OpenBLAS

Tree: 2c3cdaf74e

History

Chris Sidebottom 2c3cdaf74e Optimized BGEMV for NEOVERSEV1 target - Adds bgemv T based off of sbgemv T kernel - Adds bgemv N which is slightly alterated to not use Y as an accumulator due to the output being bf16 which results in loss of precision - Enables BGEMM_GEMV_FORWARD to proxy BGEMM to BGEMV with new kernels		2 months ago
..
KERNEL	Further rearranged the rotm kernel for the different architectures.	8 months ago

KERNEL.A64FX	Performance improvements of [SD]DOT with loop-unrolling on A64FX	3 months ago

KERNEL.AMPERE1	reduce duplicate kernel code	3 months ago

KERNEL.ARMV8	Add workaround for NVIDIA HPC mishandling of the asm DOT kernels	4 years ago

KERNEL.ARMV8SVE	Use SVE kernel for S/DGEMVN for SVE machines	5 months ago

KERNEL.ARMV9SME	Support for SME1 based sgemm_direct kernel for cblas_sgemm level 3 API	7 months ago

KERNEL.CORTEXA53	optimize cgemm on ARM cortex A53 & cortex A55	3 years ago

KERNEL.CORTEXA55	Reduce duplication in kernel definitions	1 year ago

KERNEL.CORTEXA57	Add workaround for NVIDIA HPC mishandling of the asm DOT kernels	4 years ago

KERNEL.CORTEXA72	Simplifying ARMv8 build parameters	6 years ago

KERNEL.CORTEXA73	Simplifying ARMv8 build parameters	6 years ago

KERNEL.CORTEXA76	Add support for Cortex-A76	1 year ago

KERNEL.CORTEXA510	Fix outdated SVE kernel definitions for Cortex cpus by aliasing to ARMV8SVE	1 year ago

KERNEL.CORTEXA710	Fix outdated SVE kernel definitions for Cortex cpus by aliasing to ARMV8SVE	1 year ago

KERNEL.CORTEXX1	CortexX1 is ARMV8 like A7x	3 years ago

KERNEL.CORTEXX2	Fix outdated SVE kernel definitions for Cortex cpus by aliasing to ARMV8SVE	1 year ago

KERNEL.EMAG8180	Add preliminary support for EMAG8180	5 years ago

KERNEL.FALKOR	Simplifying ARMv8 build parameters	6 years ago

KERNEL.FT2000	Add initial support for Phytium FT2000 series and ARMV9 Cortex 510/710/X1/X2	3 years ago

KERNEL.NEOVERSEN1	Merge pull request #5225 from annop-w/gemv_n	5 months ago

KERNEL.NEOVERSEN2	Use SVE kernel for S/DGEMVN for SVE machines	5 months ago

KERNEL.NEOVERSEV1	Optimized BGEMV for NEOVERSEV1 target	2 months ago

KERNEL.NEOVERSEV2	Add dispatch of SBGEMVNKERNEL for NEOVERSEN2 and NEOVERSEV2	7 months ago

KERNEL.THUNDERX	Add workaround for NVIDIA HPC	4 years ago

KERNEL.THUNDERX2T99	Add SVE implementation for sdot/ddot	2 years ago

KERNEL.THUNDERX3T110	Reduce duplication in kernel definitions	1 year ago

KERNEL.TSV110	Add workaround for NVIDIA HPC	4 years ago

KERNEL.VORTEX	Use Neoverse's current mix of ThunderX2 kernels for Vortex as well	4 years ago

KERNEL.generic	Further rearranged the rotm kernel for the different architectures.	8 months ago

