Commit Graph

8616 Commits

Author SHA1 Message Date
Martin Kroeker 89c7bbcba6
add cblas_?gemm_batch 2024-05-29 15:47:02 +02:00
Martin Kroeker 103637887e
add cblas_?gemm_batch 2024-05-29 15:46:10 +02:00
Martin Kroeker 0073affe63
Merge pull request #4693 from goplanid/locks-improvement
Lock Management Improvements for Memory Allocation Efficiency
2024-05-26 23:14:52 +02:00
Martin Kroeker 834e633d79
Merge pull request #4718 from martin-frbg/issue4713
Override Intel icx's default fp-model to ensure correct handling on NaNs
2024-05-26 16:38:18 +02:00
Martin Kroeker 3833190454
Merge pull request #4716 from martin-frbg/lapack1018
Fix a potential bounds error in ?UNHR_COL/?ORHR_COL (Reference-LAPACK PR 1018)
2024-05-26 14:01:31 +02:00
Martin Kroeker cf7e668fe8
Merge pull request #4709 from martin-frbg/docsbuildbranch
Don't try to deploy docs when PR-building in a fork
2024-05-26 14:01:05 +02:00
Martin Kroeker 8b4996a2d5
Override icx's default fast math mode to ensure correct NaN handling 2024-05-26 13:16:03 +02:00
Martin Kroeker 616cc28d82
Override icx's default fast math mode to ensure correct NaN handling 2024-05-26 12:59:11 +02:00
Martin Kroeker 772116879d
Merge pull request #4717 from bartoldeman/zscal-float-inf-fix
Replace use of FLT_MAX in x86_64 zscal.c by isinf()
2024-05-26 12:26:41 +02:00
Bart Oldeman 62f7b244ff Replace use of FLT_MAX in x86_64 zscal.c by isinf()
Commit def4996 fixed issues with inf and nan values in zscal,
but used FLT_MAX, where DBL_MAX or isinf() is more appropriate,
as FLT_MAX is for single precision only.
Using FLT_MAX caused test case failures in the LAPACK tests.

isinf() is consistent with the later fix 969601a1
2024-05-24 17:20:27 +00:00
Martin Kroeker 7ebbe3cc72
Fix potential bounds error (Reference-LAPACK PR 1018) 2024-05-23 23:12:19 +02:00
Martin Kroeker 791e015024
Fix potential bounds error (Reference-LAPACK PR 1018) 2024-05-23 23:11:14 +02:00
Martin Kroeker 4dd715d220
Fix potential bounds error (Reference-LAPACK PR 1018) 2024-05-23 23:09:55 +02:00
Martin Kroeker e2c1a1e269
Fix potential bounds error (Reference-LAPACK PR 1018) 2024-05-23 23:08:27 +02:00
Rajalakshmi Srinivasaraghavan e112191b54 POWER: Fix issues in zscal to address lapack failures
This patch fixes following lapack failures with clang compiler on POWER.
zed.out: ZVX:   18 out of  5190 tests failed to pass the threshold
zgd.out: ZGV drivers:     25 out of   1092 tests failed to pass the threshold
zgd.out: ZGV drivers:      6 out of   1092 tests failed to pass the threshold
2024-05-22 08:00:06 -05:00
Martin Kroeker 172d91846f
Don't try to deploy docs in a fork 2024-05-20 22:53:43 +02:00
Martin Kroeker 700ea74a37
Merge pull request #4705 from martin-frbg/issue4703
Fix INTERFACE64 builds on Loongarch64
2024-05-18 21:38:22 +02:00
Martin Kroeker aa259b141d
Merge pull request #4704 from amritahs-ibm/saxpy_perf_fix
Fix regression SAXPY when compiler with OpenXL compiler.
2024-05-18 19:11:25 +02:00
Martin Kroeker 25b34e67f9
Merge pull request #4678 from ev-br/codspeed
WIP: add codspeed benchmarks [skip cirrus]
2024-05-18 16:51:02 +02:00
Martin Kroeker 6494f432df
Fix INTERFACE64 builds on Loongarch64 2024-05-18 16:49:03 +02:00
Evgeni Burovski 81cf0db047 DOC: add a readme for benchmarks/pybench 2024-05-18 15:30:00 +03:00
Evgeni Burovski 9f28161837 BENCH: add benchmarks using codspeed.io 2024-05-18 15:25:16 +03:00
Martin Kroeker 5015548d18
Merge pull request #4700 from martin-frbg/fix4698
Remove spurious brace in cmake/system.cmake
2024-05-16 15:38:01 +02:00
Matthias Langer 0050a9660b Correctly detect ARM Neoverse V2 CPUs. 2024-05-16 09:59:52 +00:00
Martin Kroeker ce96e0e50f
Merge pull request #4699 from ChipKerchner/fixSwapVectorOrder
POWER: Fixing endianness issue in cswap/zswap kernel for AIX
2024-05-16 09:28:20 +02:00
Martin Kroeker a3f6b13bc9
remove spurious brace 2024-05-16 09:25:53 +02:00
Chip Kerchner 3a1417671a POWER: Fixing endianness issue in cswap/zswap kernel for AIX 2024-05-15 19:36:46 -05:00
Martin Kroeker 668f48f4fc
Use CMAKE_C_COMPILER_VERSION instead of dumpversion calls (#4698)
* Use CMAKE_C_COMPILER_VERSION throughout
2024-05-15 23:58:14 +02:00
Martin Kroeker 39c96063fb
Merge pull request #4694 from martin-frbg/issue3660
Add a minimum problem size for multithreading in GBMV
2024-05-15 22:14:41 +02:00
Martin Kroeker f5c080f083
Fix CMAKE syntax in kernel file parsing of IFNEQ conditionals (#4695)
* Fix syntax in parsing of IFNEQ
2024-05-15 20:58:31 +02:00
Martin Kroeker 9a2a6a2e52
Merge pull request #4696 from frjohnst/restore_second
Revert PRs 4515 and 4520 (restore second, dsecnd)
2024-05-15 18:35:20 +02:00
frjohnst 87026ac1b1 Revert "fix conlict between PR 4515 and AIX shared obj support"
This reverts commit bdaa6705ca.

