The International Conference for High Performance Computing, Networking, Storage, and Analysis
The International Conference for High Performance Computing, Networking, Storage, and Analysis
Workshop:WACCPD 2020: Seventh Workshop on Accelerator Programming Using Directives
Authors: Neil Mehta and Rahulkumar Gayatri (National Energy Research Scientific Computing Center (NERSC)), Yasaman Ghadar and Christopher Knight (Argonne National Laboratory (ANL)), and Jack Deslippe (National Energy Research Scientific Computing Center (NERSC))
Abstract: In this paper, we show that OpenMP 4.5-based implementation of TestSNAP, a proxy-app for the Spectral Neighbor Analysis Potential (SNAP) in LAMMPS, can be ported across the NVIDIA, Intel and AMD GPUs. Roofline analysis is employed to asses the performance of TestSNAP on each of the architectures. The main contributions of this paper are two-fold: provide OpenMP as a viable option for application portability across multiple GPU architectures, and provide a methodology based on the roofline analysis to determine the performance portability of OpenMP implementations on the target architectures. The GPUs used for this work are Intel Gen9, AMD Radeon Instinct MI60, and NVIDIA Volta V100.
Back to WACCPD 2020: Seventh Workshop on Accelerator Programming Using Directives Archive Listing