All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=801131

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=06acfe532483b67e20216db0e81390235606497105caacc4e14bd6d3439915aa


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

chromium-rel-win8-dual

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14bb0ba7040000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/14bb0ba7040000

Surface synchronization: Always allocate a new LocalSurfaceId on auto-resize
By fsamuel@chromium.org · Thu Jan 04 02:04:20 2018
chromium @ 4733241338f7408784ca6c2391c6d9c26ae04d60

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

+jbroman@ I believe you own this test? I was hoping you could help me figure out how to repro locally? Also, does anything in this test do auto-resize?

#4: it would seem odd to me that your CL would affect it (there's only one frame and it isn't resized -- not sure what auto-resize is). As far as running this benchmark:

tools/perf/run_benchmark blink_perf.bindings --story=filter=first-child

This is definitely not my bug. I'm assigning to jbroman@ for triage.

Hard to tell, but this doesn't look like a real regression (doesn't persist, doesn't occur similarly on other platforms).