Repository navigation
Publish instruction-count baselines in README with a CI freshness gate - #7
Merged
Merged
Conversation
docs/PERFORMANCE.md promised that the deterministic instruction-count numbers would get the same marker-based README treatment as the wall-clock table once the ratchet landed; deliver it. scripts/update_icount_docs.py renders the table 1:1 from bench/baselines.json, and the icount-ratchet job regenerates it and fails on any diff, so the published numbers cannot drift from the gated ones. The Platform support section now also states that Hexagon and Cortex-M55 performance is CI-gated.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes a docs-promise gap found while auditing documentation:
docs/PERFORMANCE.mdstated that the deterministic instruction-count numbers would get the same marker-based README treatment as the wall-clock table "once PR B lands, and CI regenerates them to fail on drift" — PR B (#4) landed, but neither piece existed.scripts/update_icount_docs.pyrenders a README table 1:1 frombench/baselines.jsonbetween<!-- ICOUNT:BEGIN/END -->markers (committed populated: the M55/Hexagon table including the headline 3.3× float-vs-Q15 inversion on Hexagon).icount-ratchetjob regenerates the table andgit diff --exit-codes it — since the table derives entirely from the committed baselines, the check is exact and free, and the published numbers provably cannot drift from the gated ones.docs/PERFORMANCE.md's "Docs freshness" section updated from promise to present tense.No code changes; the table content is the already-reviewed baseline data.
https://claude.ai/code/session_01HuAFfoeD5a5Xe5aGNA16M9
Generated by Claude Code