Page MenuHomePhabricator

[mlir][tensor] Move BufferizableOpInterface impl to tensor dialect
ClosedPublic

Authored by springerm on Jan 23 2022, 7:38 AM.

Details

Summary

This is in preparation of unifying the existing bufferization with One-Shot bufferization.

A subsequent commit will replace tensor-bufferize's implementation with the BufferizableOpInterface-based implementation and move over missing test cases.

This commit also introduces a new AlwaysCopyBufferizationState that will be used for partial bufferization and always copies tensor OpOperands that bufferize to memory writes. We will start using this in a subsequent commit.

Depends On D117775

Diff Detail

Event Timeline

springerm created this revision.Jan 23 2022, 7:38 AM
springerm requested review of this revision.Jan 23 2022, 7:38 AM
Herald added a project: Restricted Project. · View Herald TranscriptJan 23 2022, 7:38 AM
springerm edited the summary of this revision. (Show Details)Jan 24 2022, 1:01 AM
springerm added a reviewer: pifon2a.
mlir/lib/Dialect/Bufferization/Transforms/Bufferize.cpp
220

This seems like this should be separated and tested ?

This revision was not accepted when it landed; it landed in state Needs Review.Jan 24 2022, 6:56 AM
This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.