Research project performed as part of CSCI 6620 'Research Methods in Computer Science' at MTSU, TN, USA.

This repository contains the source code for the algorithm implementations tested as part of the research project that I performed for the 'Research Methods in Computer Science' (CSCI 6620) at Middle Tennessee State University, TN, USA, during the spring of 2018. With the exception of the Python scripts, this code is NOT MINE - I only edited it so that it would run on my machine. This code belongs to the authors of the associated papers and any inquiries should be addressed to them.

The directory names are self-explanatory (provided you have read the paper) and all contain a brief README.txt file that contains instructions on how to compile the source code and utilize the executable.

Please note: The source code was compiled and run on a machine running Debian Linux (v9 - 'Sid') using NVIDIA's nvcc compiler (v9.0.176) and (under the hood) the gcc v6.3.0 compiler with the NVIDIA CUDA v9.0.176 toolkit for a GPU with Compute Capability of 6.1 (a GP104, to be precise). If you have any trouble compiling the files, check the Makefiles and various dependencies (you may need to specify a path for an #include<>, for example) first, as certain support files may be in different locations and, particularly for Windows machines, any path names (as of writing this, I can't think of any in the source code) will almost certainly require changes. Finally, make sure that you have, at a bare minimum, CUDA v5 and a GPU with Compute Capability of at least 1.1, as the 1.0 models do not support a critical atomic function call, and then, as noted in the paper, if you use the Python-based utilities to create and convert graphs, the amount of available system RAM will dictate what you can do.