[Fwd: Success with Nutch & GCJ]

[Andrzej Bialecki]

Tom Tromey wrote:
> One thing worth trying is BC compiling your app.  If you then
> register the results with the class cache database, you can still run
> with gij but it will pick up your compiled code instead.
> Instructions here:
>
> http://gcc.gnu.org/wiki/How%20to%20BC%20compile%20with%20GCJ
>
> This is how we compile all the stuff we put into Fedora Core...
>   

I did as you suggested. Now the performance in terms of pages/s 
processed is the same as with Sun JVM. :-)

Memory consumption, though, is still about as bad as before, or rather 
the pattern for memory allocation is the same - gij allocates all heap 
up to the maximum, and keeps as much as it can resident in physical RAM 
(which hurts the rest of the system). I'm not sure what are the reasons 
that the OS decides to keep all these pages resident in this case...?

-- 
Best regards,
Andrzej Bialecki     <><
 ___. ___ ___ ___ _ _   __________________________________
[__ || __|__/|__||\/|  Information Retrieval, Semantic Web
___|||__||  \|  ||  |  Embedded Unix, System Integration
http://www.sigram.com  Contact: info at sigram dot com