a2e0 Merge pull request #5127 from Harishmcw/gesv-threshold 
Refined GESV Parallelization Logic for Windows on ARM64
 2025-02-12 22:02:37 +01:00
Martin Kroeker
453efbd103 Merge pull request #5128 from martin-frbg/issue5120 
Add -O2 to flang flags when building on WoA in Release mode
 2025-02-12 21:02:06 +01:00
Martin Kroeker
877d5a5be6 Add -O2 to flang flags when building on WoA in Release mode 2025-02-12 17:01:06 +01:00
Martin Kroeker
8d487ef6eb Merge pull request #5124 from XiWeiGu/LoongArch64-LA264-lapack-fixed 
LoongArch64: Fixed lapack test for LA264
 2025-02-12 14:58:30 +01:00
Harish-Gits
daf16b8229 Adjusted GESV threading logic for optimal performance on WoA 2025-02-12 19:25:25 +05:30
Martin Kroeker
e8b11a126b Merge pull request #5125 from martin-frbg/issue5122 
Fix SGEMV on POWER8 by reverting to the non-vectorized earlier code
 2025-02-12 12:50:44 +01:00
Martin Kroeker
9a3948df82 Merge pull request #5126 from martin-frbg/cirrusbsd4 
CirrusCI: Update FreeBSD jobs to 14.2
 2025-02-12 12:50:21 +01:00
Martin Kroeker
7f1f776f58 Update FreeBSD jobs to 14.2 2025-02-12 11:23:02 +01:00
Martin Kroeker
81eed868b6 Restore the non-vectorized code from before PR4880 for POWER8 2025-02-12 09:07:20 +01:00
Martin Kroeker
98b5ef929c Restore the non-vectorized code from before PR4880 for POWER8 2025-02-12 09:04:22 +01:00
gxw
2c4a5cc6e6 LoongArch64: Fixed snrm2_lsx.S and cnrm2_lsx.S 
When the data type is single-precision real or single-precision complex,
converting it to double precision does not prevent overflow (as exposed in LAPACK tests).
The only solution is to follow C's approach: find the maximum value in the
array and divide each element by that maximum to avoid this issue
 2025-02-12 15:48:01 +08:00
gxw
9e75d6b3d1 LoongArch64: Fixed swap_lsx.S 
Fixed the error when the stride is zero
 2025-02-12 14:57:35 +08:00
gxw
e8c740368c LoongArch64: Fixed rot_lsx.S ane crot_lsx.S 
Do not check whether the input parameters c and s are zero,
as this may cause errors with special values (same as scal).
Although OpenBLAS's own test suite doesn't catch this, it will
cause LAPACK test cases to fail.
 2025-02-12 14:52:49 +08:00
Hao Chen
c2212d0abd LoongArch64: Fixed copy_lsx.S 
Fixed incorrect store operation

Signed-off-by: gxw <guxiwei-hf@loongson.cn>
 2025-02-12 14:52:20 +08:00
Hao Chen
7f1ebc7ae6 LoongArch64: Fixed iamax_lsx.S 
Fixed index retrieval issue when there are
identical maximum absolute values

Signed-off-by: Hao Chen <chenhao@loongson.cn>
Signed-off-by: gxw <guxiwei-hf@loongson.cn>
 2025-02-12 14:44:44 +08:00
Hao Chen
31d326f895 LoongArch64: Fixed dot_lsx.S 
Fixed incorrect register usage in instructions

Signed-off-by: gxw <guxiwei-hf@loongson.cn>
 2025-02-12 14:44:11 +08:00
Hao Chen
5d6356bc16 LoongArch64: Fixed amax_lsx.S 
Fixed register zeroing operation

Signed-off-by: Hao Chen <chenhao@loongson.cn>
Signed-off-by: gxw <guxiwei-hf@loongson.cn>
 2025-02-12 14:39:29 +08:00
Martin Kroeker
f42ce7067f Merge pull request #5116 from martin-frbg/issue5110 
Handle INCX=0 in ?NRM2
 2025-02-09 23:17:20 +01:00
Martin Kroeker
7478c10268 Merge branch 'OpenMathLib:develop' into issue5110 2025-02-09 21:40:02 +01:00
Martin Kroeker
c54f5417cc Merge pull request #5118 from martin-frbg/zrot_utestext 
Disable extended utests for CSROT/ZDROT that invoke undefined behavior
 2025-02-09 21:39:30 +01:00
Martin Kroeker
57208b8bce Disable tests with incx,incy=0 (undefined behavior) 2025-02-09 20:17:29 +01:00
Martin Kroeker
3a4a9b21eb Disable tests with incx,incy=0 (undefined behavior) 2025-02-09 20:16:03 +01:00
Martin Kroeker
60d0be0e97 Update nrm2.c 2025-02-08 23:42:21 +01:00
Martin Kroeker
0fd5448b2c Handle INCX=0 2025-02-08 19:33:05 +01:00