http://gcc.gnu.org/java/done.html

Sun Jul 27 18:25:00 GMT 2003

On Sunday 27 July 2003 08:24, Erik Poupaert wrote:
> Still, the "100 times" is alarming. There is no reason to believe that the
> reporter, who ran the native and the bytecode versions did not come to this
> conclusion in good faith.

If you want I'll get the author to provide you with more data. Basic issue was 
that gcj server was using MUCH less memory, but it took a long while to 
process the irate database with it. On the other hand, the sun jvm uses more 
memory, but the tasks run 100x faster. ie a few seconds instead of minutes.
Also, I suspect that this may be due to use using a relatively ancient version 
of gcj(3.2.1 or whatever is in debian unstable).
We would be very happy to switch our server back to gcj if the perfomance 
problems are worked out. We are also in the process of redesigning it to do 
less processing, so we'll probably switch back anyways.
What sort of information do you need to nail this bug?
Taras