1754806 - 35.54% JS (Linux) regression on Mon February 7 2022

Status: RESOLVED FIXED

(Core :: Graphics: WebRender, defect)

Description

[Andrej (:aglavic)]

Perfherder has detected a awsy performance regression from push b7f88e5c537bb7d64ba9d4cb58b052650d510638. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

Ratio	Test	Platform	Options	Absolute values (old vs new)
47%	Heap Unclassified	linux1804-64-shippable-qr	tp6	217,212,336.17 -> 319,185,222.07

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)
38%	JS	linux1804-64-shippable-qr	tp6	279,847,831.90 -> 174,237,172.25
36%	JS	linux1804-64-shippable-qr	tp6	281,836,628.95 -> 181,662,092.22

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

[Glenn Watson [:gw]]

I'll investigate this today / early next week - the links above to Heap Unclassified and JS both go to a page that says to do: add test definition. Who can we ask about what these numbers represent?

Comment 2

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1749380

Comment 3

[Andrej (:aglavic)]

Hi Glenn, thanks for letting us know that the awsy docs show those todo statements. I will confirm with my team the exact values those numbers represent and let you know asap. From the awsy docs if you scroll lower down there is a section on base heap-unclassified and base content JS. It looks like the heap-unclassified section in base is something you can use as a definition while the JS one looks like too different of a definition to use.

Comment 4

[Glenn Watson [:gw]]

Fixed by backout - I will be re-landing this patch series next week with some additional changes that should resolve this regression. Please re-open if it occurs again.

Comment 5

[Marco Castelluccio [:marco]]

Given the large regression, I think we should consider this S2 and not S3.