[all-commits] [llvm/llvm-project] a75e06: [LV] Use vscale for tuning when updating profile i...

[David Sherwood via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: a75e0627f97ccc36ec222a53c6a1106157a380ac
      https://github.com/llvm/llvm-project/commit/a75e0627f97ccc36ec222a53c6a1106157a380ac
  Author: David Sherwood <david.sherwood at arm.com>
  Date:   2025-06-16 (Mon, 16 Jun 2025)

  Changed paths:
    M llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
    A llvm/test/Transforms/LoopVectorize/AArch64/check-prof-info.ll
    M llvm/test/Transforms/LoopVectorize/check-prof-info.ll

  Log Message:
  -----------
  [LV] Use vscale for tuning when updating profile information (#143690)

In fixVectorizedLoop we call setProfileInfoAfterUnrolling to update the
profile information after vectorising, however for scalable VFs we
pessimistically assume vscale=1. We can improve upon this by using the
value of vscale used for tuning, i.e. when targeting neoverse-v1 the
expected value is 2.



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