MPI Blocks

EMA3D® is fully parallelized using protocols established with the Message Passing Interface (MPI) Standard, Version 2.2. EMA3D is designed to execute on any number of processors. To implement parallel processing requires the division of the finite difference problem space into smaller volumetric regions or MPI blocks. Each block corresponds to one processor. The problem space is divided, or blocked, by specifying the number of blocks in each coordinate direction. The total number of blocks, or processors, is obtained by multiplying the three provided numbers together.