New issue
Advanced search Search tips

Issue 793226 link

Starred by 1 user
Status: WontFix
Owner:
hlundin@chromium.org
Closed: Dec 2017
Cc:
kwiberg@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Windows , Mac
Pri: 2
Type: Bug-Regression



Sign in to add a comment

3.1% regression in webrtc_perf_tests at 21141:21141

Project Member Reported by hlundin@chromium.org, Dec 8 2017 
This seems to be due to https://webrtc-review.googlesource.com/c/src/+/26580, which tries to fix  crbug.com/784216 .
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Dec 8 2017 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=793226

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=edcd731eeef06a87ac619e2ecf2b6cc163692cda5571f3e85e042faae0f027ac


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

webrtc-win-large-tests

Comment 2 by hlundin@chromium.org, Dec 8 2017 

Interesting observation is that while the performance regresses for Windows, it improves for Linux and Mac (no alert). Android seems mostly unaffected.

Comment 3 by hlundin@chromium.org, Dec 8 2017

Components: Blink>WebRTC>Audio
Labels: OS-Linux OS-Mac OS-Windows

Comment 4 by hlundin@chromium.org, Dec 8 2017

Status: WontFix (was: Assigned)
It turns out that the windows bot is 32-bit, while linux and mac are 64-bit. So, it turns out the penalty for this fix is 3.1% for this particular test on 32-bit platforms. I think this is acceptable, since it only hits for the 10% loss test. There is no impact on the non-loss test.
►

Sign in to add a comment