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

Issue 759651 link

Starred by 2 users

Issue metadata

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



Sign in to add a comment

10.2% regression in system_health.memory_desktop at 497639:497657

Project Member Reported by kraynov@chromium.org, Aug 28 2017

Issue description

See the link to graphs below.
 
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Aug 28 2017

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=759651

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=b139a9d22a4f95f30275d3586f22570c502b382d14497c444a1fed66dafaba9e


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

chromium-rel-win10
Project Member

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

Cc: kozyatinskiy@chromium.org
Owner: kozyatinskiy@chromium.org
Status: Assigned (was: Untriaged)

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

Hi kozyatinskiy@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 : Alexey Kozyatinskiy
  Commit : 0bdc7bbabc291651eedc5c2cd963af4528047f98
  Date   : Fri Aug 25 23:43:42 2017
  Subject: [inspector] improved injected-script-source

Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:v8:effective_size_avg/load_news/load_news_hackernews
  Change       : 10.22% | 5132288.0 -> 5656576.0

Revision                           Result              N
chromium@497638                    5132288 +- 0.0      6      good
chromium@497648                    5132288 +- 0.0      6      good
chromium@497651                    5132288 +- 0.0      6      good
chromium@497652                    5132288 +- 0.0      6      good
chromium@497652,v8@0bdc7bbabc      5656576 +- 0.0      6      bad       <--
chromium@497652,v8@109e8b98f5      5656576 +- 0.0      6      bad
chromium@497652,v8@24b88776aa      5656576 +- 0.0      6      bad
chromium@497653                    5656576 +- 0.0      6      bad
chromium@497657                    5656576 +- 0.0      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=release_x64 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=load.news.hackernews system_health.memory_desktop

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8970001397400868512


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

Comment 4 by 42576172...@developer.gserviceaccount.com, Aug 30 2017

Cc: u...@chromium.org
 Issue 760514  has been merged into this issue.
It's unlikely that my CL can introduce 10% memory regression.
kozyatinskiy: Two separate bisects pretty clearly identified this CL as a memory regression (this and the one in https://bugs.chromium.org/p/chromium/issues/detail?id=760514#c3). Please investigate. You can use the docs linked in the bisect comment: https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

ulan: any pointers for kozyatinskiy?
I've only one theory here, we store injected script source as a raw string in binary and new ~20 characters hits page memory limit and we need another 512 Kb page. I'll investigate how we can reduce memory required for injected-script-source.js.
Project Member

Comment 8 by bugdroid1@chromium.org, Sep 11 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/d415be61a121cdfacc226cb074b504fd03971b90

commit d415be61a121cdfacc226cb074b504fd03971b90
Author: Alexey Kozyatinskiy <kozyatinskiy@chromium.org>
Date: Mon Sep 11 22:03:38 2017

[inspector] one more array with nullified __proto__

+ little reduction of injected-script-source size.

Bug:  chromium:759651 
Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel
Change-Id: Ia5d0b31fddc9f6c6c7e547618a6a01e93564bcbc
Reviewed-on: https://chromium-review.googlesource.com/660409
Reviewed-by: Dmitry Gozman <dgozman@chromium.org>
Commit-Queue: Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47958}
[modify] https://crrev.com/d415be61a121cdfacc226cb074b504fd03971b90/src/inspector/injected-script-source.js
[modify] https://crrev.com/d415be61a121cdfacc226cb074b504fd03971b90/test/inspector/runtime/console-methods-expected.txt
[modify] https://crrev.com/d415be61a121cdfacc226cb074b504fd03971b90/test/inspector/runtime/console-methods.js

Status: WontFix (was: Assigned)
I reduced injected-script-source size so both my CLs together now reduce binary size but some other CLs was landed and increased V8 binary size and force using another memory page. So WontFix.

Sign in to add a comment