1121409 - Substantial general performance decrease on Nightly

Status: RESOLVED INCOMPLETE

(Core :: General, defect)

Description

[Samuel Leslie]

User Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/39.0.2171.95 Safari/537.36

Steps to reproduce:

Updated to the latest Firefox Nightly (v38.0a1 2015-01-13 Win64).


Actual results:

A very noticeable & substantial decrease in general browser performance. This degradation appears to coincide with the update to the 38 Branch of Firefox. The application of subsequent nightly builds in the branch has not yet resolved the issue.

Further, the issue happened at the same time on my desktop (Win 7 SP1 x64) & laptop (Win 8.1 Update x64) which are both running Firefox Nightly Win64.


Expected results:

Performance consistent or largely unchanged.

Comment 1

[Samuel Leslie]

Please note that I attempted to install the Gecko Profiler extension to obtain some more detailed data before filing this bug as per: https://developer.mozilla.org/en-US/docs/Mozilla/Performance/Reporting_a_Performance_Problem.

Unfortunately this resulted in instant crashes. I've filed a separate bug in regards to this: https://bugzilla.mozilla.org/show_bug.cgi?id=1121407

If this bug is fixed and obtaining such data would prove useful I can of course provide it :)

Comment 2

[Samuel Leslie]

Just a quick addendum that the browser seems to quite rapidly get increasingly slow as it remains open as well. It rapidly gets to the point where it's horrendously slow & laggy to the point of being unusable.

I do have quite a number of tabs open at a time, but not an enormous amount either. At time of writing I've a total of 70 across 4 windows. I'm not running E10S either for that matter.

Comment 3

[:Gijs (he/him)]

(In reply to samuel.denbigh.leslie from comment #0)
> A very noticeable & substantial decrease in general browser performance.
> This degradation appears to coincide with the update to the 38 Branch of
> Firefox.

Have you checked the last 37 nightlies and/or used mozregression to narrow this down to a specific day when things got worse? That would be quite helpful.



(In reply to samuel.denbigh.leslie from comment #1)
> Please note that I attempted to install the Gecko Profiler extension to
> obtain some more detailed data before filing this bug as per:
> https://developer.mozilla.org/en-US/docs/Mozilla/Performance/
> Reporting_a_Performance_Problem.
> 
> Unfortunately this resulted in instant crashes. I've filed a separate bug in
> regards to this: https://bugzilla.mozilla.org/show_bug.cgi?id=1121407

Ugh. :-(

You could still get a profile by using Tools > Web Developer > Browser Toolbox (you may need to turn on remote + browser/chrome debugging in the web dev 'normal' toolbox options) and using the profiler in there, maybe?

Comment 4

[Samuel Leslie]

Hey there,

The good news is that the bug in Gecko Profiler is now fixed. Further, I've used mozregression and believe I've narrowed the regression window down to the following pushlog:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=86f3f3bee063&tochange=06402d84ac3d

Of the two changesets this one seems to be the far more likely candidate for the issue:
https://hg.mozilla.org/integration/mozilla-inbound/rev/06402d84ac3d

Incidentally, you can get immediate feedback as to the issue as the per-tab page loading animation (animated green circular rotation) is very noticeably "laggy". I'll try and use the build immediately prior to this changeset if possible for a while to be doubly certain that this is the faulty changeset but my testing with mozregression seemed pretty conclusive.

Comment 5

[:Gijs (he/him)]

Aaron, any chance you know what part of your change could cause this?

Samuel, could you put up the raw data from about:support (there's a button on it to get this in the clipboard) in an attachment (you can paste text by clicking the link on this page: https://bugzilla.mozilla.org/attachment.cgi?bugid=1121409&action=enter )? Thanks!

(Asking because that change, as best I can tell, should only be affecting plugins and/or pages with them. It's surprising to me that it'd affect browser performance itself - but perhaps it does for the plugins you have?)

Comment 6

[(No longer employed by Mozilla) Aaron Klotz]

(In reply to :Gijs Kruitbosch from comment #5)
> Aaron, any chance you know what part of your change could cause this?

That change is a crash fix; when not crashing, the same code would run prior to that change as after the change.

Comment 7

[:Gijs (he/him)]

(In reply to :Gijs Kruitbosch from comment #5)
> Samuel, could you put up the raw data from about:support (there's a button
> on it to get this in the clipboard) in an attachment (you can paste text by
> clicking the link on this page:
> https://bugzilla.mozilla.org/attachment.cgi?bugid=1121409&action=enter )?
> Thanks!
> 
> (Asking because that change, as best I can tell, should only be affecting
> plugins and/or pages with them. It's surprising to me that it'd affect
> browser performance itself - but perhaps it does for the plugins you have?)

I forgot to set needinfo here. I guess my additional question would be... do you still have the mozilla-central pushlog as well as the inbound one?

Thanks!

Comment 8

[Kanchan Kumari QA]

Resolved-Incomplete due to time since last communication/update by reporter.
Please feel free to reopen if the error occurs in a current Firefox version.

Comment 9

[Firefox Bug Husbandry Bot]

Moving from Core::Untriaged to Core::General https://bugzilla.mozilla.org/show_bug.cgi?id=1407598