🚀 go-pugleaf

                                PH.D. STUDENT TALKS

                University at Albany - State University of New York

                            Computer Science Department

                                     Tsozen Yeh

            Monday, December 1, 1997, at ****3:35p.m. in LI-83*********

                Double Standards:  Bringing Task Parallelism to HPF

                         via the Message Passing Interface

                                     Authors:
            Ian Foster, David R. Kohr,Jr., Rakesh Krishnaiyer, Alok Choudhary

                                      Abstract

                 High Performance Fortran  (HPF)  provides  a  portable,
            high-level notation for expression data-parallel algorithms.
            However, HPF does not allow efficient  expression  of  mixed
            task/data-parallel   computations   or   the   coupling   of
            separately compiled data-parallel modules.  This paper shows
            how  a coordination library implementing the Message Passing
            Interface (MPI) can be used to represent these common paral-
            lel  program  structures.   The  MPI  library  allows  data-
            parallel tasks to exchange distributed data structures using
            calls to simple communication functions. The authors present
            microbenchmark results that characterize the performance  of
            this  library  and that quantify the impact of optimizations
            that allow reuse of communication schedules in common situa-
            tions. In addition, results from two-dimensional FFT, convo-
            lution, and multiblock programs demonstrate that the HPF/MPI
            library  can  provide  performance  superior to that of pure
            HPF. We conclude that this synergistic  combination  of  two
            parallel  programming standards represents a useful approach
            to task parallelism in a data-parallel framework, increasing
            the  range  of problems addressable in HPF without requiring
            complex compiler technology.