phynexis-deps/OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
9,676 Commits 60 Branches 70 Tags
e01b1094de1a94ccebb91a1c6f59456bcf3ba2a1
Commit Graph
9676 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Martin Kroeker e939c6c315 Merge pull request #5471 from quic/topic/ssymm_direct_sme1 
Support for SME1 based ssymm_direct kernel for cblas_ssymm level 3 API
 2025-10-01 06:22:36 -07:00
Martin Kroeker 03c8d1010f Merge pull request #5476 from ChipKerchner/fasterGEMVNRISCV 
Tranverse matrix data in a cache friendly manner for GEMV_N (RISCV)
 2025-10-01 00:38:36 -07:00
Martin Kroeker 3eed1883cd Merge pull request #5475 from martin-frbg/issue5424 
Fix Fortran compiler autodetection in the presence of compiler arguments
 2025-09-30 15:45:59 -07:00
Chip Kerchner 36f9cb85b1 Fix pre-RVV 1.0. 2025-09-30 22:41:31 +00:00
Chip Kerchner 2d82d144e2 Tranverse matrix data in a cache friendly manner for GEMV_N (RISCV). 2025-09-30 21:22:10 +00:00
Chip Kerchner dd7a1d6d4d Merge branch 'develop' of https://github.com/OpenMathLib/OpenBLAS into develop 2025-09-30 20:16:19 +00:00
Martin Kroeker b7c8c5c190 Fix autodetection in the presence of compiler arguments 2025-09-30 21:29:24 +02:00
Martin Kroeker aaa5c377bc Merge pull request #5465 from ChipKerchner/addRVVVectorizedFP16Packing 
Add vectorized packing for FP16 and BF16 for RISC-V.  Reactivate vector packing for FP64 transposed
 2025-09-30 09:21:15 -07:00
Chris Marsh c1f607c43c fix -lto_library filtering for apple-clang + gfortran on sdk26 (#5474) 
* ensure filter-out applies to subsequent sdk versions
 2025-09-30 16:28:37 +02:00
Rajendra Prasad Matcha 19268471cc Support for SME1 based ssymm_direct kernel for cblas_ssymm level 3 API 2025-09-30 15:05:33 +05:30
Martin Kroeker 1dcd4cb6e9 Merge pull request #5472 from pratiklp00/fix_build_issue 
Fix typos in Power10 kernel
 2025-09-30 00:52:09 -07:00
Chip Kerchner 67ddda394e Merge branch 'develop' into addRVVVectorizedFP16Packing 2025-09-29 13:49:57 +00:00
pratiklp00 d7b11605d1 fix build issue 2025-09-29 02:02:13 -05:00
Martin Kroeker b27a1184fd Merge pull request #5468 from martin-frbg/issue5464 
Fix CMake builds not passing the HFLOAT16 parameter to generated files
 2025-09-27 14:05:31 -07:00
Martin Kroeker 4ef70b490c Add support for passing the HFLOAT16 option where required 2025-09-27 22:19:23 +02:00
Martin Kroeker b537dacb34 Merge pull request #5463 from sharkcz/ppc 
fix typos in Power8 routines
 2025-09-27 04:29:15 -07:00
Dan Horák f5ec1c4e53 fix typos in Power8 routines 
Fixes: https://github.com/OpenMathLib/OpenBLAS/pull/5448
 2025-09-26 16:54:03 +02:00
Dan Horák 681af71d95 drop gcc 15 workaround 
As the assembler routines has correctly specified parameter we can drop
the previously applied workaround in https://github.com/OpenMathLib/OpenBLAS/pull/5409.
