This is an archive of the discontinued LLVM Phabricator instance.

DAGCombiner: Turn truncate of a bitcasted vector to an extract
ClosedPublic

Authored by arsenm on Feb 12 2016, 10:28 AM.

Details

Summary

On AMDGPU where operations i64 operations are often bitcasted to v2i32
and back, this pattern shows up regularly where it breaks some
expected combines on i64, such as load width reducing.

This fixes some test failures in a future commit when i64 loads
are changed to promote.

Diff Detail

Event Timeline

arsenm updated this revision to Diff 47818.Feb 12 2016, 10:28 AM
arsenm retitled this revision from to DAGCombiner: Turn truncate of a bitcasted vector to an extract.
arsenm updated this object.
arsenm added a subscriber: llvm-commits.
resistor accepted this revision.Mar 1 2016, 12:48 PM
resistor edited edge metadata.

LGTM.

This revision is now accepted and ready to land.Mar 1 2016, 12:48 PM
arsenm closed this revision.Mar 1 2016, 1:36 PM

r262397