Talos Regression :( Trace Malloc Allocs increase 6.89% on MacOSX 10.7 Mozilla-Aurora

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 11:41:15 AM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Trace Malloc Allocs increase 6.89% on MacOSX 10.7 Mozilla-Aurora
------------------------------------------------------------------------------
Previous: avg 596797.700 stddev 3388.043 of 30 runs up to revision abb6ae769136
New : avg 637930.600 stddev 4099.330 of 5 runs since revision a8e95ae10ea7
Change : +41132.900 (6.89% / z=12.141)
Graph : http://mzl.la/U8omDS

Changeset range: http://hg.mozilla.org/releases/mozilla-aurora/pushloghtml?fromchange=abb6ae769136&tochange=a8e95ae10ea7

Changesets:
* http://hg.mozilla.org/releases/mozilla-aurora/rev/9697eadafa13
: Lukas Blakk <lsb...@mozilla.com> - Merging in version bump NO BUG

* http://hg.mozilla.org/releases/mozilla-aurora/rev/6fa5491eabd5
: Lukas Blakk <lsb...@mozilla.com> - Added tag FIREFOX_AURORA_16_BASE for changeset 9697eadafa13

* http://hg.mozilla.org/releases/mozilla-aurora/rev/142f4a95eaf7
: Lukas Blakk <lsb...@mozilla.com> - Added tag FIREFOX_AURORA_16_BASE for changeset 6fdf9985acfe

* http://hg.mozilla.org/releases/mozilla-aurora/rev/c918ff2f0994
: Etienne Segonzac <eti...@segonzac.info> - Bug 774216. Can't launch apps on B2G since it's seen as XP_UNIX by WebappOSUtils. r=felipe
: http://bugzilla.mozilla.org/show_bug.cgi?id=774216

* http://hg.mozilla.org/releases/mozilla-aurora/rev/0e6039a2c90a
: L. David Baron <dba...@dbaron.org> - Make test_visited_reftests report failures instead of using timeout to indicate failure. No bug.

* http://hg.mozilla.org/releases/mozilla-aurora/rev/ac282e15dc02
: L. David Baron <dba...@dbaron.org> - Limit the hashless color quirk to the properties where it's needed, per http://simon.html5.org/specs/quirks-mode#the-hashless-hex-color-quirk . (Bug 774122, patch 1) r=bzbarsky

This changes test_property_syntax_errors.html in the following ways:
(1) removes the "known failures" handling, since there aren't any
(2) changes it to test syntax errors in both quirks mode and standards
mode, instead of only in standards mode
(3) uses a (new) quirks_values mechanism to property_database.js to
test cases that should be invalid in standards mode but accepted in
quirks mode
: http://bugzilla.mozilla.org/show_bug.cgi?id=774122

* http://hg.mozilla.org/releases/mozilla-aurora/rev/e3007422b989
: L. David Baron <dba...@dbaron.org> - Limit the unitless length quirk to the properties where it's needed, per http://simon.html5.org/specs/quirks-mode#the-unitless-length-quirk . (Bug 774122, patch 2) r=bzbarsky
: http://bugzilla.mozilla.org/show_bug.cgi?id=774122

* http://hg.mozilla.org/releases/mozilla-aurora/rev/64ff8c2d37f9
: L. David Baron <dba...@dbaron.org> - Disable the unitless length quirk inside of calc(). (Bug 774122, patch 3) r=bzbarsky

I noticed this bug because the new assertions in ParseVariant,
introduced in the previous patch, are stricter and would assert without
this fix.
: http://bugzilla.mozilla.org/show_bug.cgi?id=774122

* http://hg.mozilla.org/releases/mozilla-aurora/rev/846e8d18edbe
: Mark Hammond <mham...@skippinet.com.au> - Bug 773164: remove social frameworker load listener. r=gavin
: http://bugzilla.mozilla.org/show_bug.cgi?id=773164

