e3039fa7f6
Merge pull request #4351 from catap/cmake-old-macos
...
Use 64bit build on `CMAKE_SYSTEM_PROCESSOR=i386` on Darwin
2023-12-05 14:40:18 +01:00
4a12cf53ec
[RISC-V] Improve RVV kernel generator LMUL usage
...
The RVV kernel generation script uses the provided LMUL to increase the number of accumulator registers.
Since the effect of the LMUL is to group together the vector registers into larger ones, it actually should be used as a multiplier in the calculation of vlenmax.
At the moment, no matter what LMUL is provided, the generated kernels would only set the maximum number of vector elements equal to VLEN/SEW.
Commit changes the use of LMUL to properly adjust vlenmax. Note that an increase in LMUL results in a decrease in the number of effective vector registers.
2023-12-04 11:13:35 +00:00
e4586e81b8
[RISC-V] Add RISC-V Vector 128-bit target
...
Current RVV x280 target depends on vlen=512-bits for Level 3 operations.
Commit adds generic target that supports vlen=128-bits.
New target uses the same scalable kernels as x280 for Level 1&2 operations, and autogenerated kernels for Level 3 operations.
Functional correctness of Level 3 operations tested on vlen=128-bits using QEMU v8.1.1 for ctests and BLAS-Tester.
2023-12-04 11:02:18 +00:00
2381132ada
Darwin < 20: always write xerbla.c.o into archive
...
Write xerbla.c.o into archive regardless of timestamp by using ar -rs
instead of ar -ru.
2023-12-03 19:13:56 +01:00
89fa51d495
Revert 42b5e08 ("Allow weak linking on old macOS")
2023-12-03 19:06:49 +01:00
08fde5ebd2
Use 64bit build on `CMAKE_SYSTEM_PROCESSOR=i386` on Darwin
...
Here a bit tricky things.
A value `CMAKE_SYSTEM_PROCESSOR` is came from output of `uname -m` which
migth be 32bit with 64bit building applicaiton.
So, for that case use `CMAKE_SIZEOF_VOID_P` to detect the target.
See https://trac.macports.org/ticket/68488
2023-11-30 21:24:58 +00:00
39bf8ece20
Merge pull request #4340 from yinshiyou/la-dev
...
Add some refines and optimizations for LoongArch.
2023-11-29 08:22:25 +01:00
42b5e081d8
Merge pull request #4348 from catap/macos-undefinded-dynamic-lookup
...
Allow weak linking on old macOS
2023-11-28 22:14:53 +01:00
a1562e4bae
Allow weak linking on old macOS
2023-11-28 14:04:01 +00:00
c4a622db9e
Merge pull request #4346 from martin-frbg/issue4343
...
Fix CMAKE installation location of lapacke_mangling header
2023-11-28 14:01:14 +01:00
9fe07d82fd
loongarch: Add LSX optimization for dot.
2023-11-28 20:24:18 +08:00
13b8c44b44
loongarch: Add optimization for dsdot kernel.
2023-11-28 20:24:16 +08:00
3def6a8143
loongarch: Add LASX optimization for dot.
2023-11-28 20:24:14 +08:00
1310a0931b
loongarch: Refine build control for loongarch64.
...
1. Use getauxval instead of cpucfg to test hardware capability.
2. Remove unnecessary code and option for compiler check in c_check.
2023-11-28 20:23:55 +08:00
ff92e6e707
Fix installation location of lapacke_mangling header
2023-11-28 12:53:35 +01:00
b7a28f5e42
Merge pull request #4344 from catap/macos-always-use-ar
...
Enable overstep of too long args without DYNAMIC_ARCH
2023-11-28 12:39:45 +01:00
9beee55167
Enable overstep of too long args without DYNAMIC_ARCH
2023-11-27 23:41:56 +00:00
01c7010543
cmake/openblas.pc.in: fixed version and URL
2023-11-27 14:51:58 +00:00
fc66ecd25a
Merge pull request #4339 from martin-frbg/lapack-3-12-0
...
Update version number and documentation of Reference-LAPACK to 3.12.0
2023-11-25 23:54:05 +01:00
08be9004f8
Update version number and copyright date to Reference-LAPACK 3.12.0
2023-11-25 18:57:17 +01:00
578f0f9590
Update version number to 3.12.0
2023-11-25 18:53:16 +01:00
3d9e20f614
Update version to 3.12.0
2023-11-25 18:51:54 +01:00
f7351e493c
Update Reference-LAPACK docs to 3.12.0
2023-11-25 18:49:34 +01:00
be8661ba40
Merge pull request #4338 from martin-frbg/lapack941
...
Docu fix for Truncated QR With Pivoting (Reference-LAPACK PR 941)
2023-11-25 18:41:25 +01:00
ca5a87ff1d
Small documentation fix for Truncated QR With Pivoting (Reference-LAPACK PR 941)
2023-11-25 15:31:18 +01:00
f745f02f35
benchmark: Fix missing colons in outputs of ./strsv.goto
2023-11-24 14:55:18 +08:00
97d3c9b827
Merge pull request #4336 from martin-frbg/fix4322
...