 2025-09-26 16:52:17 +02:00
Chip Kerchner 07d0e742c2 Add vectorized packing for FP16 and BF16. Reactivate vector packing for FP64 transposed. 2025-09-26 14:50:38 +00:00
Martin Kroeker e58f6dc50d Add extensions ?GEMM_BATCH_STRIDED and CBLAS_?GEMM_BATCH_STRIDED (#5458) 
* Add ?GEMM_BATCH_STRIDED and CBLAS_?GEMM_BATCH_STRIDED
 2025-09-26 14:00:47 +02:00
Martin Kroeker c92f7f6bb2 Merge pull request #5448 from martin-frbg/issue5372-2 
Fix clobber list entries for arrays in POWER kernels that use inline asm
 2025-09-26 02:24:50 -07:00
Martin Kroeker 65c626e646 Merge pull request #5456 from ChipKerchner/riscv-bf16-fix 
Disable bf16 flags on RISC-V unless BUILD_BFLOAT16=1
 2025-09-26 00:09:33 -07:00
Martin Kroeker f2304af1d3 Merge pull request #5462 from martin-frbg/issue5460 
Use generic kernels for SCAL on POWER4/POWER5 to fix corner cases of Inf/NAN
 2025-09-25 22:41:43 -07:00
Martin Kroeker 2852171d4f Merge pull request #5461 from martin-frbg/applemlinux 
Fix Apple M cpu identification under Linux
 2025-09-25 14:10:22 -07:00
Martin Kroeker 916a47409c Merge pull request #5455 from martin-frbg/lapack1158 
Rescale Infinity input in CSRSCL/ZDRSCL early to avoid infinite loop later (Reference-LAPACK PR 1158)
 2025-09-25 12:48:23 -07:00
Martin Kroeker 14c9dcaac7 Use generic kernels for SCAL to fix corner cases of Inf/NAN 2025-09-25 20:31:12 +02:00
Martin Kroeker 6798aec462 Any Apple cpu running Linux has to be Vortex 2025-09-25 20:26:47 +02:00
Martin Kroeker 7989caabdc Merge pull request #5449 from OpenMathLib/gemm_batch 
[WIP] Add BLAS interface to ?GEMM_BATCH to complement the CBLAS one
 2025-09-24 02:42:08 -07:00
Chip Kerchner fc7d6e65a1 Change BF16 warning message. 2025-09-23 20:26:44 +00:00
Chip Kerchner 9427eaf4c4 Reduce flags for BF16 to only needed ones. 2025-09-23 20:24:40 +00:00
Chip Kerchner 3116749717 Disable bf16 flags on RISC-V unless BUILD_BFLOAT16=1 2025-09-23 15:02:20 +00:00
Martin Kroeker 99c077a3a7 Update gemm_batch.c
apple m / build (cmake, gfortran, 0, 0) (push) Canceled after 0s
apple m / build (cmake, gfortran, 0, 1) (push) Canceled after 0s
apple m / build (cmake, gfortran, 1, 0) (push) Canceled after 0s
apple m / build (cmake, gfortran, 1, 1) (push) Canceled after 0s
apple m / build (make, gfortran, 0, 0) (push) Canceled after 0s
apple m / build (make, gfortran, 0, 1) (push) Canceled after 0s
apple m / build (make, gfortran, 1, 0) (push) Canceled after 0s
apple m / build (make, gfortran, 1, 1) (push) Canceled after 0s
c910v qemu test / TEST (riscv64-linux-gnu, NO_SHARED=1 TARGET=C910V, C910V, riscv64-unknown-linux-gnu) (push) Canceled after 0s
c910v qemu test / TEST (riscv64-linux-gnu, NO_SHARED=1 TARGET=RISCV64_GENERIC, RISCV64_GENERIC, riscv64-linux-gnu) (push) Canceled after 0s
Run codspeed benchmarks / benchmarks (make, gfortran, ubuntu-22.04, 3.12) (push) Canceled after 0s
continuous build / build (cmake, flang, ubuntu-latest) (push) Canceled after 0s
continuous build / build (cmake, gfortran, macos-latest) (push) Canceled after 0s
continuous build / build (cmake, gfortran, ubuntu-latest) (push) Canceled after 0s
continuous build / build (make, flang, ubuntu-latest) (push) Canceled after 0s
continuous build / build (make, gfortran, macos-latest) (push) Canceled after 0s
continuous build / build (make, gfortran, ubuntu-latest) (push) Canceled after 0s
continuous build / msys2 (None, fc, int32, UCRT64, mingw-w64-ucrt-x86_64) (push) Canceled after 0s
continuous build / msys2 (Release, fc, int32, CLANG64, mingw-w64-clang-x86_64) (push) Canceled after 0s
continuous build / msys2 (Release, fc, int32, MINGW32, mingw-w64-i686) (push) Canceled after 0s
continuous build / msys2 (Release, fc, int32, UCRT64, mingw-w64-ucrt-x86_64) (push) Canceled after 0s
continuous build / msys2 (Release, fc, int64, -DBINARY=64 -DINTERFACE64=1, CLANG64, mingw-w64-clang-x86_64) (push) Canceled after 0s
