New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 617162 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner:
Closed: Aug 2016
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

2.1% regression in page_cycler.intl_es_fr_pt-BR at 397574:397601

Project Member Reported by rmcilroy@chromium.org, Jun 3 2016

Issue description

Looks to be specific to http___terra.es_ metric.
 
All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=617162

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgnIHirAkM


Bot(s) for this bug's original alert(s):

android-nexus6
Cc: eero.hak...@intel.com
Owner: eero.hak...@intel.com

=== Auto-CCing suspected CL author eero.hakkinen@intel.com ===

Hi eero.hakkinen@intel.com, the bisect results pointed to your CL below as possibly
causing a regression. Please have a look at this info and see whether
your CL be related.


===== BISECT JOB RESULTS =====
Status: completed


===== SUSPECTED CL(s) =====
Subject : Fix MotionEvent conversion unit test to use valid tilt angles.
Author  : eero.hakkinen
Commit description:
  
Motion event tilt angles must be within a range from 0 to PI according to
http://developer.android.com/reference/android/view/MotionEvent.html#AXIS_TILT

BUG= 514360 

Review-Url: https://codereview.chromium.org/1499873002
Cr-Commit-Position: refs/heads/master@{#397501}
Commit  : 46ca814dfe298e5179e57e620e8104d42437ca06
Date    : Thu Jun 02 20:52:07 2016


===== TESTED REVISIONS =====
Revision         Mean     Std Dev  N  Good?
chromium@397500  63421.2  1475.01  5  good
chromium@397501  74526.4  1044.39  5  bad    <--
chromium@397502  73986.4  703.896  5  bad
chromium@397504  74536.4  1024.44  5  bad
chromium@397507  74463.8  778.177  5  bad
chromium@397513  74013.8  764.83   5  bad
chromium@397526  73432.4  390.797  5  bad
chromium@397551  73985.4  1454.97  5  bad
chromium@397601  69615.8  5730.09  8  bad

Bisect job ran on: android_nexus6_perf_bisect
Bug ID: 617162

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests page_cycler.intl_es_fr_pt-BR
Test Metric: vm_private_dirty_final_renderer/http___terra.es_
Relative Change: 3.28%
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus6_perf_bisect/builds/2216
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9010864400198910272


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5801144874434560

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!
Owner: ----
Status: (was: Assigned)
My CL ("Fix MotionEvent conversion unit test to use valid tilt angles.") changed only a unit test but not the browser proper nor any performance test related code. As my CL did not change any code path tested by the perf bot, my CL cannot be the cause the detected regression.
Status: Available
Kicked off a new bisect in https://chromeperf.appspot.com/buildbucket_job_status/9008937982862585840 (terra.es page, which shows a very clear regression)
Project Member

Comment 6 by 42576172...@developer.gserviceaccount.com, Jun 25 2016

Cc: haibinlu@chromium.org
Owner: haibinlu@chromium.org

=== Auto-CCing suspected CL author haibinlu@chromium.org ===

Hi haibinlu@chromium.org, the bisect results pointed to your CL below as possibly
causing a regression. Please have a look at this info and see whether
your CL be related.


===== BISECT JOB RESULTS =====
Status: completed


===== SUSPECTED CL(s) =====
Subject : Blimp Engine is based on Linux with USE_AURA, and its client is an Android app. It is desirable for end users to see the same Chrome Android external popup menu handling.
Author  : haibinlu
Commit description:
  
Thus, add USE_EXTERNAL_POPUP_MENU flag so that Blimp Engine can follow the same code path of the external Android popup menus before it reaches WebContentsView.

BUG= 598764 
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/1992393002
Cr-Commit-Position: refs/heads/master@{#397575}
Commit  : c643d33cf668665bac4562e69a72f9ef29b5bbed
Date    : Fri Jun 03 02:23:50 2016


===== TESTED REVISIONS =====
Revision         Mean     Std Dev  N  Good?
chromium@397565  76264.9  8031.75  9  good
chromium@397570  80498.8  254.037  5  good
chromium@397573  74065.0  1350.34  5  good
chromium@397574  66349.8  287.701  5  good
chromium@397575  89113.3  7163.36  8  bad    <--
chromium@397584  85817.1  2951.23  8  bad

Bisect job ran on: android_s5_perf_bisect
Bug ID: 617162

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests page_cycler.intl_es_fr_pt-BR
Test Metric: vm_private_dirty_final_renderer/http___terra.es_
Relative Change: 25.59%
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_s5_perf_bisect/builds/738
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9008937982862585840


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6405406386028544

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!
haibinlu@ Could you comment on how likely you think it is that your CL above is related to this regression?

The bisect results are better now, but still not quite straightforward.
My CL is basically updating the code below 

  #if defined(OS_MACOSX) || defined(OS_ANDROID)
    // do something
  #endif

to 

  #if defined(USE_EXTERNAL_POPUP_MENU)
    // do something
  #endif

USE_EXTERNAL_POPUP_MENU is set to true for OS_MACOSX and OS_ANDROID

It should not affect performance or functionality.
   

Owner: ----
Project Member

Comment 10 by sheriffbot@chromium.org, Jul 7 2016

Labels: -M-53 M-54 MovedFrom-53
Moving this nonessential bug to the next milestone.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Project Member

Comment 12 by 42576172...@developer.gserviceaccount.com, Jul 16 2016

Cc: stip@chromium.org
Owner: stip@chromium.org

=== Auto-CCing suspected CL author stip@chromium.org ===

Hi stip@chromium.org, the bisect results pointed to your CL below as possibly
causing a regression. Please have a look at this info and see whether
your CL be related.


