Pooling, memory, and what not

["Peter M. Goldstein" <pe...@yahoo.com>]

All,

I wanted to give you a quick update on the status of the
memory/performance enhancements.  After analyzing our last run of data,
Noel suggested that object pooling could substantially improve James
performance.  We have implemented partial pooling (Watchdogs remain
unpooled - I hope to get to that tomorrow) and our initial tests bear
out Noel's intuition.  We have achieved early measurements of about 4300
messages/minute, with heap size remaining < 5 MB.  These results are
only preliminary.  At this point we haven't run very long tests, but
hope to start that shortly.  That will give us a clearer idea of the
performance gains.

I expect to post another updated version of the code tomorrow.
Hopefully this time I'll get it right and it will be complete.  :)  That
should allow you all to see where we are with the code.

--Peter 



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>