Page MenuHomePhabricator

AMDGPU: Raise the priority of MAD24 in instruction selection.
ClosedPublic

Authored by cfang on Jan 15 2019, 2:46 PM.

Details

Summary

We have seen performance regression when v_add3 is generated. The major reason is that the v_mad pattern
is broken when v_add3 is generated. We also see the register pressure increased. While we could not properly
estimate register pressure during instruction selection, we can give mad a higher priority.

In this work, we raise the priority for mad24 in selection and resolve the performance regression.

Diff Detail

Repository
rL LLVM

Event Timeline

cfang created this revision.Jan 15 2019, 2:46 PM
rampitec added inline comments.Jan 15 2019, 2:50 PM
test/CodeGen/AMDGPU/add3.ll
33 ↗(On Diff #181889)

There should be no unnamed variables left.

cfang updated this revision to Diff 181896.Jan 15 2019, 3:05 PM

update the test based on the following suggestion:

There should be no unnamed variables left.

cfang marked an inline comment as done.Jan 15 2019, 3:06 PM
This revision is now accepted and ready to land.Jan 15 2019, 3:07 PM
This revision was automatically updated to reflect the committed changes.