* http://hg.mozilla.org/releases/mozilla-aurora/rev/6d2002bd64c5
: Mark Hammond <mham...@skippinet.com.au> - Bug 773160: expose a reduced 'navigator' object to social frameworker workers. r=gavin
: http://bugzilla.mozilla.org/show_bug.cgi?id=773160

* http://hg.mozilla.org/releases/mozilla-aurora/rev/492da706c14c
: Shane Caraveo <mixed...@gmail.com> - Bug 774178: make some changes to provider profile/notification handling to support "logout", r=gavin
: http://bugzilla.mozilla.org/show_bug.cgi?id=774178

* http://hg.mozilla.org/releases/mozilla-aurora/rev/ca832757d4b7
: Shane Caraveo <mixed...@gmail.com> - Bug 771826: implement Social toolbar button UI, including notification icons, notification panel, and "profile" panel. Styling only for Windows/Mac for the moment, r=gavin
: http://bugzilla.mozilla.org/show_bug.cgi?id=771826

* http://hg.mozilla.org/releases/mozilla-aurora/rev/bc090e21a088
: Gavin Sharp <ga...@gavinsharp.com> - Bug 773845: lazy-load some social modules to see what effect that has on maxheap/allocs tests
: http://bugzilla.mozilla.org/show_bug.cgi?id=773845

* http://hg.mozilla.org/releases/mozilla-aurora/rev/e85863a0d897
: Gavin Sharp <ga...@gavinsharp.com> - Backed out changeset 64ff8c2d37f9 (bug 774122)
: http://bugzilla.mozilla.org/show_bug.cgi?id=774122

* http://hg.mozilla.org/releases/mozilla-aurora/rev/588ea3ebb1ef
: Gavin Sharp <ga...@gavinsharp.com> - Backed out changeset e3007422b989 (bug 774122)
: http://bugzilla.mozilla.org/show_bug.cgi?id=774122

* and 4139 more

Bugs:
* http://bugzilla.mozilla.org/show_bug.cgi?id=2
* http://bugzilla.mozilla.org/show_bug.cgi?id=3
* http://bugzilla.mozilla.org/show_bug.cgi?id=4
* http://bugzilla.mozilla.org/show_bug.cgi?id=778247 - GeckoApp.mAppContext shouldn't be a "public static" member
* http://bugzilla.mozilla.org/show_bug.cgi?id=8
* http://bugzilla.mozilla.org/show_bug.cgi?id=782338 - [B2G] Intermittent test_mobile_voice_state.js | Emulator callback still pending when finish() called
* http://bugzilla.mozilla.org/show_bug.cgi?id=15
* http://bugzilla.mozilla.org/show_bug.cgi?id=778257 - JavaScript Warning: TypeError: function sh_endSelection does not always return a value
* http://bugzilla.mozilla.org/show_bug.cgi?id=778259 - relpath bug in manifestparser.py
* http://bugzilla.mozilla.org/show_bug.cgi?id=23
* http://bugzilla.mozilla.org/show_bug.cgi?id=776825 - Separate message managers into senders and broadcasters
* http://bugzilla.mozilla.org/show_bug.cgi?id=778268 - Assertion failure: !pn->isDefn(), at ../frontend/ParseNode.h:1437
* http://bugzilla.mozilla.org/show_bug.cgi?id=761895 - [Azure] Get text in Cairo backend working for GTK
* http://bugzilla.mozilla.org/show_bug.cgi?id=778282 - image/test/browser/Makefile.in should use MOCHITEST_BROWSER_FILES
* http://bugzilla.mozilla.org/show_bug.cgi?id=783709 - Launching a web application on Nightly shows a window with no title and white content
* and 2257 more

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 3:00:24 PM8/28/12

to dev-tree-...@lists.mozilla.org

Improvement! DHTML Row Major MozAfterPaint decrease 11.2% on MacOSX 10.6 (rev4) Mozilla-Aurora
----------------------------------------------------------------------------------------------
Previous: avg 464.235 stddev 4.900 of 30 runs up to revision abb6ae769136
New : avg 412.300 stddev 4.532 of 5 runs since revision a8e95ae10ea7
Change : -51.935 (11.2% / z=10.600)
Graph : http://mzl.la/RqC7xK

