New issue
Advanced search Search tips

Issue 725008 link

Starred by 1 user

Issue metadata

Status: Assigned
Owner:
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

6.6%-8.3% regression in memory.desktop at 472679:472764

Project Member Reported by jochen@chromium.org, May 22 2017

Issue description

See the link to graphs below.
 

Comment 1 by jochen@chromium.org, May 22 2017

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=725008

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg2oW0sAsM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDguqjyrgoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg2uHvsQkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgusjorAoM


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

chromium-rel-win10
chromium-rel-win7-gpu-ati
chromium-rel-win7-gpu-intel
chromium-rel-win7-gpu-nvidia
Project Member

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

Cc: tebbi@chromium.org
Owner: tebbi@chromium.org

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

Hi tebbi@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 : tebbi
  Commit : d4f80f4c0117456e87099498512ba51100cbdc79
  Date   : Wed May 17 15:27:37 2017
  Subject: [builtins] Implement %TypedArray%.prototype.map in the CSA

Bisect Details
  Configuration: winx64ati_perf_bisect
  Benchmark    : memory.desktop
  Metric       : memory:chrome:renderer_processes:reported_by_chrome:v8:heap:effective_size_avg/TrivialBlurAnimationPageSharedPageState
  Change       : 5.81% | 6441252.57143 -> 6815744.0

Revision                           Result                 N
chromium@472687                    6441253 +- 886208      14      good
chromium@472726                    6349710 +- 494303      9       good
chromium@472745                    6478702 +- 939966      14      good
chromium@472746                    6516151 +- 970793      14      good
chromium@472746,v8@f1e82a2ee9      6291456 +- 0.0         6       good
chromium@472746,v8@d4f80f4c01      6815744 +- 0.0         14      bad       <--
chromium@472746,v8@6bd1aeee00      6815744 +- 0.0         14      bad
chromium@472747                    6815744 +- 0.0         14      bad
chromium@472748                    6815744 +- 0.0         9       bad
chromium@472750                    6815744 +- 0.0         9       bad
chromium@472755                    6815744 +- 0.0         9       bad
chromium@472764                    6815744 +- 0.0         9       bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release_x64 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=TrivialBlurAnimationPageSharedPageState memory.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8978900726482026032

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5821824266403840


| 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 tebbi@chromium.org, Aug 29 2017

This is an increase in snapshot size caused by the rather big CSA builtins for higher-order typed array functions. https://chromium-review.googlesource.com/c/v8/v8/+/512625 would solve this problem, but causes inconsistent behavior with respect to when TypedArrays operations throw TypeError.

Sign in to add a comment