Location of MPI implementation
[primaryLib,extras] = mpiLibConf
| MPI implementation library used by a communicating job. |
| Cell array of other required library names. |
[primaryLib,extras] = mpiLibConf returns
the MPI implementation library to be used by a communicating job. primaryLib is
the name of the shared library file containing the MPI entry points. extras is
a cell array of other library names required by the MPI library.
To supply an alternative MPI implementation, create a file named mpiLibConf.m,
and place it on the MATLAB® path. The recommended location is matlabroot/toolbox/distcomp/usermpiLibConf.m file must be higher on the cluster
workers' path than matlabroot/toolbox/distcomp/mpimpiLibConf.m as a file dependency for
this purpose does not work.) After your mpiLibConf.m file
is in place, update the toolbox path caching with the following command
in MATLAB:
rehash toolboxcache
Use the mpiLibConf function to view the current
MPI implementation library:
mpiLibConf
mpich2.dll
Under all circumstances, the MPI library must support all MPI-1
functions. Additionally, the MPI library must support null arguments
to MPI_Init as defined in section 4.2 of the MPI-2
standard. The library must also use an mpi.h header
file that is fully compatible with MPICH2.
When used with the MATLAB job scheduler or the local cluster, the MPI library must support the following additional MPI-2 functions:
MPI_Open_port
MPI_Comm_accept
MPI_Comm_connect
When used with any third-party scheduler, it is important to launch the workers using the version of mpiexec corresponding to the MPI library being used. Also, you might need to launch the corresponding process management daemons on the cluster before invoking mpiexec.