Issues of partitioning Finite Element Graphs are central to parallel formulations of the Finite Element Method. Due to the nature of the problem, optimal partitioning schemes should conform to three basic criteria: equal load on all processors; locality of communication; and maximum computation to communication ratio associated with each partition. Many techniques have been presented in literature which achieve these to different degrees. This paper presents scalability analysis of three partitioning strategies, namely, striped partitioning, binary decomposition, and scattered decomposition. The analysis is performed using the Isoefficiency metric, which helps in predicting performance of these schemes on a range of processors and architectures.
|Original language||English (US)|
|Title of host publication||Proceedings of the 1992 ACM/IEEE conference on Supercomputing, Supercomputing 1992|
|Publisher||Association for Computing Machinery|
|Number of pages||10|
|State||Published - Dec 1 1992|
|Event||1992 ACM/IEEE conference on Supercomputing, Supercomputing 1992 - Minneapolis, United States|
Duration: Nov 16 1992 → Nov 20 1992
|Name||Proceedings of the International Conference on Supercomputing|
|Other||1992 ACM/IEEE conference on Supercomputing, Supercomputing 1992|
|Period||11/16/92 → 11/20/92|
Bibliographical noteFunding Information:
These conditions represent the classical communication - load imbalance tradeoffs. Optimizing one of these criterion leads to a deterioration with respect to one or more of the other criteria. The mapping *This work was supported by IST/SDIO through the Army Research Office grant # 28408-MA-SDI to the University of Minnesota and by the Army High Performance Computing Research Center at the University of Minnesota.