A hierarchical partition model for adaptive finite element computation (Q1574333)

This paper presents a hierarchical partition model for adaptive finite element computations. The authors describe the design and implementation of parallel mesh structures within an adaptive framework. The most fundamental concept is a hierarchical partition model used to distribute finite elements and associated data on parallel computers. The hierarchical model represents heterogeneous processors and network speeds, and can be used to investigate the corresponding processes in any parallel computing environment, including a distributed-memory computer, a network of workstations, or some their combinations. Using this model, the authors are able to segment the computations into chunks which can fit into cache memory. This provides a potential gain in efficiency based on the increased cache hit rate, even in a single processor environment. The authors also give an information about different processor speeds, memory sizes, and the corresponding interconnection networks which can be useful for the construction of dynamic load balancing algorithms aimed to achieve a good balance with minimal interprocessor communication penalties when a slow interconnection network is involved.