A Tuple-Based Data Structure for Distributed Parallel Processing of 3D Dynamic Meshes
by Peter Hartmann
Abstract:
A data structure for the distributed representation of inhomogeneous and dynamically changing spatial structures is introduced. It is based on the concept of cellular hypergraphs (CllG), which provides us with a universal framework in this domain: a CllG is used to represent the topological relationships within a spatial structure, and the temporal dynamics are formulated with the productions of a parallel replacement system. The presented data structure is based on tuples keeping relationships between edges of the graph and components storing state information. A method for moving cells between processor nodes is described in detail, for it has a central meaning in load balancing algorithms and the distributed implementation of the parallel replacement systems.
Keywords: theory
Source:
P. Hartmann, A Tuple-Based Data Structure for Distributed Parallel Processing of 3D Dynamic Meshes. In V. Malyshkin (ed.),
Parallel Computing Technologies: Proceedings of the 4th International Conference,
Lect. Notes in Comp. Sci., Vol. 1277, Springer, 1997, pp. 25-39