[Webkit-unassigned] [Bug 244172] New: WebGL performance issue. <10fps for simple shader. Resolves with tiny tweaks to shader
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Sun Aug 21 13:04:29 PDT 2022
https://bugs.webkit.org/show_bug.cgi?id=244172
Bug ID: 244172
Summary: WebGL performance issue. <10fps for simple shader.
Resolves with tiny tweaks to shader
Product: WebKit
Version: Other
Hardware: Unspecified
OS: Unspecified
Status: NEW
Severity: Normal
Priority: P2
Component: WebGL
Assignee: webkit-unassigned at lists.webkit.org
Reporter: Ian.kettlewell at gmail.com
CC: dino at apple.com, kbr at google.com, kkinnunen at apple.com
Created attachment 461779
--> https://bugs.webkit.org/attachment.cgi?id=461779&action=review
A minimal index.html that reproduces the performance issue
Attached is an ~160 line index.html / shader that contains a reduced self-contained example of a performance issue with Safari WebGL2.
The original context where I discovered this bug is within rendering code with visuals, but in the process of reducing the repro case the visuals were removed.
When viewing this page the canvas will remain black but Safari will become extremely laggy.
Chrome and Firefox run this page at 60fps.
There are three comments in the shader that begin with "NOTE: " and call out locations
where the shader can be slightly tweaked to increase performance to 60fps.
The most interesting (to me) tweak that improves performance is simply changing the for loop from
`for (int i = 0; i < 10; i++) {` to for (int i = 0; i < 9; i++) {`
makes the page run at 60 fps again.
This repro case was reduced with help from `webgl-capture` (https://github.com/greggman/webgl-capture)
I have tested this exclusively on an M1 Macbook Air. I have tested this is Safari 15.6 and Safari Technology Preview 16.0.
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20220821/47d5f2ca/attachment-0001.htm>
More information about the webkit-unassigned
mailing list