jboss 1.0 RMI call were at 3ms
jboss 1.0 IN-VM calls were <0.1ms
jboss 2.0 uses the SAME invocation layer.
So there is something really silly going on, it will be the first we look at
as soon as we address the TestBeans.jar thing (almost there). Rickard was
tracking some class loader silliness but it seems the kid just died by the
road while climbing his own mountain ;-).... we will get the corpse to the
top, don't you worry!
regards
marc
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]]On Behalf Of Andy Lewis
> Sent: Thursday, July 27, 2000 2:33 PM
> To: jBoss Developer
> Subject: [jBoss-Dev] Performance Question....
>
>
>
> We've been doing some internal testing with EJBs using jBoss and have seen
> function calls between entity beans on the same server taking a
> long time -
> like about a second. They were slow enough that output from them could be
> "visually" timed. I know performance hasn't been really addressed yet, but
> this seemed a bit much.
>
> Are we just doing something really stupid? (my first assumption would be
> probably yes) Are all calls between beans going through RMI? (no
> in-VM optimization?)
>
> Any comments here are appreciated....
>
> Thanks....
>
>
>
