New issue
Advanced search Search tips

Issue 780872 link

Starred by 1 user

Issue metadata

Status: Untriaged
Owner: ----
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: ----
Type: ----



Sign in to add a comment

gpu_perftests failing on chromium.perf/Android One Perf

Project Member Reported by sullivan@chromium.org, Nov 2 2017

Issue description

gpu_perftests failing on chromium.perf/Android One Perf

Builders failed on: 
- Android One Perf: 
  https://build.chromium.org/p/chromium.perf/builders/Android%20One%20Perf


Sample log:
https://uberchromegw.corp.google.com/i/chromium.perf/builders/Android%20One%20Perf/builds/533
https://chromium-swarm.appspot.com/task?id=38ba8e566b030c10&refresh=10&show_raw=1

I'm having a lot of trouble understanding that output, it looks like a device error, but other tests in that build are running fine on the device. I'm going to try a bisect.
 

=== BISECT JOB RESULTS ===
NO Perf regression found, tests failed to produce values

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : gpu_perftests
  Metric       : 1024_GL_LUMINANCE/drawarrays_cpu


To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=drawarrays.cpu gpu_perftests

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964020259921711600


For feedback, file a bug with component Speed>Bisection
Cc: simonhatch@chromium.org
+simonhatch: I don't see where in the output the test was actually run? Any ideas? https://build.chromium.org/p/tryserver.chromium.perf/builders/android_one_perf_bisect/builds/2828
No benchmark named "gpu_perftests".
Do you mean any of those benchmarks below?
Available benchmarks are:
  blink_perf
  blink_perf.css
  blink_perf.events
  blink_perf.paint
  blink_perf.svg
  blink_perf.xml_http_request
  memory.dual_browser_test    Measures memory usage while interacting with two different browsers.
  skpicture_printer
  skpicture_printer_ct        Captures SKPs for Cluster Telemetry.
Pass --browser to list benchmarks for another browser.


Looks like the dashboard thinks this is a telemetry test, I'll whitelist this as a non-telemetry test and rekick bisect.

=== BISECT JOB RESULTS ===
NO Test failure found

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : ???
  Metric       : 1024_GL_LUMINANCE/drawarrays_cpu

Revision             Exit Code      N
chromium@503115      1 +- N/A       20      good
chromium@503280      1 +- N/A       20      bad

To Run This Test
  ./src/out/Release/gpu_perftests --test-launcher-print-test-stdio=always --verbose

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964017376991952064


For feedback, file a bug with component Speed>Bisection

=== BISECT JOB RESULTS ===
NO Test failure found

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : ???
  Metric       : 1024_GL_LUMINANCE/drawarrays_cpu

Revision             Exit Code      N
chromium@503115      1 +- N/A       20      good
chromium@503280      1 +- N/A       20      bad

To Run This Test
  src/build/android/test_runner.py gtest --release -s gpu_perftests --verbose

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964013514452128064


For feedback, file a bug with component Speed>Bisection
Hmmm, this fails at the start of the range, but on the perf waterfall it's passing at the start of the range. Trying a bisect on a wider range...
So looks like the issue now is the same root issue cc_perftests was having a little while ago, in that the builders don't build all targets ( crbug.com/752409 )

Fixing this involves switching the bisect bots to use different archived builds, and then switching the tryserver versions to upload to the new bucket. Was kinda hoping Pinpoint would have rolled out to android by now that we could avoid doing this.



=== BISECT JOB RESULTS ===
NO Perf regression found, tests failed to produce values

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : gpu_perftests
  Metric       : 1024_GL_LUMINANCE/drawarrays_cpu


To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=drawarrays.cpu gpu_perftests

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964011206217114976


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

Comment 13 by sheriffbot@chromium.org, Nov 5

Labels: Hotlist-Recharge-Cold
Status: Untriaged (was: Available)
This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

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

Sign in to add a comment