Investigation of Parallel Computation – MPI, CUDA and Parallel Visualization
Department of Computer Science, University of Wisconsin-La Crosse, La Crosse, Wisconsin
University of Wisconsin-La Crosse, 2011
@phdthesis{fan2011investigation,
title={Investigation of Parallel Computation-MPI, CUDA and Parallel Visualization},
author={Fan, L.},
year={2011},
school={University of Wisconsin}
}
In this manuscript, the parallel computation is investigated including reviewing different programming APIs and architectures. Two specific parallel API-MPI and CUDA C are deeply analyzed. Two sorting algorithms and a visual mathematic problem are implemented with MPI alone with performance analysis. A stable fluid dynamics simulation has been experimented with CUDA. We also present a parallel way to implement a visualization application in the cluster with integration of multi-thread, CUDA and OpenGL. The performance analysis is given and future improvements are also pointed out.
January 18, 2012 by hgpu
