Many problems decompose well in large chunks that are well done with MPI on different nodes, and tight loops that are best done locally with shared memory processors. In such problems one can use both types of parallelisms well. My understanding of non-thread safe MPI is that one should avoid having MPI commands within OpenMP optimized chunk of code, but the contrary is allowed.
Dan _______________________________________________ Beowulf mailing list, Beowulf@beowulf.org To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
