430065 - 25% Tgfx_avg regression on WINNT 5.1 mini talos trunk nochrome qm-mini-xp05

Status: RESOLVED INCOMPLETE

(Core :: Layout, defect)

Description

[Steve England [:stevee]]

Attachment: Image showing graph

Tgfx_avg has jumped from around 26.5ms to around 33ms

WINNT 5.1 mini talos trunk nochrome qm-mini-xp05

build: 2008/04/03 05:29:30 -- tgfx: 26.25
build: 2008/04/03 07:57:31 -- tgfx: 33.00

Generous range, Checkins to module PhoenixTinderbox between 2008-04-03 03:29 and 2008-04-03 09:57 : 
http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=PhoenixTinderbox&branch=HEAD&branchtype=match&dir=&file=&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=explicit&mindate=2008-04-03+03%3A29&maxdate=2008-04-03+09%3A57&cvsroot=%2Fcvsroot

I'd guess due to bug 382392.

Graph:
http://graphs.mozilla.org/graph.html#spst=range&spstart=0&spend=1208778274&bpst=cursor&bpstart=0&bpend=1208778274&m1tid=146318&m1bl=0&m1avg=0

Comment 1

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Hmm. Do we still have the detailed test results that show which tests got worse?

Comment 2

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Here are the logs:
Before:
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1207225860.1207239043.27565.gz&fulltext=1
After:
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1207234680.1207244443.9309.gz&fulltext=1

Relevant test results:
Before:
|1;text2.html;26.5;26.25;26;27;26;27;26;27;26
|2;tile-jpg.html;27;26.75;26;27;26;27;27;27;27
|3;tile-png.html;27;27;27;27;27;27;27;27;27
|4;borders-solid.html;27;27;27;28;28;27;27;27;27
|5;borders-dashed.html;27;26.75;26;27;27;27;26;27;27
|6;borders-rounded.html;26;26;26;27;27;26;26;26;26
After:
|1;text2.html;33;33;33;153;33;33;33;153;33
|2;tile-jpg.html;33;33;33;35;33;33;35;33;33
|3;tile-png.html;33;33;33;33;33;33;33;33;33
|4;borders-solid.html;33;33;33;34;34;33;33;33;33
|5;borders-dashed.html;33;33;33;33;33;33;33;33;33
|6;borders-rounded.html;33.5;33.25;33;39;39;33;33;33;34

That doesn't give us much to go on.

Given their uniformity, I'm not sure how much stock to put on them. Vlad, is this worth investigating?

Comment 3

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Not really; Tgfx needs to be pulled from the tinderboxes until we have time to fix it, essentially.  Now, it's possible that there was an actual regression, but I wouldn't put any stock in the numeric value of the regression..

Comment 4

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Alright, I'm not going to let this block unless/until we get data on a better testcase.

Comment 5

[Shawn Wilsher :sdwilsh]

I'm going through and marking old performance regression bugs as INCOMPLETE that are likely too old to be valid or get any traction on them.

Please re-open if you have more information or can demonstrate the regression still exists.