A New Approach to rCUDA
DISCA, Universitat Polit’ecnica de Val’encia (UPV)
XXII Jornadas de Paralelismo, 2011
@article{duato2011new,
title={A New Approach to rCUDA},
author={Duato, J. and Pena, A.J. and Silla, F. and Fern{‘a}ndez, J.C. and Mayo, R. and Quintana-Ort{i}, E.S.},
year={2011}
}
In this paper we propose a first step towards a general and open source approach for using GPGPU (General-Purpose Computation on GPUs) features within virtual machines (VMs). In particular, we describe the use of rCUDA, a GPGPU virtualization framework, to permit the execution of GPU-accelerated applications within VMs, thus enabling GPGPU capabilities on any virtualized environment. Our experiments with rCUDA in the context of KVM and VirtualBox on a system equipped with two NVIDIA GeForce 9800 GX2 cards illustrate the overhead introduced by the rCUDA middleware and prove the feasibility and scalability of this general virtualizing solution.
January 23, 2012 by hgpu
