[PATCH] D75566: [Matrix] Add initial tiling for load/multiply/store chains.

[Florian Hahn via Phabricator via llvm-commits]

fhahn updated this revision to Diff 252334.
fhahn marked 11 inline comments as done.
fhahn added a comment.



In D75566#1937436 <https://reviews.llvm.org/D75566#1937436>, @anemet wrote:

> I have a few specific comments below but overall it would be great if we could simplify VisitBBFusion to avoid recursion and invalidating the iterator...


Thanks Adam!

I've split the code into 3 parts:

1. Collect all instructions with shape information and fusion candidates.
2. Iterate over fusion candidates and try to fuse. Also collect set of instructions completely eliminated by fusion.
3. Iterate over all instructions with shape info, skip the ones eliminated by fusion and lower the rest.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D75566/new/

https://reviews.llvm.org/D75566

Files:
  llvm/lib/Transforms/Scalar/LowerMatrixIntrinsics.cpp
  llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-multiple-blocks.ll
  llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D75566.252334.patch
Type: text/x-patch
Size: 65255 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200324/926552cc/attachment-0001.bin>