Replace two reduces with explicit loops in SIMD implementations.
#72423
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
`reduce` has enough machinery that it pushes us over some inlining thresholds before optimizations happen, resulting in much worse codegen for String breadcrumbs. So a nice pickup there, but also generally applicable to other code as well.
|
@swift-ci smoke test |
|
@swift-ci smoke-test |
glessard
reviewed
Mar 19, 2024
| @@ -927,7 +931,13 @@ extension SIMD where Scalar: FloatingPoint { | |||
| // llvm.experimental.vector.reduce.fadd or an explicit tree-sum. Open- | |||
There was a problem hiding this comment.
Nit to past-you: The first line of the implementation note has an elided word, "this eventually" probably needs a "will" or "should" somewhere in there.
glessard
approved these changes
Mar 19, 2024
|
@swift-ci smoke test |
Catfish-Man
approved these changes
Mar 20, 2024
|
We don't want this to be always inline, because it blows out code size if inlined into contexts where it can't be specialized. We really want some sort of "inline if-and-only-if specialized" annotation for these, but I also don't want to further complicate the inlining story... |
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Outdated suggestions cannot be applied.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.
reducehas enough machinery that it pushes us over some inlining thresholds before optimizations happen, resulting in much worse codegen for String breadcrumbs. So a nice pickup there, but also generally applicable to other code as well.split out from #72205