Hi, Adam. Adam Jenkins <adamjenkinstmpredir...@yahoo.com.au> wrote on 01/25/2010 12:33:23 AM: > Over the coming months, I'm going to be messing around with the > Xalan code a bit, trying out a few things...and I want to start with > a few killer benchmarks...so I can see if my changes have any > performance impacts. > > If anyone has some really killer XSLTs...you know the ones....that > create heaps of objects, or take 10 minutes to run...that kinds > stuff...that aren't proprietary, I'd really appreciate you emailing > me a copy to use as a benchmark. > > That way I can set up some good profiling before work.
I just wanted to mention something you should keep in mind when you're doing this: the stylesheets that take many minutes to run processing a single input are interesting and important, but at least as important are stylesheets that will be running on a server, processing hundreds of requests per second. If you can find a copy of the XSLTMark benchmark tests anywhere, you might want to try them out - they aren't necessarily the best, and don't necessarily reflect typical usage on a server, but there haven't been many other XSLT benchmark tests published. Something else to consider - most of the performance work that's been carried out in the Apache Xalan-Java project, other than early on in the project's history, has been focused on the XSLTC processor rather than on the Xalan-J interpreter. Thanks, Henry ------------------------------------------------------------------ Henry Zongaro XML Transformation & Query Development IBM Canada Lab T/L 313-6044; Phone +1 905 413-6044 mailto:zong...@ca.ibm.com
