All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=883832

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=951dc9566ee1ce35b9e5ae359a6565bb7f629edf236602580c231062cd15a4d2


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

mac-10_12_laptop_low_end-perf

system_health.common_desktop - Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/123dfeb5640000

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/123dfeb5640000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/11966ceb640000

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/11966ceb640000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/157706f7640000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/157706f7640000

[LayoutNG] Sync PaintBoxDecorationBackground with BoxPainter by kojii@chromium.org
https://chromium.googlesource.com/chromium/src/+/07172170d6f94704bd47389e39251e24b905baaa
572.6 → No values

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

Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

I don't think #7 is correct, the CL changes code that runs only when LayoutNG runtime flag is enabled, and the pinpoint seemed to have failed.

How can I run it again?

Go to the graphs -> click on the regression point -> bisect
I'll run it again but it looks like the graphs are recovering

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/15edbd87640000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/15edbd87640000

Change old and new selection representation to use Node instead of LayoutObject by yoichio@chromium.org
https://chromium.googlesource.com/chromium/src/+/3cb73e6590a6c6c4b34088196b77ed492f7b77ce
657.1 → 579 (-78.11)

Disable loading.desktop/AllRecipes_cold on ChromeOS by vovoy@chromium.org
https://chromium.googlesource.com/chromium/src/+/2de8f4f0f966637c76152b54aee275dcc49fbbb9
611.7 → 701.4 (+89.75)

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

Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

It is really unclear which score is better, lower or higher?

Lower is better (I don't see this info from Pinpoint but if you click at the graphs in #1 the graphs show this information

"Disable loading.desktop/AllRecipes_cold on ChromeOS" disable a test on ChromeOS. It's very unlikyly to cause regression on Mac.

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/1527f313640000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/1527f313640000

Disable loading.desktop/AllRecipes_cold on ChromeOS by vovoy@chromium.org
https://chromium.googlesource.com/chromium/src/+/2de8f4f0f966637c76152b54aee275dcc49fbbb9
651.2 → 756.7 (+105.6)

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

Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

The value went back to normal on 591963
https://chromeperf.appspot.com/group_report?bug_id=883832

"Disable loading.desktop/AllRecipes_cold on ChromeOS" should not affect Mac.

Dispersion of the graph is big. Is the perf test really trusty?