ABSTRACT
It is often desirable to break up a large problem into parts in such a way as to distribute the workload in a balanced manner and also to optimize the volume of communication between the processes carrying out the computation. Often this gets modeled as a graph/hypergraph partitioning problem.
