1872564 - optimize DisplayPortUtils::MaybeCreateDisplayPortInFirstScrollFrameEncountered by checking for leaf frames

Status: RESOLVED FIXED

(Core :: Panning and Zooming, enhancement, P3)

Comment 1

[Timothy Nikkel (:tnikkel)]

Attachment: Bug 1872564. Optimize DisplayPortUtils::MaybeCreateDisplayPortInFirstScrollFrameEncountered by checking for leaf frames. r?botond,#layout-reviewers

On 20 runs of speedometer3 this makes MaybeCreateDisplayPortInFirstScrollFrameEncountered 30% faster.

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/9ea90dc23395
Optimize DisplayPortUtils::MaybeCreateDisplayPortInFirstScrollFrameEncountered by checking for leaf frames. r=layout-reviewers,emilio

Comment 3

[Cristian Tuns]

https://hg.mozilla.org/mozilla-central/rev/9ea90dc23395

Comment 4

[Nicolas Chevobbe [:nchevobbe]]

it looks like this caused performance regression in some DevTools tests:

== Change summary for alert #40836 (as of Thu, 04 Jan 2024 17:12:17 GMT) ==

Regressions:

Ratio	Test	Platform	Options	Absolute values (old vs new)
44%	damp console.typing	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	408.81 -> 590.60
41%	damp console.typing	linux1804-64-shippable-qr	e10s fission stylo webrender	421.14 -> 595.40
7%	damp console.typing	windows10-64-shippable-qr	e10s fission stylo webrender-sw	376.48 -> 404.18
7%	damp console.typing	windows10-64-shippable-qr	e10s fission stylo webrender	375.80 -> 400.21
6%	damp custom.netmonitor.requestsFinished.DAMP	windows10-64-shippable-qr	e10s fission stylo webrender-sw	2,307.28 -> 2,450.47
5%	damp custom.netmonitor.requestsFinished.DAMP	windows10-64-shippable-qr	e10s fission stylo webrender	2,306.01 -> 2,430.50
5%	damp custom.netmonitor.requestsFinished.DAMP	linux1804-64-shippable-qr	e10s fission stylo webrender	2,844.16 -> 2,984.38
5%	damp custom.netmonitor.requestsFinished.DAMP	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	2,793.09 -> 2,927.53
5%	damp custom.inspector.expandall.balanced	windows10-64-shippable-qr	e10s fission stylo webrender-sw	511.88 -> 535.22
5%	damp custom.netmonitor.reload.DAMP	windows10-64-shippable-qr	e10s fission stylo webrender-sw	346.95 -> 362.70
...	...	...	...	...
4%	damp custom.inspector.expandall.balanced	windows10-64-shippable-qr	e10s fission stylo webrender	520.28 -> 540.87
3%	damp custom.jsdebugger.stepOver.DAMP	windows10-64-shippable-qr	e10s fission stylo webrender	433.99 -> 445.26
2%	damp custom.inspector.deeplynestedrule.refresh	windows10-64-shippable-qr	e10s fission stylo webrender-sw	83.22 -> 85.22
2%	damp custom.jsdebugger.stepOver.DAMP	windows10-64-shippable-qr	e10s fission stylo webrender-sw	431.01 -> 441.19
2%	damp custom.inspector.deeplynestedrule.refresh	windows10-64-shippable-qr	e10s fission stylo webrender	72.46 -> 74.11

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=40836

---

the on console.typing is the most worrysome, source of the test is here: https://searchfox.org/mozilla-central/rev/f961e5f2a22f4d41733545190892296e64c06858/testing/talos/talos/tests/devtools/addon/content/tests/webconsole/typing.js

Comment 5

[Timothy Nikkel (:tnikkel)]

Where is the link to the graph which specific revisions and results?

Comment 6

[Nicolas Chevobbe [:nchevobbe]]

(In reply to Timothy Nikkel (:tnikkel) from comment #5)

Where is the link to the graph which specific revisions and results?

this one for example https://treeherder.mozilla.org/perfherder/graphs?highlightAlerts=1&highlightChangelogData=1&highlightCommonAlerts=0&selected=4769075,1819997504&series=autoland,4769075,1,12&timerange=1209600&zoom=1704281516390,1704292072516,361.76580040613305,414.09605031047914

Comment 7

[Nicolas Chevobbe [:nchevobbe]]

(note that other, more obvious graphs, like this one , points to a pushlog with more entries (https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=655286516798565db497786ca92d9a8f7f7744bf&tochange=9ea90dc23395fc3e2d9d2bd18598cc49434eeff1), so maybe this bug is not at fault here (and I didn't run any pushes to assert that this bug was the culprit)

Comment 8

[Timothy Nikkel (:tnikkel)]

Thanks. The graph is also a quick way to get to a job for which I can trigger a job with a profile.

I think I see what the problem is that would make this bug a plausible culprit. Do you want to file a new bug for the regression or should I?

Comment 9

[Nicolas Chevobbe [:nchevobbe]]

(In reply to Timothy Nikkel (:tnikkel) from comment #8)

Thanks. The graph is also a quick way to get to a job for which I can trigger a job with a profile.

I think I see what the problem is that would make this bug a plausible culprit. Do you want to file a new bug for the regression or should I?

Sure, I filed Bug 1873414