Date of Defense

4-19-2016

Date of Graduation

4-2016

Department

Computer Science

First Advisor

John Kapenga

Second Advisor

Elise de Doncker

Abstract

The best software is easy to configure and compile, is expandable and is well tested. Development of the ParInt software package for parallel integration stopped a number of years ago. When handed the software, parts of the package no longer configured or compiled consistently. Furthermore it relied on one random number generator and had no functional testing. The team tuned the auto configuration so that the program would configure and build on current systems, created expandable functionality to add new random number generators and created two functional test packs. Now ParInt can be developed with confidence knowing that it compiles on current systems, multiple random number generators can be used and regressional tests can be run when changes have been made to the source code.

Comments

co-authored by Dillon Daudert, Austin Jones, and Kendrick Cline

Access Setting

Honors Thesis-Open Access

ParInt 1.3 Power Point Presentation.pdf (1228 kB)
Thesis Presentation

Share

COinS