CUDA implementation of the Bellman-Ford algorithm

This project is the thesis work of Filippo Barbari for academic year 2020/2021.

This is the modern version of the project (work in progress), for the original code, take a look at the old branch.

Compilation

In order to compile this project, you need:

• a CUDA-capable GPU
• CMake
• a C99 compiler
• a CUDA compiler

cmake -S . -B build
cmake --build build

Testing

After compilation, you can run the test suite to check that everything is set up correctly.

cd build
ctest

Each executable represents a different implementation of the algorithm and can be used in one of two ways:

• ./bf1-aos-sh <input_file> and the result will be printed on stdout.
• ./bf1-aos-sh <input_file> <solution_file> and the result will be checked against the given solution file. The result will not be printed.

Document

The main document of this project is a tex file inside the doc folder. You can create the .pdf file running

biber "tesi"
pdflatex -synctex=1 -interaction=nonstopmode "tesi".tex

License

This program is distributed under the GNU GPL v3 license. Please check the COPYING file.