1859427 - 354.03 - 234.9% perf_reftest_singletons input-type-number-setAttribute.html / perf_reftest_singletons input-type-number-setAttribute.html + 1 more (Linux, OSX, Windows) regression on Tue October 10 2023

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Andrej (:aglavic)]

Perfherder has detected a talos performance regression from push 686073214e498a3ed6018f15d9cd093812f5ab6f. 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)
354%	perf_reftest_singletons input-type-number-setAttribute.html	windows10-64-shippable-qr	e10s fission stylo webrender	74.85 -> 339.85
353%	perf_reftest_singletons input-type-number-setAttribute.html	macosx1015-64-shippable-qr	e10s fission stylo webrender	78.28 -> 354.27
235%	perf_reftest_singletons input-type-number-setAttribute.html	linux1804-64-shippable-qr	e10s fission stylo webrender	184.03 -> 616.33

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 patch(es) may be backed out in accordance with our regression policy.

If you need the profiling jobs you can trigger them yourself from treeherder job view or ask a sheriff to do that for you.

For more information on performance sheriffing please see our FAQ.

The patch was recently uplifted to beta and we also recorded the same regression there
beta regression

Comment 1

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

Set release status flags based on info from the regressing bug 1857506

Comment 2

[Emilio Cobos Álvarez [:emilio]]

Unfortunately I think we're going to have to take this because it's a correctness issue. We're still much faster than before bug 1842027 which is when this subtest was introduced.

I'll take a look to see if there's some wins to be had here.

Comment 3

[Emilio Cobos Álvarez [:emilio]]

Attachment: Bug 1859427 - Cache number parsers and formatters per language. r=mstange,smaug

This makes number localization cheaper / halves the time in the
microbenchmark.

Change the content-language handling to use atoms. This exposes some
interesting inconsistencies but I tried not to change behavior there.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/f392a193fe58
Cache number parsers and formatters per language. r=smaug

Comment 5

[Noemi Erli[:noemi_erli]]

https://hg.mozilla.org/mozilla-central/rev/f392a193fe58

Comment 6

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

The patch landed in nightly and beta is affected.
:emilio, is this bug important enough to require an uplift?

• If yes, please nominate the patch for beta approval.
• If no, please set status-firefox119 to wontfix.

For more information, please visit BugBot documentation.

Comment 7

[Emilio Cobos Álvarez [:emilio]]

Probably not.

Comment 8

[Andrej (:aglavic)]

(In reply to Noemi Erli[:noemi_erli] from comment #5)

https://hg.mozilla.org/mozilla-central/rev/f392a193fe58

We just wanted to let you know that, from our end we see the improvements this patch made, thanks for prioritizing performance :)

== Change summary for alert #40022 (as of Mon, 23 Oct 2023 23:00:56 GMT) ==

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)
71%	perf_reftest_singletons input-type-number-setAttribute.html	macosx1015-64-shippable-qr	e10s fission stylo webrender	354.47 -> 101.89
71%	perf_reftest_singletons input-type-number-setAttribute.html	macosx1015-64-shippable-qr	e10s fission stylo webrender	354.31 -> 103.05
69%	perf_reftest_singletons input-type-number-setAttribute.html	windows10-64-shippable-qr	e10s fission stylo webrender	338.11 -> 105.46
62%	perf_reftest_singletons input-type-number-setAttribute.html	linux1804-64-shippable-qr	e10s fission stylo webrender	612.47 -> 232.92

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=40022