Open
Conversation
azecevicTT
reviewed
Feb 12, 2026
Comment on lines
+14
to
+17
| // Skip if compute config is already set | ||
| if (srcOp.getComputeConfigAttr()) { | ||
| return failure(); | ||
| } |
Contributor
There was a problem hiding this comment.
Shouldn't the exit condition be HiFi4?
Comment on lines
+32
to
+39
| auto newOp = rewriter.create<ReduceScatterOp>( | ||
| srcOp.getLoc(), srcOp.getResult().getType(), srcOp.getInput(), | ||
| srcOp.getReduceTypeAttr(), srcOp.getScatterDimAttr(), | ||
| srcOp.getClusterAxisAttr(), srcOp.getSubDeviceIdAttr(), | ||
| srcOp.getMemoryConfigAttr(), srcOp.getNumLinksAttr(), | ||
| srcOp.getTopologyAttr(), computeConfig); | ||
|
|
||
| rewriter.replaceOp(srcOp, newOp.getResult()); |
Contributor
There was a problem hiding this comment.
Nit: create followed by replaceOp can be replaced with replaceOpWithNewOp
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Ticket
Link to Github Issue
Problem description
reduce scatter was producing bad results compared to all_gather followed by local reduce. found out it's due to the compute config. Update RS to always use fp_32 accumulation (like local reduce does by default)
What's changed
add WA for reduce scatter to use fp32 accum, pipe it through to runtime. I'll remove change in metal repo once my branch there gets merged and uplifted.
Checklist