[llvm] ed90f15 - Revert "[InstSimplify][EarlyCSE] Try to CSE PHI nodes in the same basic block"

Owen Anderson via llvm-commits llvm-commits at lists.llvm.org
Sat Aug 29 00:10:02 PDT 2020



> On Aug 28, 2020, at 11:24 PM, Roman Lebedev <lebedev.ri at gmail.com> wrote:
> 
> On Sat, Aug 29, 2020 at 2:43 AM Owen Anderson via llvm-commits
> <llvm-commits at lists.llvm.org <mailto:llvm-commits at lists.llvm.org>> wrote:
>> 
>> 
>> Author: Owen Anderson
>> Date: 2020-08-28T23:43:42Z
>> New Revision: ed90f15efb40d26b5d3ead3bb8e9e284218e0186
>> 
>> URL: https://github.com/llvm/llvm-project/commit/ed90f15efb40d26b5d3ead3bb8e9e284218e0186
>> DIFF: https://github.com/llvm/llvm-project/commit/ed90f15efb40d26b5d3ead3bb8e9e284218e0186.diff
>> 
>> LOG: Revert "[InstSimplify][EarlyCSE] Try to CSE PHI nodes in the same basic block"
>> 
>> This reverts commit 6102310d814ad73eab60a88b21dd70874f7a056f.  It
>> appears to cause compilation non-determinism and caused stage3
>> mismatches.
> Thank you, this is very interesting.
> 
> I don't suppose there is any actuable reproducer?

I don’t have a small one, no.  I was building clang stage3 and getting a mismatch with stage2.

—Owen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200829/6e0acb51/attachment.html>


More information about the llvm-commits mailing list