Calling jemalloc_purge_freed_pages() to purge freed pages on OS X slow?

RESOLVED INCOMPLETE

Status

()

RESOLVED INCOMPLETE
6 years ago
2 years ago

People

(Reporter: u462496, Unassigned)

Tracking

Firefox Tracking Flags

(Not tracked)

Details

(Reporter)

Description

6 years ago
This bug is in response to comments 25 - 29 in Bug 852520.

(In reply to Justin Lebar [:jlebar] from comment #26)
> Yes: It can be very slow, for reasons we don't understand.  That's what bug
> 789975 was about.

Is 44 ms slow to release 565 MB of memory?  That is the average of 3 tests I made using this code:

var t1 = Date.now();

let res = Cc["@mozilla.org/memory-reporter-manager;1"].
                getService(Ci.nsIMemoryReporterManager).resident;
                
dump("    et : "+(Date.now() - t1)+"\n");

I saw cpu in Activity Monitor spike momentarily 2%, then go back to 0.

Is this test naive?  Is there stuff that runs in the background?
(Reporter)

Updated

6 years ago
Summary: Calling jemalloc_purge_freed_pages() to release memory to system slow? → Calling jemalloc_purge_freed_pages() to free pages slow?
(Reporter)

Updated

6 years ago
Summary: Calling jemalloc_purge_freed_pages() to free pages slow? → Calling jemalloc_purge_freed_pages() to free pages on OS X slow?
(Reporter)

Updated

6 years ago
Summary: Calling jemalloc_purge_freed_pages() to free pages on OS X slow? → Calling jemalloc_purge_freed_pages() to purge freed pages on OS X slow?
> Is 44 ms slow to release 565 MB of memory?

44ms would be too slow for us to call while the system was active, as it's on the order of a GC pause.

But also Ehsan saw much longer pauses than this under some circumstances.

> Is this test naive?

Not really.  It is a bit naive in that it's measuring the time to get the RSS as well as the time to purgge.

> Is there stuff that runs in the background?

I don't know what you mean.
(Reporter)

Comment 2

6 years ago
> I don't know what you mean.

I mean, is there another process spawned by this call that would run in the background and would not be reflected in the 44ms figure, but I can see by your answer that must not be the case.
Hi reporter,

Can you please provide more details about this issue? A link where you encounter this bug? A more detailed test case?

Also, please try to reproduce this on the latest release(43.0.4) and latest Nightly(46.0a1) and provide the results. When doing this, please try to reproduce with a new clean Firefox profile, maybe even in safe mode, as some of this issues may be caused by third party installed add-ons or custom settings(https://support.mozilla.org/en-US/kb/troubleshoot-and-diagnose-firefox-problems).

Thanks,
Paul.
Flags: needinfo?(allassopraise)
Since the reporter didn't provide the requested information, I will mark this issue as RESOLVED INCOMPLETE. If you still encounter this problem, please feel free to reopen this bug, or file a new one.
Status: UNCONFIRMED → RESOLVED
Last Resolved: 3 years ago
Flags: needinfo?(allassopraise)
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.