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 3:01:20 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Paint increase 14.3% on MacOSX 10.7 Mozilla-Aurora
----------------------------------------------------------------
Previous: avg 371.690 stddev 7.454 of 30 runs up to revision abb6ae769136
New : avg 425.000 stddev 8.031 of 5 runs since revision a8e95ae10ea7
Change : +53.310 (14.3% / z=7.152)
Graph : http://mzl.la/RqCgkN

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 3:01:30 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( DHTML Row Major MozAfterPaint increase 9.67% on Linux x64 Mozilla-Aurora
--------------------------------------------------------------------------------------
Previous: avg 301.345 stddev 2.035 of 30 runs up to revision abb6ae769136
New : avg 330.488 stddev 3.695 of 5 runs since revision a8e95ae10ea7
Change : +29.143 (9.67% / z=14.323)
Graph : http://mzl.la/RqCjNJ

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 3:02:10 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Paint increase 13.9% on MacOSX 10.6 (rev4) Mozilla-Aurora
-----------------------------------------------------------------------
Previous: avg 300.021 stddev 6.179 of 30 runs up to revision abb6ae769136
New : avg 341.600 stddev 9.127 of 5 runs since revision a8e95ae10ea7
Change : +41.579 (13.9% / z=6.729)
Graph : http://mzl.la/RqClVC

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 3:02:16 PM8/28/12

to dev-tree-...@lists.mozilla.org

Improvement! tscroll Row Major decrease 1.54% on MacOSX 10.6 (rev4) Mozilla-Aurora
----------------------------------------------------------------------------------
Previous: avg 13705.920 stddev 48.080 of 30 runs up to revision abb6ae769136
New : avg 13495.120 stddev 29.394 of 5 runs since revision a8e95ae10ea7
Change : -210.800 (1.54% / z=4.384)
Graph : http://mzl.la/RqCmck

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 3:20:26 PM8/28/12

to dev-tree-...@lists.mozilla.org

Improvement! Tp5 No Network Row Major Responsiveness MozAfterPaint decrease 51.7% on MacOSX 10.6 (rev4) Mozilla-Aurora
----------------------------------------------------------------------------------------------------------------------
Previous: avg 218.367 stddev 31.517 of 30 runs up to revision abb6ae769136
New : avg 105.400 stddev 11.393 of 5 runs since revision a8e95ae10ea7
Change : -112.967 (51.7% / z=3.584)
Graph : http://mzl.la/RqEOj6

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 3:21:42 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( DHTML Row Major MozAfterPaint increase 5.39% on MacOSX 10.7 Mozilla-Aurora
----------------------------------------------------------------------------------------
Previous: avg 386.894 stddev 3.766 of 30 runs up to revision abb6ae769136
New : avg 407.735 stddev 1.404 of 5 runs since revision a8e95ae10ea7
Change : +20.841 (5.39% / z=5.533)
Graph : http://mzl.la/RqEXTq

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 3:21:58 PM8/28/12

to dev-tree-...@lists.mozilla.org

Improvement! Tp5 No Network Row Major Responsiveness MozAfterPaint decrease 33.6% on MacOSX 10.7 Mozilla-Aurora
---------------------------------------------------------------------------------------------------------------
Previous: avg 237.000 stddev 5.736 of 30 runs up to revision abb6ae769136
New : avg 157.400 stddev 16.772 of 5 runs since revision a8e95ae10ea7
Change : -79.600 (33.6% / z=13.878)
Graph : http://mzl.la/RqEVer

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 3:22:03 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Tp5 No Network Row Major MozAfterPaint increase 11.9% on MacOSX 10.7 Mozilla-Aurora
-------------------------------------------------------------------------------------------------
Previous: avg 233.992 stddev 0.740 of 30 runs up to revision abb6ae769136
New : avg 261.882 stddev 6.656 of 5 runs since revision a8e95ae10ea7
Change : +27.890 (11.9% / z=37.695)
Graph : http://mzl.la/RqEVuR

