Date of Award


Document Type


Degree Name

Master of Science in Computer Engineering


Department of Electrical and Computer Engineering

First Advisor

William C. Hobart Jr., PhD


This research addressed the development of a parallel computational fluid dynamics unstructured grid generator using Delaunay triangulation. The generator is applied to simple elliptical and cylindrical two-dimensional bodies. The methodologies used included Watsons point insertion algorithm, Holmes and Snyders point creation algorithm, a discretized surface definition, Andersons clustering function, and a Laplacian smoother. The first version of the software involved a processor boundary exchange at the end of each iteration with no inter-processor communications during the iterations The second version used inter-processor communication during each iteration instead of the boundary exchange. Version 1 demonstrated a speedup of 1.8 for some portions of the code, but proved to be unscalable for more than two nodes due to the interdependency of the triangular elements. The results of Version 2 were similar. Two distribution methodologies, a simple 360-degree distribution and recursive spectral bisection RSB, were examined. For the initial grid distribution, the distribution generated by the RSB code would be similar to the distribution generated by the 360-degree methodology and would require significantly more time to execute.

AFIT Designator


DTIC Accession Number



The author's Vita page is omitted.