Benchmark of 6 XML parsers on Solaris

[David Brownell]

"Mark D. Anderson" wrote:
> 
> >1. Java is faster than Perl/Python for _all_ test cases
> 
> there are substantial variations in VM performance across
> vendor and OS; see:
> http://www.javaworld.com/jw-03-1999/jw-03-volanomark.html?022399ibd
> 
> Also, for reasons that I can only attribute to laziness on
> the part of implementors, java VMs seem to take a ridiculous
> amount of time to start up. Beats me what they are doing.

I suspect that nobody's ever placed aggressive tuning targets
on the startup time.  There's operating system overhead for
setting up processes (including setting up shared libraries
and intitializing data structures), and often some JIT costs
(since JVMs don't yet cache JIT results AFAIK).

That's not to disagree that startup costs shouldn't be cut,
only to point out that there are reasons for them that don't
have a thing to do with laziness.  Many apps have startup times
that are substantial, so they run faster when they're used.
 

> Judging from the level of discussion on various other lists I'm
> on, none of java, perl, or python were designed with lightweight,
> accurate, and meaningful profiling of their own run time built in.

Not so.  With Java 2, check out "java -Xrunhprof:help" output to
see the built in profiler.  Works nicely on SPARC out of the box,
but the Win32 version at one point needed a JIT update to get CPU
profiling to work (by default it only does memory profiling).

- Dave

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/ and on CD-ROM/ISBN 981-02-3594-1
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)