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:00:40 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Ts, MAX Dirty Profile increase 5.98% on XP Mozilla-Aurora
-----------------------------------------------------------------------
Previous: avg 446.190 stddev 3.718 of 30 runs up to revision abb6ae769136
New : avg 472.853 stddev 6.321 of 5 runs since revision a8e95ae10ea7
Change : +26.663 (5.98% / z=7.171)
Graph : http://mzl.la/OrG54W

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:01:10 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Tp5 No Network Row Major MozAfterPaint (Main RSS) increase 2.11e+08% on Win7 Mozilla-Aurora
---------------------------------------------------------------------------------------------------------
Previous: avg 65.302 stddev 0.466 of 30 runs up to revision abb6ae769136
New : avg 137568800.000 stddev 8266425.056 of 5 runs since revision a8e95ae10ea7
Change : +137568734.698 (2.11e+08% / z=295177088.989)
Graph : http://mzl.la/OrGapu

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:01:14 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Tp5 No Network Row Major MozAfterPaint (Content RSS) increase 1.91% on Win7 Mozilla-Aurora

--------------------------------------------------------------------------------------------------------
Previous: avg 65.302 stddev 0.466 of 30 runs up to revision abb6ae769136

New : avg 66.551 stddev 0.241 of 5 runs since revision a8e95ae10ea7
Change : +1.249 (1.91% / z=2.680)
Graph : http://mzl.la/OrGd4D

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:01:18 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Tp5 No Network Row Major MozAfterPaint increase 20.3% on Win7 Mozilla-Aurora
------------------------------------------------------------------------------------------
Previous: avg 206.291 stddev 2.775 of 30 runs up to revision abb6ae769136
New : avg 248.089 stddev 1.802 of 5 runs since revision a8e95ae10ea7
Change : +41.798 (20.3% / z=15.064)
Graph : http://mzl.la/OrGdS1

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:01:22 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( SVG, Opacity Row Major increase 57.6% on Win7 Mozilla-Aurora
--------------------------------------------------------------------------
Previous: avg 90.600 stddev 0.803 of 30 runs up to revision abb6ae769136
New : avg 142.800 stddev 2.683 of 5 runs since revision a8e95ae10ea7
Change : +52.200 (57.6% / z=65.005)
Graph : http://mzl.la/OrGep8

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:01:47 PM8/28/12

to dev-tree-...@lists.mozilla.org

Improvement! Tp5 No Network Row Major Responsiveness MozAfterPaint decrease 15.2% on XP Mozilla-Aurora
------------------------------------------------------------------------------------------------------
Previous: avg 1499.100 stddev 61.290 of 30 runs up to revision abb6ae769136
New : avg 1271.600 stddev 24.069 of 5 runs since revision a8e95ae10ea7
Change : -227.500 (15.2% / z=3.712)
Graph : http://mzl.la/OrGgxg

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:01:51 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Tp5 No Network Row Major MozAfterPaint (Private Bytes) increase 7.71% on XP Mozilla-Aurora
--------------------------------------------------------------------------------------------------------
Previous: avg 105248366.667 stddev 1061838.532 of 30 runs up to revision abb6ae769136
New : avg 113364600.000 stddev 1552992.048 of 5 runs since revision a8e95ae10ea7
Change : +8116233.333 (7.71% / z=7.644)
Graph : http://mzl.la/OrGjJk

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:01:55 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Tp5 No Network Row Major MozAfterPaint (Main RSS) increase 1.94e+08% on XP Mozilla-Aurora
-------------------------------------------------------------------------------------------------------
Previous: avg 57.782 stddev 0.258 of 30 runs up to revision abb6ae769136
New : avg 112101800.000 stddev 953772.090 of 5 runs since revision a8e95ae10ea7
Change : +112101742.218 (1.94e+08% / z=434800224.882)
Graph : http://mzl.la/OrGh4j

