A Package for OpenCL Based Heterogeneous Computing on Clusters with Many GPU Devices
Department of Computer Science, The Hebrew University of Jerusalem, Jerusalem 91904, Israel
IEEE International Conference on Cluster Computing Workshops and Posters (CLUSTER WORKSHOPS), 2010
@conference{barak2010package,
title={A package for OpenCL based heterogeneous computing on clusters with many GPU devices},
author={Barak, A. and Ben-Nun, T. and Levy, E. and Shiloh, A.},
booktitle={Cluster Computing Workshops and Posters (CLUSTER WORKSHOPS), 2010 IEEE International Conference on},
pages={1–7},
organization={IEEE}
}
Heterogeneous systems provide new opportunities to increase the performance of parallel applications on clusters with CPU and GPU architectures. Currently, applications that utilize GPU devices run their device-executable code on local devices in their respective hosting-nodes. This paper presents a package for running OpenMP, C++ and unmodified OpenCL applications on clusters with many GPU devices. This Many GPUs Package (MGP) includes an implementation of the OpenCL specifications and extensions of the OpenMP API that allow applications on one hosting-node to transparently utilize cluster-wide devices (CPUs and/or GPUs). MGP provides means for reducing the complexity of programming and running parallel applications on clusters, including scheduling based on task dependencies and buffer management. The paper presents MGP and the performance of its internals.
March 15, 2011 by hgpu
