Parallel Computing with Mathematica

The MathLink communication protocol can be used to control several Mathematica kernel processes from within Mathematica. This feature allows the implementation of an environment for parallel programming, the Parallel Computing Toolkit, using processes with distributed memory. The toolkit is written completely in Mathematica in a machine-independent way, allowing its use in heterogeneous networks without common file systems, as well as on multi-processor machines. All library and application code is distributed through MathLink. Mathematica’s high-level language allows for the easy and natural expression of parallel programming paradigms and for the parallelization of other than purely numeric computations. Examples presented include virtual shared memory, fault-tolerance, and automatic parallelization of functional programming constructs.