Re: [xslt] is xsltproc really this fast?

[Tom Moog <tmoog polhode com>]

The non-validating piccolo sax parser (written in java) is
very nearly as fast as expat (once the jit has warmed up).
expat claims to be the fastest C/C++ sax parser available
(although it does have high initialization cost).


On Wed, 18 Sep 2002, Daniel Veillard wrote:

> On Tue, Sep 17, 2002 at 06:13:38PM -0400, Paul Tremblay wrote:
> > I hope this comment isn't too off topic, but is xsltproc really 25 times
> > faster than xalan?
>
>   Honnestly I don't think so... Faster in average, yes but not by such
> a margin. I don't try to do benchmark, but I try to keep libxml2/libxslt
> relatively fast and lightweight.
>
> > I have a small stylesheet, which I processed firts with xsltproc and
> > then with xalan. xsltproc took .75 seconds to process. xalan took 26
> > seconds! Other tests have yielded similar results.
> >
> > I should point out that I have a pentium one box, so maybe java just
> > can't run efficiently on a machine this slow.
>
> Well running a single processing possibly small document really defavor a
> Java based processor since it has to set-up the Java environment,
> and a single execution doesn't allow to take the benefits of JIT
> or HotSpot optimizations. That said, I think Java is a pig ressource
> wise, and in cases like yours C - while being harder to program - really
> really make a difference, and it shows.
>
> Daniel
>
> --
> Daniel Veillard      | Red Hat Network https://rhn.redhat.com/
> veillard@redhat.com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
> http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/
> _______________________________________________
> xslt mailing list, project page http://xmlsoft.org/XSLT/
> xslt@gnome.org
> http://mail.gnome.org/mailman/listinfo/xslt
>