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

Issue 602306 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Last visit > 30 days ago
Closed: Apr 2016
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

1.5%-2% regression in webrtc.datachannel at 386129:386239

Project Member Reported by rsch...@chromium.org, Apr 11 2016

Issue description

See the link to graphs below.
 
Project Member

Comment 2 by 42576172...@developer.gserviceaccount.com, Apr 11 2016

Cc: cco3@chromium.org
Owner: cco3@chromium.org

=== Auto-CCing suspected CL author cco3@chromium.org ===

Hi cco3@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 : Record UMA for Physical Web state
Author  : cco3
Commit description:
  
The "Physical Web state" refers to:
* Data connection
* BT permission/enablement
* Location permission/provider

We record this state:
* At chrome startup
* When launching the ListUrlsActivity from the diagnostics page
* When launching the ListUrlsActivity from the preferences page

BUG= 594194 

Review URL: https://codereview.chromium.org/1814213004

Cr-Commit-Position: refs/heads/master@{#386215}
Commit  : 10cbc3ad60388fc48ebfbbd16c8d3db9872c009b
Date    : Fri Apr 08 21:58:38 2016


===== TESTED REVISIONS =====
Revision                Mean Value  Std. Dev.   Num Values  Good?
chromium@386189         24059.2     58.062036   5           good
chromium@386208         23992.8     81.260076   5           good
chromium@386213         24039.2     42.839234   5           good
chromium@386214         24060.8     21.052316   5           good
chromium@386215         24447.2     24.23221    5           bad         <-
chromium@386216         24399.2     41.318277   5           bad
chromium@386218         24445.6     52.199617   5           bad
chromium@386227         24488.8     48.282502   5           bad

Bisect job ran on: android_nexus6_perf_bisect
Bug ID: 602306

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests webrtc.datachannel
Test Metric: vm_private_dirty_final_browser/vm_private_dirty_final_browser
Relative Change: 1.79%
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus6_perf_bisect/builds/2120
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9015662003381421408


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=602306

| 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!
This change doesn't seem like it could be related, but at the same time the bisect looks pretty clear. cco3, any idea what's going on?

Comment 4 by cco3@chromium.org, Apr 11 2016

No idea...what is the "mean value" measuring?

I could imagine this having an impact on startup time, but "webrtc" seems totally out of the scope of this change.
Cc: phoglund@chromium.org
Mean value is just the average value of the test over 5 runs, in this case measuring memory: vm_private_dirty_final_browser.

Agreed that it doesn't make much sense for BT UMA to be impacting webrtc. If you were adding some crazy overhead into the browser process we should be seeing it on more than just webrtc tests. Submitted some more bisects to confirm.

In the meantime, adding test owner phoglund: any way you could see this impacting webrtc tests?
Project Member

Comment 6 by 42576172...@developer.gserviceaccount.com, Apr 11 2016


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


===== SUSPECTED CL(s) =====
Subject : Record UMA for Physical Web state
Author  : cco3
Commit description:
  
The "Physical Web state" refers to:
* Data connection
* BT permission/enablement
* Location permission/provider

We record this state:
* At chrome startup
* When launching the ListUrlsActivity from the diagnostics page
* When launching the ListUrlsActivity from the preferences page

BUG= 594194 

Review URL: https://codereview.chromium.org/1814213004

Cr-Commit-Position: refs/heads/master@{#386215}
Commit  : 10cbc3ad60388fc48ebfbbd16c8d3db9872c009b
Date    : Fri Apr 08 21:58:38 2016


===== TESTED REVISIONS =====
Revision                Mean Value  Std. Dev.   Num Values  Good?
chromium@386189         26701.33333356.038082   6           good
chromium@386214         26629.6     84.645142   5           good
chromium@386215         27046.4     57.313175   5           bad         <-
chromium@386216         27105.6     36.397802   5           bad
chromium@386217         27071.2     64.274412   5           bad
chromium@386220         27084.0     92.9516     5           bad
chromium@386226         27060.0     63.560994   5           bad
chromium@386238         27090.4     49.767459   5           bad

Bisect job ran on: android_nexus6_perf_bisect
Bug ID: 602306

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests webrtc.getusermedia
Test Metric: vm_private_dirty_final_browser/vm_private_dirty_final_browser
Relative Change: 1.47%
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus6_perf_bisect/builds/2124
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9015655918526853536


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=602306

| 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!
Project Member

Comment 7 by 42576172...@developer.gserviceaccount.com, Apr 11 2016


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


===== SUSPECTED CL(s) =====
Subject : Record UMA for Physical Web state
Author  : cco3
Commit description:
  
The "Physical Web state" refers to:
* Data connection
* BT permission/enablement
* Location permission/provider

We record this state:
* At chrome startup
* When launching the ListUrlsActivity from the diagnostics page
* When launching the ListUrlsActivity from the preferences page

BUG= 594194 

Review URL: https://codereview.chromium.org/1814213004

Cr-Commit-Position: refs/heads/master@{#386215}
Commit  : 10cbc3ad60388fc48ebfbbd16c8d3db9872c009b
Date    : Fri Apr 08 21:58:38 2016


===== TESTED REVISIONS =====
Revision                Mean Value  Std. Dev.   Num Values  Good?
chromium@386128         24006.4     81.613724   5           good
chromium@386178         24000.8     37.459311   5           good
chromium@386203         23980.8     39.939955   5           good
chromium@386209         24013.6     49.363954   5           good
chromium@386212         24048.8     46.337889   5           good
chromium@386214         24016.8     41.703717   5           good
chromium@386215         26411.2     4557.842823 5           bad         <-
chromium@386227         24420.8     70.05141    5           bad

Bisect job ran on: android_nexus6_perf_bisect
Bug ID: 602306

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests webrtc.datachannel
Test Metric: vm_private_dirty_final_browser/vm_private_dirty_final_browser
Relative Change: 1.73%
Score: 0.0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus6_perf_bisect/builds/2125
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9015655914450471248


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=602306

| 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!
Alright, looks like we're sure something bad happened here - an unrelated test also regressed on memory.

cco3, can you revert your patch?

Comment 9 by cco3@chromium.org, Apr 11 2016

OK, will do.
Status: Fixed (was: Assigned)
The revert seems to have made the graph recover.

Sign in to add a comment