Joseph Shraibman wrote:
> I've had similar experiences, though on a 2.2.12 kernel.
> I get the dumps when using either javacomp or sunwjit (Blackdown
> 1.2.2RC4) - maybe after minutes, maybe after hours.
>
> I have been able to run long-term high-volume tests (48 hours,
> constant hammering, max-throughput testing) without problems when
> running with no jit at all - that's the only configuration in which
> I've had consistent repeatable results.
>
> Haven't done any 1.3 SMP testing yet.

What about tya?

We're considering buying a dual processor machine to run our java
programs but if we can't get something running with a jdk on an smp
kernel there probably won't be much point.

Haven't tried tya on SMP, so can't report on this. (Anyone else?)
Only reasonable results I got were with no jit (java -Djava.compiler=none etc.). For our testing (JMS environment) we got better throughput on 2 CPU SMP server with no jit than on non-SMP w/jit.
I did see on the JDK 1.3 Linux beta readme that "Sun does support or recommend running the Java 2 SDK on SMP kernels."

Can anyone from Sun / Blackdown comment on future SMP support, and what the underlying issues are?

--
David Marshall       mailto:[EMAIL PROTECTED]
VM Systems, Inc.           http://www.vmguys.com
 


Reply via email to