Closed Bug 777039 Opened 13 years ago Closed 2 years ago

IE9 performs significantly better at WebVizBench

Categories

(Core :: Graphics, defect)

Product:
Core
Component:
Graphics
Platform:
x86_64
Windows 7
Type:
defect

Tracking

()

Status:
RESOLVED INVALID

People

(Reporter: bas.schouten, Unassigned)

Details

IE9 does a lot better at webvizbench. I don't remember this used to be the case, I wonder what regressed this.
Alice, do you find performance regressions? :-)
(In reply to Robert O'Callahan (:roc) (Mozilla Corporation) from comment #1) > Alice, do you find performance regressions? :-) Is url the following? http://webvizbench.com/Default.htm?skipdetect I cannot any regression in Firefox10-17... And Firefox is much better than IE9... *Note: Firefox does not play background video. your score 3180 / 7.18 fps 1018 x 775 your browser Internet Explorer 9 your score 5580 / 21.2 fps 1018 x 775 your browser Firefox 17 http://hg.mozilla.org/mozilla-central/rev/a26e751bfb54 Mozilla/5.0 (Windows NT 6.1; WOW64; rv:17.0) Gecko/17.0 Firefox/17.0 ID:20120724071408 your score 5460 / 18.67 fps 1018 x 775 your browser Firefox 16 your score 5400 / 20.41 fps 1018 x 775 your browser Firefox 15 your score 5270 / 19.52 fps 1018 x 775 your browser Firefox 14 your score 5300 / 19.44 fps 1018 x 777 your browser Firefox 13 your score 5060 / 18.21 fps 1018 x 777 your browser Firefox 12 your score 5020 / 18.4 fps 1018 x 774 your browser Firefox 11 your score 5080 / 18.6 fps 1018 x 774 your browser Firefox 10
Interesting, I wonder if this is related to hardware differences, I'll test myself on some more machines.
So on my laptop I get: 4130/13.15 1680x947 Firefox 14 5070/23.16 fps 1680x965 IE9 On my desktop I get: 4830 / 17.15 fps 2560x1268 Firefox 14 4840 / 20.61 fps 2560x1304 IE9 So in both cases IE9 did better, however on my desktop with a very powerful GPU compared to CPU the difference isn't very big. On my laptop with a poor GPU compared to CPU (running on the integrated GPU), IE9 does a lot better. I guess that means there might not be a regression, we might've just always done poorly compared to IE on this hardware.
I am not sure but the benchmark may be sensitive to window size? And Strange thing here... IE9 is better than Firefox in *Browsing mode* of http://webvizbench.com/Default.htm?skipdetect IE9: 54-56fps @ 1018 x 775 Firefox17:34-36 fps @ 1018 x 775 Firefox16:34-36 fps @ 1018 x 775 Firefox15:32-34 fps @ 1018 x 775 Firefox14:34-36 fps @ 1018 x 775
(In reply to Alice0775 White from comment #5) > I am not sure but the benchmark may be sensitive to window size? > > And Strange thing here... > IE9 is better than Firefox in *Browsing mode* of > http://webvizbench.com/Default.htm?skipdetect > > IE9: 54-56fps @ 1018 x 775 > Firefox17:34-36 fps @ 1018 x 775 > Firefox16:34-36 fps @ 1018 x 775 > Firefox15:32-34 fps @ 1018 x 775 > Firefox14:34-36 fps @ 1018 x 775 Forgive my ignorance, but what exactly does 'browsing mode' mean in this context? :)
Yes it is.
Ah, right, now I understand what you mean.
Funny story, this benchmark thinks that Edge is Chrome 42. That said, I get the following results on my W530 w/ Win10: Fx44 trunk: 7020 / 42.27 fps @ 1920 x 1094 Chrome 45: 6000 / 34.72 fps @ 1920 x 1100 Edge: 7580 / 50.15 fps @ 1920 x 1084 IE11: 7300 / 47.65 fps @ 1920 x 1106 Looks like there's still some room for improvement, but looking for a regression range seems pretty futile given the other comments in this bug.
Keywords: regressionwindow-wanted
Severity: normal → S3

Unable to find a working version of this benchmark.

Status: NEW → RESOLVED
Closed: 2 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.