[all-commits] [llvm/llvm-project] 34a321: [LV][NFC] Remove instcombine pass from RUN lines i...

[David Sherwood via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 34a321a74fdf87c3e3f5e80d95d581629513441d
      https://github.com/llvm/llvm-project/commit/34a321a74fdf87c3e3f5e80d95d581629513441d
  Author: David Sherwood <david.sherwood at arm.com>
  Date:   2026-06-11 (Thu, 11 Jun 2026)

  Changed paths:
    M llvm/test/Transforms/LoopVectorize/ARM/mve-qabs.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-reduction-predselect.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-reductions-interleave.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-reductions.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-saddsatcost.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-selectandorcost.ll
    M llvm/test/Transforms/LoopVectorize/ARM/pointer_iv.ll
    M llvm/test/Transforms/LoopVectorize/ARM/tail-fold-multiple-icmps.ll

  Log Message:
  -----------
  [LV][NFC] Remove instcombine pass from RUN lines in ARM tests (#202913)

Following on from PR #197448 I've now removed the instcombine pass from
RUN lines in the ARM test directory, which exposes some potential
missing optimisations in vplan:

1. We could be folding IR into saturating math intrinsic calls to better
reflect the cost.
2. Masked load + select -> masked load with different passthru.
3. icmp + select -> smin/smax.

Some of these were already observed in #197448



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