projects / platform / upstream / openblas.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 00dc09a) | patch
Use sgemm_ncopy_4_skylakex.c also for Haswell
authorArjan van de Ven <arjan@linux.intel.com>
Sat, 15 Dec 2018 13:43:07 +0000 (13:43 +0000)
committerArjan van de Ven <arjan@linux.intel.com>
Sat, 15 Dec 2018 13:49:19 +0000 (13:49 +0000)
commit0586899a10b97bf1baf50e4988d18b4268317420
tree9d1136ce5e63e5fb1abd3f0827e35387a1ccce3dtree | snapshot
parent00dc09ad198aedec53fd05ea1b13d72d7a9a517acommit | diff
Use sgemm_ncopy_4_skylakex.c also for Haswell

sgemm_ncopy_4_skylakex.c uses SSE transpose operations where the
real perf win happens; this also works great for Haswell.

This gives double digit percentage gains on small and skinny matrices
kernel/x86_64/KERNEL.HASWELL diff | blob | history
kernel/x86_64/sgemm_ncopy_4_skylakex.c diff | blob | history
Domain: Machine Learning / ML Framework;