Early Application Experiences on a Modern GPU-Accelerated Arm-based HPC Platform
Wael Elwasif, Sergei Bastrakov, Spencer H. Bryngelson, Michael Bussmann, Sunita Chandrasekaran, Florina Ciorba, M. A. Clark, Alexander Debus, William Godoy, Nick Hagerty, Jeff Hammond, David Hardy, J. Austin Harris, Oscar Hernandez, Balint Joo, Sebastian Keller, Paul Kent, Henry Le Berre, Damien Lebrun-Grandie, Elijah MacCarthy, Verónica G. Melesse Vergara, Bronson Messer, Ross Miller, Sarp Oral, Jean-Guillaume Piccinali, Anand Radhakrishnan, Osman Simsek, Filippo Spiga, Klaus Steiniger, Jan Stephan, John E. Stone, Christian Trott, René Widera, Jeffrey Young
author={Elwasif, Wael and Bastrakov, Sergei and Bryngelson, Spencer H. and Bussmann, Michael and Chandrasekaran, Sunita and Ciorba, Florina and Clark, M. A. and Debus, Alexander and Godoy, William and Hagerty, Nick and Hammond, Jeff and Hardy, David and Harris, J. Austin and Hernandez, Oscar and Joo, Balint and Keller, Sebastian and Kent, Paul and Berre, Henry Le and Lebrun-Grandie, Damien and MacCarthy, Elijah and Vergara, Verónica G. Melesse and Messer, Bronson and Miller, Ross and Oral, Sarp and Piccinali, Jean-Guillaume and Radhakrishnan, Anand and Simsek, Osman and Spiga, Filippo and Steiniger, Klaus and Stephan, Jan and Stone, John E. and Trott, Christian and Widera, René and Young, Jeffrey},
keywords={Distributed, Parallel, and Cluster Computing (cs.DC), Hardware Architecture (cs.AR), FOS: Computer and information sciences, FOS: Computer and information sciences},
title={Early Application Experiences on a Modern GPU-Accelerated Arm-based HPC Platform},
This paper assesses and reports the experience of eleven application teams working to build, validate, and benchmark several HPC applications on a novel GPU-accelerated Arm testbed. The testbed consists of the latest, at time of writing, Arm Devkits from NVIDIA with server-class Arm CPUs and NVIDIA A100 GPUs. The applications and mini-apps are written using multiple parallel programming models, including C++, C, CUDA, Fortran, OpenACC, and OpenMP. Each application builds extensively on the other tools available in the programming environment, including scientific libraries, compilers, and other tooling. Our goal is to evaluate application readiness for the next generation of Arm and GPU-based HPC systems and determine the tooling readiness for future application developers. On both accounts, the reported case studies demonstrate that the diversity of software and tools available for GPU-accelerated Arm systems are prepared for production, even before NVIDIA deploys their next-generation such platform: Grace.