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:01:59 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Tp5 No Network Row Major MozAfterPaint increase 30.5% on XP Mozilla-Aurora
----------------------------------------------------------------------------------------
Previous: avg 170.042 stddev 2.087 of 30 runs up to revision abb6ae769136
New : avg 221.854 stddev 1.286 of 5 runs since revision a8e95ae10ea7
Change : +51.812 (30.5% / z=24.821)
Graph : http://mzl.la/OrGkgC

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:02:10 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Ts, MAX Dirty Profile increase 7.41% on Win7 Mozilla-Aurora
-------------------------------------------------------------------------
Previous: avg 583.637 stddev 5.738 of 30 runs up to revision abb6ae769136
New : avg 626.905 stddev 9.839 of 5 runs since revision a8e95ae10ea7
Change : +43.268 (7.41% / z=7.541)
Graph : http://mzl.la/OrGlkE

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:02:15 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Ts, MED Dirty Profile increase 7.76% on Win7 Mozilla-Aurora
-------------------------------------------------------------------------
Previous: avg 585.507 stddev 6.757 of 30 runs up to revision abb6ae769136
New : avg 630.968 stddev 10.635 of 5 runs since revision a8e95ae10ea7
Change : +45.461 (7.76% / z=6.728)
Graph : http://mzl.la/OrGogj

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 6:40:21 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( a11y Row Major MozAfterPaint increase 7.72% on XP Mozilla-Aurora
------------------------------------------------------------------------------
Previous: avg 308.683 stddev 2.863 of 30 runs up to revision abb6ae769136
New : avg 332.500 stddev 4.717 of 5 runs since revision a8e95ae10ea7
Change : +23.817 (7.72% / z=8.318)
Graph : http://mzl.la/PP2Gdg

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 7:00:31 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Paint increase 21.9% on Win7 Mozilla-Aurora
---------------------------------------------------------
Previous: avg 167.507 stddev 7.515 of 30 runs up to revision abb6ae769136
New : avg 204.200 stddev 3.834 of 5 runs since revision a8e95ae10ea7
Change : +36.693 (21.9% / z=4.883)
Graph : http://mzl.la/PP4xhW

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 7:00:37 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( a11y Row Major MozAfterPaint increase 9.54% on Win7 Mozilla-Aurora
--------------------------------------------------------------------------------
Previous: avg 327.383 stddev 3.360 of 30 runs up to revision abb6ae769136
New : avg 358.600 stddev 3.305 of 5 runs since revision a8e95ae10ea7
Change : +31.217 (9.54% / z=9.291)
Graph : http://mzl.la/Rr0fAw

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 7:00:41 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( tscroll Row Major increase 7.47% on Win7 Mozilla-Aurora
---------------------------------------------------------------------
Previous: avg 10099.237 stddev 1.596 of 30 runs up to revision abb6ae769136
New : avg 10853.500 stddev 9.701 of 5 runs since revision a8e95ae10ea7
Change : +754.263 (7.47% / z=472.659)
Graph : http://mzl.la/Rr0ifE

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 8:21:43 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( DHTML Row Major MozAfterPaint increase 6.83% on Win7 Mozilla-Aurora
---------------------------------------------------------------------------------
Previous: avg 307.558 stddev 0.977 of 30 runs up to revision abb6ae769136
New : avg 328.577 stddev 3.431 of 5 runs since revision a8e95ae10ea7
Change : +21.019 (6.83% / z=21.519)
Graph : http://mzl.la/PPbM9S

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 10:00:56 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Trace Malloc Allocs increase 3.26% on CentOS release 5 (Final) Mozilla-Aurora
-------------------------------------------------------------------------------------------
Previous: avg 471891.033 stddev 711.403 of 30 runs up to revision abb6ae769136
New : avg 487281.400 stddev 3362.925 of 5 runs since revision a8e95ae10ea7
Change : +15390.367 (3.26% / z=21.634)
Graph : http://mzl.la/PZWy3N

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 10:01:01 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Trace Malloc MaxHeap increase 2.86% on CentOS release 5 (Final) Mozilla-Aurora
--------------------------------------------------------------------------------------------
Previous: avg 39158906.667 stddev 132678.374 of 30 runs up to revision abb6ae769136
New : avg 40278800.000 stddev 8955.445 of 5 runs since revision a8e95ae10ea7
Change : +1119893.333 (2.86% / z=8.441)
Graph : http://mzl.la/PZWwZC

