[SimplifyCFG] Handle trunc condition in foldBranchToCommonDest. #135490
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-llvm-transforms Author: Andreas Jonson (andjo403) ChangesNoticed regressions due to this missing fold when trying to add the fold proof: https://alive2.llvm.org/ce/z/v32Aof Full diff: https://github.com/llvm/llvm-project/pull/135490.diff 2 Files Affected:
diff --git a/llvm/lib/Transforms/Utils/SimplifyCFG.cpp b/llvm/lib/Transforms/Utils/SimplifyCFG.cpp
index eac7e7c209c95..7f53aa7d4f73d 100644
--- a/llvm/lib/Transforms/Utils/SimplifyCFG.cpp
+++ b/llvm/lib/Transforms/Utils/SimplifyCFG.cpp
@@ -4087,9 +4087,7 @@ bool llvm::foldBranchToCommonDest(BranchInst *BI, DomTreeUpdater *DTU,
Instruction *Cond = dyn_cast<Instruction>(BI->getCondition());
- if (!Cond ||
- (!isa<CmpInst>(Cond) && !isa<BinaryOperator>(Cond) &&
- !isa<SelectInst>(Cond)) ||
+ if (!Cond || !isa<CmpInst, BinaryOperator, SelectInst, TruncInst>(Cond) ||
Cond->getParent() != BB || !Cond->hasOneUse())
return false;
diff --git a/llvm/test/Transforms/SimplifyCFG/fold-branch-to-common-dest.ll b/llvm/test/Transforms/SimplifyCFG/fold-branch-to-common-dest.ll
index 37f48a9a7e03a..7b88ec338cf5e 100644
--- a/llvm/test/Transforms/SimplifyCFG/fold-branch-to-common-dest.ll
+++ b/llvm/test/Transforms/SimplifyCFG/fold-branch-to-common-dest.ll
@@ -88,10 +88,9 @@ define void @one_pred_trunc_cond(i8 %v0, i8 %v1) {
; CHECK-LABEL: @one_pred_trunc_cond(
; CHECK-NEXT: pred:
; CHECK-NEXT: [[C0:%.*]] = icmp eq i8 [[V0:%.*]], 0
-; CHECK-NEXT: br i1 [[C0]], label [[DISPATCH:%.*]], label [[FINAL_RIGHT:%.*]]
-; CHECK: dispatch:
; CHECK-NEXT: [[C1:%.*]] = trunc i8 [[V1:%.*]] to i1
-; CHECK-NEXT: br i1 [[C1]], label [[FINAL_LEFT:%.*]], label [[FINAL_RIGHT]]
+; CHECK-NEXT: [[OR_COND:%.*]] = select i1 [[C0]], i1 [[C1]], i1 false
+; CHECK-NEXT: br i1 [[OR_COND]], label [[FINAL_LEFT:%.*]], label [[FINAL_RIGHT:%.*]]
; CHECK: common.ret:
; CHECK-NEXT: ret void
; CHECK: final_left:
|
|
There are test failures. |
andjo403
force-pushed
the
truncInSimplifyCfg
branch
from
f2bf8e2 to
ca29cef
Compare
|
sorry about that the tests are fixed now |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/55/builds/9813 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/94/builds/6162 Here is the relevant piece of the build log for the reference |
|
Looks like this has negative impact on clang stage2: https://llvm-compile-time-tracker.com/compare.php?from=0f607f3df54e22896b484510f0c1ccfb718de67a&to=ed43207306f7351f2b4f8284710b028df973d74e&stat=instructions:u |
|
So shall I revert it? do I understand correct that it is not the optimization that is slow but it generates a slower binary |
Yes, it generates a slower binary, at least in terms of instructions retired. I think part of the problem here may be that a lot of targets seem to treat all truncates as free? https://llvm.godbolt.org/z/49Tq8qEqe So they don't count towards the bonus inst threshold. A trunc i1 without nuw should have the same cost as an and 1, I think. |
I thought this patch would reduce the number of branches. Perhaps we can adjust |
|
The condition is not included in the cost comparison see: llvm-project/llvm/lib/Transforms/Utils/SimplifyCFG.cpp Lines 4150 to 4152 in da17ced so I do not know if there will be some change to execution time except that less branches will be removed. |
var-const
pushed a commit
to ldionne/llvm-project
that referenced
this pull request
Apr 17, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Noticed regressions due to this missing fold when trying to add the fold
icmp ne (and X, 1), 0 --> trunc X to i1proof: https://alive2.llvm.org/ce/z/v32Aof