Dual browser benchmark should have same number of samples for all browsers |
|||||
Issue descriptionSince Chrome is not alive during the first iteration of the benchmark, when WebView is foreground, after running 10 pages 5 times each we end up with 50 memory samples for WebView but only 49 for Chrome. To prevent complexities introduced by this asymmetry, it's better to make sure that both Chrome and WebView are alive on all iterations of the benchmark.
,
Jun 23 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/9a647d91e3999a0ba99963eefd6ce7460483116f commit 9a647d91e3999a0ba99963eefd6ce7460483116f Author: perezju <perezju@chromium.org> Date: Thu Jun 23 14:04:29 2016 [tools/perf] Keep both browsers open for the whole of the dual browser test Ensure that all browsers needed for the story set remain open for the whole duration of the test. This makes sure that, when taking memory dumps, we always get a dump for all browsers all the time. Previous attempt from https://codereview.chromium.org/2073713003/ did not work because the browsers need to be created *after* WPR has been started. BUG= 616043 Review-Url: https://codereview.chromium.org/2097493002 Cr-Commit-Position: refs/heads/master@{#401598} [modify] https://crrev.com/9a647d91e3999a0ba99963eefd6ce7460483116f/tools/perf/page_sets/dual_browser_story.py
,
Jun 30 2016
,
Jun 30 2016
,
Jun 30 2016
,
Jul 27 2016
The fix landed in #2, although the test is still sometimes failing for some other reasons. |
|||||
►
Sign in to add a comment |
|||||
Comment 1 by bugdroid1@chromium.org
, Jun 22 2016