If we use on processor, and if a large number of calculations is performed at each point in the domain, then it will take a long time to perform the computation.  Instead we can split up this region and assign different parts to different machines.

Suppose we have 5 machines.  The region could be split into four separate regions so the number of computations performed by each machine is reduced.  Information about the boundaries of these regions will need to be shared between the different machines.

Talk about reason for doing so. Why use MATLAB Instead of FORTRAN or C ?
Easy to program. Easier to edit and modify.  MATLAB come with many toolboxes that are useful for solving engineering problems. MATLAB is optimized for performing matrix calculations and it has many valuable tools to solve problems so this it will become the number one choice for engineers as it continues to improve.

Some of these toolboxes are based upon MPI and others are based upon TCP/IP.  The first three enable MPI calls inside Matlab And the other three are based on TCP/IP Communication.  Up to this point in my research, MPITB and Paramatlab have been installed.  Two nodes have been configured to run MATLAB in Parallel and it seems that MPITB is almost ready to function properly.