0dbe61a612
s390x: choose SIMD kernels at run-time based on OS and compiler support
...
Extend and simplify the run-time detection for dynamic architecture support for z
to check HW_CAP and only use SIMD features if advertised by the OS.
While at it, also honor the env variable LD_HWCAP_MASK and do not use
the CPU features masked there.
Note that we can only use the SIMD features on z13 or newer (i.e.,
Vector Facility or Vector-Enhancements Facilities) when the operating
system supports properly context-switching the vector registers. The OS
advertises that support as a bit in the HW_CAP value in the auxiliary
vector. While all recent Linux kernels have that support, we should
maintain compatibility with older versions that may still be in use.
Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
2020-05-12 11:01:16 +02:00
62cf391cbb
s390x: only build kernels supported by gcc with dynamic arch support
...
When building with dynamic arch support, only build kernels for
architectures that are supported by the gcc we are building with.
Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
2020-05-12 11:01:16 +02:00
8c338616f9
s390x: gate dynamic arch detection on gcc version and add generic
...
When building OpenBLAS with DYNAMIC_ARCH=1 on s390x (aka zarch), make
sure to include support for systems without the facilities introduced
with z13 (i.e., zarch_generic). Adjust runtime detection to fallback to
that generic code when running on a unknown platform other than Z13
through Z15.
When detecting a Z13 or newer system, add a check for gcc support for
the architecture-specific features before selecting the respective
kernel. Fallback to Z13 or generic code, in case.
Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
2020-05-12 11:01:16 +02:00
f94c53ec0a
Merge pull request #2612 from RajalakshmiSR/testshgemm
...
Improve shgemm test
2020-05-12 08:34:02 +02:00
8efba9b7c0
Improve shgemm test
...
This patch adds another check to test shgemm results.
2020-05-11 17:15:10 -05:00
4fffa556d8
Merge pull request #2611 from RajalakshmiSR/bench_half
...
Include shgemm in benchtest
2020-05-11 21:08:41 +02:00
ce90e2bd3f
Include shgemm in benchtest
...
This patch is to enable benchtest for half precision gemm
when BUILD_HALF is set during make.
2020-05-11 09:57:46 -05:00
948b6712ba
Merge pull request #2610 from martin-frbg/issue2552-3
...
Temporary workaround for excessive LAPACK test failures with COMPLEX on Skylake-X
2020-05-10 13:10:31 +02:00
2271c3506b
Work around excessive LAPACK test failures on Skylake-X
...
Something in the plain C parts of x86_64 cscal.c and zscal.c appears to be miscompiled by both gfortran9 and ifort when compiling for skylakex-avx512, even when the optimized Haswell microkernel is not in use.
2020-05-09 23:49:18 +02:00
db00b21445
Merge pull request #2609 from martin-frbg/issue2552-2
...
Correct ifort options
2020-05-09 21:33:02 +02:00
58d26b4448
Correct ifort options
...
to same as suggested by reference-lapack
2020-05-09 17:15:36 +02:00
8e47d14053
Merge pull request #2608 from martin-frbg/issue2604
...
Handle trailing whitespace and empty variables in KERNEL files
2020-05-09 16:36:14 +02:00
cd10b35fe9
Handle trailing spaces and empty condition variables
2020-05-09 13:42:33 +02:00
9472dd99cd
Merge pull request #57 from xianyi/develop
...
rebase
2020-05-09 13:20:44 +02:00
7181665452
Merge pull request #2605 from RajalakshmiSR/cmake-power
...
Fix cmake compilation issue - POWER9
2020-05-09 11:29:28 +02:00
bd9ff820bc
Fix cmake compilation issue - POWER9
...
This patch removes extra space in the sgemmotcopy filename
thereby allowing it to create entry in kernel/Makefile
created by cmake.
2020-05-08 20:31:56 -05:00
63e45def70
Merge pull request #2603 from martin-frbg/issue2552
...
Add FFLAGS_DRV entry to the generated make.inc to fix lapack-test failure with Intel compilers
2020-05-08 22:08:39 +02:00
ec0f228632
Add FFLAGS_DRV to the generated make.inc to fix lapack-test on x86_64 with icc/ifort
...
fixes #2552
2020-05-08 18:06:12 +02:00
90e2941c61
Merge pull request #56 from xianyi/develop
...
rebase
2020-05-07 22:43:48 +02:00
10d5f3c87b
Merge pull request #2602 from ashwinyes/thunderx2_develop
...
DAXPY Optimizations for ThunderX2
2020-05-07 22:06:41 +02:00
8353cb245a
ARM64: Improve DAXPY for ThunderX2
...
Improve performance of DAXPY for ThunderX2
when the vector fits in L1 Cache.
2020-05-07 09:22:50 -07:00
ec2dd7b875
Merge pull request #2601 from martin-frbg/issue818
...
