[all-commits] [llvm/llvm-project] f2ccf8: [LICM] Fold associative binary ops to promote cod...

[Ricardo Jesus via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: f2ccf80136a01ca69f766becafb329db6c54c0c8
      https://github.com/llvm/llvm-project/commit/f2ccf80136a01ca69f766becafb329db6c54c0c8
  Author: Ricardo Jesus <rjj at nvidia.com>
  Date:   2024-07-23 (Tue, 23 Jul 2024)

  Changed paths:
    M llvm/lib/Transforms/Scalar/LICM.cpp
    M llvm/test/CodeGen/PowerPC/common-chain.ll
    M llvm/test/CodeGen/PowerPC/p10-spill-crlt.ll
    A llvm/test/Transforms/LICM/hoist-binop.ll
    M llvm/test/Transforms/LICM/sink-foldable.ll
    M llvm/test/Transforms/LICM/update-scev-after-hoist.ll

  Log Message:
  -----------
   [LICM] Fold associative binary ops to promote code hoisting  (#81608)

Perform the transformation

  "(LV op C1) op C2" ==> "LV op (C1 op C2)"

where op is an associative binary op, LV is a loop variant, and C1 and
C2 are loop invariants to hoist.

Similar patterns could be folded (left in comment) but this one seems to
be the most impactful.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications