[all-commits] [llvm/llvm-project] 5b96b1: [SCEV] Improve AddRecs' range computation in Expen...

[Max Kazantsev via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 5b96b13fdffebc1e81cdca3466e75b0d0e584c27
      https://github.com/llvm/llvm-project/commit/5b96b13fdffebc1e81cdca3466e75b0d0e584c27
  Author: Max Kazantsev <mkazantsev at azul.com>
  Date:   2023-04-10 (Mon, 10 Apr 2023)

  Changed paths:
    M llvm/lib/Analysis/ScalarEvolution.cpp
    M llvm/test/Analysis/ScalarEvolution/decrementing_addrecs.ll

  Log Message:
  -----------
  [SCEV] Improve AddRecs' range computation in Expensive Range Sharpening mode

Apply loop guards to AddRec's start in range computation for
non-self-wrapping AddRecs.

According to CT measurements, this has a wide negative compile time impact,
so we hold it in expensive range sharpening mode where it's not so critical.
However, we need to find a way to share benefits of this mode with default mode.

Patch by Aleksandr Popov!

Differential Revision: https://reviews.llvm.org/D147557
Reviewed By: mkazantsev