International Conference on Lightning Protection · 1988 · 41 citations · 0 references
ComputingComputer EngineeringComputer ScienceComputer ArchitectureParallel AlgorithmsParallel ComputingParallel Algorithm DesignParallel ArchitecturesParallel ComputationParallel Computer ArchitectureShared-memory MultiprocessingShared-memory ParallelismAbstract MachinesConcurrent Logic ProgrammingDependency AnalysisExecution ModelParallel Logic ProgrammingParallel Speedup AnalysisShared MemoryShared Memory Multiprocessors
This paper presents the implementation of an AND-parallel execution model of logic programs on a shared-memory multiprocessor. The major features of the implementation are (i) dependency analysis between literals of a clause is done dynamically without incurring excessive run-time overhead; (ii) backtracking is done intelligently at the clause level without incurring any extra cost for the determination of the backtrack literal; (iii) the implementation is based upon the Warren Abstract Machine (WAM), hence retains most of the efficiency of the WAM for sequential segments of logic programs. Performance results on Sequent Balance 21000 show that our parallel implementation can achieve reasonable speedup on dozens of processors.