New issue
Advanced search Search tips

Issue 812708 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner: ----
Closed: Feb 2018
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression

Blocked on:
issue 813883



Sign in to add a comment

19.3% regression in media_perftests at 536224:536251

Project Member Reported by wolenetz@chromium.org, Feb 15 2018

Issue description

This trace looks like a valid alert regression (caveat: media_perftests have no ref trace).

Perhaps useful to investigating is that two other media_perftests' metrics *improved* in the same range as this alert's original range. Those improvements are convert_fifo_only on chromium-rel-win7-gpu-{ati,nvidia}: https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICQnN6eoQoM


 
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Feb 15 2018

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=812708

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=15576f0cde23a271fd1b654c1c4560971ffeaaed44ea014c5ebc271aa9c22516


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

chromium-rel-win7-gpu-ati
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Feb 15 2018

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/12d1a743840000
#4 is across a wider range than #2 (extended earlier).
Hmm. I'm not sure why #4 still hasn't completed. I'll try starting another shortly.
Project Member

Comment 8 by 42576172...@developer.gserviceaccount.com, Feb 20 2018

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/14ef42f7840000
Blockedon: 813883
I'll try an even wider bisect. It's really hard to tell without a ref trace.
Project Member

Comment 12 by 42576172...@developer.gserviceaccount.com, Feb 20 2018

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/14e37c8f840000
Status: WontFix (was: Untriaged)
Given no repro multiple times and with widened bisects, my best guess is this isn't a Chrome regression, rather a change in infra. Lack of ref trace (crbug 714812) makes it hard to verify this assumption further. Closing as no-repro.

Sign in to add a comment