All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=905190

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=78c3b388516e68a470ca4ce61ee7ee80ccb5e37f9aca7cd9d0b066cf3fa941a1


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

mac-10_12_laptop_low_end-perf

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

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/13285b63e40000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/13285b63e40000

Increase padding in layer count based benchmark pages. by vmiura@chromium.org
https://chromium.googlesource.com/chromium/src/+/85f78938e304061d5dea7a5455ffaab5d7881adb
thread_total_all_cpu_time_per_frame: 7.111 → 7.895 (+0.7839)

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

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

This changed the performance test, and is an expected change.  Both latest Chrome and Reference times increased at the same time.

The reason the CPU time went up is that layers were moved such that they are no longer part inside the "SOON" tile priority bucket, which means they are never rasterized.  CC time goes up for layers that still need raster, and would be improved by solving Issue 880186.