[PATCH] D90681: [CostModel] fix cost calc bug for sadd/ssub with overflow
Sanjay Patel via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Nov 3 06:48:05 PST 2020
spatel added a comment.
In D90681#2371131 <https://reviews.llvm.org/D90681#2371131>, @samparker wrote:
> Thanks for doing this. But now I'm surprised to not see any Arm changes?
I double-checked, and nothing there is wiggling. So I partly stepped through the same v4i32 example using `-mtriple=thumbv8m.main -mattr=mve`, and there's apparently no difference in the underlying throughput costs of the (mis-specified type) cmp and sel. Bugs on top of bugs?
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D90681/new/
https://reviews.llvm.org/D90681
More information about the llvm-commits
mailing list