1918441 - 3.64 - 2.01% pdfpaint issue9367.pdf / pdfpaint bug1795263.pdf + 17 more (Windows) regression on Mon August 26 2024

Status: REOPENED

(Core :: Graphics: Canvas2D, defect)

Description

[Treeherder Bug Filer]

Perfherder has detected a talos performance regression from push 1b51a0386c493d4e12543c9ae6c6ee198a709949. 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)
4%	pdfpaint issue9367.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	173.39 -> 179.70
3%	pdfpaint issue3188.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	161.86 -> 167.18
3%	pdfpaint issue9285.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	173.78 -> 179.47
3%	pdfpaint issue9915_reduced.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	133.24 -> 137.59
3%	pdfpaint issue9084.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	131.31 -> 135.45
3%	pdfpaint issue8097_reduced.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	138.62 -> 142.68
3%	pdfpaint issue8117.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	131.28 -> 134.90
3%	pdfpaint jbig2_huffman_1.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	126.14 -> 129.54
3%	pdfpaint mmtype1.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	129.42 -> 132.90
3%	pdfpaint issue7598.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	133.46 -> 137.04
...	...	...	...	...
2%	pdfpaint issue3061.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	131.83 -> 134.53
2%	pdfpaint javauninstall-7r.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	198.20 -> 202.24
2%	pdfpaint issue8019.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	158.49 -> 161.70
2%	pdfpaint issue8613.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	189.80 -> 193.62
2%	pdfpaint bug1795263.pdf	windows11-64-shippable-qr	e10s fission stylo webrender	174.79 -> 178.30

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.

You can run these tests on try with ./mach try perf --alert 1951

For more information on performance sheriffing please see our FAQ.

Comment 1

[Emilio Cobos Álvarez (:emilio)]

Hmmm... Fairly unlikely? My patch is a front-end patch, how can it affect pdfpaint? Calixte, any idea? I'm not familiar with this test...

Comment 2

[Calixte Denizet (:calixte)]

We're very likely somewhere around the margin of error.
And looking at the patch, it's for sure not related to those small variations.

Comment 3

[Mark Banner (:standard8)]

I think this is miss-classified. Looking at the zoomed-in performance graphs, I think it is either:

https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=5a8d6ce1655d114207f0129b163b98b49de3b656&tochange=74b2859d958214928b202808ccef33e19fadeb5b

or:

https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=74b2859d958214928b202808ccef33e19fadeb5b&tochange=d794404ec2a245802e16f814fb29c5882c7b088a

The change in comment 1 seems to be after the main shift has been made.

Comment 4

[Mark Banner (:standard8)]

Reopening & moving across to untriaged, since I think there was potentially a regression here and it was miss-classified - it at least should get a second look wrt the graphs.

Comment 5

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

The Bugbug bot thinks this bug should belong to the 'Core::Graphics: Canvas2D' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Comment 6

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

This bug has been marked as a regression. Setting status flag for Nightly to affected.

Comment 7

[Andrej (:aglavic)]

If you all think this is misclassified, I will close the alert, we're confident that the alert is the right culprit as all the alerts are windows11 and the regressing bug in question(1914798) is windows11 specific as all the alerts are

Comment 8

[Andrej (:aglavic)]

I apologize I did not see that :standard8 had re-openned
I'll take another look as well and double check

Comment 9

[Ryan VanderMeulen [:RyanVM]]

Emilio, can you please take a look at the ranges in comment 3 and re-evaluate?

Comment 10

[Calixte Denizet (:calixte)]

If I can help with something please tell me.
Could the change slightly impact the rendering of the UI ?