Skip to content

Revert #4778 (row-level decode cache): no gain over #4777's window cache - #4779

Merged
trissim merged 1 commit into
mainfrom
claude-laptop/revert-row-cache
Nov 15, 2026
Merged

trissim merged 1 commit into
mainfrom
claude-laptop/revert-row-cache

Conversation

@trissim

@trissim trissim commented Nov 15, 2026

Copy link
Copy Markdown
Contributor

Measured the row-level cache installed: observer child ping median ~5.5% across 6 runs, same as #4777's window-level cache (~5.1%) within noise. The row refinement bought nothing — the remaining ~5% is not decode (both caches handle decode; neither gets under 5%), it's the other per-ping work (SQL scan, consumption/responses reads, sidebar rebuild, IPC). So #4778 was added complexity (per-row LRU + a loosened test) for zero measured gain; reverting to #4777's simpler window cache, same result with less machinery. Getting under 5% needs profiling the non-decode per-ping cost — a separate task.

🤖 Generated with Claude Code

https://claude.ai/code/session_01XbFMSnLEc51gw21bLFFRK1

@trissim
trissim merged commit 2dd64b7 into main Nov 15, 2026
1 of 3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant