New issue
Advanced search Search tips

Issue 661523 link

Starred by 1 user

Issue metadata

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



Sign in to add a comment

A zero-to-nonzero regression in thread_times.simple_mobile_sites at 426910:426987

Project Member Reported by primiano@chromium.org, Nov 2 2016

Issue description

See the link to graphs below.
 
Cc: danakj@chromium.org
Owner: danakj@chromium.org

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

Hi danakj@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 : cc: Add back a benchmark category trace for counting renderer frames
Author  : danakj
Commit description:
  
When deleting DelegatingRenderer::SwapBuffers I failed to notice that
the TRACE_EVENT in there had the benchmark category. This category
is used by perf tests, and without that event they had no way to count
the number of renderer frames produced.

This adds benchmark to the DrawLayers trace event in LayerTreeHostImpl
instead and points the perf tests at that. Also removes the benchmark
category from the GLRenderer and SoftwareRenderer classes as they are
never used in the renderer and are not used for perf tests (tools/ no
longer refers to the name "SwapBuffers" for perf tests).

R=enne@chromium.org
TBR=dtu
BUG= 656947 
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Review-Url: https://chromiumcodereview.appspot.com/2442453003
Cr-Commit-Position: refs/heads/master@{#426952}
Commit  : 7607802c76a4a0b826d22807213e981025ec734b
Date    : Sat Oct 22 02:26:00 2016


===== TESTED REVISIONS =====
Revision         Mean     Std Dev    N  Good?
chromium@426937  0.0      0.0        5  good
chromium@426946  0.0      0.0        5  good
chromium@426950  0.0      0.0        5  good
chromium@426951  0.0      0.0        5  good
chromium@426952  2.20074  0.0210644  5  bad    <--
chromium@426954  2.17229  0.0265345  5  bad

Bisect job ran on: android_nexus5_perf_bisect
Bug ID: 661523

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests thread_times.key_mobile_sites_smooth
Test Metric: thread_renderer_compositor_cpu_time_per_frame/thread_renderer_compositor_cpu_time_per_frame
Relative Change: Zero to non-zero
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus5_perf_bisect/builds/4313
Job details: https://chromeperf.appspot.com/buildbucket_job_status/8997109633166444496


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5546601169289216

| 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!
Status: WontFix (was: Untriaged)
ok the commit message explains everything.

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


===== SUSPECTED CL(s) =====
Subject : cc: Add back a benchmark category trace for counting renderer frames
Author  : danakj
Commit description:
  
When deleting DelegatingRenderer::SwapBuffers I failed to notice that
the TRACE_EVENT in there had the benchmark category. This category
is used by perf tests, and without that event they had no way to count
the number of renderer frames produced.

This adds benchmark to the DrawLayers trace event in LayerTreeHostImpl
instead and points the perf tests at that. Also removes the benchmark
category from the GLRenderer and SoftwareRenderer classes as they are
never used in the renderer and are not used for perf tests (tools/ no
longer refers to the name "SwapBuffers" for perf tests).

R=enne@chromium.org
TBR=dtu
BUG= 656947 
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Review-Url: https://chromiumcodereview.appspot.com/2442453003
Cr-Commit-Position: refs/heads/master@{#426952}
Commit  : 7607802c76a4a0b826d22807213e981025ec734b
Date    : Sat Oct 22 02:26:00 2016


===== TESTED REVISIONS =====
Revision         Mean     Std Dev    N  Good?
chromium@426937  0.0      0.0        5  good
chromium@426950  0.0      0.0        5  good
chromium@426951  0.0      0.0        5  good
chromium@426952  5.6663   0.118397   5  bad    <--
chromium@426953  5.71261  0.252634   5  bad
chromium@426956  5.71032  0.0919683  5  bad
chromium@426962  5.57613  0.095446   5  bad
chromium@426987  5.47176  0.0799473  5  bad

Bisect job ran on: android_nexus5_perf_bisect
Bug ID: 661523

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests thread_times.key_mobile_sites_smooth
Test Metric: thread_renderer_main_cpu_time_per_frame/thread_renderer_main_cpu_time_per_frame
Relative Change: Zero to non-zero
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus5_perf_bisect/builds/4314
Job details: https://chromeperf.appspot.com/buildbucket_job_status/8997109588126753344


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6125087562924032

| 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!

Sign in to add a comment