Revert unintentional change to gmake linking rule in LAPACK TESTING/LIN
2023-11-22 22:44:21 +01:00
c883abf838
Revert unintentional change to linking rule from PR 4322
2023-11-22 22:41:53 +01:00
8138999cd0
Merge pull request #4333 from codeworm96/update_dynamic_core_readme
...
Update the list of default dynamic targets for x86_64 in the README to be consistent with the Makefile
2023-11-21 13:50:37 +01:00
a938e48fa2
Merge pull request #4334 from RajalakshmiSR/Makefile_power
...
POWER: Fixing Makefile error
2023-11-21 10:24:25 +01:00
47da601a2d
POWER: Fixing Makefile error
...
Recent commit d99aad8ee3
2023-11-20 17:24:22 -06:00
54be8f4d67
Update the list of default dynamic targets for x86_64 in the README to be consistent with the Makefile
...
Signed-off-by: Yuning Zhang <codeworm96@outlook.com>
2023-11-20 13:28:25 -08:00
d526c4306f
Merge pull request #4329 from isuruf/sbgemm
...
Fix building test_sbgemm
2023-11-20 15:30:02 +01:00
2ea65bacd0
Merge pull request #4330 from bartoldeman/asum-init-mask
...
Use _mm_set1_epi{32,64x} to init mask in x86-64 [cz]asum
2023-11-20 05:38:39 +01:00
c34e2cf380
Use _mm_set1_epi{32,64x} to init mask in x86-64 [cz]asum
...
for skylake kernels. This is the same method as used in [sd]asum.
_mm_set1_epi64x was commented out for zasum, but has the advantage
of avoiding possible undefined behaviour (using an uninitialized
variable), optimized out by NVHPC and icx. The new code works
fine with those compilers.
For GCC 12.3 the generated code is identical; no matter what method
you use, the compiler optimizes the code into a compile-time
constant, there is no performance benefit using mm_cmpeq_epi8
since the corresponding instruction (VPCMPEQB) isn't actually
generated!
2023-11-19 21:28:35 +00:00
864c65b526
Merge pull request #4328 from martin-frbg/4239-3
...
Copy XCode15-specific workaround for Apple M to Fortran flags to fix build of tests
2023-11-19 22:06:17 +01:00
6b2651ece3
Fix building test_sbgemm
2023-11-19 02:57:13 -06:00
22aa401656
Temporarily disable the AVX512 CASUM/ZASUM microkernels for any version of NVIDIA HPC ( #4327 )
...
* Temporarily disable the C/ZASUM microkernels for any version of NVHPC
2023-11-19 00:04:31 +01:00
47b03fd4b4
Copy XCode15-specific workaround to Fortran flags to fix build of tests
2023-11-18 23:45:02 +01:00
df4cd7e82c
Merge pull request #4326 from bartoldeman/fix-casum-backup-kernel
...
Fix casum fallback kernel for x86_64
2023-11-18 19:06:06 +01:00
f8ad5344c2
Fix casum fallback kernel.
...
This kernel is only used on Skylake+ if the kernel with AVX512
intrinsics can't be used, but used the variable x1 incorrectly
in the tail end of the loop, as it is still at the initial
value instead of where x points to.
This caused 55 "other error"s in the LAPACK tests
(https://github.com/OpenMathLib/OpenBLAS/issues/4282 )
This change makes casum.c as similar as possible as zasum.c,
because zasum.c does this correctly.
2023-11-17 23:53:56 +00:00
cb2950709f
Merge pull request #4322 from martin-frbg/lapack891
...
Add truncated QR with pivoting (Reference-LAPACK PR 891)
2023-11-16 08:40:17 +01:00
cc622f2406
restore OpenBLAS-specific target_link_libraries
2023-11-15 22:51:09 +01:00
ee47e4e494
run m1/llvm/cmake buid on all 4 cores
2023-11-15 15:21:32 +01:00
8b2a956890
Implement truncated QR with pivot (Reference-LAPACK PR 891)
2023-11-15 14:20:12 +01:00
20a2a83f49
Implement truncated QR with pivoting (Reference-LAPACK PR 891)
2023-11-15 12:18:15 +01:00
f437339130
Implement truncated QR with pivoting (Reference-LAPACK PR 891)
2023-11-15 12:12:26 +01:00
5bf87c86f5
Implement truncated QR with pivoting (Reference-LAPACK PR 891)
2023-11-15 12:10:20 +01:00
0eb8a87977
Implement truncated QR with pivoting (Reference-LAPACK PR 891)
2023-11-15 09:56:37 +01:00
387830b9d5
Implement truncated QR with pivoting (Reference-LAPACK PR 891)
2023-11-15 09:53:06 +01:00
Martin Kroeker
Octavian Maghiar
Erik Bråthen Solem
Kirill A. Korinsky
Shiyou Yin
Rajalakshmi Srinivasaraghavan
Yuning Zhang
Bart Oldeman
Isuru Fernando