All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=847476

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=73843ecce6019ce3263a7efb23633f321d18c99bf78ccaabecb41df374a3edb6


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

win-high-dpi

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12fec09a240000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/12fec09a240000

[RLS] Ignore fixed when applying root scroll offset for interest rect by pdr@chromium.org
https://chromium.googlesource.com/chromium/src/+/cfebac612b10fdf961dca9df2bb28581577a612a

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

 Issue 847477  has been merged into this issue.

 Issue 847478  has been merged into this issue.

Interesting! The patch that clearly regressed has been rolled out but the regression persists. I double-checked and the revert looks correct. We do not see this regression on other platforms. My guess is that we were on the precipice of a memory regression (maybe we were just barely doing a GC), my patch regressed performance which pushed us over the cliff, but subsequent patches further regressed so that when my patch was rolled out we were still over the cliff.

I think we should close this as wontfix.