In the (cmp eq/ne (uadd.sat X, Y),0) case, we where missing a
hasOneUse check.
Details
Details
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Paths
| Differential D149425
[InstCombine] Don't fold `uadd.sat` to `or` if it increase instruction count ClosedPublic Authored by goldstein.w.n on Apr 27 2023, 11:23 PM.
Details Summary In the (cmp eq/ne (uadd.sat X, Y),0) case, we where missing a
Diff Detail
Event Timelinegoldstein.w.n added a parent revision: D149424: [InstCombine] Add tests `(cmp eq/ne (umax/uadd.sat X, Y), 0)`; NFC.Apr 27 2023, 11:28 PM goldstein.w.n added a child revision: D149426: [InstCombine] Fold `(cmp eq/ne (umax X, Y),0)` -> `(cmp eq/ne (or X, Y),0)`. This revision is now accepted and ready to land.Apr 28 2023, 12:27 AM This revision was landed with ongoing or failed builds.Apr 29 2023, 10:39 AM Closed by commit rGecad53c3f442: [InstCombine] Don't fold `uadd.sat` to `or` if it increase instruction count (authored by goldstein.w.n). · Explain Why This revision was automatically updated to reflect the committed changes.
Revision Contents
Diff 517809 llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
llvm/test/Transforms/InstCombine/cmp-intrinsic.ll
|