Re: [xml] xmlbench Benchark results... (fwd)
Luca Padovani <[email protected]> Wed, 19 Mar 2003 10:25:16 +0100 (CET)
| Newsgroups | gmane.comp.gnome.gdome |
|---|---|
| Message-ID | <[email protected]> |
In case some of you are not subscribed to the libxml2 mailing list, there is a very intersting bechmark page where both libxml and libxml + Gdome are compared against other DOM implementations. It is noticeable that libxml + Gdome is much slower than libxml2 for DOM manipulation. It would be interesting to understand how much of the extra time required is due to the memory management overhead and how much is due to the DOM event mechanism (which, and Paolo is aware of that, can be improved quite a lot). -- luca ---------- Forwarded message ---------- Date: Wed, 19 Mar 2003 03:13:09 -0500 From: Daniel Veillard <[email protected]> To: Sean Chittenden <[email protected]> Cc: [email protected] Subject: Re: [xml] xmlbench Benchark results... On Tue, Mar 18, 2003 at 04:23:45PM -0800, Sean Chittenden wrote: > I apologize if others have seen this, but: > > http://xmlbench.sourceforge.net/results/benchmark/index.html > > libxml performs quite well in the stack up. Congrats to the libxml > developers. I wonder how Oracle can be orders of magnitude faster > though in several of the benchmarks. They must be doing something Which ones ? It's seems they are a bit faster on DOM processing, but I don't see orders of magnitude. > fundamentally different under the hood when it comes to generating > messages, DOM manipulation, or validation. Then again, things are > really quick at that level and libxslt is the fastest out there it > looks like. In all fairness it doesn't include any Java nor MS stack. > How does libxml stack up to MS's XML library? -sc I can't tell, but from what I gathered: - libxml might be faster for parsing - MS generate more compatc documents - MS XSLT engine seems faster overall - but libxslt is slightly faster for DocBook XSLt processing but I have no precise idea. I think I could speed up parsing speed and a number of small operations by using alloca for some tasks and possibly interning some strings at the parser level. I will probably go over another profiling work if I have some time after Relax NG and Schemas work. Daniel -- Daniel Veillard | Red Hat Network https://rhn.redhat.com/ [email protected] | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/ _______________________________________________ xml mailing list, project page http://xmlsoft.org/ [email protected] http://mail.gnome.org/mailman/listinfo/xml