Tree-based parallel load-balancing methods for solution-adaptive unstructured finite element models on distributed memory multicomputers

Abstract

In this paper, we propose three tree-based parallel load-balancing methods, the MCSTPLB method, the BTPLB method, and the CBTPLB method, to deal with the load unbalancing problems of solution-adaptive finite element application programs. To evaluate the performance of the proposed methods, we have implemented those methods along with three mapping methods, the AE/ORB method, the AE/MC method, and the MLkP method, on an SP2 parallel machine. The experimental results show that (1) ifa mapping method is used for the initial partitioning and this mapping method or a load-balancing method is used in each refinement, the execution time of an application program under a load-balancing method is always shorter than that of the mapping method. (2) The execution time of an application program under the CBTPLB method is better than that of the BTPLB method that is better than that of the MCSTPLB method.