All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=913500

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=b0bb7a3927d6999935a2cf0d5f7f2a682f2642d36b46c579b7ce7cf3d7048167


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

Win 7 Nvidia GPU Perf

loading.desktop - Benchmark documentation link:
  https://bit.ly/loading-benchmarks

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/143960d4140000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/143960d4140000

Start a field trial for Isolated code cache by mythria@chromium.org
https://chromium.googlesource.com/chromium/src/+/142ae7f8b7a7426ca426524a9593a39cc3e4f9ac
timeToFirstContentfulPaint: 262.1 → 159.4 (-102.7)

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

Benchmark documentation link:
  https://bit.ly/loading-benchmarks

This is only a regression on a couple of bots on one page. My change could have impacted a bit, but I don't think it should have such a large regression. I started a couple of more bisects and will also look into traces to see what's happening there.

 Issue 911469  has been merged into this issue.

Looking at the bisect job, looks like my cl has actually improved things. I am trying to find a windows machine to try and reproduce this locally. Though on Windows 10 bots there is no regression. So I am not very sure if I would be able to try this locally.

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14e8d5e9140000