JS speed - Konq vs. Mozilla

[Germain Garand]

Le Mercredi 21 Août 2002 16:01, Vadim Plessky a écrit :
> I don't have latest Konq from CVS, and testing with KDE 3.0.2 doesn't sound
> as a good idea...
> Anyone want to check URL below and statements about Mozilla vs. Konq speed?
>
> :-)
>

This test just benchmarks "using local variables in tight inner loops"...
I don't see how this could reflect the overall speed of JS engines, but 
anyway... results for this test are positively awful for a +/- 12 days old 
CVS Konq (enable-debug = yes [not full])  vs. a rather old Moz (0.9.8 - 
stripped).

Local/Global variables test :
Konq: 3.8 sec. / 4.6 sec
Moz:  0.1 sec / 0.41 sec

Yuck. Hope it's just the debugging overhead.




> ----------  Forwarded Message  ----------
>
> Subject: Re: Discussion about futur of DHTML in Mozilla
> Date: Wed, 21 Aug 2002 02:50:42 +0100
> From: Scott <scott at toktik.demon.co.uk.org.net.us.eu.netscape.com>
> To: mozilla-layout at mozilla.org
>
> On 20 Aug 2002 15:08:20 GMT, lucy-ples at mtu-net.ru (Vadim Plessky) wrote:
> >On Tuesday 20 August 2002 4:14 am, Scott wrote:
> >|  Actually, I've run benchmarks on the browsers javascript engines, and
> >| (at the time of testing, about 4 months ago) Mozilla's javascript engine
> >| was about the same speed as IE (IE was slightly faster) and MUCH faster
> >| than Konqueror's (I like Konq, BTW, but it's not a good DHTML browser).
> >| At that
> >
> >Any testcase/URL?
> >My testing shows that Konqueror is about 4 times faster than Mozilla, but
> >still slower than MS IE :-(
> >// it has better standards-compliance, though :-)
>
> Yep, http://www.javascript-games.org/articles/local_global_bench.html .
> It's a very simple benchmark, it was a demonstration I used during a
> discussion on my forum about the advantages of using local over global
> variables in tight inner loops with the various browsers. As you can see,
> Moz was 3 to 16 times faster than Konq, and NS4 3 to 10 times as fast under
> linux. I haven't tested with the latest version of Moz...
>
> -------------------------------------------------------