Using the Parallel Research Kernels to Study PGAS Models

A subset of the Parallel Research Kernels (PRK),simplified parallel application patterns, are used to studythe behavior of different runtimes implementing the PGASprogramming model. The goal of this paper is to show thatsuch an approach is practical and effective as we approachthe exascale era. Our experimental results indicate that forthe kernels we selected, MPI with two-sided communicationsoutperforms the PGAS runtimes SHMEM, UPC, Grappa, andMPI-3 with RMA extensions.