An Efficient Implementation of GPU Virtualization in High Performance Clusters
Departamento de Informatica de Sistemas y Computadores, Universidad Politecnica de Valencia (UPV), 46022 – Valencia, Spain
In Euro-Par 2009, Parallel Processing Workshops, Vol. 6043 (2010), pp. 385-394
@conference{duato2010efficient,
title={An efficient implementation of GPU virtualization in high performance clusters},
author={Duato, J. and Igual, F. and Mayo, R. and Pe{~n}a, A. and Quintana-Ort{‘i}, E. and Silla, F.},
booktitle={Euro-Par 2009–Parallel Processing Workshops},
pages={385–394},
year={2010},
organization={Springer}
}
Current high performance clusters are equipped with high bandwidth/low latency networks, lots of processors and nodes, very fast storage systems, etc. However, due to economical and/or power related constraints, in general it is not feasible to provide an accelerating co-processor – such as a graphics processor (GPU) – per node. To overcome this, in this paper we present a GPU virtualization middleware, which makes remote CUDA-compatible GPUs available to all the cluster nodes. The software is implemented on top of the sockets application programming interface, ensuring portability over commodity networks, but it can also be easily adapted to high performance networks.
November 22, 2010 by hgpu