Makefile	added experimental support for ARMV8	12 years ago

amax.S	ARM64: Convert all labels to local labels	8 years ago

asum.S	ARM64: Convert all labels to local labels	8 years ago

axpy.S	ARM64: Convert all labels to local labels	8 years ago

bgemm_beta_neon.c	Add optimized BGEMM kernel for NEOVERSEV1 target	2 months ago

bgemm_kernel_4x4_neoversev1.c	Add optimized BGEMM kernel for NEOVERSEV1 target	2 months ago

bgemm_kernel_4x4_neoversev1_impl.c	Add optimized BGEMM kernel for NEOVERSEV1 target	2 months ago

bgemv_n_sve_v3x4.c	Optimized BGEMV for NEOVERSEV1 target	2 months ago

casum.S	ARM64: Convert all labels to local labels	8 years ago

casum_thunderx2t99.c	Fixed a few more unnecessary calls to num_cpu_avail.	7 years ago

cgemm_kernel_4x4.S	ARM64: Convert all labels to local labels	8 years ago

cgemm_kernel_8x4.S	move ALPHA_I out of register 18 (reserved on OSX)	2 years ago

cgemm_kernel_8x4_cortexa53.c	optimize cgemm on ARM cortex A53 & cortex A55	3 years ago

cgemm_kernel_8x4_thunderx2t99.S	Move ALPHA_I out of register 18 (reserved on OSX)	2 years ago

cgemm_kernel_sve_v1x4.S	Use SVE zgemm/cgemm on Arm(R) Neoverse(TM) V1 core	2 years ago

cgemm_ncopy_sve_v1.c	Disambiguate whilelt	2 years ago

cgemm_tcopy_sve_v1.c	Disambiguate whilelt	2 years ago

copy.S	ARM64: Convert all labels to local labels	8 years ago

copy_thunderx2t99.c	[WIP] Work around assembler limitations in current LLVM for Windows on Arm (#5076)	8 months ago

csum.S	Add ARM64 implementations of ?sum	6 years ago

csum_thunderx2t99.c	add csum/zsum kernels (trivially derived from the asum ones)s)	1 year ago

ctrmm_kernel_4x4.S	ARM64: Convert all labels to local labels	8 years ago

ctrmm_kernel_8x4.S	Move ALPHA_I out of register 18 (reserved on OSX)	2 years ago

ctrmm_kernel_sve_v1x4.S	add cgemm ctrmm sve kernels	3 years ago

dasum_thunderx2t99.c	[WIP] Work around assembler limitations in current LLVM for Windows on Arm (#5076)	8 months ago

daxpy_thunderx.c	aarch64 fix std=c18 compilation	5 years ago

daxpy_thunderx2t99.S	ARM64: Improve DAXPY for ThunderX2	5 years ago

ddot_thunderx.c	ARM64: Rename kernel files to have consistent naming	8 years ago

dgemm_beta.S	Fix zero initialization for beta=0 case	5 years ago

dgemm_kernel_4x4.S	ARM64: Convert all labels to local labels	8 years ago

dgemm_kernel_4x4_cortexa53.c	MOD: optimize normal DGEMM on ARMV8 cortex-A53 & cortex-A55	3 years ago

dgemm_kernel_4x8.S	ARM64: Convert all labels to local labels	8 years ago

dgemm_kernel_8x4.S	ARM64: Convert all labels to local labels	8 years ago

dgemm_kernel_8x4_thunderx2t99.S	ARM64: Move parameters from parameter.c to param.h	7 years ago

dgemm_kernel_sve_v1x8.S	some clean-up & commentary	3 years ago

dgemm_kernel_sve_v2x8.S	Remove prefetches from SVE kernels	2 years ago

dgemm_ncopy_4.S	ARM64: Convert all labels to local labels	8 years ago

dgemm_ncopy_8.S	ARM64: Convert all labels to local labels	8 years ago

dgemm_small_kernel_nn_sve.c	Better header guard around bridge	1 year ago

dgemm_small_kernel_nt_sve.c	Better header guard around bridge	1 year ago

dgemm_small_kernel_tn_sve.c	small gemm kernel packing modifications	8 months ago

dgemm_small_kernel_tt_sve.c	small gemm kernel packing modifications	8 months ago

dgemm_tcopy_4.S	ARM64: Convert all labels to local labels	8 years ago

dgemm_tcopy_8.S	Remove unused TEMP2 and reshuffle to leave x18 unused (reserved on OSX)	4 years ago

