[PATCH] D126772: [LoopVectorize] Add support for invariant stores of ordered reductions

[David Sherwood via Phabricator via llvm-commits]

david-arm accepted this revision.
david-arm added a comment.
This revision is now accepted and ready to land.

LGTM! From private discussions I also understand you've run the LLVM test suite using scalable vectorisation on SVE hardware and all tests passed. Can you address my test comment before merging please?



================
Comment at: llvm/test/Transforms/LoopVectorize/AArch64/strict-fadd.ll:1371
+; CHECK-ORDERED: middle.block
+; CHECK-ORDERED: store float %4, float* %[[DEST_PTR]]
+; CHECK-ORDERED: for.body
----------------
I think this should be something like

  store float %[[RDX]], float* %[[DEST_PTR]]

because otherwise the tests might break if the numbering changes in future.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D126772/new/

https://reviews.llvm.org/D126772