[Paraview] opinion on new machine for visualization

[Ricardo Reis]

On Mon, 7 Apr 2008, Moreland, Kenneth wrote:

> We use OpenMPI on our cluster, and a command that seems to work is the following:
>
>  mpirun -bynode -np 4 /path/to/paraview -display localhost:0 :
>     -np 4 /path/to/paraview -display localhost:1
>
> The -bynode flag tells mpirun to assign processes in a round robin fashion.  The next part says that the first 4 processes should be launched as paraview with the display set to the first graphics card.  The colon (:) sets of the next partition of processors, another 4 instances of paraview with their display set to the second graphics card.
>
> I don't know if other implementations of MPI such as mpich support the -bynode flag and the use of the : to specify different command lines for different blocks of processes, but I am sure that there is some convention for launching heterogeneous jobs.

Thanks for the info, it should prove usefull. I'm using openMPI to.

  greets

  Ricardo Reis

  'Non Serviam'

  PhD student @ Lasef
  Computational Fluid Dynamics, High Performance Computing, Turbulence
  http://www.lasef.ist.utl.pt

  &

  Cultural Instigator @ Rádio Zero
  http://www.radiozero.pt

  http://www.flickr.com/photos/rreis/