This is an archive of the discontinued LLVM Phabricator instance.

AMDGPU/GlobalISel: Improve lowering of G_SEXT_INREG
ClosedPublic

Authored by arsenm on Jan 13 2020, 5:41 AM.

Details

Reviewers
nhaehnle
kerbowa
Summary

Clamping the scalar is much better than lowering with superwide shifts
for types > s64.

Diff Detail

Event Timeline

arsenm created this revision.Jan 13 2020, 5:41 AM
kerbowa accepted this revision.Jan 14 2020, 6:21 AM

LGTM

This revision is now accepted and ready to land.Jan 14 2020, 6:21 AM