Abstract:
The geometric parallelism is often used for the numerical solution of problems
in the field of mathematical physics on distributed memory systems. As a result,
there arises the problem of balanced mesh distribution among processors.
This problem can be reduced to the graph partitioning problem. The parallel
decomposition of large triangular and tetrahedral meshes is the aim of this paper.
A parallel mesh partitioning tool is developed on the basis of the incremental
algorithm for graph partitioning and the recursive coordinate bisection algorithm.
The work was supported by the Russian Foundation for Basic Research (projects
05-01-00750, 08-07-00458, and 09-01-12022).
Fulltext:
PDF file (1349 kB)