[all-commits] [llvm/llvm-project] 33de41: Partial revert "[NVPTX] Enhance mul.wide and mad.w...

[Justin Fargnoli via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 33de41988fcac8c3e1475fbe9a04ddb0375ba67d
      https://github.com/llvm/llvm-project/commit/33de41988fcac8c3e1475fbe9a04ddb0375ba67d
  Author: Justin Fargnoli <jfargnoli at nvidia.com>
  Date:   2025-08-25 (Mon, 25 Aug 2025)

  Changed paths:
    M llvm/lib/Target/NVPTX/NVPTXInstrInfo.td
    M llvm/test/CodeGen/NVPTX/bug26185-2.ll
    M llvm/test/CodeGen/NVPTX/combine-wide.ll
    M llvm/test/CodeGen/NVPTX/local-stack-frame.ll
    M llvm/test/CodeGen/NVPTX/vector-loads.ll

  Log Message:
  -----------
  Partial revert "[NVPTX] Enhance mul.wide and mad.wide peepholes #150477" (#155024)

Fix
https://github.com/llvm/llvm-project/pull/150477#issuecomment-3191367837

Undo `add (mul.wide a, b), c` -> `mad.wide a, b, c` instruction
selection pattern, but preserve `mul.wide` DAG combine.



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