All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=856044

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=b0f5bd75b76e3302fa219fe4d798d1c74bc8e07f9d34e9e5d5aa9a77c81bc1fe


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

android-nexus5
android-nexus5X

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12c47ed3240000

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/12c47ed3240000

The swarming task expired. The bots are likely overloaded, dead, or misconfigured.

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/1487d161a40000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/1487d161a40000

Reland "paint timing: Use presentation-time instead of swap-time." by sadrul@chromium.org
https://chromium.googlesource.com/chromium/src/+/992979a662b28e7b9a62419f87b1839557611efc
582.1 → 759 (+177)

presentation feedback: Use a GLFence fallback. by sadrul@chromium.org
https://chromium.googlesource.com/chromium/src/+/3821fdf5c54add24d02adef248a311c293ea0bd0
774.7 → 805.5 (+30.82)

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

The CL changes to using the presentation-time, rather than gpu-swap-time, to report the time-to-first-paint metrics. So the new results are more precise, and are expected to be a bit larger than the old metrics.