{{ message }}
Repository navigation
Conversation
Open
3 of 4 tasks
Manfredss
force-pushed
the
fix/session-lazy-mamba-slots
branch
from
September 27, 2026 00:00
f2e8786 to
568cc37
Compare
Contributor
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Lazy Mamba buffers contain unused entries marked
-1. Closing or aborting a streaming session returned these entries to the pool, so 8 slots could become 9 free slots, including an invalid one.Related #39838. The report's private 8-GPU workload has not been reproduced.
Modifications
Skip unused entries when counting and releasing session buffers. Rebased on current main and updated the regression tests to use the current request release path.
Accuracy Tests
22 related CPU tests and 8 subtests pass. On current main without the fix, all 4 lazy-buffer regression cases fail with
9 != 8; regular buffers pass.Speed Tests and Profiling
Not applicable to this buffer cleanup change; model calculations are unchanged.
Checklist
CI States
Latest PR Test (Base): ❌ Run #36281197970
Latest PR Test (Extra): ❌ Run #36281197906
Latest PR Test (AMD ROCm 10): ❌ Run #36281198004