[all-commits] [llvm/llvm-project] 337936: [LV] Ignore some costs when loop gets fully unroll...

[Igor Kirillov via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 337936a83bc9ffaba67658245a8c9bef159584a3
      https://github.com/llvm/llvm-project/commit/337936a83bc9ffaba67658245a8c9bef159584a3
  Author: Igor Kirillov <igor.kirillov at arm.com>
  Date:   2024-12-09 (Mon, 09 Dec 2024)

  Changed paths:
    M llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
    M llvm/test/Transforms/LoopVectorize/AArch64/fully-unrolled-cost.ll

  Log Message:
  -----------
  [LV] Ignore some costs when loop gets fully unrolled (#106699)

When VF has a fixed width and equals the number of iterations, and we are not
tail folding by masking, comparison instruction and induction operation will be DCEed later.
Ignoring the costs of these instructions improves the cost model.



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