Balance-Enforced Multi-Level Algorithm for Multi-Criteria Mesh Partitioning

1 Context Nowadays, numerical simulations model increasingly complex phenomena. They require deeply coupled multi-physics codes that are designed to run on large distributed memory computers. On these kinds of architecture, data decomposition is critical to achieve good performance. While distributing the data between the processes, two challenges must be addressed: balancing the workload between all processes, while minimizing the communications between processes. In multi-physics simulations, one iteration gathers the computations for several phases (see algorithm 1). The computation costs depend on the physical phase, so a partition well-balanced for one phase can be imbalanced for another. Computing one partition per phase must update the data distribution after each phase, leading to a great amount of communications, as well as one partition computation per phase. Finding a partition well-balanced for all phases garantees a correct parallel efficiency.