821188 - Firefox fails SVG Performance test at 0 frames per second

Status: RESOLVED FIXED

(Core :: SVG, defect)

Description

[Eric Seidel]

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_2) AppleWebKit/537.21 (KHTML, like Gecko) Chrome/25.0.1349.2 Safari/537.21

Steps to reproduce:

I'm using FF 17.0.1 (Release channel)

http://philbit.com/SvgImagePerformance/backgroundImage.html
FF renders that at 0 fps.  WebKit does about 4 fps, but pdr@ is about to fix us to do 60fps.

WARNING: That link will hang FF so hard you can't interact with it.

http://philbit.com/SvgImagePerformance/ has some other perf tests from pdr.

Comment 1

[Eric Seidel]

I misspoke, WebKit does 53fps on the backgroundImage.html test, but does poorly on others in that directory.

Comment 2

[(mostly gone) XtC4UaLL [:xtc4uall]]

Attachment: animatedtiger.svg

Comment 3

[(mostly gone) XtC4UaLL [:xtc4uall]]

Attachment: stats.min.js

Comment 4

[(mostly gone) XtC4UaLL [:xtc4uall]]

Attachment: Reporter's Testcase

Comment 5

[(mostly gone) XtC4UaLL [:xtc4uall]]

SPS Profile against Mozilla/5.0 (Windows NT 6.1; WOW64; rv:22.0) Gecko/20130308 Firefox/22.0 ID:20130308031005 CSet: cb432984d5ce:

http://people.mozilla.com/~bgirard/cleopatra/#report=34b661e1775b74e2ce1a4a41fbe7fc7c8fae09bc

Comment 6

[(mostly gone) XtC4UaLL [:xtc4uall]]

FWIW, I see Attachment 692358 [details] somewhat WFM against Aurora (24)/Nightly (25).

Comment 7

[Loic]

Working range:
bad=2013-06-14
good=2013-06-15
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=b197bed90a98&tochange=3d16d59c9317

Any idea about the bug which fixed that? Bug 678037?

Comment 8

[Alice0775 White]

Working range:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=59181cb244fd&tochange=57e0a788b159

Fixed by Bug 600207

Comment 9

[Guilherme Lima]

Windows 7 - Nightly 27.0a1 (2013-10-23) I'm getting 20fps only.
On Firefox 25 Beta 11 I'm getting 60fps.

Comment 10

[(mostly gone) XtC4UaLL [:xtc4uall]]

(In reply to Guilherme Lima from comment #9)
> Windows 7 - Nightly 27.0a1 (2013-10-23) I'm getting 20fps only.
> On Firefox 25 Beta 11 I'm getting 60fps.

Yep, I can confirm that Attachment 692358 [details] regressed on Trunk/Nightly (cset 5d5591687e67); still WFM against Aurora.

Comment 11

[Alice0775 White]

STR
Open http://philbit.com/SvgImagePerformance/backgroundImage.html

Regression window(m-i)
Good 53fps:
http://hg.mozilla.org/integration/mozilla-inbound/rev/9c8ab7e9ae41
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:27.0) Gecko/20100101 Firefox/27.0 ID:20131017025216
Bad 15fps:
http://hg.mozilla.org/integration/mozilla-inbound/rev/45d9e6cd3473
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:27.0) Gecko/20100101 Firefox/27.0 ID:20131017030414
Pushlog:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=9c8ab7e9ae41&tochange=45d9e6cd3473

regressed by:
a417424f9213	Robert O'Callahan — Bug 923193. Make transform-origin on SVG elements use the SVG bbox as the reference rectangle. r=heycam

Comment 12

[bhavana bajaj [:bajaj]]

Needinfo on Roc here to help understand if this would have any user facing impact ?

Comment 13

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Fixed by 929021?

Comment 14

[Alice0775 White]

I can still see the performance regression in 
http://hg.mozilla.org/mozilla-central/rev/2f2a45f04e7c
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:27.0) Gecko/20100101 Firefox/27.0 ID:20131025100746

Small improved by Bug 929021.

Before landing Bug 923193, 53fps
After landing Bug 923193, 15fps
After landing Bug 929021, 22fps

Comment 15

[bhavana bajaj [:bajaj]]

(In reply to Robert O'Callahan (:roc) (Mozilla Corporation) from comment #13)
> Fixed by 929021?

:roc, is comment #14 satisfactory with the expectation ?

Comment 16

[bhavana bajaj [:bajaj]]

NI on :roc to see if any new landings have fixed this and also help shed light on if this is an important perf regression to track from a blocker perpective.

Comment 17

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Should have been fixed via this backout:
https://hg.mozilla.org/integration/mozilla-inbound/rev/3fb151446ec5

If there is any other regression it's probably bug 922942.

Comment 18

[Jonathan Watt [:jwatt]]

Fixed by bug 600207.