We don't really have this kind of fine-grained processor affinity
control in Open MPI yet.
Is there a reason you want to oversubscribe cores this way? Open MPI
assumes that each process should be as aggressive as possible in
terms of performance -- spinning heavily until progress can be ma
I assumed you meant processes on a host, but I noticed that you
wrote "cores". I'm not sure what the answer is if you really meant cores.
I run different number of processes per "node" using mpirun
-hostfile=hosts ,
where hosts file contains:
host0 slots=5
host1 slots=5
host2 slots=5
host
Thanks for the answer. I have been using LAM MPI. I'm using FDS (Fire
Dynamic Simulator) and in my model I have 7 threads and want to tie them to
two cores which means I don't want the 7 threads to use all four cores on
the cluster that we have. Have you done some thing similar to this?
Thanks,
S
I do this using the hostfile. There might be a more sophisticated way too.
Siamak Riahi wrote:
I have a question about using the open mpi.
I want to tie "N" number of processes to one core and "M" number of
processes to another core. I want to know if open mpi is capable of
doing that.
Tha
I have a question about using the open mpi.
I want to tie "N" number of processes to one core and "M" number of
processes to another core. I want to know if open mpi is capable of doing
that.
Thanks,
Siamak
