projects / platform / upstream / openblas.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: bc37284) | patch
Neoverse N2 sbgemm:
authorHonglin Zhu <zhuhonglin.zhl@alibaba-inc.com>
Wed, 22 Jun 2022 15:00:40 +0000 (23:00 +0800)
committerHonglin Zhu <zhuhonglin.zhl@alibaba-inc.com>
Wed, 29 Jun 2022 02:14:21 +0000 (10:14 +0800)
commit123e0dfb62b21f2468c19be6c8415331faa56fd5
tree1692bc238f70ef5910b99c3fcaf6399e2a582863tree | snapshot
parentbc3728475fc329eb29adfb5954864d7763c37284commit | diff
Neoverse N2 sbgemm:

    1. Modify the algorithm to resolve multithreading failures
    2. No memory allocation in sbgemm kernel
    3. Optimize when alpha == 1.0f
kernel/arm64/sbgemm_kernel_8x4_neoversen2.c diff | blob | history
kernel/arm64/sbgemm_kernel_8x4_neoversen2_impl.c [new file with mode: 0644] blob
kernel/arm64/sbgemm_ncopy_neoversen2.c diff | blob | history
kernel/arm64/sbgemm_tcopy_neoversen2.c diff | blob | history
param.h diff | blob | history
Domain: Machine Learning / ML Framework;