Repository navigation
Viewer: automatic quality steps down sooner, and on Retina ends with fewer pixels - #83
Merged
Merged
Conversation
…fewer pixels A Retina display draws four times the pixels, and Low there costs more than High at one pixel per CSS pixel (integrated Radeon, cove: 175 ms vs 111 ms). Automatic quality now settles a frame rate under half its target in two reports, skips one report rather than four after each step, and on a display of ratio over 1 takes a last step: Low at one canvas pixel per CSS pixel. Presets chosen by hand still draw at full density. Co-Authored-By: Claude Opus 5.5 <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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What and why
Someone reported Solstice is really slow on an Apple M3 Pro MacBook. There is no Mac here, so this works from measurement on stand-ins. A MacBook's Retina display (devicePixelRatio 2) draws about 4x the pixels of the same window at ratio 1. Even at Low it costs more than High at 1x: on the integrated Radeon, cove Low at 3024x1800 takes 175 ms, against 111 ms for High at 1512x900. Automatic quality also took 7 s to reach Low, and stopped there.
apps/viewer/web/auto-quality.js(copied bybuild-web.sh) and is tested ineditor/test.mjs.Findings, ranked, plus what needs a Mac to confirm:
m3-pro-perf.mdin the coordinator's notes folder. In short: the Retina pixel count first, screen GI as the largest pass, then the slow step-down. The visibility buffer's atomics without a depth test may cost more on Apple's tile-based GPUs, but only a Mac capture can tell.Verified
--stream --frames 120 --day-length 2), total GPU ms. RTX 5080: cove High 2.55 at 1920x1080 vs 5.01 at 3024x1800; forest High 3.80 vs 7.73. Integrated Radeon: cove High 111 at 1512x900 vs 621 at 3024x1800; Low 53 vs 175.bash scripts/check-changed.sh --all: All checks passed for: all.node apps/viewer/web/editor/test.mjs: all passed (10 new checks for automatic quality).Authorship
Written by: Claude Opus 5.5 through Claude Code
Reviewed by: Claude Sonnet 5.5 (no correctness bugs found; noted that the ratio is fixed when the page starts, as before)
🤖 Generated with Claude Code