projects / platform / upstream / openblas.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 87315e8) | patch
POWER10: Improve copy performance
authorRajalakshmi Srinivasaraghavan <rajis@linux.ibm.com>
Sun, 13 Dec 2020 16:41:45 +0000 (10:41 -0600)
committerRajalakshmi Srinivasaraghavan <rajis@linux.ibm.com>
Sun, 13 Dec 2020 16:41:45 +0000 (10:41 -0600)
commit2fb11f873bfb5d690cbe096d81a837ede4cfa63f
tree8995d985da3a8a8b817a45c0e63d463f177ee179tree | snapshot
parent87315e8a8d1f27684d886c31742d95d98886aa8acommit | diff
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.
kernel/power/ccopy_microk_power10.c [new file with mode: 0644] blob
kernel/power/ccopy_power10.c diff | blob | history
kernel/power/copy_microk_power10.c diff | blob | history
kernel/power/dcopy_power10.c diff | blob | history
kernel/power/scopy_power10.c diff | blob | history
Domain: Machine Learning / ML Framework;