All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=886681

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=1abe4b32a580fa8369a107f9a493eecaf99223b7d202e44ea151a31866d11ac8


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

android-nexus5x-perf

rendering.mobile - Benchmark documentation link:
  https://bit.ly/rendering-benchmarks

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/166da23f640000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/166da23f640000

Roll AFDO from 71.0.3554.0_rc-r1 to 71.0.3555.0_rc-r1 by afdo-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com
https://chromium.googlesource.com/chromium/src/+/4b8803051fc24556622b6c8532a7b7ba73f64a8d
22.09 → 22.56 (+0.468)

Improve EOF handling in the headless REPL by fs@opera.com
https://chromium.googlesource.com/chromium/src/+/67fecf88ed445726d770d6e7e6b0cf746a995efb
22.56 → 22.38 (-0.1749)

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

Benchmark documentation link:
  https://bit.ly/rendering-benchmarks

My change was to headless - which I don't think would be used (or even built) in this case? The numbers also seem to point at the AFDO roll.

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/15629670e40000

Lets see what the second bisect finds.

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/15629670e40000

Roll AFDO from 71.0.3554.0_rc-r1 to 71.0.3555.0_rc-r1 by afdo-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com
https://chromium.googlesource.com/chromium/src/+/4b8803051fc24556622b6c8532a7b7ba73f64a8d
22.12 → 22.52 (+0.3982)

Assigning to sheriff gbiv@chromium.org because "Roll AFDO from 71.0.3554.0_rc-r1 to 71.0.3555.0_rc-r1" is a roll.

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

Benchmark documentation link:
  https://bit.ly/rendering-benchmarks

Yeah, I'm totally willing to buy that this is AFDO's fault

(Incoming copy-paste ;) )

At this point, I've received five perf bugs for this single 3555 profile roll, all of which have recovered with the 3556 roll.

It's rare (< monthly), but we will occasionally have a profile that's just way off. I'll look into whether there's any obvious reason for this particular flake (we use Chrome ToT; if that crashes or glitches out a benchmark, or profile gets really skewed), but given that everything's back to normal and there's ~no signal we can realistically get from these flakes, ...

Hopefully our swap to field-focused profiles (Q4 maybe?) will make this issue go away, but only time will tell.

Added a note of this all in the denoising bug, issue 849881.