[meta] High CPU load (40% - 100%) caused by the top extensions loading animation on https://code.visualstudio.com/docs/
Categories
(Core :: Graphics: WebRender, defect, P3)
Tracking
()
| Performance Impact | medium |
People
(Reporter: whimboo, Unassigned)
References
(Depends on 2 open bugs, Blocks 1 open bug)
Details
(4 keywords)
Attachments
(3 files, 1 obsolete file)
Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:86.0) Gecko/20100101 Firefox/86.0 ID:20201217214927
Having the page https://code.visualstudio.com/docs/?dv=osx open in the currently selected tab causes a permanent high CPU load (40 - 100%) of the MainThread.
This is actually a new regression and I can see this first with the Firefox 84 release. Using 83.0 the CPU load is around 7% only.
Let me run mozregression to find the causing change.
| Reporter | ||
Comment 1•5 years ago
|
||
The problem here is actually the loading animation under Top Extensions. As it looks like whatever should be displayed there does not finish loading, and causes this high CPU load. Disabling JavaScript for that domain via NoScript makes it go away.
| Reporter | ||
Comment 2•5 years ago
|
||
If you are not seeing this problem because the top extensions list will be loaded please just add the following node to the DOM:
<div class="loading"></div>
Doing the same in Safari I see no CPU load.
| Reporter | ||
Updated•5 years ago
|
| Reporter | ||
Comment 3•5 years ago
|
||
With Firefox 78 the CPU load is only around 20% but still high. So not sure if that is a regression per se.
Steps:
- Load https://code.visualstudio.com/docs/?dv=osx
- Open devtools and run the following code in the console:
var main = document.getElementById("main");
var div = document.createElement("div");
div.className = "loading";
main.appendChild(div);
- Scroll down until the load animation is visible.
Moving bug over to the Performance component for triaging.
| Reporter | ||
Comment 4•5 years ago
|
||
Here a gecko profile with all threads and IPC messages enabled: https://share.firefox.dev/3h17cJo
| Reporter | ||
Comment 5•5 years ago
|
||
As it looks like Emilio and Olli cannot see this on Linux. Markus, given that you are on MacOS do you have a chance to check? Thanks.
Comment 6•5 years ago
|
||
I can reproduce this.
Comment 7•5 years ago
|
||
Comment 8•5 years ago
|
||
Comment 9•5 years ago
|
||
Comment 10•5 years ago
|
||
Comment 11•5 years ago
|
||
Instruments profile: https://drive.google.com/file/d/1K823Ql4Dv99Pm1Uf0CkQ_3M4UyADcCkU/view?usp=sharing
I'm planning to file bugs for the individual pieces I found in there, on Monday.
| Reporter | ||
Updated•5 years ago
|
Comment 12•5 years ago
|
||
Markus: Could you please link the bugs that you mentioned in comment #11 and set a priority/severity for this bug?
Comment 13•5 years ago
|
||
Actually I'll move this into the Graphics component, IIRC all the issues here were gfx issues.
I'll have to look at the profiles and testcases again to remember what I found.
Comment 14•4 years ago
|
||
Ticking needinfo to remind you to do that. :)
| Reporter | ||
Comment 15•4 years ago
|
||
The one dependent bug has been fixed. Markus, did you already file the other ones but maybe missed to add the dependency?
Comment 17•4 years ago
|
||
I have now filed everything I could find. In order of priority, I think bug 1688086 was the most important (and is now fixed), bug 1690621 is the next most important, and then bug 1689215.
Updated•4 years ago
|
Updated•4 years ago
|
| Reporter | ||
Updated•4 years ago
|
Updated•4 years ago
|
Updated•3 years ago
|
Updated•1 year ago
|
Description
•