Regression in Event.Latency.TouchToFirstScrollUpdateSwapBegin in 55 |
|||||
Issue descriptionThere seems to be a significant regression for Event.Latency.TouchToFirstScrollUpdateSwapBegin in 55.0.2860.0 looking at https://uma.googleplex.com/timeline_v2?sid=f8a49e8f066600697b1bb72e99a250cd
,
Oct 14 2016
I ran a bisect for first_gesture_scroll_update_latency/http___games.yahoo.com in the dashboard https://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus5X_perf_bisect/builds/764/steps/Post%20bisect%20results/logs/stdio and it seems that we got a regression after this rollout https://codereview.chromium.org/2310733002 I'm not sure how to dig deeper here. nednguyen@ do you have any suggestions?
,
Oct 14 2016
A catapult change wouldn't have effected the UMA metric.
,
Oct 17 2016
Looking at this https://uma.googleplex.com/timeline_v2?sid=457c6918723aff01fc3b96010a0b6294 which is splitting based on different devices, it does seem that the regression happened on only a subset of devices. I'm not able to see what the affected devices have in common or why Nexus 5 for example is completely unchanged. skyostil@, brianderson@, alexclarke@, do you have any opinion on this issue and how to proceed?
,
Oct 18 2016
Should we just close this? The regression has improved. I looked at the change lists for the regression and following improvement and nothing jumped out at me as being a cause.
,
Oct 18 2016
Yup. Makes sense. The regression is fixed in the next release. Let's mark it as fixed then. |
|||||
►
Sign in to add a comment |
|||||
Comment 1 by nzolghadr@chromium.org
, Sep 22 2016