feat(artifact): add wait_for_upload_headroom producer-side backpressure hook#2721
Open
Farhad-Shabani wants to merge 1 commit into
Open
feat(artifact): add wait_for_upload_headroom producer-side backpressure hook#2721Farhad-Shabani wants to merge 1 commit into
wait_for_upload_headroom producer-side backpressure hook#2721Farhad-Shabani wants to merge 1 commit into
Conversation
Farhad-Shabani
force-pushed
the
farhad/redis-artifact-backpressure
branch
from
d1ef1ee to
fc92022
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Adds one defaulted trait method to
ArtifactClientand one call to it increate_core_proving_task, right before the trace-chunkupload().Why
External provers running sp1-cluster with
NODE_ARTIFACT_STORE=redisare hitting Redis OOM on large requests (recent repro: 278 B cycles / 266 B PGUs, ~8 k shards, 512 GB Redis)Root cause
the local node's implicit backpressure (
ProverSemaphore) is lost when sp1-cluster distributes producer and consumer across a gRPC boundary.SendSpliceWorkeruploads a ~200 MB trace chunk to Redis, callssubmit_task, the coordinator accepts the row in milliseconds and returns, the producer loops. Redis fills faster than GPUs drain it.Fix
Restores backpressure by giving the artifact store a way to say "I'm full, slow down." The executor asks before every upload; the store answers based on its own state.
Companion PR: succinctlabs/sp1-cluster#84