more on XSLT processor performance

[Sebastian Rahtz <sebastian.rahtz@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>]

People who like these things may like to re-examine
http://users.ox.ac.uk/~rahtz/xsltest/Report.html, as I have redone all
my tests. I  added Transformiix into the set, and the sieve of
erathosthenes, and also ran my largest data file (4227978 elements).

I have run the tests with MSXML3, but the computer's clock was
broken:-} I cannot run it on the same computer as the other tests, cos
MS ie5setup has left my Windows setup in an unuseable state, and I
gave up after wasting 1/2 hours rebooting the machine 50
times. Anyway, the July MS thing seemed to scoot along nicely, so far
as I can tell, on most of the tests. Someone with more patience than
me can retry...

My conclusions, again
:
 - XT: efficient and fast, but not complete or entirely conformant
 - Saxon: efficient, fast and conformant (a few, a very few, bugs)
 - Oracle: fast, conformant, but does not scale as well as XT and
     Saxon *in my environment*. much greater memory use, and slows down
     in my memory setup with larger files. bug in ora:output
 - Xalan: slows down (or dies) considerably on larger jobs, but delivers the
     goods otherwise.
 - Sablotron: incomplete (work in progress), and gets locked in bigger
     jobs for ever, but looking promising
 - Transformiix: has a good go, but cannot cope with larger files
 - MSXML3 (July): looking good, but not there yet

For my type of work, in my computer setup, Saxon wins pretty
easily. Others will have different criteria to choose from what is
really a very rich field of contenders.

Sebastian Rahtz


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