RE: [ANN] XSDBench XML Schema Benchmark 1.0.0 released

> 
> The problem with large documents is that it becomes hard to 
> cache them in memory. 

I don't think you mentioned that you cached the document in memory. Is that
as unparsed text, as a DOM, or what?
> 
> > (Though I can't complain, because this one test did find a 
> bug in my 
> > product which none of the 3000 test cases in the W3C test suite had 
> > shown up!)
> 
> And did you get any performance numbers? 
> 

About 2.5Mb/s, without taking any special pains (running in the IDE with
debugging enabled, other applications on the machine, etc). That's parsing
and validating directly from the file on disk, excluding schema loading
time, and running the validation 1000 times to get the Java VM fully warmed
up (the first 500 runs took 55% of the time). I've no idea how the machine
spec compares with the ones you were using: it's a Centrino laptop.

I've never done any intensive performance tuning or comparative measurement
on the schema processor, so I'm quite happy with that result.

Michael Kay
http://www.saxonica.com/

Received on Wednesday, 18 October 2006 13:08:34 UTC