Pavan Balaji, a computer scientist in Argonne’s Mathematics and Computer Science Division, delivered a keynote address at the13th International Workshop on OpenMP (IWOMP) held at Stony Brook, New York, in September 2017.
Balaji’s presentation began with the challenging statement “Interoperability between MPI and OpenMP is a myth.” He then rephrased his statement by explaining that although the MPI standard does not define any form of interoperability with OpenMP, it does define interoperability with threads. When the mapping of computation to threads is well defined, interoperability between MPI and OpenMP is possible; but as the semantics and usage models of MPI and OpenMP become more complex, the challenges are far greater.
Balaji discussed the status of the hybrid MPI+OpenMP programming model and compared it with competitors such as MPI+MPI programming that uses process shared memory within the node instead of OpenMP.
IWOMP is the premier forum to present and discuss issues, trends, recent research ideas, and results related to parallel programming with OpenMP. For further information, see the website http://openmpcon.org/conf2017/program/.