continuous build / msys2 (Release, fc, int64, -DBINARY=64 -DINTERFACE64=1, UCRT64, mingw-w64-ucrt-x86_64) (push) Canceled after 0s
continuous build / cross_build (DYNAMIC_ARCH=1 TARGET=GENERIC, mips64el, mips64el-linux-gnuabi64) (push) Canceled after 0s
continuous build / cross_build (TARGET=EV4, alpha, alpha-linux-gnu) (push) Canceled after 0s
continuous build / cross_build (TARGET=MIPS1004K, mipsel, mipsel-linux-gnu) (push) Canceled after 0s
continuous build / cross_build (TARGET=RISCV64_GENERIC, riscv64, riscv64-linux-gnu) (push) Canceled after 0s
continuous build / neoverse_build (push) Canceled after 0s
harmonyos / build (push) Canceled after 0s
loongarch64 qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=GENERIC, DYNAMIC_ARCH, loongarch64-linux-gnu) (push) Canceled after 0s
loongarch64 qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA264, LA264, loongarch64-linux-gnu) (push) Canceled after 0s
loongarch64 qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA464, LA464, loongarch64-linux-gnu) (push) Canceled after 0s
loongarch64 qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA64_GENERIC, LA64_GENERIC, loongarch64-linux-gnu) (push) Canceled after 0s
loongarch64 qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSON2K1000, LOONGSON2K1000, loongarch64-linux-gnu) (push) Canceled after 0s
loongarch64 qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSON3R5, LOONGSON3R5, loongarch64-linux-gnu) (push) Canceled after 0s
loongarch64 qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSONGENERIC, LOONGSONGENERIC, loongarch64-linux-gnu) (push) Canceled after 0s
loongarch64 clang qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=GENERIC, DYNAMIC_ARCH) (push) Canceled after 0s
loongarch64 clang qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA264, LA264) (push) Canceled after 0s
loongarch64 clang qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA464, LA464) (push) Canceled after 0s
loongarch64 clang qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA64_GENERIC, LA64_GENERIC) (push) Canceled after 0s
loongarch64 clang qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSON2K1000, LOONGSON2K1000) (push) Canceled after 0s
loongarch64 clang qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSON3R5, LOONGSON3R5) (push) Canceled after 0s
loongarch64 clang qemu test / TEST (NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSONGENERIC, LOONGSONGENERIC) (push) Canceled after 0s
mips64 qemu test / TEST (NO_SHARED=1 TARGET=I6400, I6400, mipsisa64r6el-linux-gnuabi64) (push) Canceled after 0s
mips64 qemu test / TEST (NO_SHARED=1 TARGET=I6500, I6500, mipsisa64r6el-linux-gnuabi64) (push) Canceled after 0s
mips64 qemu test / TEST (NO_SHARED=1 TARGET=MIPS64_GENERIC, MIPS64_GENERIC, mips64el-linux-gnuabi64) (push) Canceled after 0s
mips64 qemu test / TEST (NO_SHARED=1 TARGET=P6600, P6600, mipsisa64r6el-linux-gnuabi64) (push) Canceled after 0s
mips64 qemu test / TEST (NO_SHARED=1 TARGET=SICORTEX, SICORTEX, mips64el-linux-gnuabi64) (push) Canceled after 0s
riscv64 zvl256b qemu test / TEST (TARGET=RISCV64_GENERIC BINARY=64 ARCH=riscv64 DYNAMIC_ARCH=1, rv64,g=true,c=true,v=true,vext_spec=v1.0,vlen=256,elen=64, DYNAMIC_ARCH=1) (push) Canceled after 0s
riscv64 zvl256b qemu test / TEST (TARGET=RISCV64_ZVL128B BINARY=64 ARCH=riscv64, rv64,g=true,c=true,v=true,vext_spec=v1.0,vlen=128,elen=64, RISCV64_ZVL128B) (push) Canceled after 0s
riscv64 zvl256b qemu test / TEST (TARGET=RISCV64_ZVL256B BINARY=64 ARCH=riscv64, rv64,g=true,c=true,v=true,vext_spec=v1.0,vlen=256,elen=64, RISCV64_ZVL256B) (push) Canceled after 0s
2025-09-23 15:43:33 +02:00
Martin Kroeker fe5402d82b Merge pull request #5454 from ChipKerchner/addRISCV64_BFLOAT16 
Add support for BF16 sbgemm on RISCV.
