Set SWT-bench cache-mode default to max#547
Merged
simonrosenberg merged 1 commit intomainfrom Mar 20, 2026
Merged
Conversation
Once SDK PR software-agent-sdk#2522 lands (ARG ordering fix), registry cache will provide cross-SDK-bump cache hits for apt-get and npm layers, cutting per-image build time by ~50%. Refs: #544
all-hands-bot
approved these changes
Mar 20, 2026
Collaborator
all-hands-bot
left a comment
all-hands-bot
left a comment
There was a problem hiding this comment.
🟢 Good taste - This is textbook pragmatic engineering.
What's Right:
- Solving a real, measured problem (2.1x build speedup: 154s vs 322s)
- Minimal change (2 lines)
- Dependency verified (SDK PR #2522 merged 6 minutes ago)
- A/B tested before changing default
VERDICT: ✅ Worth merging immediately
KEY INSIGHT: This is what good engineering looks like - measure the problem, fix the root cause upstream (ARG ordering), validate the improvement, then flip the default. No complexity, no over-engineering, just results.
This was referenced Mar 20, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Change the default
cache-modefor SWT-bench image builds fromofftomax.Why
With the ARG ordering fix in SDK PR #2522, registry cache will provide cross-SDK-bump cache hits for
apt-getandnpm installlayers. But this only works if cache is actually exported (cache-mode=max).The
offdefault was set in PR #541 because registry cache had a 100% miss rate — the ARG before apt-get made every layer hash SDK-specific. With the ARG moved after the expensive layers, the cache tags are stable across SDK bumps and registry cache becomes effective.Validated
A/B test (#544):
cache-mode=max+ ARG fix → apt-get CACHED from registry, 2.1x faster per image (154s vs 322s).Depends on
SDK PR #2522 (ARG ordering fix). Without it,
cache-mode=maxadds export overhead with no cache benefit.