Skip to content

frame-cost todo: option 3 is done, and what it did not fix - #44

Merged
rossry merged 1 commit into
mainfrom
docs/frame-cost-after-candles
Aug 28, 2026
Merged

rossry merged 1 commit into
mainfrom
docs/frame-cost-after-candles

Conversation

@rossry

@rossry rossry commented Aug 28, 2026

Copy link
Copy Markdown
Owner

Bookkeeping I said I'd do once the merge order settled. The file landed with #42 quoting pre-Candles numbers and listing option 3 as future work; #43 then did option 3. Both facts re-measured on merged main.

show before either change now worst seam is
koln 31.8 ms 31.9 ms coming-to-rest (NoiseGlow + Motif)
apollo 34.5 ms 32.9 ms matta (AuroraVeils)
promises 54.1 ms 50.2 ms answer (AuroraVeils into AuroraVeils)
nocturne 42–46 ms 41.7 ms deep-sea (NoiseGlow + Starfield)

koln and apollo now fit inside a 30 fps frame, apollo only barely. promises and nocturne don't, and both are bounded by seams no candle work touches — option 1 is the only remaining one that covers promises, because the problem there is two expensive scenes of the same kind blended together, not one wasteful primitive.

Two honesty notes added to the file: repeat runs vary by a few ms (the seam ranking is stable, the individual figures less so), and the two shows that now fit a frame fit it with nothing else running — no codec, no serial, no viewers, no six boards attached. That is not the condition they will be played in.

Docs only; no code touched.

The table in this file quoted numbers from before the Candles gather
landed, and its option 3 was still written as future work. Both are now
true again, and the numbers are re-measured on merged main.

The headline worth keeping: koln (31.9 ms) and apollo (32.9 ms) now fit
a 30 fps frame; promises (50.2 ms) and nocturne (41.7 ms) do not, and
both are bounded by AuroraVeils/NoiseGlow seams that no candle work
touches. Option 1 is the only remaining one that covers promises,
because the problem there is two expensive scenes of the same kind
blended together rather than one wasteful primitive.

Also noted: repeat runs vary by a few ms — the ranking of seams is
stable, the individual figures less so — and the two shows that now fit
a frame fit it with nothing else running, which is not the condition
they will be played in.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FbzmvdzTNSUxDUjVraRZMx
@rossry
rossry merged commit 4d935cb into main Aug 28, 2026
1 check 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