This is an archive of the discontinued LLVM Phabricator instance.

AMDGPU: Restrict extract_vector_elt combine to loads
ClosedPublic

Authored by arsenm on Aug 31 2018, 7:24 AM.

Details

Reviewers
rampitec
Summary

The intention is to enable the extract_vector_elt load combine,
and doing this for other operations interferes with more
useful optimizations on vectors.

Handle any type of load since in principle we should do the
same combine for the various load intrinsics.

Diff Detail

Event Timeline

arsenm created this revision.Aug 31 2018, 7:24 AM
This revision is now accepted and ready to land.Aug 31 2018, 8:21 AM
arsenm closed this revision.Aug 31 2018, 8:42 AM

r341219