[Allocator 2/5 — after #71] Model GPU fabric, NUMA, MIG, and transfer contention - #73
Open
deehw wants to merge 2 commits into
Open
Conversation
deehw
marked this pull request as draft
September 3, 2026 00:57
deehw
changed the title
deehw
marked this pull request as ready for review
September 3, 2026 03:11
Contributor
Author
|
Qualification update: physically validated on C (2x RTX 4090) and D (1x RTX 4090). This found and fixed idle PCIe generation underestimation; C now reports the native PHB topology at 23.628 GB/s rather than 3.0 GB/s. Focused tests, lint, CI, and the 4,164-test combined suite all pass. Ready for review. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Stacked follow-up to #71. Adds schedulable GPU/MIG device and peer-link wire models; best-effort NVIDIA UUID, MIG, NUMA, PCIe, NVLink, and NIC discovery; joint GPU subset feasibility; new model-profile fabric/NUMA/MIG constraints; and cold-load ranking based on artifact size, transfer bandwidth, and active transfer contention. Advanced requirements fail closed on unknown topology. Includes 522 focused passing tests and operator docs. Base will be retargeted to main after #71 merges.