[PATCH] D98958: [NewPM] Disable non-trivial loop-unswitch on targets with divergence

Thu Mar 25 21:59:00 PDT 2021

sameerds added a subscriber: rsmith.
sameerds added inline comments.

================
Comment at: llvm/test/Transforms/LoopUnswitch/AMDGPU/uniform-unswitch.ll:2
+; RUN: opt -mtriple=amdgcn-- -O3 -S %s | FileCheck %s
+; XFAIL: *

----------------
aeubanks wrote:
> RKSimon wrote:
> > tra wrote:
> > > Please add a comment/TODO explaining that the unswitch is expected to happen in principle, but we still need <something> to make it work.
> > @sameerds This is breaking a windows buildbot as its actually passing.....
> > 
> > http://lab.llvm.org:8011/#/builders/123/builds/3602
> If I had to guess, some bots are still using the legacy PM.
> This test should either use `-passes=default<O3>` (new PM) or `-O3 -enable-new-pm=0` (legacy PM).
Thanks to Richard Smith (@rsmith ?) for fixing this. The test was updated to force -enable-new-pm=1.

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98958/new/

https://reviews.llvm.org/D98958