PARCS: An MPP-Oriented CLP Language

: PARCS is a declarative parallel constraint logic programming (CLP) language designed for efficient execution on distributed memory massively parallel processors (MPPs). As a language model, PARCS offers efficient parallel execution control via priority-based goal and constraint execution, powerful pruning mechanisms, and implicit OR-parallelism. Its MPPoriented implementation techniques include efficient load balancing, branching method and static compile-time optimization techniques such as mode-specific code generation, static ordering of deterministic constraints and determinacy analysis. Actual experiments have shown good speedups with respect to the number of processors for various kinds of programs, such as N-queens and paraffin problem. Keywords: constraint logic programming, massively parallel processors, load balancing, compilation techniques 1 Introduction Although massively parallel processors (MPPs) are available from various vendors these days, it is often cumbersome t..