New issue
Advanced search Search tips

Issue 683680 link

Starred by 1 user
Status: WontFix
Owner:
pdfium-deps-roller@chromium.org
Closed: Jan 2017
Cc:
simonhatch@chromium.org
geoffl...@chromium.org
oysteine@chromium.org
pdfium-deps-roller@chromium.org
jmad...@chromium.org
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression

Blocked on:
issue 683998
issue 683999


Sign in to add a comment

35.1% regression in angle_perftests at 444805:444889

Project Member Reported by jmad...@chromium.org, Jan 22 2017 
See the link to graphs below.

Comment 1 by jmad...@chromium.org, Jan 22 2017 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=683680

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgkPmXrQoM


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

chromium-rel-win7-gpu-nvidia
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Jan 22 2017 


=== BISECT JOB RESULTS ===
Perf regression found but unable to continue

Bisect was stopped because a commit couldn't be classified as either
good or bad.


Bisect Details
  Configuration: winx64nvidia_perf_bisect
  Benchmark    : angle_perftests
  Metric       : Uniforms_gl_400_vec4_repeating/score


To Run This Test
  .\src\out\Release_x64\angle_perftests.exe --test-launcher-print-test-stdio=always --test-launcher-jobs=1

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989755723196399680

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6634346225598464


| 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 5 by 42576172...@developer.gserviceaccount.com, Jan 22 2017 


=== BISECT JOB RESULTS ===
Perf regression found but unable to continue

Bisect was stopped because a commit couldn't be classified as either
good or bad.


Bisect Details
  Configuration: winx64nvidia_perf_bisect
  Benchmark    : angle_perftests
  Metric       : Uniforms_gl_400_vec4_repeating/score


To Run This Test
  .\src\out\Release_x64\angle_perftests.exe --test-launcher-print-test-stdio=always --test-launcher-jobs=1

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989752412962063264

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5262877268639744


| 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!

Comment 6 by jmad...@chromium.org, Jan 23 2017

Cc: oysteine@chromium.org simonhatch@chromium.org
Hey Simon/Oysteine - any help with this failing bisect?

Comment 7 by simonhatch@chromium.org, Jan 23 2017

Blockedon: 683998

Comment 8 by simonhatch@chromium.org, Jan 23 2017

Blockedon: 683999

Comment 9 by jmad...@chromium.org, Jan 23 2017 

Thanks Simon.

Comment 10 by jmad...@chromium.org, Jan 24 2017 

Simon, with  issue 683999  fixed, will the bisect work? Or is  issue 683998  still preventing the job from finishing?

Comment 12 by simonhatch@chromium.org, Jan 24 2017 

Hopefully everything is fixed, kicking it again!
Project Member

Comment 13 by 42576172...@developer.gserviceaccount.com, Jan 25 2017

Cc: geoffl...@chromium.org
Owner: geoffl...@chromium.org

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

Hi geofflang@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 : Geoff Lang
  Commit : 2cf7c51ec26016edaedeba03bd1d126e67d46f8a
  Date   : Wed Jan 18 16:42:38 2017
  Subject: Add a perf test for binding buffers.

Bisect Details
  Configuration: winx64nvidia_perf_bisect
  Benchmark    : angle_perftests
  Metric       : Uniforms_gl_400_vec4_repeating/score
  Change       : 35.23% | 7855.0 -> 5087.5

Revision                              Result                  N
chromium@444804                       7855.0 +- 466.223       12      good
chromium@444826                       7776.17 +- 733.598      12      good
chromium@444837                       8209.17 +- 965.166      12      good
chromium@444842                       7803.67 +- 876.682      12      good
chromium@444843                       8165.83 +- 766.852      12      good
chromium@444843,angle@ab9f9c31f7      7803.83 +- 972.575      12      good
chromium@444843,angle@2cf7c51ec2      5220.83 +- 250.725      6       bad       <--
chromium@444843,angle@65603eb814      5295.5 +- 180.714       6       bad
chromium@444844                       5196.17 +- 241.654      6       bad
chromium@444845                       4953.5 +- 639.149       6       bad
chromium@444847                       5083.67 +- 240.502      6       bad
chromium@444889                       5087.5 +- 533.477       6       bad

