P5: placement breadth — artifact locality, load, NVLink affinity - #41
Merged
Conversation
…finity select_placement now consumes the full capacity signals nodes report: prefers nodes holding the artifact (locality) > lower running-engine load > more headroom, and for count>1 requires the devices to share an NVLink group when groups are declared. POST /admin/place accepts count and artifact_digest. Completes the Phase-5 placement decision on the control side. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Completes the Phase-5 placement decision on the control side. Pairs with meridian-node PR #12 (the node reports these signals).
What
select_placement(required_vram_bytes, count=1, artifact_digest=None)now consumes the full capacity signals:held_artifactsinclude the digest.running_engines.count>1, the devices must share an NVLink group when groups are declared (_pick_devices).Preference order: locality > lower load > more headroom.
POST /admin/placeacceptscountandartifact_digest.Tests
pytest tests/control -q→ 34 passed (+3: locality beats headroom, load tie-break, multi-GPU requires a shared NVLink group); existing placement tests unchanged; ruff + mypy clean.🤖 Generated with Claude Code