RE: ANN: XSLBench - XSLT Processor Benchmark

["Kevin Jones" <kjouk@xxxxxxxxxxx>]

Andy,

I must apologise for this. I actually issued the wrong driver file in the
zip. The real one is, I believe, using a 'compiled' style sheet. I have
mailed you directly the driver I used for the test. I would appreciate any
help people can offer to ensure that drivers give the best for a given
processor.

I like the idea you mentioned earlier of allowing 'optimised' versions of
style sheet to be used. Its clear that in XSLT there is normally more than
one way to skin a cat and this approach may result in a fairer benchmark. I
am going to re-run the tests to update some processors and try some
different JVM options over the next day or so. Perhaps the next stage is a
XSLT performance competition?

Regards,
Kev.

-----Original Message-----
From: owner-xsl-list@xxxxxxxxxxxxxxxx
[mailto:owner-xsl-list@xxxxxxxxxxxxxxxx]On Behalf Of Andrew Kimball
Sent: 04 October 2000 00:24
To: 'xsl-list@xxxxxxxxxxxxxxxx'
Subject: RE: ANN: XSLBench - XSLT Processor Benchmark


Kevin,

Looking at your MSXSL driver code, I notice that you are not using compiled
stylesheets.  Are you using compiled stylesheets with the other processors
(well, except for Sablotron, which you say doesn't support them)?  If so, I
would be happy to send you some C++ code that shows how to use compiled
stylesheets with MSXSL.

~Andy Kimball
MSXSL Dev


-----Original Message-----
From: Kevin Jones [mailto:kjouk@xxxxxxxxxxx]
Sent: Tuesday, October 03, 2000 2:52 PM
To: xsl-list@xxxxxxxxxxxxxxxx
Subject: RE: ANN: XSLBench - XSLT Processor Benchmark



I just used the IBM JDK 1.1.8 without any memory arguments. From my work in
Java this has always proved to be the quickest and most reliable VM I have
found. I will try playing with the memory settings to see how that effects
the results.

The Sablotron processor is indeed a very good performer on the these tests
but I must say the surprise for me was MSXML. Despite their early poor
conformance record the results are good.

It is puzzling to me that Sablotron does not appear to support any form of
compiled style sheets. It's such an obvious way of speeding up the transform
that I can't believe they have not thought about this?

Regards,
Kev.

-----Original Message-----
From: owner-xsl-list@xxxxxxxxxxxxxxxx
[mailto:owner-xsl-list@xxxxxxxxxxxxxxxx]On Behalf Of Sebastian Rahtz
Sent: 03 October 2000 19:11
To: xsl-list@xxxxxxxxxxxxxxxx
Subject: Re: ANN: XSLBench - XSLT Processor Benchmark


Kevin Jones writes:
 >
 > To answer questions I was having over XSLT processor performance I have
 > built a small benchmark of some of the main XSLT processors. You can find
 > the results at,
 >
 > 	http://www.tfi-technology.com/xml/xslbench.html

Thats quite interesting. I found when testing processors that  the
Java-based ones vary *widely* depending on the Java setup used, the
memory allocated, etc. So I would take your results with  a pinch of
salt at present.

The consistent feature to my eye of this test, the Perl-based
reported yesterday, and my own is the excellent performance of
Sablotron. I just hope they can keep up the work as they add the
remaining missing features.

Sebastian


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list