1720071 - 4.21 - 3.27% amazon fcp / amazon fnbpaint + 2 more (Linux) regression on Tue July 6 2021

Status: RESOLVED INVALID

(Core :: Layout, defect)

Description

[Florin Strugariu [:Bebe]]

Perfherder has detected a browsertime performance regression from push a83914c4bef76a513c2036911355389f7e9edae8. 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)
4%	amazon	fcp	linux1804-64-shippable-qr	warm webrender	372.04 -> 387.71
4%	amazon	fnbpaint	linux1804-64-shippable-qr	warm webrender	379.12 -> 393.50
4%	amazon	ContentfulSpeedIndex	linux1804-64-shippable-qr	warm webrender	427.62 -> 443.17
3%	amazon	fnbpaint	linux1804-64-shippable-qr	warm webrender	378.31 -> 390.67

Improvements:

Ratio	Suite	Test	Platform	Options	Absolute values (old vs new)
4%	amazon	loadtime	linux1804-64-shippable-qr	warm webrender	571.75 -> 550.50

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

[Timothy Nikkel (:tnikkel)]

Can you have a second look at the perf numbers? The linked push has a test change and a debug only assert change that shouldn't affect opt builds.

Comment 2

[Hiroyuki Ikezoe (:hiro)]

How confident did bug 1717726 regress the performance? The change for bug 1717726 is just adding a debug only assertion (MOZ_ASSERT), if it regressed release builds, compilers did something silly for this do {} while(false), I believe compilers are smart enough to handle it properly.

(Timothy beats me. :))

Comment 3

[Florin Strugariu [:Bebe]]

you are right here after reviewing this alert it it scaused by Bug 1635451 - Limit number of client connections to X for fission to avoid xclient exhaustion (fission)

Comment 4

[Hiroyuki Ikezoe (:hiro)]

Thank you Florin for the clarification!