Undefine NAME/CNAME etc in Makefile.system before defining them
2020-05-07 10:12:33 +02:00
4e82eb9f8a
Undefine ASMNAME/NAME/CNAME before defining them
...
to avoid redefinition warning when environment variables like CFLAGS are being used (fixes #818 )
2020-05-07 00:31:32 +02:00
61300bb735
Merge pull request #55 from xianyi/develop
...
rebase
2020-05-07 00:27:14 +02:00
33e9b12464
Merge pull request #2597 from martin-frbg/appleclang
...
Use Clang 9.0.0 miscompilation fix for corresponding AppleClang version as well
2020-05-05 13:55:08 +02:00
90dba9f716
Duplicate earlier Clang 9.0.0 workaround for corresponding Apple Clang version
...
As discussed on the original PR #2329 , the "Apple Clang 11.0.3" that appears to be based the same LLVM release produces the same miscompilation of this file.
2020-05-05 10:44:50 +02:00
424d551e01
Merge pull request #53 from xianyi/develop
...
rebase
2020-05-01 15:18:46 +02:00
596f5df9e8
Merge pull request #2591 from RajalakshmiSR/testhalf
...
Add test for shgemm
2020-05-01 09:59:39 +02:00
5dd14e3d48
Make building the bfloat16 functions conditional on option BUILD_HALF ( #2590 )
...
* make building the bfloat16 BLAS functions conditional on BUILD_HALF
* pass the BUILD_HALF option to gensymbol
* Pass BUILD_HALF as a compiler define for dynamic_arch builds
2020-05-01 09:58:30 +02:00
a54e35e780
Merge pull request #2586 from martin-frbg/miscfixes
...
Trivial fix for compiler warnings
2020-04-29 22:01:41 +02:00
564b0d39ef
Add test for shgemm
...
This patch has Makefile changes to add test for shgemm which
compares sgemm and shgemm result.
2020-04-29 13:40:34 -05:00
5d58b11101
Merge pull request #52 from xianyi/develop
...
rebase
2020-04-29 14:36:15 +02:00
d394d4e677
Merge pull request #2585 from martin-frbg/mips64fix
...
Increase default BUFFER_SIZE on MIPS64
2020-04-28 19:47:55 +02:00
f4248af26e
Fix compiler warnings
2020-04-28 10:43:12 +02:00
2d89603e9d
Increase BUFFER_SIZE on mips64 to match SGEMM parameters
2020-04-28 10:40:40 +02:00
26bc15258a
Merge pull request #51 from xianyi/develop
...
rebase
2020-04-28 10:38:50 +02:00
141998dce2
Merge pull request #2584 from martin-frbg/issue2583
...
[WIP] Have CMAKE parse conditional lines in KERNEL files
2020-04-28 10:35:12 +02:00
3bd56846bb
Silence a debug message
2020-04-27 16:27:09 +02:00
e7bbdfdf84
Have CMAKE parse conditional lines in KERNEL files
...
Supports ifeq and ifneq, but requires both to have an else branch
2020-04-27 15:20:03 +02:00
b6795db731
Merge pull request #2582 from martin-frbg/mips32fix
...
Increase BUFFER_SIZE on MIPS32 to accomodate SGEMM requirements
2020-04-27 09:18:34 +02:00
5e0dbf8dfe
Increase default BUFFER_SIZE to accomodate SGEMM parameters
...
in response to compile-time warning from #2551
2020-04-26 22:21:05 +02:00
955d73127f
Merge pull request #50 from xianyi/develop
...
rebase
2020-04-26 22:17:56 +02:00
a8c1bea7ae
Merge pull request #2581 from martin-frbg/raji
...
Fix travis configuration and update CONTRIBUTORS.md
2020-04-25 19:57:10 +02:00
e43b49e064
Drop the set -e from travis scripts
2020-04-25 16:18:54 +02:00
3e28db7f38
Update CONTRIBUTORS.md
2020-04-25 13:51:44 +02:00
4b69ee31af
Merge pull request #2580 from martin-frbg/issue2538-3
...
Increase POWER8 ZGEMM_R and use same R values for POWER9
2020-04-25 00:28:18 +02:00
03ff213c51
Increase POWER8 ZGEMM_R and use same R values for POWER9
...
fixes lapack-test zger failures seen in #2299 after application of my PR #2551
2020-04-24 21:46:54 +02:00
299d1c8de0
Merge pull request #2578 from martin-frbg/issue2576
...
Quote getarch include paths in prebuild.cmake
2020-04-24 14:32:46 +02:00
70869d571f
Quote include paths for getarch to protect any embedded spaces
2020-04-24 10:30:44 +02:00
cba87222b2
Merge pull request #49 from xianyi/develop
...
rebase
2020-04-24 10:21:48 +02:00
Marius Hillenbrand
Martin Kroeker
Rajalakshmi Srinivasaraghavan
Ashwin Sekhar T K
Rajalakshmi Srinivasaraghavan