Parallel implementation of the Floyd-Warshall algorithm.
The Floyd-Warshall algorithm has been implemented with MPI, OpenMP, and CUDA. Please check the corresponding directories for source code.
Almost linear speedup for MPI and OpenMP (tested with < 10 processors). 600+ speed up for CUDA with a graph of size 4096 ^ 2 (tested with NVIDIA GTX 770).