Re: xsl speed tests
| From: | Gabor Hojtsy | Date: | Fri, 08 Feb 2002 14:53:20 +0000 |
| Subject: | Re: xsl speed tests | ||
| References: | 1 2 3 4 | Groups: | php.doc |
| Request: | Send a blank email to phpdoc+get-969342100@lists.php.net to get a copy of this message | ||
> > Instant Saxon uses MS JVM, XSLTProc is exe... My testmachine was
> > a 850Mhz AMD Duron.
>
> Saxon with Sun's or IBM's JVM is much more faster than with MS one.
Is there any difference on Windows between IBM and SUN JVMs?
I have the latest SUN JSDK, but if IBMs quicker, I'll download
and install it.
Also what you think about style sheet complecity? We actually only
use a small subset of the DocBook DTD. It is possible to increase
the speed significantly, with dropping out unsused parts from the
style sheet. More precisely with creating a custom modules and a
driver with only the templates we use?
We use 97 different tags (exact value computed by our tag usage
stat program). I don't know how many tags are included in the
DocBook DTD, but there must be much-much more. This means many
unsuccessful matches for templates while processing...
The DocBook XSL docs also mentions, that creating a custom driver
can increase speed:
| Another approach to customizing the stylesheets is to write your
| own driver file. Instead of using <xsl:import href="docbook.xsl"/>,
| you copy that file to a new name and rewrite any of the
| <xsl:include/> instructions to assemble a custom collection of
| stylesheet modules. One reason to do this is to speed up processing
| by reducing the size of the stylesheet. If you are using a customized
| DocBook DTD that omits many elements you never use, you might be
| able to omit those modules of the stylesheet.
How much speed increase we can gain with this method? Is it significant?
Goba