dot.S	ARM64: Fix utest dsdot errors	7 years ago

dot.c	Performance improvements of [SD]DOT with loop-unrolling on A64FX	3 months ago

dot_kernel_asimd.c	Accumulate results in output register explicitly	3 months ago

dot_kernel_sve.c	add clobber list	1 year ago

dot_kernel_sve_v8.c	Performance improvements of [SD]DOT with loop-unrolling on A64FX	3 months ago

dot_sve_v8.c	Performance improvements of [SD]DOT with loop-unrolling on A64FX	3 months ago

dot_thunderx.c	ARM64: Rename kernel files to have consistent naming	8 years ago

dtrmm_kernel_4x4.S	ARM64: Convert all labels to local labels	8 years ago

dtrmm_kernel_4x8.S	ARM64: Convert all labels to local labels	8 years ago

dtrmm_kernel_8x4.S	Move temp to x21 to leave x18 unused (reserved on OSX)	4 years ago

dtrmm_kernel_sve_v1x8.S	some clean-up & commentary	3 years ago

dznrm2_thunderx2t99.c	remove another early exit for incx < 0	1 year ago

dznrm2_thunderx2t99_fast.c	Fixed a few more unnecessary calls to num_cpu_avail.	7 years ago

gemm_ncopy_complex_sve_v1x4.c	Use SVE zgemm/cgemm on Arm(R) Neoverse(TM) V1 core	2 years ago

gemm_ncopy_sve_v1x8.c	Work around Clang failing to disambiguate SVE intrinsics and add AppleClang crossbuild to MacOS/arm64 DYNAMIC_ARCH in AzureCI (#4140)	2 years ago

gemm_small_kernel_permit_sve.c	Remove k2 loop from DGEMM TN and use a more conservative heuristic for SGEMM	1 year ago

gemm_tcopy_complex_sve_v1x4.c	Use SVE zgemm/cgemm on Arm(R) Neoverse(TM) V1 core	2 years ago

gemm_tcopy_sve_v1x8.c	Work around Clang failing to disambiguate SVE intrinsics and add AppleClang crossbuild to MacOS/arm64 DYNAMIC_ARCH in AzureCI (#4140)	2 years ago

gemv_n.S	ARM64: Convert all labels to local labels	8 years ago

gemv_n_sve.c	Optimize gemv_n_sve kernel	6 months ago

gemv_n_sve_v1x3.c	Optimize gemv_n_sve_v1x3 kernel	3 months ago

gemv_n_sve_v4x3.c	fixed a potential out-of-bounds on gemv.	5 months ago

gemv_t.S	Add accumulators to AArch64 GEMV Kernels	1 year ago

gemv_t_sve.c	Add accumulators to AArch64 GEMV Kernels	1 year ago

gemv_t_sve_v1x3.c	Simplify gemv_t_sve_v1x3 kernel	8 months ago

gemv_t_sve_v4x3.c	Loop-unrolled transposed [SD]GEMV kernels for A64FX and Neoverse V1	10 months ago

iamax.S	ARM64: Convert all labels to local labels	8 years ago

iamax_thunderx2t99.c	Fixed a few more unnecessary calls to num_cpu_avail.	7 years ago

izamax.S	ARM64: Convert all labels to local labels	8 years ago

izamax_thunderx2t99.c	Fixed a few more unnecessary calls to num_cpu_avail.	7 years ago

nrm2.S	Fix accidental duplication of jump instruction	6 years ago

rot.S	ARM64: Convert all labels to local labels	8 years ago

rot.c	Added Updated swap and rot sve kernels.	9 months ago

rot_kernel_c.c	Added Updated swap and rot sve kernels.	9 months ago

rot_kernel_sve.c	Added Updated swap and rot sve kernels.	9 months ago

sasum_thunderx2t99.c	[WIP] Work around assembler limitations in current LLVM for Windows on Arm (#5076)	8 months ago