It turns out that PRs 4515 and 4520 break the tests under
lapack-netlib/TESTING which require SECOND and DSECND. IBM
has decided this is a bigger biger problem than the conflict
between lapack second_ and the xlf run time.
2024-05-15 09:45:17 -04:00
frjohnst 56d3d1039c Revert "resolve second_ conflict which breaks xlf timef"
This reverts commit 9b24b31419.

It turns out that PRs 4515 and 4520 break the tests under
lapack-netlib/TESTING which require SECOND and DSECND. IBM
has decided this is a bigger biger problem than the conflict
between lapack second_ and the xlf run time.
2024-05-15 09:44:29 -04:00
Martin Kroeker 2957281275
Introduce a lower limit for multithreading 2024-05-14 18:59:21 +02:00
Martin Kroeker 5fd871d7ea
Introduce a lower limit for multithreading 2024-05-14 18:48:03 +02:00
Martin Kroeker 6ca9ffa7f5
Merge pull request #4655 from yamazakimitsufumi/update_2d_thread_distribution
Expanding the scope of 2D thread distribution to improve multi-threaded DGEMM performance
2024-05-14 18:12:43 +02:00
Deeksha Goplani 0dc80a5c8d locks improvement 2024-05-13 22:17:23 +05:30
Martin Kroeker b45a78c6e9
fix zdotu argument passing in utest_ext on windows (#4691)
* fix passing of results on windows
2024-05-13 14:50:50 +02:00
Martin Kroeker 1ab9f50561
Merge pull request #4690 from mattip/blasint
use blasint instead of int to quiet warnings
2024-05-13 11:00:33 +02:00
Amrita H S 87b3d9054f Fix regression SAXPY when compiler with OpenXL compiler.
SAXPY built with OpenXL regresses when compared to SAXPY
built with gcc. OpenXL compiler doesn't know that the
SAXPY inner kernel assembly is a 64 element loop and
to it the remainder loop is the main loop. It vectorizes
and interleaves the remainder to be a 48 elements per
iteration loop. With a max of 63 iterations, a 48 element
loop is mostly not going to get executed, so the 1 element
scalar loop that is the remainder after that is probably
mostly what gets executed.

This can be fixed by adding a pragma, loop interleave_count(2)
which will result in 8 element loop.

Signed-off-by: Amrita H S <amritahs@linux.vnet.ibm.com>
2024-05-12 23:27:55 -05:00
Matti Picus 243640c354 use blasint instead of int to quiet warnings 2024-05-12 10:24:16 +03:00
Martin Kroeker f0560f906f
Merge pull request #4689 from martin-frbg/issue4684
Fix compilation of the BLAS extension utests for NO_CBLAS=1
2024-05-11 14:39:54 +02:00
Martin Kroeker e1e0d9a2ae
Merge pull request #4688 from XiWeiGu/loongarch64_fixed_gcc14_compilation
loongarch64: Fixed GCC14 compilation issue
2024-05-11 13:38:45 +02:00
Martin Kroeker d8baf2f2ea
Support compilation without CBLAS 2024-05-11 13:10:54 +02:00
Martin Kroeker a6c184d150
forward NO_CFLAGS to the CFLAGS, if set 2024-05-11 13:07:30 +02:00
gxw ecf8b588a9 loongarch64: Fixed GCC14 compilation issue 2024-05-11 16:14:18 +08:00
Martin Kroeker 8da6f7e5f2
Merge pull request #4686 from XiWeiGu/loongarch64_dgemm_kernel_16x6
Loongarch64: Improving the Performance and Stability of dgemm
2024-05-10 11:29:12 +02:00
gxw f9a26240a7 loongarch64: Fixed icamax_lsx 2024-05-10 14:16:40 +08:00
gxw cb0f707409 loongarch64: Fixed utest fork:safety 2024-05-10 14:16:36 +08:00
gxw 637c650f4f loongarch64: Add buffer offset for target LOONGSON3R5 2024-05-10 11:42:53 +08:00