This is an archive of the discontinued LLVM Phabricator instance.

[mlir][bufferize] Improve resolveConflicts for ExtractSliceOp
ClosedPublic

Authored by springerm on Jun 1 2022, 1:24 PM.

Details

Summary

It is sometimes better to make a copy of the OpResult instead of making a copy of the OpOperand. E.g., when bufferizing tensor.extract_slice.

This implementation will eventually make parts of extract_slice's bufferize implementation obsolete (and simplify it). It will only need to handle in-place OpOperands.

Depends On D126817

Diff Detail

Event Timeline

springerm created this revision.Jun 1 2022, 1:24 PM
Herald added a project: Restricted Project. · View Herald TranscriptJun 1 2022, 1:24 PM
springerm requested review of this revision.Jun 1 2022, 1:24 PM
Herald added a project: Restricted Project. · View Herald TranscriptJun 1 2022, 1:24 PM
springerm retitled this revision from [mlir][tensor][bufferize] Implement resolveConflicts for ExtractSliceOp to [mlir][bufferize] Improve resolveConflicts for ExtractSliceOp.Jun 8 2022, 7:03 AM
springerm edited the summary of this revision. (Show Details)
aartbik accepted this revision.Jun 9 2022, 8:53 AM
This revision is now accepted and ready to land.Jun 9 2022, 8:53 AM
This revision was landed with ongoing or failed builds.Jun 9 2022, 1:19 PM
This revision was automatically updated to reflect the committed changes.