sbgemm_beta_neoversen2.c	neoverse n2 sbgemm: init file	3 years ago

sbgemm_beta_neoversev1.c	* checkpoint sbgemm for SVE-256	8 months ago

sbgemm_kernel_4x4_neoversev1.c	optimized sbgemm kernel for neoverse-v1 (sve-256)	7 months ago

sbgemm_kernel_4x4_neoversev1_impl.c	optimized sbgemm kernel for neoverse-v1 (sve-256)	7 months ago

sbgemm_kernel_8x4_neoversen2.c	Change file name to match the norm and delete useless code.	2 years ago

sbgemm_kernel_8x4_neoversen2_impl.c	Change file name to match the norm and delete useless code.	2 years ago

sbgemm_ncopy_4_neoversen2.c	Change file name to match the norm and delete useless code.	2 years ago

sbgemm_ncopy_4_neoversev1.c	optimized sbgemm kernel for neoverse-v1 (sve-256)	7 months ago

sbgemm_ncopy_8_neoversen2.c	bugfix for sbgemm_ncopy_8_neoversen2	2 years ago

sbgemm_tcopy_4_neoversen2.c	Add sbgemm_ncopy_8 and sbgemm_tcopy_4	2 years ago

sbgemm_tcopy_4_neoversev1.c	Update sbgemm_tcopy_4_neoversev1 kernel to use standard C types	3 months ago

sbgemm_tcopy_8_neoversen2.c	Improve the performance of sbgemm_tcopy on neoversen2	2 years ago

sbgemv_n_neon.c	fix bugs in aarch64 sbgemv_n kernel	6 months ago

sbgemv_t_bfdot.c	Optimized BGEMV for NEOVERSEV1 target	2 months ago

scal.S	make NAN handling depend on the dummy2 parameter	1 year ago

scnrm2_thunderx2t99.c	remove another early exit for incx < 0	1 year ago

sgemm_beta.S	fix initialization to zero in the NEON SGEMM_BETA kernel as well	5 years ago

sgemm_direct_alpha_beta_arm64_sme1.c	SME1 based direct kernel with alpha and beta for cblas_sgemm level 3 API.	2 months ago

sgemm_direct_arm64_sme1.c	Fix ARMV9SME target in DYNAMIC_ARCH and add SME query code for MacOS (#5222)	4 months ago

sgemm_direct_sme1.S	Support for SME1 based sgemm_direct kernel for cblas_sgemm level 3 API	7 months ago

sgemm_direct_sme1_preprocess.S	Support for SME1 based sgemm_direct kernel for cblas_sgemm level 3 API	7 months ago

sgemm_kernel_4x4.S	ARM64: Convert all labels to local labels	8 years ago

sgemm_kernel_8x8.S	ARM64: Convert all labels to local labels	8 years ago

sgemm_kernel_8x8_cortexa53.S	fix INIT8x4	5 years ago

sgemm_kernel_16x4.S	ARM64: Convert all labels to local labels	8 years ago

sgemm_kernel_16x4_thunderx2t99.S	ARM64: Convert all labels to local labels	8 years ago

sgemm_kernel_sve_v1x8.S	add sgemm kernel and copy functions for sgemm and ssymm	3 years ago

sgemm_kernel_sve_v2x8.S	Remove prefetches from SVE kernels	2 years ago

sgemm_ncopy_4.S	Optimize aarch64 sgemm_ncopy	6 months ago

sgemm_ncopy_8.S	Optimize aarch64 sgemm_ncopy	6 months ago

sgemm_small_kernel_nn_sve.c	Better header guard around bridge	1 year ago

sgemm_small_kernel_nt_sve.c	Better header guard around bridge	1 year ago

sgemm_small_kernel_tn_sve.c	small gemm kernel packing modifications	8 months ago

sgemm_small_kernel_tt_sve.c	small gemm kernel packing modifications	8 months ago

sgemm_tcopy_8.S	sgemm copy source init	5 years ago

sgemm_tcopy_16.S	change line endings from CRLF to LF	2 years ago

sgemv_n_neon.c	Improve performance for SGEMVN on NEONVERSEN1	5 months ago

sme_abi.h	SME1 based direct kernel with alpha and beta for cblas_sgemm level 3 API.	2 months ago

