You are right that technically increasing application
startup time by one second might be acceptable, but it
bothers me when processing a 200 line file takes one
second on a 2 GHz machine. That simply isn't right.
Anyway, I solved the problem by using Sun's JVM
instead of IBM's. The time came down to 160 ms, once
all classes were already loaded (compared to 500 ms
with IBM). Very interesting. In all benchmarks IBM JVM
has always been significantly faster than Sun - this
is the first time that I see such a significant
difference in the other direction.
Technically, 160 ms is also a very long time to
process a 200 line file - that is almost a millisecond
per line - but I will have to make some more precise
measurements to determine what part of the time to
attribute to the parser, and what to Castor.
Do you expect that using compile-time descriptors
could speed things up significantly ?
best regards,
Tzvetan
--- Keith Visco wrote:
>
>
> I can't tell you why it's slow on your system, or
> configuration, but if
> loading the mapping file is the only thing that's
> slow...then I don't
> really see a big problem.
>
> The mapping file should only be loaded once and
> should be saved in
> memory for successive calls to the
> Unmarshaller/Marshaller. So it should
> be a one-time penalaty that should be done outside
> any loops.
>
> --Keith
__________________________________
Do you Yahoo!?
Meet the all-new My Yahoo! - Try it today!
http://my.yahoo.com
-----------------------------------------------------------
If you wish to unsubscribe from this mailing, send mail to
[EMAIL PROTECTED] with a subject of:
unsubscribe castor-user
