chore(configs): router replay perf repro#2614
Draft
mikasenghaas wants to merge 14 commits into
Draft
Conversation
Configs used to reproduce and iterate on the R3 orchestrator event-loop lag investigation (router replay, Qwen3-30B-A3B, 2 train / 1 infer / 4 replicas). Inline comments capture the per-round findings (v5 process pool, v6 gather_chunk_size sweep, v7 wandb samples disable, v9 raw rollout dump disable). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Removes use_token_client, use_process_pool, dump_raw_rollouts, and gather_chunk_size — these were experimental knobs used during the R3 lag investigation that never landed in OrchestratorConfig on main. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
mikasenghaas
changed the title
…er_node Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…uffer Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…epro Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Shared [model] section already propagates name to inference; declaring inference.model.name as well now fails the RLConfig shared/sub-config collision validator. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…is the default #2615 made X-Session-ID = trajectory_id the default sticky-routing header, so the explicit override here is redundant. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Required by configs/r3_perf/debug.toml — the orchestrator crashes at load_environment without it being installed. Co-Authored-By: Claude Opus 4.7 (1M context) <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
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.
Summary
configs/r3_perf/debug.toml, the repro config used to investigate the R3 orchestrator event-loop lag on Qwen3-30B-A3B with router replay (2 train / 1 infer / 4 replicas).🤖 Generated with Claude Code