strmm_kernel_4x4.S	ARM64: Convert all labels to local labels	8 years ago

strmm_kernel_8x8.S	ARM64: Convert all labels to local labels	8 years ago

strmm_kernel_8x8_cortexa53.S	use general register to speedup	5 years ago

strmm_kernel_16x4.S	Move temp to x21 to leave x18 unused (reserved on OSX)	4 years ago

strmm_kernel_sve_v1x8.S	strmm sve v1x8 kernel	3 years ago

sum.S	Add ARM64 implementations of ?sum	6 years ago

swap.S	ARM64: Convert all labels to local labels	8 years ago

swap.c	Added Updated swap and rot sve kernels.	9 months ago

swap_kernel_c.c	Added Updated swap and rot sve kernels.	9 months ago

swap_kernel_sve.c	Update swap_kernel_sve.c	9 months ago

swap_thunderx2t99.S	THUNDERX2T99: Add optimized S/D/C/Z SWAP Implementations	8 years ago

symm_lcopy_sve.c	Disambiguate whilelt	2 years ago

symm_ucopy_sve.c	Disambiguate whilelt	2 years ago

symv_L_asimd_4x4.c	Add symv kernels for arm64	5 months ago

symv_L_sve_v1x4.c	Add symv kernels for arm64	5 months ago

symv_U_asimd_4x4.c	Add symv kernels for arm64	5 months ago

symv_U_sve_v1x4.c	Add symv kernels for arm64	5 months ago

symv_microk_asimd_4x4.c	Add symv kernels for arm64	5 months ago

symv_microk_sve_v1x4.c	Add symv kernels for arm64	5 months ago

trmm_lncopy_sve_v1.c	Work around Clang failing to disambiguate SVE intrinsics and add AppleClang crossbuild to MacOS/arm64 DYNAMIC_ARCH in AzureCI (#4140)	2 years ago

trmm_ltcopy_sve_v1.c	Work around Clang failing to disambiguate SVE intrinsics and add AppleClang crossbuild to MacOS/arm64 DYNAMIC_ARCH in AzureCI (#4140)	2 years ago

trmm_uncopy_sve_v1.c	Work around Clang failing to disambiguate SVE intrinsics and add AppleClang crossbuild to MacOS/arm64 DYNAMIC_ARCH in AzureCI (#4140)	2 years ago

trmm_utcopy_sve_v1.c	Work around Clang failing to disambiguate SVE intrinsics and add AppleClang crossbuild to MacOS/arm64 DYNAMIC_ARCH in AzureCI (#4140)	2 years ago

trsm_kernel_LN_sve.c	add sve ztrsm	3 years ago

trsm_kernel_LT_sve.c	add sve ztrsm	3 years ago

trsm_kernel_RN_sve.c	add sve ztrsm	3 years ago

trsm_kernel_RT_sve.c	add sve ztrsm	3 years ago

trsm_lncopy_sve.c	Work around Clang failing to disambiguate SVE intrinsics and add AppleClang crossbuild to MacOS/arm64 DYNAMIC_ARCH in AzureCI (#4140)	2 years ago

trsm_ltcopy_sve.c	Disambiguate whilelt	2 years ago

trsm_uncopy_sve.c	Disambiguate whilelt	2 years ago

trsm_utcopy_sve.c	Disambiguate whilelt	2 years ago

zamax.S	Fix the functional bugs for zamax.	5 years ago

zasum.S	ARM64: Convert all labels to local labels	8 years ago

zasum_thunderx2t99.c	[WIP] Work around assembler limitations in current LLVM for Windows on Arm (#5076)	8 months ago