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 10:01:34 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Trace Malloc Allocs increase 3.48% on CentOS (x86_64) release 5 (Final) Mozilla-Aurora
----------------------------------------------------------------------------------------------------
Previous: avg 462628.467 stddev 765.441 of 30 runs up to revision abb6ae769136
New : avg 478729.400 stddev 2298.902 of 5 runs since revision a8e95ae10ea7
Change : +16100.933 (3.48% / z=21.035)
Graph : http://mzl.la/PZWBMY

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 10:01:40 PM8/28/12

to dev-tree-...@lists.mozilla.org

Improvement! Trace Malloc MaxHeap decrease 14.4% on CentOS (x86_64) release 5 (Final) Mozilla-Aurora
----------------------------------------------------------------------------------------------------
Previous: avg 51511870.000 stddev 372759.256 of 30 runs up to revision abb6ae769136
New : avg 44097400.000 stddev 8331.566 of 5 runs since revision a8e95ae10ea7
Change : -7414470.000 (14.4% / z=19.891)
Graph : http://mzl.la/PZWDEC

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 10:40:56 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( DHTML Row Major MozAfterPaint increase 10.5% on Linux Mozilla-Aurora
----------------------------------------------------------------------------------
Previous: avg 318.757 stddev 1.980 of 30 runs up to revision abb6ae769136
New : avg 352.076 stddev 3.225 of 5 runs since revision 380fdb7249aa
Change : +33.320 (10.5% / z=16.824)
Graph : http://mzl.la/Q00v8z

Changeset range: http://hg.mozilla.org/releases/mozilla-aurora/pushloghtml?fromchange=abb6ae769136&tochange=380fdb7249aa

* and 4141 more

* and 2258 more

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 11:01:05 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( SVG, Opacity Row Major increase 13.8% on Linux Mozilla-Aurora
---------------------------------------------------------------------------
Previous: avg 66.967 stddev 1.090 of 30 runs up to revision abb6ae769136
New : avg 76.200 stddev 1.095 of 5 runs since revision 380fdb7249aa
Change : +9.233 (13.8% / z=8.470)
Graph : http://mzl.la/OIipLC

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 11:20:56 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Trace Malloc Allocs increase 5.41% on WINNT 5.2 Mozilla-Aurora
----------------------------------------------------------------------------
Previous: avg 373022.533 stddev 473.594 of 30 runs up to revision abb6ae769136
New : avg 393188.400 stddev 649.932 of 5 runs since revision a8e95ae10ea7
Change : +20165.867 (5.41% / z=42.580)
Graph : http://mzl.la/OIkTtJ

Changeset range: http://hg.mozilla.org/releases/mozilla-aurora/pushloghtml?fromchange=abb6ae769136&tochange=a8e95ae10ea7

* and 4139 more

* and 2257 more

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 11:21:01 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Trace Malloc MaxHeap increase 3.3% on WINNT 5.2 Mozilla-Aurora
----------------------------------------------------------------------------
Previous: avg 39158710.000 stddev 131659.973 of 30 runs up to revision abb6ae769136
New : avg 40449200.000 stddev 9427.089 of 5 runs since revision a8e95ae10ea7
Change : +1290490.000 (3.3% / z=9.802)
Graph : http://mzl.la/OIkW8J

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 11:40:52 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Tp5 No Network Row Major MozAfterPaint increase 25.1% on Linux Mozilla-Aurora
-------------------------------------------------------------------------------------------
Previous: avg 231.296 stddev 3.175 of 30 runs up to revision abb6ae769136
New : avg 289.351 stddev 2.324 of 5 runs since revision 380fdb7249aa
Change : +58.055 (25.1% / z=18.283)
Graph : http://mzl.la/OIndRa

Changeset range: http://hg.mozilla.org/releases/mozilla-aurora/pushloghtml?fromchange=abb6ae769136&tochange=380fdb7249aa

