UCP/DEVICE: Fix lane selection - #11815
Open
ofirfarjun7 wants to merge 1 commit into
Open
Conversation
Co-authored-by: OpenAI Codex <codex@openai.com>
ofirfarjun7
changed the title
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What?
Fix device memory-list lane selection to use only transport resources that are both present on the element’s endpoint and compatible with its remote key.
Why?
The device transport bitmap is collected from the worker context and can contain resources not selected for a specific endpoint. Previously, one unavailable resource caused UCX to reject the entire LKEY or no-LKEY lane class, even when another compatible lane was available. This resulted in lane not found for element errors.
How?
Build a per-element selected transport bitmap by filtering the context candidates:
Skip resources that are not endpoint lanes.
Skip endpoint lanes whose destination MD is absent from the element’s remote-key MD map.
Pack only the remaining lanes.
Stop packing once the allocated number of lanes has been filled.