1690967 - 0.13 - 0.56% installer size (osx-aarch64-shippable, osx-shippable) regression on push f71aafc4e84564120b0357deb419ccab1afb09a2 (Tue February 2 2021)

Status: RESOLVED WONTFIX

(Core :: Graphics: WebRender, defect)

Description

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

Perfherder has detected a build_metrics performance regression from push f71aafc4e84564120b0357deb419ccab1afb09a2. 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)
1%	installer size		osx-shippable	nightly	85,372,788.50 -> 85,849,150.08
0.13%	installer size		osx-aarch64-shippable	aarch64 nightly	79,397,469.92 -> 79,501,547.67
0.13%	installer size		osx-shippable	nightly	85,207,478.06 -> 85,314,230.75

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

[Lee Salzman [:lsalzman]]

This is expected due to the addition of new critical code on heavily inlined paths that yielded some very substantial performance gains. We will accept the regression here.

Comment 2

[Lee Salzman [:lsalzman]]

Attachment: Bug 1690967 - remove some calls to commit_span. r?jrmuizel

This removes some calls to commit_span from inside the draw_span specializers.
Instead it relies upon the span rasterizer loop to do some of the work, which
will incur a function pointer call in the rare case we actually return out
of a specializer early. This shouldn't be that performance critical and will
remove some inliner bloat.

Comment 3

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

Set release status flags based on info from the regressing bug 1674524

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/30a57c252ef0
remove some calls to commit_span. r=bradwerth

Comment 5

[Andreea Pavel [:apavel]]

https://hg.mozilla.org/mozilla-central/rev/30a57c252ef0

Comment 6

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

The graph doesn't seem like fixed. https://treeherder.mozilla.org/perfherder/graphs?highlightAlerts=1&highlightChangelogData=1&highlightedRevisions=1e0238293dea9dd83644ff46dca31c9046339dcb&series=autoland,1921012,1,2&timerange=1209600

Comment 7

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

== Change summary for alert #28622 (as of Thu, 04 Feb 2021 10:34:33 GMT) ==

Regressions:

Ratio	Suite	Test	Platform	Options	Absolute values (old vs new)
7%	perf_reftest_singletons	id-getter-1.html	macosx1014-64-shippable-qr	e10s stylo webrender	460.27 -> 492.01

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=28622

Comment 8

[Lee Salzman [:lsalzman]]

The remaining regressions are acceptable, as stated above. I did what I could to mitigate the size increases somewhat, and there is nothing more that can be done without impacting performance significantly.

Comment 9

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

== Change summary for alert #28664 (as of Sun, 07 Feb 2021 20:29:36 GMT) ==

Improvements:

Ratio	Suite	Test	Platform	Options	Absolute values (old vs new)
6%	perf_reftest_singletons	id-getter-1.html	macosx1014-64-shippable-qr	e10s stylo webrender	491.95 -> 462.17

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=28664