New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 729759 link

Starred by 2 users

Issue metadata

Status: WontFix
Owner:
Last visit > 30 days ago
Closed: Jun 2017
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

1.5%-3.2% regression in memory.top_10_mobile at 475537:475803

Project Member Reported by m...@chromium.org, Jun 5 2017

Issue description

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

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

Hi fsamuel@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 : fsamuel
  Commit : 505078044cfca8e67070b4daed6e6157d7641774
  Date   : Tue May 30 18:21:01 2017
  Subject: SurfaceDependencyTracker: Only observe BeginFrames if a deadline is needed

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:private_dirty_size_avg/foreground/https_m_facebook_com_rihanna
  Change       : 3.18% | 65077176.0 -> 67148386.6667

Revision             Result                   N
chromium@475536      65077176 +- 558920       6      good
chromium@475572      65592589 +- 798151       6      good
chromium@475590      65170701 +- 877986       6      good
chromium@475599      65549581 +- 524149       6      good
chromium@475600      65120184 +- 758407       6      good
chromium@475601      67717731 +- 1157186      6      bad       <--
chromium@475602      67522488 +- 387197       6      bad
chromium@475604      67404387 +- 477159       6      bad
chromium@475608      67347725 +- 288102       6      bad
chromium@475680      67148387 +- 795588       6      bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8977590958727132864

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5852799528599552


| 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!
Issue 729223 has been merged into this issue.
It seems highly unlikely that my change would cause a memory regression. Surface synchronization is not yet turned on on Android. 
Note that field trials are turned on via this config on the perfbots: https://cs.chromium.org/chromium/src/testing/variations/fieldtrial_testing_config.json
Is it in that file?

The bisect results in #3 and https://bugs.chromium.org/p/chromium/issues/detail?id=729223#c6 both look like a pretty clear jump at your CL.

I'll kick off a few more bisects, but please investigate using the documentation linked on the bisect bug: https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:private_dirty_size_avg/background/after_http_m_intl_taobao_com_group_purchase_html

Revision             Result                  N
chromium@475679      26290341 +- 914525      21      good
chromium@475803      26275127 +- 873369      21      bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8977442251561663968

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5538624549945344


| 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!

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

Suspected Commit
  Author : fsamuel
  Commit : 505078044cfca8e67070b4daed6e6157d7641774
  Date   : Tue May 30 18:21:01 2017
  Subject: SurfaceDependencyTracker: Only observe BeginFrames if a deadline is needed

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:private_dirty_size_avg/foreground/https_m_facebook_com_rihanna
  Change       : 3.23% | 65019149.3333 -> 67121762.6667

Revision             Result                   N
chromium@475536      65019149 +- 577417       6      good
chromium@475572      65286755 +- 699623       6      good
chromium@475590      65373453 +- 1024977      6      good
chromium@475599      65053283 +- 404299       6      good
chromium@475600      65052600 +- 616009       6      good
chromium@475601      67471288 +- 562623       6      bad       <--
chromium@475602      67536824 +- 467545       6      bad
chromium@475604      67305400 +- 541168       6      bad
chromium@475608      67086947 +- 848429       6      bad
chromium@475680      67121763 +- 482960       6      bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8977442297842346448

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5262034561662976


| 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!

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

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:private_dirty_size_avg/background/after_http_m_intl_taobao_com_group_purchase_html

Revision             Result                   N
chromium@475679      26284880 +- 1002982      21      good
chromium@475803      26278248 +- 777947       21      bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8977408250555026560

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5538624549945344


| 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!

=== BISECT JOB RESULTS ===
Bisect was unable to run to completion

Error: INFRA_FAILURE

The bisect was able to narrow the range, you can try running with:
  good_revision: e146085ef2bed9d808fc90288f72fcaa2e3a0788
  bad_revision : 0529413d84c06f3720bf4a56b28bb2c7358cefd1

If failures persist contact the team (see below) and report the error.


Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:private_dirty_size_avg/background/after_https_m_facebook_com_rihanna

Revision             Result                   N
chromium@475536      53951761 +- 2978411      21       good
chromium@475572      53796406 +- 3095029      14       good
chromium@475591      53954394 +- 2621880      14       good
chromium@475599      53641050 +- 4014980      14       good
chromium@475600      ---                      ---      build failure
chromium@475608      54343222 +- 3342752      14       bad
chromium@475680      54134520 +- 4862212      21       bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8977442280139746656

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5343472577413120


| 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!
Cc: jasontiller@chromium.org
 Issue 728375  has been merged into this issue.

=== BISECT JOB RESULTS ===
Perf regression found but unable to narrow commit range

Build failures prevented the bisect from narrowing the range further.


Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:private_dirty_size_avg/background/after_https_m_facebook_com_rihanna
  Change       : 1.66% | 53778266.2857 -> 54693722.8571

Suspected Commit Range
  2 commits in range
  https://chromium.googlesource.com/chromium/src/+log/e146085ef2bed9d808fc90288f72fcaa2e3a0788..505078044cfca8e67070b4daed6e6157d7641774


Revision             Result                   N
chromium@475536      53778266 +- 2964202      14       good
chromium@475572      53690202 +- 2716139      14       good
chromium@475590      53829271 +- 4207866      21       good
chromium@475599      53974290 +- 1036445      14       good
chromium@475600      ---                      ---      build failure
chromium@475601      54126426 +- 4089552      14       bad
chromium@475602      54558685 +- 648138       9        bad
chromium@475604      54512719 +- 856165       6        bad
chromium@475608      54127889 +- 5168947      21       bad
chromium@475680      54693723 +- 1201809      14       bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 memory.top_10_mobile

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8977332755629525488

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5343472577413120


| 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!
Status: WontFix (was: Untriaged)
Should be back to normal levels ToT after I landed https://codereview.chromium.org/2917743002/

Sign in to add a comment