Parallelization strategy (distributed memory)

Mesh computations are readily parallelized for distributed-memory architectures using the following approach:

• The grid (N-dimensional) is distributed over an N-dimensional grid of processes.
• A separate (optional) host process is used for reads/writes involving a whole array.
• Non-grid variables -- global constants and reduction variables, e.g. -- are duplicated in each process.

• Distributing the data
• Computing new values for grid points
• Performing I/O
• Structuring the parallel program