All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=613181

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjO-8twoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgzLDc9AoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjMWppwoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjIqopgkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjJaSpwoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjOb6rQsM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjP23qwoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgzLTmvgoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjKa58AgM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjI_qqwkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjOb67QgM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjKGgrQoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgjOHotgkM


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

chromium-rel-mac-hdd
chromium-rel-mac-retina
chromium-rel-mac10
chromium-rel-mac11

It looks like two of the regressions linked to this bug are explained by the GPU raster revert. We're planning on turning GPU raster back on, and have it in a 50% finch experiment at the moment, so no action is needed on those. The final issue attached to this bug, for the regression of "thread_times.tough_scrolling_cases", does not seem to be related to GPU raster - I don't see an improvement when GPU raster was originally committed. That issue probably needs a new bug.

Thanks for the information. The metrics have all recovered which I believe is attributed to enabling GPU raster back on on Mac waterfall [1]. \o/

The last regression seems to also be related. I see improvement when GPU raster was originally submitted and they have recovered now. It also makes sense, because without GPU raster CPU has to do more work. :)


[1]https://chromium.googlesource.com/chromium/src/+/f91789afe1568b37a4e93dfe0580c96df24137f6

Deleted: Screenshot from 2016-05-25 10:41:22.png 69.1 KB

	Screenshot from 2016-05-25 10:41:22.png 69.1 KB View Download