On Jun 5, 2006, at 2:58 PM, Josh Aune wrote:
> On 4/5/06, Jeff Squyres (jsquyres) <jsquyres_at_[hidden]> wrote:
>> This is going to be influenced by how many processes bproc tells Open
>> MPI can be launched on each node.
>> Check out the FAQ for the -bynode and -byslot arguments to mpirun for
>> more details:
> I have tried these arguments several times (up through 1.0.2a4) and I
> always get the same ordering.
I think (and I could be wrong here) that this is because of some code
we have that only works with bproc (that allows you to not give a
hostfile) that (at least on my small machine) doesn't set the number
of cpus to anything greater than 1, so -bynode and -byslot will give
the same answer.
Your best bet is to specify a hostfile or use a batch scheduler.
Since this will limit the number of nodes available and you can set
the available number of slots, this should allow you to schedule more
than one process per processor.
Open MPI developer