Jeff,
Thanks, I tried that after I sent out my question and it didn't work
(it couldn't make a connection or something like that). But now it does
(I lamcleaned and lamhalted since). Actually, both tcp and usysv now
work. Now, which is better for dual processor machines (linux with SMP
kernel and DP Macs)? I'm not doing interslave communication, only
communication is from the slaves to master node.
Thanks
Marc
On Jun 23, 2004, at 7:36 AM, Jeff Squyres wrote:
> On Mon, 21 Jun 2004, Marc Colosimo wrote:
>
>> I'm running code between Macs and Linux (Java bytecode) and I got
>> this message. rank 0 is the Mac and rank 7 is the linux (RH9) box.
>> How do I fix this?
>
> Normally, LAM tries to select the "best" RPI from all those available.
>
> Probably what is happening here is that usysv is not available on rank
> 0, and so it is falling back to tcp. In this case, you can just force
> all ranks to use tcp with:
>
> mpirun -ssi rpi tcp ...rest of your mpirun args...
>
>> ----------------------------------------------------------------------
>> -------
>> It seems that [at least] one of the processes that was started with
>> mpirun chose a different RPI than its peers. For example, at least
>> the following two processes mismatched in their RPI selections:
>>
>> MPI_COMM_WORLD rank 0: tcp (v7.0.0)
>> MPI_COMM_WORLD rank 7: usysv (v7.0.0)
>>
>> All MPI processes must choose the same RPI module and version when
>> they start. Check your SSI settings and/or the local environment
>> variables on each node.
>> ----------------------------------------------------------------------
>> -------
>>
>>
>> Thanks
>> Marc
>>
>>
>> _______________________________________________
>> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>>
>
> --
> {+} Jeff Squyres
> {+} jsquyres_at_[hidden]
> {+} http://www.lam-mpi.org/
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
|
