Disable JITs for chrome JS in unimportant iframe content processes
Categories
(Core :: DOM: Content Processes, task, P3)
Tracking
()
Tracking | Status | |
---|---|---|
firefox-esr68 | --- | wontfix |
firefox75 | --- | wontfix |
firefox76 | --- | wontfix |
firefox77 | --- | fix-optional |
People
(Reporter: mccr8, Unassigned)
References
(Blocks 1 open bug)
Details
This is just a wild idea I had, but I wonder how much memory we could save by disabling JITs for ad iframes, given that we'll probably not be able to disable it entirely.
Reporter | ||
Comment 1•5 years ago
|
||
Maybe the JITs are adaptive enough that they'll drop whatever extra memory they use in a cold process.
Reporter | ||
Comment 2•5 years ago
|
||
Do you think this would save much memory?
Updated•5 years ago
|
Comment 3•5 years ago
|
||
(In reply to Andrew McCreight [:mccr8] from comment #2)
Do you think this would save much memory?
As far as I know we're pretty good at discarding JIT code periodically. Releasing the JitScript will discard all JIT code + IC data + StackTypeSets for the script.
What does take up quite a lot of space is TI data for property types. As part of WarpBuilder work it's now possible (and necessary) to turn off a lot of TI, but it's mostly an untested configuration and it doesn't give you all the memory wins yet so I'm hesitant about using that.
WarpBuilder is progressing well so unless this is critical I'd suggest waiting for that to mature. Maybe we could enable WarpBuilder first in the parent process + these iframe processes to save memory. Once WarpBuilder is the default everywhere, we can remove ObjectGroups entirely, and that will save even more memory.
Comment 4•4 years ago
|
||
The severity field is not set for this bug.
:neha, could you have a look please?
For more information, please visit auto_nag documentation.
Reporter | ||
Updated•4 years ago
|
Updated•4 years ago
|
Comment hidden (advocacy) |
Description
•