Closed Bug 1733957 Opened 3 years ago Closed 2 years ago

11.32% perf_reftest_singletons inline-style-cache-1.html (Linux) regression on Wed September 29 2021

Categories

(Core :: DOM: Content Processes, defect, P3)

defect

Tracking

()

RESOLVED FIXED
96 Branch
Fission Milestone Future
Tracking Status
firefox-esr78 --- unaffected
firefox-esr91 --- unaffected
firefox92 --- unaffected
firefox93 --- unaffected
firefox94 --- wontfix
firefox95 --- wontfix
firefox96 --- fixed

People

(Reporter: aesanu, Assigned: smaug)

References

(Blocks 1 open bug, Regression)

Details

(4 keywords)

Attachments

(1 file)

Perfherder has detected a talos performance regression from push 8f2f2cfb962094db37c4a08ca10c181000699930. 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)
11% perf_reftest_singletons inline-style-cache-1.html linux1804-64-shippable-qr e10s fission stylo webrender 1,034.14 -> 1,151.25

Improvements:

Ratio Test Platform Options Absolute values (old vs new)
7% tsvgr_opacity macosx1014-64-shippable-qr e10s fission stylo webrender 152.22 -> 141.51
4% tp5o_scroll macosx1015-64-shippable-qr e10s fission stylo webrender 1.63 -> 1.57
3% tp5o linux1804-64-shippable-qr e10s fission stylo webrender 260.42 -> 251.46
3% tp5o windows10-64-shippable-qr e10s fission stylo webrender 273.18 -> 264.81
3% tsvg_static macosx1014-64-shippable-qr e10s fission stylo webrender 67.67 -> 65.62

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.

Flags: needinfo?(nika)

I think the performance change is coming from bug 1731132. Bug 1731792 just made that change effective on Fission.

But inline-style-cache-1.html seems to be a bit bogus. It depends on whether or not paint happens before load event.

emilio, would it make sense to wait for the page load in the test and only then create all the elements?

Flags: needinfo?(nika) → needinfo?(emilio)

Yeah, that'd make sense to me.

Flags: needinfo?(emilio)

Setting status-firefox94=wontfix because we don't need to uplift a fix to Beta 94. This bug is a problem with the inline-style-cache-1.html Talos test.

Fission Milestone: ? → ---
Regressed by: 1731132
No longer regressed by: 1731792
Has Regression Range: --- → yes

Olli says fixing this test doesn't need to block Fission MVP.

Fission Milestone: --- → Future
Comment 3 is private: false
Priority: -- → P3
Assignee: nobody → bugs
Status: NEW → ASSIGNED
Pushed by opettay@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/e2cae4b15039
tweak inline-style-cache-1.html to depend less on how the page is loaded (perf numbers are expected to change a bit), r=emilio,perftest-reviewers,sparky
Status: ASSIGNED → RESOLVED
Closed: 2 years ago
Resolution: --- → FIXED
Target Milestone: --- → 96 Branch
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: