Parser benchmark result

[Michael Kay]

At the risk of making myself unpopular by looking gift
horses in their mouths, I report the results of running a
simple SAX 1.0 application with five different parsers.

They were all run using the default SAXON Renderer
application (which
simply reconstitutes the XML input file supplied) against an
XML file containing 291089 bytes, or 11500 elements, each on
the same machine
(a reasonably powerful Windows NT server, with the SUN Java
VM), which was otherwise idle (as far as one can tell). I
measured elapsed time by
calling java.util.Date#getTime() at the start and end of the
run.

Each was run twice, I report both results.
Elapsed time in milliseconds:

AElfred: 8203, 8215
Lark: 10422, 10453
MSXML: 13250, 12250
xml4j: 18125, 18313
xp: 8156, 7907

These were obtained after some performance tuning on SAXON:
the issued version takes about twice as long.

Of course, there are many attributes to a piece of software
other than its raw speed, and the results cannot necessarily
be extrapolated to a different application or a different
data file.

The really good news is that (with the one caveat noted in
an earlier message) the application worked unchanged on all
five parsers.

Mike Kay, ICL



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)