754267 - Memory usage increase on all Nightly builds since May 5th

Status: RESOLVED WONTFIX

(Mozilla QA Graveyard :: Mozmill Tests, defect)

Description

[Mihaela Velimiroviciu (:mihaelav)]

Memory usage increased significantly since May 7 on all Nightly builds, compared to May 4(May 5 and 6 results are not available)

Explicit memory (MB):
Ubuntu 11.10 x86: 65->100
Ubuntu 11.10 x86_64: 90->160
Mac 10.6.8: 120->140
Mac 1.7.3: 110->130
Win 5.1: 100->140
Win 6.0: 90->120
Win 6.1 x86: 90->125
Win 6.1 x86_64: 85->134

Resident memory (MB):
Ubuntu 11.10 x86: 95->120
Ubuntu 11.10 x86_64: 130->180
Mac 10.6.8: 150->170
Mac 1.7.3: 200->210
Win 5.1: 130->150
Win 6.0: constant at 140
Win 6.1 x86: 120->140
Win 6.1 x86_64: 140->170

Charts: http://mozmill-ci.blargon7.com/#/endurance/charts?branch=15.0&platform=All&from=2012-04-24&to=2012-05-11

This may be related to bug #752887 since the tests which use more memory are the same (/testTabbedBrowsing_OpenNewTab/test1.js testOpenNewTab and /testBookmarks_OpenAllInTabs/test1.js testOpenAllBookmarksInTabs) and the spike appeared at the same time (May 7)

Comment 1

[u279076]

Yes, let's please keep this on one bug.

Comment 2

[Dave Hunt [:davehunt] [he/him] ⌚BST]

Moved investigation from bug 752887.

I have replicated this using our Mozmill CI system with the only difference being the Firefox build, and therefore suspect a regression. The memory increase looks like it is across the entire testsuite, but /testTabbedBrowsing_OpenNewTab/test1.js:testOpenNewTab appears to be the most noticeably worse.

Firefox 15.0a1 (15.0a1, en-US, 20120504030509)
Average explicit: 55MB
Average resident: 84MB
http://mozmill-crowd.blargon7.com/#/endurance/report/f87375a634b1a5ba746e5f763a13b569

Firefox 15.0a1 (15.0a1, en-US, 20120508030517)
Average explicit: 71MB
Average resident: 96MB
http://mozmill-crowd.blargon7.com/#/endurance/report/f87375a634b1a5ba746e5f763a13c4da

Comment 3

[Nicholas Nethercote [inactive]]

From comment 0:

> Memory usage increased significantly since May 7 on all Nightly builds, compared
> to May 4(May 5 and 6 results are not available)

Compartment-per-global (bug 650353) landed on mozilla-central on 2012-05-04.  It's almost certainly the cause;  it caused a spike on areweslimyet.com too.  This was expected.

Assuming that's correct, we can WONTFIX this.  People are looking into ways to reduce the CPG memory hit, but that'll take a while.  Until then, this is the new baseline.

Comment 4

[Henrik Skupin [:whimboo][⌚️UTC+1]]

(In reply to Nicholas Nethercote [:njn] from comment #3)
> Compartment-per-global (bug 650353) landed on mozilla-central on 2012-05-04.
> It's almost certainly the cause;  it caused a spike on areweslimyet.com too.
> This was expected.

First I would like to see results from 2012-05-05 so we can make sure that this bump is only related to the GPC fix. Dave would you mind running another test against such a build? Thanks.

Comment 5

[Dave Hunt [:davehunt] [he/him] ⌚BST]

Henrik: I did this the first time you asked and reported on the other bug: https://bugzilla.mozilla.org/show_bug.cgi?id=752887#c14

Quoted below:

Done. It looks like it regressed in 20120505030510.

Here's the pushlog: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=2db9df42823d&tochange=0a48e6561534

Comment 6

[Henrik Skupin [:whimboo][⌚️UTC+1]]

Ah gotcha. Thanks Dave. In that case I agree that we can close the bug as WONTFIX and use the current memory usage as baseline.