Martin Kroeker
85e5165e98
Merge pull request #3046 from martin-frbg/nvidiasdk-ppc
Support NVIDIA HPC SDK on POWERPC
4 years ago
Martin Kroeker
17c16f2a71
Implement builtin_cpu_is and limit cpu choices to P8 and P9 for NVIDIA compilers
5 years ago
Martin Kroeker
91c3f86c2b
NVIDIA compiler does not yet support POWER10
5 years ago
Martin Kroeker
75b1f3becc
Limit POWERPC DYNAMIC_CORE list to P8 and P9 for NVIDIA compilers
5 years ago
Martin Kroeker
07c5e549b2
Merge pull request #3045 from martin-frbg/nvidiasdk
Support NVIDIA HPC SDK 20.11 compilers on x86_64
5 years ago
Martin Kroeker
005cce5507
Amend SkylakeX options to support the NVIDIA compiler
5 years ago
Martin Kroeker
b859b6e79d
Add nvfortran
5 years ago
Martin Kroeker
b212a2fb9f
Add/modify "PGI" compiler options for NVIDIA SDK 20.11
5 years ago
Martin Kroeker
e40416567a
Add version printout for PGI/NVIDIA compiler
5 years ago
Martin Kroeker
b37e5fa2f8
Merge pull request #5 from xianyi/develop
rebase
5 years ago
Martin Kroeker
326469ef4a
Merge pull request #3042 from martin-frbg/develop
Move FMA3 option setting to the kernel makefile
5 years ago
Martin Kroeker
c73d8ee40d
Conditionally add -mfma to compiler options where needed
5 years ago
Martin Kroeker
abef2ea770
Move -fma option setting to kernel/Makefile.L1
5 years ago
Martin Kroeker
b26e32c3af
Merge pull request #3040 from martin-frbg/fixfcheck
Fix undefined CC variable in check for clang+gfortran combo
5 years ago
Martin Kroeker
7822eff936
Merge pull request #3038 from martin-frbg/issue3037
Fix spurious assumption of cross-compilation on some architectures
5 years ago
Martin Kroeker
b03dc011be
Fix undefined CC variable in clang check
5 years ago
Martin Kroeker
00ce35336e
Fix spurious removal of a trailing character from the hostarch string on x86_64
5 years ago
Martin Kroeker
723776ddf7
Merge pull request #4 from xianyi/develop
rebase
5 years ago
Martin Kroeker
5a77ec7f1c
Merge pull request #3036 from RajalakshmiSR/p10copyalign
POWER10: Improve copy performance
5 years ago
Rajalakshmi Srinivasaraghavan
2fb11f873b
POWER10: Improve copy performance
This patch aligns the stores to 32 byte boundary for scopy and dcopy
before entering into vector pair loop. For ccopy, changed the store
instructions to stxv to improve performance of unaligned cases.
5 years ago
Martin Kroeker
87315e8a8d
Update version to 0.3.13.dev
5 years ago
Martin Kroeker
9031ebd7d5
Update version to 0.3.13.dev
5 years ago
Martin Kroeker
12b41d5598
Merge pull request #3034 from xianyi/release-0.3.0
Merge back the release branch into develop to copy tag
5 years ago
Martin Kroeker
d2b11c4777
Merge pull request #3033 from xianyi/develop
Update branch from develop to release 0.3.13
5 years ago
Martin Kroeker
7bc0e4a2e0
Update version to 0.3.13 for release
5 years ago
Martin Kroeker
d3ec787f77
Update version to 0.3.13 for release
5 years ago
Martin Kroeker
2c309c235d
Merge pull request #3031 from martin-frbg/changelog13
Update Changelog.txt
5 years ago
Martin Kroeker
3dec81200c
Update Changelog.txt
Co-authored-by: h-vetinari <h.vetinari@gmx.com>
5 years ago
Martin Kroeker
737724607f
Merge pull request #3030 from martin-frbg/fix2994
Make fallback from POWER10 to POWER9 depend on new enough compiler
5 years ago
Martin Kroeker
77edf82c7f
Update Changelog.txt for 0.3.13
5 years ago
Martin Kroeker
6232237dba
Make fallback from P10 to P9 conditional on suitable compiler
5 years ago
Martin Kroeker
7d81acc762
Merge pull request #3 from xianyi/develop
rebase
5 years ago
Martin Kroeker
18d8a67485
Merge pull request #2994 from antonblanchard/power10-fixes
Power10 fixes
5 years ago
Martin Kroeker
043128cbe5
Merge pull request #3029 from RajalakshmiSR/axpyp10
POWER10: Improve axpy performance
5 years ago
Martin Kroeker
3331ca492d
Merge pull request #3021 from austinpagan/trsm_p10
POWER: Added special unrolled vectorized versions of "Solve" for specific si…
5 years ago
Rajalakshmi Srinivasaraghavan
346e30a46a
POWER10: Improve axpy performance
This patch aligns the stores to 32 byte boundary for saxpy and daxpy
before entering into vector pair loop. Fox caxpy, changed the store
instructions to stxv to improve performance of unaligned cases.
5 years ago
Martin Kroeker
83de62c20d
Merge pull request #3026 from martin-frbg/revert747
Revert PR747 - SYRK parameter changes for Haswell and related targets
5 years ago
Martin Kroeker
658da9a769
Merge pull request #3027 from gxw-loongson/develop
Add msa support for loongson
5 years ago
gxw
be24c66a7c
Keep LOONGSON3A and LOONGSON3B for loongson
5 years ago
gxw
4b548857d6
Add msa support for loongson
1. Using core loongson3r3 and loongson3r4 for loongson
2. Add DYNAMIC_ARCH for loongson
Change-Id: I1c6b54dbeca3a0cc31d1222af36a7e9bd6ab54c1
5 years ago
Martin Kroeker
d71fe4ed4e
Remove GEMM_DEFAULT_UNROLL_MN parameters for Haswell and ZEN (introduced in PR747)
5 years ago
Martin Kroeker
a554712439
remove extra/intermediate size step for min_jj introduced in PR747
5 years ago
Martin Kroeker
5d26223f4a
remove extra/intermediate size step of min_jj from PR747
5 years ago
Martin Kroeker
980ab349bc
Merge pull request #2 from xianyi/develop
rebase
5 years ago
gxw
d67babf345
Remove gcc unrecognized option '-msched-weight' when check msa
5 years ago
Martin Kroeker
7f11e33e8d
Merge pull request #3025 from TiredNotTear/develop
MIPS: Fix two bugs
5 years ago
Xianyi Zhang
7834c10e2f
Add PingTouGe contribution credit.
5 years ago
Martin Kroeker
53e0837809
Merge pull request #3022 from jinboson/develop
Fix test errors reported by cblas_cgemm & cblas_ctrmm
5 years ago
Hao Chen
ad38bd0e89
Fix failed cgemv and zgemv test case after using msa optimization
The cgemv and zgemv test case will call cgemv_n/t_msa.c zgemv_n/t_msa.c files in MIPS environment.
When the macro CONJ is defined, the calculation result will be wrong due to the wrong definition of OP2.
This patch updates the value of OP2 and passes the corresponding test.
5 years ago
Hao Chen
47b639cc9b
Fix failed sswap and dswap case by using msa optimization
The swap test case will call sswap_msa.c and dswap_msa.c files in MIPS environmnet.
When inc_x or inc_y is equal to zero, the calculation result of the two functions will be wrong.
This patch adds the processing of inc_x or inc_y equal to zero, and the swap test case has passed.
5 years ago