All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=830543

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=654a28e31ba0da44755c6e747f1269d6b69d95b44c60bd43d6785bf7ad05c98f


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

chromium-rel-win7-gpu-nvidia

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/15e4f29cc40000

At least my patch "Collect static use information during parsing" included in the commit range could be expected to regress compiler performance a bit, but is required for correct operation. See:

https://chromium-review.googlesource.com/c/angle/angle/+/977964

I don't think there are any easy optimizations we could make to it. I think it's safe to just mark this as WontFix.

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/15e4f29cc40000

Collect static use information during parsing by oetuaho@nvidia.com
https://chromium.googlesource.com/angle/angle/+/94bbed1e2387130e58b3d59d489343b291129e57

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

The slight regression was expected, back to WontFix.