JS demo (https://www.fxhash.xyz/generative/slug/sweet-dreams ) periodically spends 100ms+ in GCSlice
Categories
(Core :: JavaScript: GC, task, P3)
Tracking
()
People
(Reporter: mayankleoboy1, Unassigned)
References
(Blocks 1 open bug, )
Details
Go to https://www.fxhash.xyz/generative/slug/sweet-dreams
Click on run
Profile: https://share.firefox.dev/3ZJ0z3f
Chrome doesnt have any visual janks.
Maybe some things to improve here?
|
Reporter | |
Updated•2 years ago
|
|
|
Reporter | |
Comment 1•2 years ago
|
||
This is the "regression" range i came up with: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=578715b6d3fcb21c0b61996347260ef9b119d0d2&tochange=8508c35e49793fbe402ad2a706a57fabd1c2b0c4
Profiles before this regression range: https://share.firefox.dev/3zFzNy8
|
||
Updated•2 years ago
|
|
||
Comment 2•2 years ago
|
||
(In reply to Mayank Bansal from comment #1)
It's not in the regression range but close by - could this be due to bug 1434542?
|
|
Reporter | |
Updated•1 year ago
|
|
|
Reporter | |
Comment 3•1 year ago
•
|
||
Latest profile. Still periodic janks
https://share.firefox.dev/3OfpEQq
And profile from the try build of bug 1875455: https://share.firefox.dev/3UdRTTd
|
|
Reporter | |
Comment 4•10 months ago
|
||
Profile with latest Nightly: https://share.firefox.dev/3whOl8O
|
|
||
Comment 5•10 months ago
|
||
In the latest profile 38% of GC time is spent in jemalloc free, 75% of that poisoning memory with memset.
What's happening is that we're finishing the GC non-incrementally because of the volume of malloc allocations. What's supposed to happen is that we trigger slices more aggressively before we get to this point. I don't know why that's not happening.
|
|
Reporter | |
Comment 6•1 month ago
|
||
This is what I get now: https://share.firefox.dev/40pDi8G
Calling this fixed.
Description
•