All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=830665

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=847c2e7566d050f7febe0e62feaa18ec20c920aef68053a2aa2f433db762a87f


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

chromium-rel-win7-gpu-ati

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/114d5b78c40000

Looking through the ChangeLog:
https://chromium.googlesource.com/chromium/src/+log/d22bdd95acd389c555257634a1e74132338b94a9%5E..0f2b5cdd3a4386de2f44efbaf248b4a367bd377e?pretty=fuller&n=1000

nothing obvious turns up. There is an ANGLE roll which looks unrelated, one change to DrawingBuffer teardown which should not affect runtime.

There are some changes related to viz, surfaces, etc. which I also don't think should have affected performance; or, if they did, the regression would show up in other benchmarks too.

Let's see what the Pinpoint job says.

Unless there are multiple causes, this is because of an ANGLE change. I have a patch already landed that should fix some of the regression.

Merging into existing issue.

Also feel free to un-dupe if multiple causes show up.

Aha. Very good Jamie. Thanks.