 2025-09-23 01:02:50 -07:00
Martin Kroeker 5a8f9aa8fc Remove LABAD (Reference-LAPACK PR 805) and fix infinite loop scaling Inf (PR 1158) 2025-09-22 13:27:59 -07:00
Chip Kerchner 92f09a6a98 Add BF16 sbgemm on RISCV. 2025-09-22 14:32:43 +00:00
Martin Kroeker fc042d928b fix symbol naming (underscoring) 2025-09-21 19:01:11 +02:00
Martin Kroeker e5c836136d Merge pull request #5451 from ChipKerchner/fixRVVSHGEMM 
Fix _Float16 casting issue and reduce LMUL for certain vector instruction from m2 to m1.
 2025-09-20 03:07:14 -07:00
Chip Kerchner a4abf7828e Fix _Float16 casting issue and reduce LMUL for certain vector instruction from m2 to m1. 2025-09-18 21:30:22 +00:00
Martin Kroeker cdeac0a3ae Build non-CBLAS ?gemm_batch too 2025-09-17 11:42:11 -07:00
Martin Kroeker eb931deb22 Add BLAS interface to ?GEMM_BATCH 2025-09-17 10:23:48 -07:00
Martin Kroeker 1d5279fd29 Fix clobber list entries for arrays in inline asm 2025-09-17 07:02:18 -07:00
Martin Kroeker 79a1f38770 Merge pull request #5445 from Dayuxiaoshui/develop 
Add RVV vectorized optimization for omatcopy function on RISC-V64 architecture
 2025-09-17 06:38:41 -07:00
Dayuxiaoshui becda2faa4 Merge branch 'develop' of https://github.com/Dayuxiaoshui/OpenBLAS into develop 2025-09-15 18:50:22 +08:00
Dayuxiaoshuiandgong-flying b61a2d7648 Clean up RISC-V RVV performance test files 
Removed temporary test files and benchmarks created for omatcopy RVV optimization validation:
- Removed benchmark scripts and result files
- Removed test executables (scalar vs RVV comparison)
- Removed temporary test source files

Previous testing on sg2044 showed RVV implementation achieved ~30% performance improvement
over scalar version for 3000x4000 matrix operations (0.384s vs 0.552s per iteration).

Co-authored-by: gong-flying <gongxiaofei24@iscas.ac.cn>
 2025-09-15 18:49:23 +08:00
学习中的牛马 8b7e4c2b5c Merge branch 'OpenMathLib:develop' into develop 2025-09-15 12:08:17 +08:00
Dayuxiaoshui 2265318d3e Optimize RISC-V RVV omatcopy implementation with latest RVV API\n\nCo-authored-by: gong-flying <gongxiaofei24@iscas.ac.cn> 2025-09-15 11:46:50 +08:00
Martin Kroeker 43fdff7f14 Merge pull request #5444 from yuanjia111/develop 
riscv：Remove  the manual unrolling in gemv_t_vector.c  to performance better
 2025-09-12 23:58:13 -07:00
yuanjia 826cb4588f remove unused variable 2025-09-13 11:35:49 +08:00
yuanjia 53d7452cdf riscv: gemv_t_vector.c optimize 2025-09-13 11:24:49 +08:00
Dayuxiaoshuiandgong-flying bd45b82ed0 Optimize RISC-V RVV omatcopy_ct implementation with advanced vectorization 
- Implement block-based memory access optimization (64x64 blocks)
- Add 4-way loop unrolling to reduce loop overhead
- Optimize VSETVL calls to improve vectorization efficiency
- Add software prefetching for better memory access patterns
- Implement fast path for small matrices (<64x64)
- Add cross-compilation script for RISC-V testing
- Improve boundary handling with separate main/tail loops

Co-authored-by: gong-flying <gongxiaofei24@iscas.ac.cn>
 2025-09-11 20:01:39 +08:00