New issue
Advanced search Search tips

Issue 593045 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Mar 2016
Cc:
EstimatedDays: ----
NextAction: ----
OS: Android
Pri: 1
Type: Bug-Regression



Sign in to add a comment

Many smoothness failures on chromium.perf on android

Project Member Reported by simonhatch@chromium.org, Mar 8 2016

Issue description


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


=== Bisection aborted ===
The bisect was aborted because The metric values for the initial "good" and "bad" revisions match the expected direction of improvement. Thus, likely represent an improvement and not a regression.
Please contact the the team (see below) if you believe this is in error.

=== Warnings ===
The following warnings were raised by the bisect job:

 * The initial regression range for return code appears to show NO sign of a regression.

===== TESTED REVISIONS =====
Revision                Exit Code   Std. Dev.   Num Values  Good?
chromium@379568         0           N/A         5           good
chromium@379603         0           N/A         5           bad

Bisect job ran on: android_nexus7_perf_bisect
Bug ID: 593045

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests smoothness.gpu_rasterization.polymer
Test Metric: frame_times/frame_times
Relative Change: 0.00%
Score: 0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus7_perf_bisect/builds/2836
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9018741244087363696


| 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 label Cr-Tests-AutoBisect.  Thank you!

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


=== Bisection aborted ===
The bisect was aborted because The metric values for the initial "good" and "bad" revisions match the expected direction of improvement. Thus, likely represent an improvement and not a regression.
Please contact the the team (see below) if you believe this is in error.

=== Warnings ===
The following warnings were raised by the bisect job:

 * The initial regression range for return code appears to show NO sign of a regression.

===== TESTED REVISIONS =====
Revision                Exit Code   Std. Dev.   Num Values  Good?
chromium@379589         0           N/A         5           good
chromium@379625         0           N/A         5           bad

Bisect job ran on: android_nexus5_perf_bisect
Bug ID: 593045

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests rasterize_and_record_micro.key_silk_cases
Test Metric: record_time/record_time
Relative Change: 0.00%
Score: 0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus5_perf_bisect/builds/3473
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9018741713162945840


| 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 label Cr-Tests-AutoBisect.  Thank you!

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


=== Bisection aborted ===
The bisect was aborted because The metric values for the initial "good" and "bad" revisions match the expected direction of improvement. Thus, likely represent an improvement and not a regression.
Please contact the the team (see below) if you believe this is in error.

=== Warnings ===
The following warnings were raised by the bisect job:

 * The initial regression range for return code appears to show NO sign of a regression.

===== TESTED REVISIONS =====
Revision                Exit Code   Std. Dev.   Num Values  Good?
chromium@379608         0           N/A         5           good
chromium@379641         0           N/A         5           bad

Bisect job ran on: android_nexus9_perf_bisect
Bug ID: 593045

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests smoothness.key_mobile_sites_smooth
Test Metric: queueing_durations/queueing_durations
Relative Change: 0.00%
Score: 0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus9_perf_bisect/builds/1669
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9018741537071798720


| 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 label Cr-Tests-AutoBisect.  Thank you!
Cc: fgor...@chromium.org
Owner: fgor...@chromium.org

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

Hi fgorski@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 : [Offline pages] Reduce the icon set for offline bolt, update colors
Author  : fgorski
Commit description:
  
This patch:
* removes the assets for light offline bolt
* applies tinting to draw the bolt using the right color
* updates the colors based on UX recommendation

BUG= 591105 
R=tedchoc@chromium.org

Review URL: https://codereview.chromium.org/1753573005

Cr-Commit-Position: refs/heads/master@{#379595}
Commit  : a85ab8e4d88a1e32a0db232cfd946533dd49ef16
Date    : Mon Mar 07 18:53:12 2016


===== TESTED REVISIONS =====
Revision                Exit Code   Std. Dev.   Num Values  Good?
chromium@379594         0           N/A         5           good
chromium@379595         1           N/A         5           bad
chromium@379596         1           N/A         5           bad
chromium@379598         1           N/A         5           bad
chromium@379601         1           N/A         5           bad
chromium@379607         1           N/A         5           bad
chromium@379620         1           N/A         5           bad
chromium@379646         1           N/A         5           bad
chromium@379698         1           N/A         5           bad

Bisect job ran on: android_nexus6_perf_bisect
Bug ID: 593045

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests service_worker.service_worker
Test Metric: page_load_1st_speed_index/page_load_1st_speed_index
Relative Change: Zero to non-zero
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus6_perf_bisect/builds/2008
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9018732455827457824


| 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 label Cr-Tests-AutoBisect.  Thank you!
Cc: -fgor...@chromium.org
Owner: simonhatch@chromium.org
Ignore that, bad bisect.
Cc: penghuang@chromium.org
Owner: penghuang@chromium.org

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

Hi penghuang@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 : Get rid of gpu related switches by passing gpu::GpuPreferences via IPC.
Author  : penghuang
Commit description:
  
BUG= 590825 
CQ_INCLUDE_TRYBOTS=tryserver.chromium.win:win_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel

Review URL: https://codereview.chromium.org/1753383003

Cr-Commit-Position: refs/heads/master@{#379651}
Commit  : 248b114c9d3034964c7259222712dfcccfad8f48
Date    : Mon Mar 07 21:52:11 2016


