[all-commits] [llvm/llvm-project] b64150: [LoopUnroll] Introduce parallel accumulators when ...

[Julian Nagele via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: b64150963733d04e524f501cdd736963a7a3ba6d
      https://github.com/llvm/llvm-project/commit/b64150963733d04e524f501cdd736963a7a3ba6d
  Author: Julian Nagele <j.nagele at apple.com>
  Date:   2025-11-27 (Thu, 27 Nov 2025)

  Changed paths:
    M llvm/lib/Transforms/Utils/LoopUnroll.cpp
    M llvm/test/Transforms/LoopUnroll/partial-unroll-reductions.ll
    M llvm/test/Transforms/LoopUnroll/runtime-unroll-reductions.ll

  Log Message:
  -----------
  [LoopUnroll] Introduce parallel accumulators when unrolling FP reductions. (#166630)

This is building on top of
https://github.com/llvm/llvm-project/pull/149470, also introducing
parallel accumulator PHIs when the reduction is for floating points,
provided we have the reassoc flag. See also
https://github.com/llvm/llvm-project/pull/166353, which aims to
introduce parallel accumulators for reductions with vector instructions.



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