Each hardware vendor may have different way to execute the MPP program.
Here is a summary for all the platforms.

	# is number of processor used for the job

Fujitsu VPP5000:

	It needs to setup a special queue to run the parallel program.
        User needs to submit the MPP job to that queue.  The number of
        processor used is control by the "-lPv #" option in the nqs
	system. 

        mpp960_executable
Fujitsu PrimePower:

        mpiexec -n # mpp970_executable

HP MPI:
        Please download software from
        http://www.software.hp.com

HP Compaq Alpha:

	dmpirun -np # mpp970_executable

HP PA8000 and Itanium2:
        
        To run the job within a single node (machine), do:

	mpirun -np # mpp970_executable <LS-DYNA command options>

        To avoid performance penalty due to slow I/O, it is preferable that
        the above command be run on a local file system.        

        To run the job across a cluster, first determine a working directory
        on each node with the same name to be used as the MPP DYNA (or MPI) 
        working directory. To avoid performance penalty due to slow I/O,
        it is preferable that this working directory be local to each node.
        Let the  name of this working directory be represented as <workdir>.
        Then invoke the following command
	
	mpirun -f appfile

        where appfile contains

	-h hostname_1st -np #1 -e MPI_WORKDIR=<workdir> mpp970_executable \
		<LS-DYNA command options>
	-h hostname_2nd -np #2 -e MPI_WORKDIR=<workdir> mpp970_executable
	-h hostname_3rd -np #3 -e MPI_WORKDIR=<workdir> mpp970_executable
	-h hostname_4th -np #4 -e MPI_WORKDIR=<workdir> mpp970_executable
        .......

        hostname_1st is the hostname of the node where your submit the "mpirun" 
        command (root node), and hostname_2nd, hostname_3rd, hostname_4th and so on
        are hostnames for other nodes. #1, #2, #3, #4 are the number of processors 
        used from each node. Different number of processors can be used from 
        each node. 


IBM pwr3 and ppc:
	poe mpp960_executable -procs # -hfile node.list

	node.list contains the hostnames list for the MPP job to run. The
      	options can also given as environment variables, i.e.

	setenv MP_PROCS #
	setenv MP_HOSTFILE node.list

	Use "man poe" to get more detail information for IBM systems

Linux cluster:
  LAM-MPI:
	LAM/MPI can be freely downloaded from "www.lam-mpi.org".  
        Follow the instruction to install the package and start the 
        "lamd" deamon, and run the MPP job

	mpirun -np # mpp960_executable

  7.0.3:
        For LAM/MPI 7.0.3, MPI share libraries are needed for execution.  Please
        create libraries with single underscore for all symbols. (Linux default
        with g77 will give double underscore, i.e. __) Otherwise, the exe will
        not find out the proper external references. 
        OR, you can download the premade libraries from
        
        ftp://user@ftp.lstc.com/mpp-dyna/LAM-MPI/

        which matches your hardware to replace your local libraries.

  MPICH:
	MPICH can be freely downloaded from "www-unix.mcs.anl.gov/mpi/mpich".
	Follow the instruction to install the package and run the mpp job

	mpirun -np # mpp960_executable

NEC SX5:
	mpirun -np # mpp960_executable

SGI Origin:
	mpirun -np # mpp960_executable

SUN SPARC:
	mpirun -np # mpp960_executable

Windows 2000/NT/XP:
        MPICH can be freely downloaded from "www-unix.mcs.anl.gov/mpi/mpich".
        Follow the instruction to install the package and configurate the 
        clusters using the tool provided.  After the clusters are setup, run 
        the mpp job under DOS emulation window

        mpirun -np # mpp960_executable