All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=822812

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=aeca56b94d9516094f4849773dbb28acc339099f4ef44c2a76e3b07e35a3e658


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

android-nexus5
android-nexus5X
android-nexus6
android-nexus7v2
android-one
android-webview-nexus5X
android-webview-nexus6
chromium-rel-mac11-air
chromium-rel-mac11-pro
chromium-rel-win7-dual
chromium-rel-win7-gpu-nvidia
chromium-rel-win7-x64-dual
chromium-rel-win8-dual

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/122746c1440000

This seems to be a regression of an earlier improvement.

The earlier improvement is in [1] and current regression is in [2]. These ranges coincide with enabling "SlimmingPaintV175 by default" which seems suspect.

bisect job will find out more but adding that CL owner FYI.

wangxianzhu@: where you aware of regressions in rasterize_and_record the first time SPV175 landed?


[1] https://chromium.googlesource.com/chromium/src/+log/0865f9bd076062390ba5d6c5a3191fdeca1e1cc3%5E..5381e7250d90e33ed29d7ce3df188c401bc4ad67?pretty=fuller&n=1000

[2]https://chromium.googlesource.com/chromium/src/+log/2f862dc336e98b1ddda51bb9eed3950964be9ffb%5E..1bb11747dabe7331ab9043399d12c5b57c457348?pretty=fuller&n=1000

[3]  https://chromium.googlesource.com/chromium/src/+/3a3c78a924a686ed0d3f90d765b00cdd78453e11

Yes. I'm aware of the regressions. SPv175 performance of had been improved a lot before we enabled it again. Some regressions of record_time are expected because we moved some work load from PrePaint stage (which should have progression) into Paint. We also have many progressions for other tests.

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/122746c1440000

Reland "[SPv175] Enable SlimmingPaintV175 by default" by wangxianzhu@chromium.org
https://chromium.googlesource.com/chromium/src/+/3a3c78a924a686ed0d3f90d765b00cdd78453e11

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

FYI, here is list of metrics that have regressions due to this:

rasterize_and_record_micro.top_25,
smoothness.tough_path_rendering_cases, smoothness.gpu_rasterization.tough_filters_cases,
smoothness.tough_path_rendering_cases,
smoothness.tough_animation_cases,
smoothness.key_silk_cases


I also see some memory regressions that coincide but not yet confirmed
if they are related.

https://chromeperf.appspot.com/group_report?rev=543291

 Issue 823464  has been merged into this issue.

 Issue 823455  has been merged into this issue.

Record_time regression is expected because we moved raster invalidation from PrePaint to Paint. Let's focus on overall performance (frame_time etc.) and PrePaint+Paint time which are tracked by other bugs.

 Issue 830781  has been merged into this issue.