projects / platform / upstream / openblas.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: aa21cb5) | patch
POWER10: Change dgemm unroll factors
authorRajalakshmi Srinivasaraghavan <rajis@linux.ibm.com>
Sat, 31 Oct 2020 23:28:57 +0000 (18:28 -0500)
committerRajalakshmi Srinivasaraghavan <rajis@linux.ibm.com>
Sat, 31 Oct 2020 23:28:57 +0000 (18:28 -0500)
commitdd7a9cc5bf6b926a44b38d13366743691fd6e604
tree1d23f4eec0ca6be02731e2e1963dd172e7d53939tree | snapshot
parentaa21cb52179b86b00f7ac52a4e41ed712836f2d1commit | diff
POWER10:  Change dgemm unroll factors

Changing the unroll factors for dgemm to 8 shows improved performance with
POWER10 MMA feature.   Also made some minor changes in sgemm for edge cases.
kernel/power/KERNEL.POWER10 diff | blob | history
kernel/power/dgemm_kernel_power10.c diff | blob | history
kernel/power/dgemm_ncopy_8_power10.c [new file with mode: 0644] blob
kernel/power/sgemm_kernel_power10.c diff | blob | history
param.h diff | blob | history
Domain: Machine Learning / ML Framework;