Simultaneous Computational and Data Load Balancing in Distributed-Memory Setting

Several successful partitioning models and methods have been proposed used for computational load balancing of irregularly sparse applications in a distributed-memory setting. However, the literature lacks that encode both data balancing. In this article, we try to close gap by proposing two hypergraph (HP) which simultaneously Both utilize two-constraint formulation, where first constraint encodes loads second loads. model, introduce explicit vertices encoding replicate those at each recursive bipartitioning (RB) step replication. weight distribution scheme update weights RB step. The nice property is they do not necessitate developing new partitioner from scratch. can easily be implemented invoking any HP tool supports multiconstraint as two-way validity are tested on widely applications: parallel mesh simulations matrix multiplication. achieve significant improvement over baseline model.