[rocjitsu] Fix gfx1250 flat private scratch decoding#7781
Merged
Conversation
0608edb to
6537c76
Compare
atgutier
approved these changes
Jun 24, 2026
atgutier
left a comment
Contributor
There was a problem hiding this comment.
LGMT, but some comments that may need to be addressed.
6537c76 to
3abd3cb
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
Fix gfx1250 flat private scratch address decoding.
Technical Details
LLVM lowers private-to-flat scratch addresses using gfx1250 lane bits in the flat scratch address encoding. We were treating those addresses as ordinary global addresses, so nonzero lanes read from the wrong backing location and the HIP repro returned incorrect values.
Add gfx1250 flat private address translation that strips the encoded lane bits and maps the private offset into RocJITsu’s per-lane scratch backing layout, with a regression test covering the lane-bit decoding path.
JIRA ID
N/A
Test Plan
Unit tests + repro in https://gist.github.com/lua1235/62fa1506c584b454e3135406eea75c42 + fpsan suite.
No generator changes.
Test Result
All good.
Submission Checklist