zaxpy.S	ARM64: Convert all labels to local labels	8 years ago

zdot.S	ARM64: Convert all labels to local labels	8 years ago

zdot_thunderx2t99.c	Add a clobber list to fix utest errors seen with gcc13 on Apple M	1 year ago

zgemm_kernel_4x4.S	move alpha to x19/x20 to leave x18 unused for OSX	4 years ago

zgemm_kernel_4x4_cortexa53.c	MOD: add comments to a53 zgemm kernel	3 years ago

zgemm_kernel_4x4_thunderx2t99.S	ARM64: Convert all labels to local labels	8 years ago

zgemm_kernel_sve_v1x4.S	Use SVE zgemm/cgemm on Arm(R) Neoverse(TM) V1 core	2 years ago

zgemm_ncopy_sve_v1.c	Disambiguate whilelt	2 years ago

zgemm_tcopy_sve_v1.c	Disambiguate whilelt	2 years ago

zgemv_n.S	ARM64: Convert all labels to local labels	8 years ago

zgemv_t.S	ARM64: Convert all labels to local labels	8 years ago

zhemm_ltcopy_sve.c	Fix ZHEMM copy for SVE	2 years ago

zhemm_utcopy_sve.c	Fix ZHEMM copy for SVE	2 years ago

znrm2.S	Remove automatic label postfixes from macro included only once	6 years ago

zrot.S	ARM64: Convert all labels to local labels	8 years ago

zscal.S	fix handling of dummy2	3 months ago

zsum.S	Add ARM64 implementations of ?sum	6 years ago

zsum_thunderx2t99.c	add csum/zsum kernels (trivially derived from the asum ones)s)	1 year ago

zsymm_lcopy_sve.c	Disambiguate whilelt	2 years ago

zsymm_ucopy_sve.c	Disambiguate whilelt	2 years ago

ztrmm_kernel_4x4.S	Move alphaI to x22 to leave x18 unused (reserved on OSX)	4 years ago

ztrmm_kernel_sve_v1x4.S	fix sve ztrmm kernel	3 years ago

ztrmm_lncopy_sve_v1.c	Disambiguate whilelt	2 years ago

ztrmm_ltcopy_sve_v1.c	Disambiguate whilelt	2 years ago

ztrmm_uncopy_sve_v1.c	Disambiguate whilelt	2 years ago

ztrmm_utcopy_sve_v1.c	Disambiguate whilelt	2 years ago

ztrsm_lncopy_sve.c	Disambiguate whilelt	2 years ago

ztrsm_ltcopy_sve.c	Disambiguate whilelt	2 years ago

ztrsm_uncopy_sve.c	Disambiguate whilelt	2 years ago

ztrsm_utcopy_sve.c	Disambiguate whilelt	2 years ago

OpenBLAS is an optimized BLAS library based on GotoBLAS2 1.13 BSD version.

C Fortran Motorola 68K Assembly Makefile Unix Assembly other

martin@ruby.chemie.uni-freiburg.de traits.zhang@gmail.com wernsaar@googlemail.com xianyi@iscas.ac.cn hank.p.anderson@gmail.com 52632443+wjc404@users.noreply.github.com guxiwei-hf@loongson.cn wangyang.guo@intel.com ashwin.sekhar@cavium.com isuruf@gmail.com chip.kerchner@ibm.com rajis@linux.ibm.com binebrank@gmail.com chris.sidebottom@arm.com wangqian10@iscas.ac.cn arjan@linux.intel.com mhillen@linux.ibm.com m.tk1@hotmail.com 16061801+brada4@users.noreply.github.com ashwin@broadcom.com ralf.gommers@gmail.com bradatajs@yahoo.com xianyi@perfxlab.com jeromerobert@gmx.com zaheer.chothia@gmail.com

Contributors (25+) All

Contributors (25+)
All