[PATCH] D93411: [GlobalISel] Transform sext (cmp pred, x, y) -> select (cmp pred, x, y) tval, 0
Matt Arsenault via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Wed Dec 16 19:14:53 PST 2020
- Previous message: [PATCH] D93411: [GlobalISel] Transform sext (cmp pred, x, y) -> select (cmp pred, x, y) tval, 0
- Next message: [PATCH] D93411: [GlobalISel] Transform sext (cmp pred, x, y) -> select (cmp pred, x, y) tval, 0
- Messages sorted by:
[ date ]
[ thread ]
[ subject ]
[ author ]
arsenm added inline comments.
================
Comment at: llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp:3151
+ LLT ExtTy = MRI.getType(MI.getOperand(0).getReg());
+ if (ExtTy.isVector() ||
+ getTargetLowering().convertSelectOfConstantsToMath(ExtTy))
----------------
Vectors should be easy to handle?
================
Comment at: llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp:3152
+ if (ExtTy.isVector() ||
+ getTargetLowering().convertSelectOfConstantsToMath(ExtTy))
+ return false;
----------------
This doesn't check the inputs are constants? The name seems wrong
================
Comment at: llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp:3163-3173
+ // Depending on the result type of the compare, we may need a different
+ // true value.
+ // E.g. let's say we have
+ //
+ // %cmp:_(s32) = G_ICMP pred, %x, %y
+ // %sext:_(s64) = G_SEXT %cmp
+ //
----------------
This comment is pretty wordy for what it's explaining. I would expect this to defer to somewhere else to explain the concept of the bool true value
================
Comment at: llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp:3173
+ //
+ // If %cmp produced a s1, however, then %sext would be -1.
+ MatchInfo = (CmpTy == LLT::scalar(1))
----------------
s/a/an/
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D93411/new/
https://reviews.llvm.org/D93411
- Previous message: [PATCH] D93411: [GlobalISel] Transform sext (cmp pred, x, y) -> select (cmp pred, x, y) tval, 0
- Next message: [PATCH] D93411: [GlobalISel] Transform sext (cmp pred, x, y) -> select (cmp pred, x, y) tval, 0
- Messages sorted by:
[ date ]
[ thread ]
[ subject ]
[ author ]
More information about the llvm-commits
mailing list