All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=775306

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=32fce3c475b4d320456528462346d238c8999dc81008a49bba7928485aeb2500


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

chromium-rel-mac11-air

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/16dad00b780000

📍 Found significant differences after each of 9 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/16dad00b780000

Enable the new Cleaner prompt by default
By ftirelo@chromium.org · Fri Oct 13 22:56:03 2017
chromium @ 0cc3fa5dd07df13f6a3c5806a86806a469811b76

<b>[bigint] Support BigInts in <<,>>,>>>,&,|,^ binary ops</b>
By jkummerow@chromium.org · Wed Oct 11 17:36:53 2017
v8 @ 1560988249c97c5cc1763aa364a0423f9e318c13

MIPS[64]: Skip regress-752764 test
By sreten.kovacevic@imgtec.com · Fri Oct 13 08:23:27 2017
v8 @ ad38a4e32391784333e893018f8b723a84e7bb49

[test] Improve printing for investigating hung tests
By machenbach@chromium.org · Fri Oct 13 08:24:36 2017
v8 @ c1f7694b5d784aa3d33d77bf04e6d4d4e50a6508

Reland "[runtime] Make JSFunction::prototype_or_initial_map field optional."
By ishell@chromium.org · Fri Oct 13 09:09:28 2017
v8 @ 36860509fbc5c63f5351aa45e66418a79fa26758

[cleanup] Remove leftover tail-call check.
By bmeurer@chromium.org · Fri Oct 13 09:28:27 2017
v8 @ e8f6d0c176c84176b171c2cd5dcf2d8ff6722ca9

[macro-assembler] Delete unused SmiToDouble
By verwaest@chromium.org · Fri Oct 13 09:40:50 2017
v8 @ 80607c40011ff479c9c74c3a09065233444eecc0

[macro-assembler] Delete unused SmiOperationConstraint
By verwaest@chromium.org · Fri Oct 13 09:52:37 2017
v8 @ 68a16b637d0382bcd7da6dcbf953bdd2c93876ba

Add support for discarding video frames post-decode.
By dalecurtis@chromium.org · Sat Oct 14 01:01:26 2017
chromium @ 5341599c0b2b5a74f62a7734d592d7faae9983c8

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

Am I correct to assume that this is Mac only? I only see mac bots here: https://chromeperf.appspot.com/group_report?sid=32fce3c475b4d320456528462346d238c8999dc81008a49bba7928485aeb2500

If so, I don't see how https://crrev.com/0cc3fa5dd07df13f6a3c5806a86806a469811b76 (Enable the new Cleaner prompt by default) would be causing the regression, since it's a Windows only change. The code depending on the flag that we enabled is only executed if we can register the Software Reporter component, which only happens on Windows (https://cs.chromium.org/chromium/src/chrome/browser/chrome_browser_main.cc?l=543-548). If the software reporter isn't installed, the code affected by the enabled feature is never exercised.

In addition, this feature has been enabled for months in the field, we are currently running an A/B test on stable at 50%, and no significant difference in performance has been observed. Details for this launch are private to Google, but please feel free to ping me if you need more info.

Seems fixed?

https://chromeperf.appspot.com/report?sid=58fdfc8e7f07d50e6ee71c8cb1d2a81d1f7e01dc4cf7f0b1629299dd3e175e2f&rev=508923

As with issue 775309, no clear revert -- probably it's the same issue. There is a v8 roll in the fix, but it doesn't have much.

It might be the same as  issue 775299  which is timing related, but maybe my change helped and another of these was the initial regression?

Any of the other cc'd folk see anything?

It may or may not be mac only.