New issue
Advanced search Search tips

Issue 710139 link

Starred by 1 user
Status: WontFix
Owner:
dalecur...@chromium.org
Closed: Jul 2017
Cc:
simonhatch@chromium.org
dalecur...@chromium.org
crouleau@chromium.org
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

18.3% regression in media.tough_video_cases at 460042:460054

Project Member Reported by wolenetz@chromium.org, Apr 10 2017 
This alert seems valid.

Comment 1 by wolenetz@chromium.org, Apr 10 2017 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=710139

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg7I3gswoM


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

chromium-rel-mac11
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Apr 10 2017 


=== BISECT JOB RESULTS ===
Bisect failed for unknown reasons

Please contact the team (see below) and report the error.


Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : media.tough_video_cases
  Metric       : time_to_play/video.html?src_tulip2.mp4


To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=video.html.src.tulip2.mp4 media.tough_video_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8982670475288066928

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5902659996876800


| 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 Speed>Bisection.  Thank you!

Comment 4 by wolenetz@chromium.org, Apr 10 2017

Cc: simonhatch@chromium.org
Project Member

Comment 6 by 42576172...@developer.gserviceaccount.com, Apr 10 2017 


=== BISECT JOB RESULTS ===
Bisect failed for unknown reasons

Please contact the team (see below) and report the error.


Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : media.tough_video_cases
  Metric       : time_to_play/video.html?src_tulip2.mp4


To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=video.html.src.tulip2.mp4 media.tough_video_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8982664805248060352

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6406681644236800


| 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 Speed>Bisection.  Thank you!
Project Member

Comment 8 by 42576172...@developer.gserviceaccount.com, Jun 29 2017 


=== BISECT JOB RESULTS ===
NO Perf regression found

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : media.tough_video_cases
  Metric       : time_to_play/video.html?src_tulip2.mp4

Revision             Result                  N
chromium@460041      77.3476 +- 20.7555      21      good
chromium@460054      76.9798 +- 19.5571      21      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=video.html.src.tulip2.mp4 media.tough_video_cases

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8975428399179842560


For feedback, file a bug with component Speed>Bisection

Comment 11 by wolenetz@chromium.org, Jun 29 2017 

#10 is a wider bisection range.
Project Member

Comment 12 by 42576172...@developer.gserviceaccount.com, Jun 29 2017 


=== BISECT JOB RESULTS ===
NO Perf regression found

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : media.tough_video_cases
  Metric       : time_to_play/video.html?src_tulip2.mp4

Revision             Result                  N
chromium@460041      76.6067 +- 29.1416      21      good
chromium@460054      77.6162 +- 15.8455      21      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=video.html.src.tulip2.mp4 media.tough_video_cases

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8975422476562650800


For feedback, file a bug with component Speed>Bisection
Project Member

Comment 13 by 42576172...@developer.gserviceaccount.com, Jun 29 2017

Cc: mdjones@chromium.org
Owner: mdjones@chromium.org

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

Hi mdjones@chromium.org, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : mdjones
  Commit : 92cb8c04c342bb35a96ee77c055f7cec8ff3ee41
  Date   : Tue Mar 28 20:33:50 2017
  Subject: Reland of Set HexSSID in network config before matching it against policies (patchset #1 id:1 of https://codereview.chromium.org/2776403002/ )

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : media.tough_video_cases
  Metric       : time_to_play/video.html?src_tulip2.mp4
  Change       : 16.57% | 78.3016666667 -> 91.2758333333

Revision             Result                  N
chromium@460041      78.3017 +- 5.15309      6      good
chromium@460152      77.6317 +- 1.48788      6      good
chromium@460180      78.8225 +- 1.90473      6      good
chromium@460194      75.8217 +- 12.0653      6      good
chromium@460201      74.2175 +- 17.8134      6      good
chromium@460204      77.79 +- 2.1241         6      good
chromium@460206      79.0139 +- 5.96267      9      good
chromium@460207      88.5417 +- 19.9468      6      bad       <--
chromium@460262      91.2758 +- 31.9077      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=video.html.src.tulip2.mp4 media.tough_video_cases

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8975418531004012912


For feedback, file a bug with component Speed>Bisection

Comment 14 by mdjones@chromium.org, Jun 30 2017

Cc: -mdjones@chromium.org
Owner: tbarzic@chromium.org
Assigning to original author.

Comment 15 by wolenetz@chromium.org, Jul 5 2017 

Interesting bisect result. Would https://crrev.com/92cb8c04c342bb35a96ee77c055f7cec8ff3ee41 impact mac builds' performance (that change looks mostly related to chromeos, not mac)?

Comment 16 by tbarzic@chromium.org, Jul 5 2017 

no, that change should not impact mac build, it's chromeos only

Comment 18 by wolenetz@chromium.org, Jul 5 2017 

I'll retry bisect then. There was clearly some persistent change in perf that's still occurring even now (see https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg7I3gswoM)
Project Member

Comment 19 by 42576172...@developer.gserviceaccount.com, Jul 5 2017

Cc: dalecur...@chromium.org
Owner: dalecur...@chromium.org

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

Hi dalecurtis@chromium.org, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : dalecurtis
  Commit : 5285d00f561827da882c8faaa7bda1d571c4fc60
  Date   : Wed Mar 29 00:27:30 2017
  Subject: Use AudioUnit instead of AudioDevice properties to get channel count.

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : media.tough_video_cases
  Metric       : time_to_play/video.html?src_tulip2.mp4
  Change       : 17.97% | 79.8144444444 -> 94.1544444444

Revision             Result                  N
chromium@460041      79.8144 +- 9.22755      9      good
chromium@460152      78.7933 +- 1.47903      6      good
chromium@460207      79.4308 +- 3.59755      6      good
chromium@460235      77.6 +- 8.34367         6      good
chromium@460249      77.4358 +- 9.65734      6      good
chromium@460250      93.2517 +- 30.8615      6      bad       <--
chromium@460251      93.4958 +- 31.2627      6      bad
chromium@460253      93.0017 +- 30.6637      6      bad
chromium@460256      92.5717 +- 26.0278      6      bad
chromium@460262      94.1544 +- 47.8073      9      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=video.html.src.tulip2.mp4 media.tough_video_cases

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8974890629697927264


For feedback, file a bug with component Speed>Bisection

Comment 20 by dalecur...@chromium.org, Jul 5 2017

Status: WontFix (was: Assigned)
Not much we can do here. The old way of querying channel layout/counts was inaccurate and breaking audio for some users. There's no real way to query if this will be an issue without creating the AudioUnit; which is what is taking the extra time here.

In practice this delta should be much smaller since we cache these parameters on the renderer side.

Comment 21 by wolenetz@chromium.org, Jul 5 2017 

SGTM. Thanks for rapid-followup on this old perf regression, Dale.
►

Sign in to add a comment