[Float2Int] Fix pessimization in the MinBW calculation. #86051
Merged
Conversation
This file contains 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
The MinBW was being calculated using the significant bits of the upper and lower bounds. The upper bound is 1 past the last value in the range so I don't think it should be included. Instead get the min and max signed values from the range and use the significant bits of those. I'm still not sure if the +1 is needed after the std::max. I haven't investigated if its possible to add a test for this. We only look at whether the MinBW is > 32 right now. So the ranges need to be carefully crafted to see a difference.
|
@llvm/pr-subscribers-llvm-transforms Author: Craig Topper (topperc) ChangesThe MinBW was being calculated using the significant bits of the upper and lower bounds. The upper bound is 1 past the last value in the range so I don't think it should be included. Instead get the min and max signed values from the range and use the significant bits of those. I'm still not sure if the +1 is needed after the std::max. I haven't investigated if its possible to add a test for this. We only look at whether the MinBW is > 32 right now. So the ranges need to be carefully crafted to see a difference. CC: @AtariDreams Full diff: https://github.com/llvm/llvm-project/pull/86051.diff 1 Files Affected:
diff --git a/llvm/lib/Transforms/Scalar/Float2Int.cpp b/llvm/lib/Transforms/Scalar/Float2Int.cpp
index ccca8bcc1a56ac..47042d96b46571 100644
--- a/llvm/lib/Transforms/Scalar/Float2Int.cpp
+++ b/llvm/lib/Transforms/Scalar/Float2Int.cpp
@@ -359,8 +359,8 @@ bool Float2IntPass::validateAndTransform() {
// The number of bits required is the maximum of the upper and
// lower limits, plus one so it can be signed.
- unsigned MinBW = std::max(R.getLower().getSignificantBits(),
- R.getUpper().getSignificantBits()) +
+ unsigned MinBW = std::max(R.getSignedMin().getSignificantBits(),
+ R.getSignedMax().getSignificantBits()) +
1;
LLVM_DEBUG(dbgs() << "F2I: MinBitwidth=" << MinBW << ", R: " << R << "\n");
|
nikic
reviewed
Mar 21, 2024
| unsigned MinBW = std::max(R.getLower().getSignificantBits(), | ||
| R.getUpper().getSignificantBits()) + | ||
| unsigned MinBW = std::max(R.getSignedMin().getSignificantBits(), | ||
| R.getSignedMax().getSignificantBits()) + |
There was a problem hiding this comment.
Use R.getMinSignedBits() instead?
nikic
approved these changes
Mar 21, 2024
chencha3
pushed a commit
to chencha3/llvm-project
that referenced
this pull request
Mar 23, 2024
The MinBW was being calculated using the significant bits of the upper and lower bounds. The upper bound is 1 past the last value in the range so I don't think it should be included. Instead use ConstantRange::getMinSignedBits. I'm still not sure if the +1 is needed after the getMinSignedBits call.
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.
The MinBW was being calculated using the significant bits of the upper and lower bounds. The upper bound is 1 past the last value in the range so I don't think it should be included. Instead get the min and max signed values from the range and use the significant bits of those.
I'm still not sure if the +1 is needed after the std::max.
I haven't investigated if its possible to add a test for this. We only look at whether the MinBW is > 32 right now. So the ranges need to be carefully crafted to see a difference.
CC: @AtariDreams