===== TESTED REVISIONS =====
Revision                Exit Code   Std. Dev.   Num Values  Good?
chromium@379547         0           N/A         5           good
chromium@379631         0           N/A         5           good
chromium@379642         0           N/A         5           good
chromium@379647         0           N/A         5           good
chromium@379650         0           N/A         5           good
chromium@379651         1           N/A         5           bad
chromium@379652         1           N/A         5           bad
chromium@379673         1           N/A         5           bad
chromium@379714         1           N/A         5           bad

Bisect job ran on: android_nexus7_perf_bisect
Bug ID: 593045

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests smoothness.gpu_rasterization.polymer
Test Metric: mean_frame_time/mean_frame_time
Relative Change: Zero to non-zero
Score: 98.0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus7_perf_bisect/builds/2837
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9018738947960287360


| 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 label Cr-Tests-AutoBisect.  Thank you!
Hey penghuang, bisect says your cl might be the cause of all this redness. Mind rolling it out and seeing if things clear up?

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


===== SUSPECTED CL(s) =====
Subject : Get rid of gpu related switches by passing gpu::GpuPreferences via IPC.
Author  : penghuang
Commit description:
  
BUG= 590825 
CQ_INCLUDE_TRYBOTS=tryserver.chromium.win:win_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel

Review URL: https://codereview.chromium.org/1753383003

Cr-Commit-Position: refs/heads/master@{#379651}
Commit  : 248b114c9d3034964c7259222712dfcccfad8f48
Date    : Mon Mar 07 21:52:11 2016


===== TESTED REVISIONS =====
Revision                Exit Code   Std. Dev.   Num Values  Good?
chromium@379566         0           N/A         5           good
chromium@379623         0           N/A         5           good
chromium@379637         0           N/A         5           good
chromium@379644         0           N/A         5           good
chromium@379648         0           N/A         5           good
chromium@379650         0           N/A         5           good
chromium@379651         1           N/A         5           bad
chromium@379679         1           N/A         5           bad
chromium@379792         1           N/A         5           bad

Bisect job ran on: android_nexus5_perf_bisect
Bug ID: 593045

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --also-run-disabled-tests smoothness.tough_webgl_ad_cases
Test Metric: frame_times/frame_times
Relative Change: Zero to non-zero
Score: 80.0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus5_perf_bisect/builds/3474
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9018737369829193920


| 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 label Cr-Tests-AutoBisect.  Thank you!
Cc: palmer@chromium.org boliu@chromium.org
So.. some switch that was removed was actually used by perf bots maybe?
Project Member

Comment 12 by bugdroid1@chromium.org, Mar 9 2016

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7a1173688cc0e8ff83260231c66fc3e118219c75

commit 7a1173688cc0e8ff83260231c66fc3e118219c75
Author: simonhatch <simonhatch@chromium.org>
Date: Wed Mar 09 00:28:26 2016

Revert of Get rid of gpu related switches by passing gpu::GpuPreferences via IPC. (patchset #8 id:140001 of https://codereview.chromium.org/1753383003/ )

Reason for revert:
Suspecting this is breaking a lot of perf bots.

BUG= 593045 

Original issue's description:
> Get rid of gpu related switches by passing gpu::GpuPreferences via IPC.
>
> BUG= 590825 
> CQ_INCLUDE_TRYBOTS=tryserver.chromium.win:win_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel
>
> Committed: https://crrev.com/248b114c9d3034964c7259222712dfcccfad8f48
> Cr-Commit-Position: refs/heads/master@{#379651}

TBR=piman@chromium.org,boliu@chromium.org,palmer@chromium.org,penghuang@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG= 590825 

Review URL: https://codereview.chromium.org/1778693002

Cr-Commit-Position: refs/heads/master@{#379993}

[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/android_webview/browser/deferred_gpu_command_service.cc
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/content/browser/android/synchronous_compositor_base.cc
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/content/browser/gpu/gpu_process_host.cc
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/content/browser/gpu/gpu_process_host.h
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/content/common/gpu/gpu_host_messages.h
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/content/content_browser.gypi
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/content/gpu/gpu_child_thread.cc
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/content/gpu/gpu_child_thread.h
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/content/gpu/in_process_gpu_thread.cc
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/content/gpu/in_process_gpu_thread.h
[delete] https://crrev.com/692a9e9ca5bb6d48b1d3c1e565be7205034e9a39/content/public/browser/gpu_utils.cc
[delete] https://crrev.com/692a9e9ca5bb6d48b1d3c1e565be7205034e9a39/content/public/browser/gpu_utils.h
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/gpu/command_buffer/service/gpu_preferences.h
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/gpu/command_buffer/service/gpu_switches.cc
[modify] https://crrev.com/7a1173688cc0e8ff83260231c66fc3e118219c75/gpu/command_buffer/service/gpu_switches.h

Status: Started (was: Untriaged)
I think I found the problem and I am working on the fix. 
Cc: -palmer@chromium.org
Cc: palmer@chromium.org
Status: Fixed (was: Started)
Updated the CL to fix the issue. Tried with https://codereview.chromium.org/1778493006/ &
https://codereview.chromium.org/1779913002/
.

And the updated CL has been re-landed.

Sign in to add a comment