All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=821408

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=5d98502017c35d17ae2b1c24b9177e6f09e64b1ecb9912cc3ed1c4ef042b3793


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

android-nexus5

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14d5ec96440000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/14d5ec96440000

ResourceLoadScheduler: remove stale configurations by toyoshim@chromium.org
https://chromium.googlesource.com/chromium/src/+/a3d458c7e9c8ebf21833bbaa48ee4e4159a516dc

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

See also  crbug.com/821409 

The change just stopped memory reporting, and does not change to run any extra code. So, my possible guess is that memory walk for the report incidentally kept current renderer's memory on CPU memory cache, and that have a good effect on smoothness by chance.

In the same smoothness.tough_canvas_cases Test Suite, I found four improved cases.

- frame_times/http___ie.microsoft.com_testdrive_Performance_LetItSnow_ (by 9.3%)
- frame_times/tough_canvas_cases_canvas-animation-no-clear.html (by 21.8%)
- frame_times/http___www.craftymind.com_factory_guimark2_HTML5ChartingTest.html (by 7.6%)
- frame_times/tough_canvas_cases_canvas-animation-no-clear.html (by 13.8%)

So, I'd close this as WontFix.