===== BISECT JOB RESULTS =====
Status: completed


===== SUSPECTED CL(s) =====
Subject : Turn on swarming for all Android tests on linux_android_rel_ng.
Author  : stip
Commit description:
  
BUG= 525873 
NOTRY=true
NOTREECHECKS=true

Review-Url: https://codereview.chromium.org/2032263002
Cr-Commit-Position: refs/heads/master@{#397574}
Commit  : d9ef1b9a56f465253b9d71815cfb9a6e34e569c4
Date    : Fri Jun 03 01:28:28 2016


===== TESTED REVISIONS =====
Revision         Mean     Std Dev  N  Good?
chromium@397573  62409.4  2051.3   5  good
chromium@397574  75398.6  690.781  5  bad    <--
chromium@397575  74775.4  678.483  5  bad
chromium@397577  74841.8  822.986  5  bad
chromium@397580  79370.6  712.256  5  bad
chromium@397587  78748.4  442.644  5  bad
chromium@397601  74735.3  6165.44  8  bad

Bisect job ran on: android_nexus6_perf_bisect
Bug ID: 617162

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests page_cycler.intl_es_fr_pt-BR
Test Metric: vm_private_dirty_final_renderer/http___terra.es_
Relative Change: 15.98%
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus6_perf_bisect/builds/2318
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9007037085936723760


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5236845458227200

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!
Owner: ----
#12 seems pretty impossible, reported as a bad bisect and trying again...
Also removed several page load time regressions where the ref build jumped at the same time as the ToT. There are some real memory regressions here, but not PLT regressions.
Project Member

Comment 17 by 42576172...@developer.gserviceaccount.com, Jul 19 2016

Cc: shaktisahu@chromium.org
Owner: shaktisahu@chromium.org

=== Auto-CCing suspected CL author shaktisahu@chromium.org ===

Hi shaktisahu@chromium.org, the bisect results pointed to your CL below as possibly
causing a regression. Please have a look at this info and see whether
your CL be related.


===== BISECT JOB RESULTS =====
Status: completed


===== SUSPECTED CL(s) =====
Subject : Blimp: Reduced logging serverity for assignment error
Author  : shaktisahu
Commit description:
  
Blimp client currently crashes with a fatal error message when there is
a client-engine version mismatch. The crash is so fast that it doesn't
give a chance for the user to try connecting with a different assigner.
Reduced the logging level from FATAL to ERROR for assigner response.

BUG=615205

Review-Url: https://codereview.chromium.org/2033273002
Cr-Commit-Position: refs/heads/master@{#397584}
Commit  : b0061b1242d40d9b5dd7d84ee6fb1d65b1eddad6
Date    : Fri Jun 03 02:31:53 2016


===== TESTED REVISIONS =====
Revision         Mean     Std Dev  N  Good?
chromium@397565  46816.8  8121.59  5  good
chromium@397575  51005.6  393.217  5  good
chromium@397580  34150.8  204.292  5  good
chromium@397582  34053.0  180.971  5  good
chromium@397583  33946.8  170.034  5  good
chromium@397584  52661.0  162.388  5  bad    <--

Bisect job ran on: android_s5_perf_bisect
Bug ID: 617162

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests page_cycler.intl_es_fr_pt-BR
Test Metric: vm_private_dirty_final_renderer/vm_private_dirty_final_renderer
Relative Change: 12.48%
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_s5_perf_bisect/builds/807
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9006793381784714464


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5850037066661888

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!
Owner: ----
My CL just changed the error logs. It seems pretty impossible that this would cause performance regression.
Lets try another bisect.
Project Member

Comment 21 by 42576172...@developer.gserviceaccount.com, Jul 30 2016

Cc: blink-rebaseline-bot@chromium.org
Owner: blink-rebaseline-bot@chromium.org

=== Auto-CCing suspected CL author blink-rebaseline-bot@chromium.org ===

Hi blink-rebaseline-bot@chromium.org, the bisect results pointed to your CL below as possibly
causing a regression. Please have a look at this info and see whether
your CL be related.


===== BISECT JOB RESULTS =====
Status: completed


===== SUSPECTED CL(s) =====
Subject : Auto-rebaseline for r397515
Author  : Rebaseline Bot
Commit description:
  
https://chromium.googlesource.com/chromium/src/+/d88f5fb93

BUG= 616287 
TBR=chrishtr@chromium.org

Review URL: https://codereview.chromium.org/2031053003 .

Cr-Commit-Position: refs/heads/master@{#397548}
Commit  : 92dd81a5d31258a67aebff15dfbeaee7b47fc3b7
Date    : Fri Jun 03 00:03:20 2016


===== TESTED REVISIONS =====
Revision         Mean     Std Dev  N  Good?
chromium@397500  36418.0  6073.24  5  good
chromium@397525  40771.2  2222.53  5  good
chromium@397538  33920.2  462.359  5  good
chromium@397545  35768.4  4878.71  5  good
chromium@397547  43269.4  185.504  5  good
chromium@397548  56255.8  5503.51  5  bad    <--
chromium@397550  48856.6  49.105   5  bad
chromium@397600  49568.2  162.606  5  bad

Bisect job ran on: android_s5_perf_bisect
Bug ID: 617162

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests page_cycler.intl_es_fr_pt-BR
Test Metric: vm_private_dirty_final_renderer/vm_private_dirty_final_renderer
Relative Change: 36.11%
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_s5_perf_bisect/builds/848
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9005789969219126880


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5810408267448320

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!

Comment 22 by stip@chromium.org, Aug 1 2016

Cc: -stip@chromium.org
Owner: rmcilroy@chromium.org
Status: Assigned (was: Available)
Status: WontFix (was: Assigned)
Marking as wontfix. The bisects haven't had any success, and there was a reduction of roughtly equal magnitude in memory usage around r406251.

Sign in to add a comment