diff options
author | Simon Pilgrim <llvm-dev@redking.me.uk> | 2017-05-13 13:42:35 +0000 |
---|---|---|
committer | Simon Pilgrim <llvm-dev@redking.me.uk> | 2017-05-13 13:42:35 +0000 |
commit | ef46c2762acc8f0af01b42b2151dba0d6089aa4e (patch) | |
tree | 66e412db18b482a0223bd48fbf9dbf3de5c92de5 /llvm/lib/CodeGen | |
parent | 7d62e4b455cf181d1657cf8e3be18fd5c0d2eb27 (diff) | |
download | bcm5719-llvm-ef46c2762acc8f0af01b42b2151dba0d6089aa4e.tar.gz bcm5719-llvm-ef46c2762acc8f0af01b42b2151dba0d6089aa4e.zip |
[x86, SSE] AVX1 PR28129 (256-bit all-ones rematerialization)
Further perf tests on Jaguar indicate that:
vxorps %ymm0, %ymm0, %ymm0
vcmpps $15, %ymm0, %ymm0, %ymm0
is consistently faster (by about 9%) than:
vpcmpeqd %xmm0, %xmm0, %xmm0
vinsertf128 $1, %xmm0, %ymm0, %ymm0
Testing equivalent code on a SandyBridge (E5-2640) puts it slightly (~3%) faster as well.
Committed on behalf of @dtemirbulatov
Differential Revision: https://reviews.llvm.org/D32416
llvm-svn: 302989
Diffstat (limited to 'llvm/lib/CodeGen')
0 files changed, 0 insertions, 0 deletions