[PATCH] D79547: [MLIR] Fix affine fusion bug/efficiency issue / enable more fusion

[Andy Davis via Phabricator via llvm-commits]

andydavis1 accepted this revision.
andydavis1 added a comment.
This revision is now accepted and ready to land.

Thanks Uday! Looks good...



================
Comment at: mlir/lib/Transforms/LoopFusion.cpp:1628
               auto loadMemRef = cast<AffineLoadOp>(loadOpInst).getMemRef();
-              if (visitedMemrefs.count(loadMemRef) == 0)
+              // NOTE: Change 'loads' to a hash set in case efficiency is an
+              // issue. We still use a vector since it's expected to be small.
----------------
Is this a TODO: consider using a hash set if performance becomes an issue?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D79547/new/

https://reviews.llvm.org/D79547