All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=600650

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgyPDCrgoM


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

chromium-rel-win7-gpu-intel

Sent another bisect.

Bisect got it down to this range: https://chromium.googlesource.com/chromium/src/+log/fc3524a0efece0f2b55b08980b74c23be65ed8dd%5E..bc4545b1289c74b44ed6f86c124391d5366eb4e3?pretty=fuller

Not a ton better, but smaller. Trying another bisect.

Looking at the results for this test for a wider revision range, it is clear that this regression is following an improvement of the same magnitude (an improvement-revert pattern):

https://chromeperf.appspot.com/report?sid=f17f71bcfe596d4951b1b4a3d3e11d36fa5feccc44d775a495df8920ce7e00da&start_rev=379792&end_rev=390625

Therefore, this should be closed.

Also, interestingly, this change is dwarfed by a much larger regression-revert pattern in the past several months.