To Run This Test
  .\src\out\Release_x64\angle_perftests.exe --test-launcher-print-test-stdio=always --test-launcher-jobs=1

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989566546706160144

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5909289375367168


| 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 15 by 42576172...@developer.gserviceaccount.com, Jan 25 2017 


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

Suspected Commit
  Author : Geoff Lang
  Commit : 2cf7c51ec26016edaedeba03bd1d126e67d46f8a
  Date   : Wed Jan 18 16:42:38 2017
  Subject: Add a perf test for binding buffers.

Bisect Details
  Configuration: winx64nvidia_perf_bisect
  Benchmark    : angle_perftests
  Metric       : Uniforms_gl_400_vec4_repeating/score
  Change       : 35.69% | 7948.83333333 -> 5111.5

Revision                              Result                  N
chromium@444804                       7948.83 +- 121.036      12      good
chromium@444826                       7833.33 +- 300.524      12      good
chromium@444837                       8310.0 +- 505.996       12      good
chromium@444842                       7835.5 +- 582.775       12      good
chromium@444843                       8302.33 +- 637.331      12      good
chromium@444843,angle@ab9f9c31f7      7895.5 +- 458.379       12      good
chromium@444843,angle@2cf7c51ec2      5021.33 +- 586.353      6       bad       <--
chromium@444843,angle@65603eb814      5318.83 +- 121.88       6       bad
chromium@444844                       5189.83 +- 324.202      6       bad
chromium@444845                       5063.17 +- 282.614      6       bad
chromium@444847                       4992.0 +- 314.398       6       bad
chromium@444889                       5111.5 +- 426.389       6       bad

To Run This Test
  .\src\out\Release_x64\angle_perftests.exe --test-launcher-print-test-stdio=always --test-launcher-jobs=1

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8989516568319009760

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5817961526329344


| 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!

Comment 16 by jmad...@chromium.org, Jan 25 2017

Status: WontFix (was: Untriaged)
I'm not sure what to do here. Maybe the perf test became sensitive to the order of the tests now? It's kind of discouraging, but I don't see how this could be a real regression.

Thanks for the help Simon in getting the bisect working.
Project Member

Comment 18 by 42576172...@developer.gserviceaccount.com, Apr 11 2017

Cc: pdfium-deps-roller@chromium.org
Owner: pdfium-deps-roller@chromium.org

=== Auto-CCing suspected CL author pdfium-deps-roller@chromium.org ===

Hi pdfium-deps-roller@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 : pdfium-deps-roller
  Commit : b9d2a237108c08e574e608fd18608cae3b1b323c
  Date   : Thu Jan 19 20:48:47 2017
  Subject: Roll src/third_party/pdfium/ 95bec8046..341b5c2c1 (1 commit).

Bisect Details
  Configuration: winx64nvidia_perf_bisect
  Benchmark    : angle_perftests
  Metric       : Uniforms_gl_400_vec4_repeating/score
  Change       : 32.60% | 7641.0 -> 5150.0

Revision             Result                  N
chromium@444804      7641.0 +- 1147.87       12      good
chromium@444826      7711.67 +- 771.305      12      good
chromium@444837      8222.83 +- 668.585      12      good
chromium@444838      7576.5 +- 1160.57       12      bad       <--
chromium@444839      7786.5 +- 506.88        12      bad
chromium@444840      7800.5 +- 508.011       12      bad
chromium@444842      7239.33 +- 3701.12      12      bad
chromium@444847      5074.33 +- 422.16       6       bad
chromium@444889      5150.0 +- 570.74        6       bad

To Run This Test
  .\src\out\Release_x64\angle_perftests.exe --test-launcher-print-test-stdio=always --test-launcher-jobs=1

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8982632318934611168

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5262877268639744


| 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!
►

Sign in to add a comment