Repository navigation
docs: record the first performance baseline measured on this box - #2586
Merged
Merged
Conversation
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #2585
Owner: Dev
Reviewer: Reviewer
The first-baseline section said Not measured — this host cannot build the deepseek-harness checkout, and its table was all
—. Both halves of that blocker are gone (#2571 for the bench, #2577 for the script default message), so this replaces the placeholder with the table the script actually prints.What changed:
docs/performance-baseline.mdonly — the placeholder paragraph and its "table to fill" block are replaced by the measured table plus the machine / commit / node / date line, and the earlier different-prompt run is kept as a one-line reference so nobody compares two prompts as one series.Ran at this tip:
core/dev/bench.sh perf 3098 8798 reset→PERF_COMMIT=87b2e08b… node dev/perf-baseline.mjs perf 3098 --out /tmp/perf87b2.json→ exit 0, 20/20 measured turns each produced a tool call →dev/bench.sh perf 3098 8798 stop. No test suite is affected by a docs-only change;node .github/scripts/changelog.mjs checknot needed (no fragment, docs-only).Docs-not-needed: docs/performance-baseline.md is the document being corrected; no subsystem page owns it and no source path changes.
Checkpoint
{
"pr": 2586,
"claim": "Dev",
"issue": 2585,
"summary": "Replaces the "not measured" placeholder in docs/performance-baseline.md with the baseline the bench now prints: 20-turn table for turn latency, tool-call latency, resident RSS and plugin load, plus the machine, commit, node and date line. The earlier different-prompt run stays as one reference line so two prompts are not read as one series. Docs only: no code, no threshold, no gate.",
"files": [
"docs/performance-baseline.md"
],
"checks": {
"ran": "core/dev/bench.sh perf 3098 8798 reset; PERF_COMMIT=87b2e08b node dev/perf-baseline.mjs perf 3098 --out /tmp/perf87b2.json; dev/bench.sh perf 3098 8798 stop",
"result": "green",
"at": "bf359c857eb1f056b173f9a755877d560e349131"
},
"next": "none — ready for review"
}