* and 4141 more

* and 2258 more

nob...@cruncher.build.mozilla.org

unread,

Aug 28, 2012, 11:40:45 PM8/28/12

to dev-tree-...@lists.mozilla.org

Regression :( Tp5 No Network Row Major MozAfterPaint (Private Bytes) increase 6.43% on Linux Mozilla-Aurora
-----------------------------------------------------------------------------------------------------------
Previous: avg 381101700.000 stddev 2274427.206 of 30 runs up to revision abb6ae769136
New : avg 405622000.000 stddev 2424660.698 of 5 runs since revision 380fdb7249aa
Change : +24520300.000 (6.43% / z=10.781)
Graph : http://mzl.la/OInbsy

nob...@cruncher.build.mozilla.org

unread,

Aug 29, 2012, 12:01:23 AM8/29/12

to dev-tree-...@lists.mozilla.org

Regression :( Ts, Paint increase 12.9% on Linux Mozilla-Aurora
--------------------------------------------------------------
Previous: avg 582.416 stddev 6.357 of 30 runs up to revision abb6ae769136
New : avg 657.600 stddev 15.169 of 5 runs since revision 380fdb7249aa
Change : +75.184 (12.9% / z=11.827)
Graph : http://mzl.la/OIpsEg

Robert Kaiser

unread,

Aug 29, 2012, 4:41:09 PM8/29/12

to

nob...@cruncher.build.mozilla.org schrieb:
> Regression :( Trace Malloc Allocs increase 6.89% on MacOSX 10.7 Mozilla-Aurora
> ------------------------------------------------------------------------------

Is it just my impression or do we really have a long list of regressions
on almost every uplift (from central to aurora as well as from aurora to
beta) in recent times, at least for Firefox desktop?

Also, I see that mbrubeck now has posted again on a few of the reports
in here but I rarely see anyone reacting to the messages in here, how
much are people actually watching here and reacting to all this stuff?

Robert Kaiser

Jeff Hammel

unread,

Aug 29, 2012, 5:17:58 PM8/29/12

to dev-tree-...@lists.mozilla.org, chri...@mozilla.com, je...@mozilla.com

I would guess blindly that not many people react or read this
newsgroup. There are > 50 messages a day for talos "improvements" and
"regressions". Most of them are probably noise, not real improvements or
regressions. There's no clear indication of what one should do about it,
outside of hunting down the bugs responsible (which to echo mbrubeck
seems to be the only person really looking at). Talos noise is
considerable:
https://wiki.mozilla.org/Auto-tools/Projects/Signal_From_Noise ,
https://wiki.mozilla.org/images/c/c0/Larres-thesis.pdf . Graphserver is
not wonderful at helping me figure out what has regressed or even what
tests are what.

I'm not trying to be a complainer or doom-sayer here. I am trying to
draw attention to the fact that A. Firefox performance testing is hard,
complex, and has lots of moving pieces; and B. consequently, since we
don't have anyone who's job it is to analyze all of this madness, we
mostly just ignore it (except for mbrubeck). We (=jmaher, Jonathan
Eads, Christina Choi, and myself) have been working on making this
better, end to end, but it is a huge job, much much bigger than the
single quarter for which the project was originally planned.

In my opinion, to make people care you need a system that is easy (maybe
even fun) to use, and you need to have a person who's job it is to sort
through this (huge amount of) data. Kudos to mbrubeck for stepping up
to the plate. He does a huge amount of work. But I'd love to see
someone really mining this data and figuring out what should be done.

> _______________________________________________
> dev-tree-management mailing list
> dev-tree-...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-tree-management

Ehsan Akhgari

unread,

Aug 29, 2012, 7:25:30 PM8/29/12

to Jeff Hammel, je...@mozilla.com, dev-tree-...@lists.mozilla.org, chri...@mozilla.com

Interestingly enough, I have also been frustrated by this trend. I
started a discussion on dev-platform about this, where more people would
have a chance to see it.

Ehsan