fix(bench): grid compression cache stores the promise — one call per key#247
Merged
Conversation
retask() compresses N holdout tasks concurrently; the check-then-set value cache raced, firing N duplicate compression calls per cell, and any single blip (one empty response among twelve) killed the run via the degenerate- output guard. The cache now stores the promise (exactly one compression per unique prompt, shared by all concurrent callers) with one named retry; artifact prompt persistence awaits the promises.
tangletools
approved these changes
Jun 10, 2026
tangletools
left a comment
tangletools
left a comment
Contributor
There was a problem hiding this comment.
✅ Auto-approved PR — 4107e7f9
Blanket team auto-approval is enabled for this reviewer service.
The full PR reviewer audit still runs separately and will publish findings if it detects issues.
tangletools · auto-approval · reason: blanket_auto_approve · 2026-06-10T22:13:37Z
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Grid run #2 failed loud exactly as designed (#246's guard) — and the autopsy of the failure found the real bug in MY code, not the model:
retask()compresses 12 holdout tasks viaPromise.all, the check-then-set value cache raced, 12 duplicate compression calls fired per κ cell, and one blipped empty → guard threw. (Probe verified flash compresses this prompt fine in isolation.)The cache now stores the promise: exactly one compression call per unique (γ, κ, prompt) key, shared by every concurrent caller, with one named retry on a degenerate first attempt. Artifact persistence awaits the cached promises.
Bench typecheck ✓. Run #2 relaunched on this code.