Fix lock acquisition - #24
Open
colincoleman wants to merge 1 commit into
Open
Conversation
compute_update() hill-climbs the analysis-window size: from the last locked window it moves to a larger (more accurate) one while the lock holds, and drops to a smaller one when it stops. But from a cold start it began at the smallest (2s) window and, if that failed, dropped to -1 and gave up - it never tried the larger windows. A weak or jittery watch produces a cleaner, lower-jitter period in a larger window (more cycles to average), so it often can only lock at 8s or 16s, never at 2s. Such a watch therefore "usually doesn't register", yet "once locked, stays locked" - because acquisition needs the 2s window but tracking, once established at a larger window, sustains fine. Separate acquisition from tracking: while unlocked, probe one window per cycle, cycling through the sizes, until one locks; then the existing hill-climb tracks it. This acquires a normal watch on the first probe (every window locks) and a marginal watch on whichever window works. It does not add a new CPU cost regime: probing cycles through the sizes (the 16s FFT runs at most every fourth idle cycle), which is less work than tracking a locked watch already does (that runs the largest locked window every cycle). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
compute_update() hill-climbs the analysis-window size: from the last locked window it moves to a larger (more accurate) one while the lock holds, and drops to a smaller one when it stops. But from a cold start it began at the smallest (2s) window and, if that failed, dropped to -1 and gave up - it never tried the larger windows.
A weak or jittery watch produces a cleaner, lower-jitter period in a larger window (more cycles to average), so it often can only lock at 8s or 16s, never at 2s. Such a watch therefore "usually doesn't register", yet "once locked, stays locked" - because acquisition needs the 2s window but tracking, once established at a larger window, sustains fine.
Separate acquisition from tracking: while unlocked, probe one window per cycle, cycling through the sizes, until one locks; then the existing hill-climb tracks it. This acquires a normal watch on the first probe (every window locks) and a marginal watch on whichever window works. It does not add a new CPU cost regime: probing cycles through the sizes (the 16s FFT runs at most every fourth idle cycle), which is less work than tracking a locked watch already does (that runs the largest locked window every cycle).