All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=890897

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=d91332cedad11be35a4ff77a74c530d8bb677825049a88ded1c3be0100b89f0e


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

linux-perf

blink_perf.css - Benchmark documentation link:
  https://bit.ly/blink-perf-benchmarks

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12a9b18f640000

📍 Found significant differences after each of 3 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/12a9b18f640000

[Extension SW] Fix a race in EventAckData by lazyboy@chromium.org
https://chromium.googlesource.com/chromium/src/+/f571252a3f381cc41032a1d09df41b085878a1d1
8693 → 8652 (-41.01)

applist: Fix flash when swiping up app to show home launcher. by sammiequon@google.com
https://chromium.googlesource.com/chromium/src/+/876e61a74532d7a569ad71dd7bc92c9c98ce9994
8652 → No values

Roll AFDO from 71.0.3555.0_rc-r1 to 71.0.3556.0_rc-r1 by afdo-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com
https://chromium.googlesource.com/chromium/src/+/aa1fffa25aab1c9d4c8d2ac3b5a3de67399b17c5
No values → 7569

Assigning to sheriff gbiv@chromium.org because "Roll AFDO from 71.0.3555.0_rc-r1 to 71.0.3556.0_rc-r1" is a roll.

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

Benchmark documentation link:
  https://bit.ly/blink-perf-benchmarks

Yeah, it's almost definitely the AFDO roll.

The history of this benchmark looks like it has a lot of AFDO-induced noise, too. Issue 889742 tracks disabling AFDO for some benchmarks (where blink_perf.css falls under "some"), so hopefully that'll smooth things out here.

Keeping this open either until it recovers for a while, or until AFDO gets turned off for this benchmark.

Thanks!

AFDO rolls in all of the later changes, both positive and negative, on that graph.