1683605 - 7.91 - 7.93% tp5o_scroll (windows10-64-shippable-qr) regression on push 756db199e7a1d1e8b5a4199ea2f866cf3c579f22 (Wed December 16 2020)

Status: RESOLVED INVALID

(Core :: Graphics: WebRender, defect)

Description

[Alexandru Ionescu (needinfo me) [:alexandrui]]

Perfherder has detected a talos performance regression from push 756db199e7a1d1e8b5a4199ea2f866cf3c579f22. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

Ratio	Suite	Test	Platform	Options	Absolute values (old vs new)
8%	tp5o_scroll		windows10-64-shippable-qr	e10s stylo webrender	1.81 -> 1.96
8%	tp5o_scroll		windows10-64-shippable-qr	e10s stylo webrender	1.81 -> 1.96

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests. Please follow our guide to handling regression bugs and let us know your plans within 3 business days, or the offending patch(es) will be backed out in accordance with our regression policy.

For more information on performance sheriffing please see our FAQ.

Comment 1

[Alexandru Ionescu (needinfo me) [:alexandrui]]

The bug that cause this is Bug 1678545 but for some reason I'm not allowed to see it or to add it in the regressed by field.

Comment 2

[Dimi Lee [:dimi]]

The patch in Bug 1678545 changed how we handle referrer when a redirect occurs, so I don't know how this is related to the tp5o_scoll test here.
The only possible scenario I can imagine is that those websites load different content based on the referrer, which affects the test, but I am not sure this is the reason...

I pushed a patch which reverts the change made, the performance doesn't seem to be changed, see here:
https://treeherder.mozilla.org/perfherder/comparesubtest?originalProject=autoland&newProject=try&newRevision=8a3a8f5bc36a9fec007c2ea31fffd0bba3cef7ea&originalSignature=2815278&newSignature=2824805&framework=1&originalRevision=756db199e7a1d1e8b5a4199ea2f866cf3c579f22

Hi Alexandru, Is it possible this is misclassified? Do you have any suggestion on how to test it? Thanks!

Comment 3

[Alexandru Ionescu (needinfo me) [:alexandrui]]

Bebe, could you please look into this?

Comment 4

[Alexandru Ionescu (needinfo me) [:alexandrui]]

(In reply to Dimi Lee [:dimi][:dlee] from comment #2)

Hi Alexandru, Is it possible this is misclassified? Do you have any suggestion on how to test it? Thanks!

Sorry for the delay, PTOs and stuff.
Yes, is very possible that it was missclassified, the graph is pretty noisy and unstable. I did some retriggers, will update as I have new data.

Comment 5

[Alexandru Ionescu (needinfo me) [:alexandrui]]

You're right, there's another culprit. I reassigned the alert. Closing as invalid.