Fine-grained parallel boundary elements

摘要

The early parallel application of the boundary element method in the mid-1980s were exclusively of the fine-grained class on large SIMD machines. The method itself has a very clear inherent parallelism which maps directly on to such architectures. Such implementation were very successful but were being developed by only a small group of researchers. The boundary element community really started taking notice of parallel computers with the availability of much cheaper MOMD architectures and this would appear to be the future for most parallel Boundary element applications.