[
Lists Home |
Date Index |
Thread Index
]
Hey Mukul,
As I have learned from creating similar tests and posting the results
there are too many outside factors that go into these one off
measurements to put any sort of weight on the result. If interested,
after I discovered the details of how to properly pressure cook a
processor I created a quick little test suite that takes any given
piece of XML, creates a range of file sizes (dynamically specified in
the config file) and combines this with a structured series of
commandline sequences to properly put a processor to the test in sets
of 1, 3, and 9 iterations through the same XML file, outputting the
results into seperate files that can then be used to develop a much
more qualified understanding of what your dealing with...
I actually have several iterations of this original test-suite (it was
a quick hack... its "grown-up" quite a bit since and I plan to release
the new and improved version around the same time Saxon.NET goes from
RC1 > RC2 > 1.0 Final in the next week or so) but for now if you want
to play around with it you can access the file and a quick explanation
of how it worls here >
http://www.xsltblog.com/archives/2004/12/next_portion_of.html
Have fun :D
<M:D/>
On Mon, 28 Mar 2005 01:25:35 -0800 (PST), Mukul Gandhi
<mukul_gandhi@yahoo.com> wrote:
> I did a small performance comparison between Xalan-J
> 2.6.0 and Saxon 8.3 . I ran a simple identity
> transform on a small XML file.
>
> The XSLT is -
>
> <?xml version="1.0"?>
> <xsl:stylesheet
> xmlns:xsl="http://www.w3.org/1999/XSL/Transform";
> version="1.0">
>
> <xsl:output method="xml" indent="yes" />
>
> <xsl:template match="@* | node()">
> <xsl:copy>
> <xsl:apply-templates select="@* | node()" />
> </xsl:copy>
> </xsl:template>
>
> </xsl:stylesheet>
>
> Average execution time recorded were:
> -------------------------------------
> Xalan-J 2.6.0 :
> with -DIAG option 220 ms
>
> Saxon 8.3 :
> with -TP option 47 ms(Saxon's performance analysis
> tool)
> with -t option 63 ms
>
> It seems Saxon is faster than Xalan. It seems upto
> about 4-5 times.. I am assuming both Xalan and Saxon
> are measuring the same things!
>
> If "GNU JAXP XSLT processor" is 2.8 times faster than
> Xalan , then Saxon is approximately 2 times faster
> than GNU XSLT processor ..
>
> I read in an article by Michael Kay that Saxon uses an
> efficient tree structure suited for XPath data model..
> DOM and XPath tree models don't map 100%. But I am not
> sure what Xalan uses.. I guess it is DOM or a slight
> variant of it..
>
> Regards,
> Mukul
>
> --- Chris Burdess <dog@bluezoo.org> wrote:
> > Michael Kay wrote:
> > >> What kind of parser is best to use for XSLT
> > transformations ?
> > >> SAX or DOM
> > >
> > > XSLT processors will in general build a tree
> > representation of the
> > > source
> > > document in memory. And in general, many of them
> > will build a tree
> > > representation that is much more efficient than
> > using a
> > > general-purpose DOM.
> > > So there's no point building an inefficient DOM
> > tree rather than
> > > letting the
> > > XSLT processor build its own. But this advice may
> > depend on the XSLT
> > > processor you are using.
> >
> > For what it's worth, the GNU JAXP XSLT processor
> > uses DOM internally
> > for both source and result trees, and is about 2.8
> > times as fast as
> > Xalan on a wide range of transformations (the OASIS
> > XSLT/XPath
> > conformance suite). I don't have figures for memory
> > usage or
> > comparisons with Saxon though.
> > --
> > Chris Burdess
> >
> >
> >
> -----------------------------------------------------------------
> > The xml-dev list is sponsored by XML.org
> > <http://www.xml.org>, an
> > initiative of OASIS <http://www.oasis-open.org>
> >
> > The list archives are at
> > http://lists.xml.org/archives/xml-dev/
> >
> > To subscribe or unsubscribe from this list use the
> > subscription
> > manager:
> > <http://www.oasis-open.org/mlmanage/index.php>
> >
> >
>
> __________________________________
> Do you Yahoo!?
> Yahoo! Mail - now with 250MB free storage. Learn more.
> http://info.mail.yahoo.com/mail_250
>
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
>
> The list archives are at http://lists.xml.org/archives/xml-dev/
>
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://www.oasis-open.org/mlmanage/index.php>
>
>
--
<M:D/>
:: M. David Peterson ::
XML & XML Transformations, C#, .NET, and Functional Languages Specialist
|
