Some frequently employed algorithms in geosciences are parallel by nature (embarrasingly parallel algorithms) and some others can be parallelized via data parallelization. Algorithms like probability analysis, linear homotopy continuation method, Gauss-Jacobi combinatorial technique are belonging to the first group, while others like algorithms for digital image processing as well as Dixon resultant's application to solving systems of polynomial equations fall into the other category.
In this case study we will illustrate how Mathematica can manage to evaluate such algorithms parallel on a multicore machine equipped with Intel Nehalem i7 (Bloomfiled) 940 processor. The analysis of the efficiency of the computation and the net reduction of the execution time are presented as well as some useful tips are given to avoid pitfalls and to utilize the advantages of parallel processing. Download
ParallelCompOnNehalem.nb (236 MB) - Mathematica Notebook (updated 11/05/09)