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

Issue 715942 link

Starred by 2 users
Status: Fixed
Owner:
█ dominicc@chromium.org
Last visit > 30 days ago
Closed: Jun 2017
Cc:
mlippautz@chromium.org
█ dominicc@chromium.org
haraken@chromium.org
rmcilroy@chromium.org
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression

Blocked on:
issue 722645


Sign in to add a comment

29.5%-36.7% regression in blink_perf.bindings at 467198:467245

Project Member Reported by rmcilroy@chromium.org, Apr 27 2017 
Status at 2017-05-16 JST: dominicc is working on this but is slowed down by perf try not working. See  Issue 722645 .

Comment 1 by rmcilroy@chromium.org, Apr 27 2017 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=715942

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg0tL2qAkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg0pKfvgoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg0s_vpgsM


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

android-nexus5
android-nexus6
android-webview-nexus6
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Apr 28 2017

Cc: dominicc@chromium.org
Owner: dominicc@chromium.org

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

Hi dominicc@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 : dominicc
  Commit : cb56b4ff751b3fef6ada2ae547a2c374caca8361
  Date   : Wed Apr 26 03:07:36 2017
  Subject: Count element name validity per DOM versus HTML parsing.

Bisect Details
  Configuration: android_webview_nexus6_aosp_perf_bisect
  Benchmark    : blink_perf.bindings
  Metric       : create-element/create-element
  Change       : 39.99% | 115.505442055 -> 69.3167998762

Revision             Result                  N
chromium@467213      115.505 +- 5.02771      6      good
chromium@467218      110.431 +- 2.82851      6      good
chromium@467219      72.9679 +- 3.46792      6      bad       <--
chromium@467220      73.0945 +- 1.94014      6      bad
chromium@467222      74.3577 +- 1.11085      6      bad
chromium@467230      69.3168 +- 1.0244       6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-webview --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.bindings

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8981169700284107792

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6754184138850304


| 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!

Comment 4 by dominicc@chromium.org, May 1 2017

Status: Started (was: Untriaged)
I'll take a look. We're trying to add counters to see whether we can change these rules. Maybe this is simply too expensive to measure.

Comment 5 by dominicc@chromium.org, May 16 2017

Blockedon: 722645

Comment 6 by dominicc@chromium.org, May 16 2017

Description: Show this description

Comment 7 by dominicc@chromium.org, May 16 2017 

I have tried adding a fast path for the common case of all ASCII alpha names in <https://chromium-review.googlesource.com/c/505671>.

Comment 8 by dominicc@chromium.org, May 16 2017

Cc: haraken@chromium.org mlippautz@chromium.org
I have been looking at this more closely and I wonder if this is contributing to this:

https://chromium.googlesource.com/chromium/src/+/79f22e60f857a83a592216c33a47035268c2cda7%5E%21/#F0
Project Member

Comment 11 by 42576172...@developer.gserviceaccount.com, May 16 2017 


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

Please try rerunning the bisect.


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


Bisect Details
  Configuration: android_nexus6_perf_bisect
  Benchmark    : blink_perf.bindings
  Metric       : create-element/create-element

Revision             Result                  N
chromium@467197      110.174 +- 3.53716      6      good
chromium@467222      68.1219 +- 2.3579       6      bad

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 blink_perf.bindings

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8979477539964057040

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5684194019115008


| 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!
Project Member

Comment 13 by 42576172...@developer.gserviceaccount.com, May 16 2017 


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

Suspected Commit
  Author : dominicc
  Commit : cb56b4ff751b3fef6ada2ae547a2c374caca8361
  Date   : Wed Apr 26 03:07:36 2017
  Subject: Count element name validity per DOM versus HTML parsing.

Bisect Details
  Configuration: android_nexus6_perf_bisect
  Benchmark    : blink_perf.bindings
  Metric       : create-element/create-element
  Change       : 38.48% | 109.729066114 -> 67.5063937539

Revision             Result                  N
chromium@467197      109.729 +- 9.80505      5      good
chromium@467210      106.319 +- 6.71071      6      good
chromium@467218      107.829 +- 1.17976      6      good
chromium@467219      69.1498 +- 1.60573      6      bad       <--
chromium@467220      68.3648 +- 1.02844      6      bad
chromium@467222      67.5064 +- 3.10137      6      bad

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 blink_perf.bindings

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8979458011370967120

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5684194019115008


| 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!
Project Member

Comment 14 by 42576172...@developer.gserviceaccount.com, May 16 2017 


=== BISECT JOB RESULTS ===
Bisect failed for unknown reasons

Please contact the team (see below) and report the error.


Bisect Details
  Configuration: android_webview_nexus6_aosp_perf_bisect
  Benchmark    : blink_perf.bindings
  Metric       : create-element/create-element


To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-webview --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.bindings

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8979477556951666480

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6070451433897984


| 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!
Project Member

Comment 16 by 42576172...@developer.gserviceaccount.com, May 16 2017 


=== BISECT JOB RESULTS ===
Bisect failed for unknown reasons

Please contact the team (see below) and report the error.


Bisect Details
  Configuration: android_webview_nexus6_aosp_perf_bisect
  Benchmark    : blink_perf.bindings
  Metric       : create-element/create-element


To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-webview --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.bindings

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8979397594699209264

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6070451433897984


| 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!
Project Member

Comment 17 by bugdroid1@chromium.org, May 30 2017 

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/fc0e2af278fff9417db7a458f170044d782523df

commit fc0e2af278fff9417db7a458f170044d782523df
Author: Dominic Cooney <dominicc@chromium.org>
Date: Tue May 30 06:28:42 2017

Remove the entered window from Document.createElement

r467219 regressed blink_perf.bindings create-element. Somewhat
surprisingly it is not the extra checks, but asking the bindings to
provide the entered window which slowed things down.

Bug:  715942 
Change-Id: I5498832ad3f6d5c1a9e3a810b60daabefaca5cd5
Reviewed-on: https://chromium-review.googlesource.com/515186
Commit-Queue: Dominic Cooney <dominicc@chromium.org>
Reviewed-by: Kentaro Hara <haraken@chromium.org>
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Cr-Commit-Position: refs/heads/master@{#475441}
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/bindings/core/v8/V0CustomElementConstructorBuilder.cpp
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/DOMImplementation.cpp
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/DOMImplementation.h
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/DOMImplementation.idl
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/Document.cpp
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/Document.h
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/Document.idl
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/TreeScopeTest.cpp
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/custom/CustomElementTest.cpp
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/custom/CustomElementTestHelpers.h
[modify] https://crrev.com/fc0e2af278fff9417db7a458f170044d782523df/third_party/WebKit/Source/core/dom/custom/CustomElementUpgradeSorterTest.cpp

Comment 18 by dominicc@chromium.org, Jun 1 2017

Status: Fixed (was: Started)
►

Sign in to add a comment