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

Issue 806827 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 806239
Owner:
Closed: May 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

A zero-to-nonzero to 13.8% regression in performance_browser_tests at 531684:531967

Project Member Reported by primiano@chromium.org, Jan 29 2018

Issue description

See the link to graphs below.
 
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Jan 29 2018

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=806827

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=9bdbade0716655dc50522356e28ded4d7a9310dcd08d18ed6ba7ea5ea7cd327d


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

chromium-rel-mac11-pro
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Jan 29 2018

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/12bf0c06840000
Kicking off a few more bisects; looks like the first one failed due to swarming task timeouts.
😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/178b559c440000
Cc: x...@chromium.org chfremer@chromium.org m...@chromium.org piman@chromium.org
Owner: m...@chromium.org
Status: Assigned (was: Untriaged)
📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/12945ddc440000

Rewrite WebContentsVideoCaptureDevice, to use new VIZ-based capture impl by miu@chromium.org
https://chromium.googlesource.com/chromium/src/+/77e19b91361effda2f4eb9f278d9e038aca58d7c

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions
Cc: s...@chromium.org sandv@chromium.org rharrison@chromium.org halliwell@chromium.org rsesek@chromium.org
📍 Found significant differences after each of 3 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/12d93c1c440000

Fix broken XFA builds with new V8 by rharrison@chromium.org
https://pdfium.googlesource.com/pdfium/+/4ad801bdc86405888290037065de5d3a3b447167

[Chromecast] Added AIDL to get assistant device logs by sandv@chromium.org
https://chromium.googlesource.com/chromium/src/+/4ddbad7f38c2b690f6bf2d7de250cc9deb812ced

Rewrite WebContentsVideoCaptureDevice, to use new VIZ-based capture impl by miu@chromium.org
https://chromium.googlesource.com/chromium/src/+/77e19b91361effda2f4eb9f278d9e038aca58d7c

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions
Cc: -rharrison@chromium.org
This shouldn't be my CL. XFA is only enabled for fuzzers and not any browser builds.
Cc: -halliwell@chromium.org

Comment 14 by m...@chromium.org, Mar 6 2018

Status: Started (was: Assigned)
No need to keep running pinpoint jobs. I think it's pretty clear it was my "Rewrite WebContentsVideoCaptureDevice" change. :)

On the whole, we replaced the legacy impl with a new one that does have increased latency in frame delivery (because it goes over mojo and process boundaries). The regression of <2 ms is totally acceptable, since delivery to remote displays is on the order of 400 ms.

My only concern is the "frame_drop_rate" metric, which shot up to 100% on Mac Pro. I'll need to investigate that, since it doesn't jive with our testing lab results in-office. It could be the test was broken and needs fixing.
Status: Assigned (was: Started)
📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/1289c2a2440000

Rewrite WebContentsVideoCaptureDevice, to use new VIZ-based capture impl by miu@chromium.org
https://chromium.googlesource.com/chromium/src/+/77e19b91361effda2f4eb9f278d9e038aca58d7c

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions
Cc: -rsesek@chromium.org

Comment 17 by m...@chromium.org, May 15 2018

Components: Internals>Media>ScreenCapture

Comment 18 by m...@chromium.org, May 15 2018

Mergedinto: 806239
Status: Duplicate (was: Assigned)

Sign in to add a comment