Experiments with Massively Parallel Matrix Multiplication
Dept. of Computer Science, The University of Akron, Akron OH 44325 USA
CATA 2014, March 24-26, 2014, Las Vegas, Nevada, USA, 135
@article{oexperiments,
title={Experiments with Massively Parallel Matrix Multiplication},
author={O’Neil, Timothy W}
}
This paper presents initial experiments in implementing two notable matrix multiplication algorithms – the DNS algorithm and Cannon’s algorithm – using NVIDIA’s general-purpose graphics processing units (GPGPUs) and CUDA development platform. We demonstrate that these implementations are comparable with traditional methods in terms of computational expense and may scale better than traditional techniques.
April 3, 2014 by hgpu
