This project provides an parallel implementation of spectral partitioning for very large sparse graphs with MPI.
Graphs are fun! This project started in SF2568 Parallel Computations for Large-Scale Problems at KTH Royal Institute of Technology.
The implemented algorithm require OpenMPI and ressources for parallel computing. The code was compiled, run and tested on PDC. The input graph files must have METIS format.
The compilation command in the makefile uses mpiicc. To compile, run the command
$> make compile
The execution command in the makefile uses mpirun. To execute, run the command
$> make run DATA=fileDirectory INPUT=file.graph