All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=825001

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=ad216f98d69eee051345f979828285f6a18749cd8da746f684215eb04afbd92f


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

android-nexus5X

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/13a66ef9440000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/13a66ef9440000

Reland: Refactor of Canvas2DLayerBridge resource model by junov@chromium.org
https://chromium.googlesource.com/chromium/src/+/5b07996e178c5f4e153e771d48ab839867fa7440

move //device/hid into //services/device/hid. by ke.he@intel.com
https://chromium.googlesource.com/chromium/src/+/624b52d729b1f1cead14c7c54df22e375199429f

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Assigning junov@. We have an assertion in //services/device/hid which ensures that this code is not compiled on Android.

pinging this bug...

flackr@ (I picked you from third_party/blink/renderer/platform/graphics/OWNERS :) ), pinpoint shows that crrev.com/c/758562/ regressed this metric, but the CL author is not in Chromium anymore. Could you please assign a proper owner for this bug to take a look at it?

Thanks a lot!

I suspect Fernando would know more about this area. Fernando can you triage this?

We already did a bunch of changes to bring back performance due to this change.


And you can see on https://chromeperf.appspot.com/group_report?bug_id=825001 that performance is even better than before.