A Paradigm of Extensions of Parallel Execution Model

In this paper, the conception extension of parallel execution model in layered idea for parallel computation model research is proposed, in which not only compiling and optimal running are included, but also partial parallel programming for parallelizing the program execution. Breaking through the thinking set of parallel algorithm design and parallel programming, an actual problem, the numerical weather prediction post-processing system, is parallelized by using a job-level parallel method. By parallelizing the execution of program, the wall-time of post-processing system is decreased evidently and ideal parallel efficiency is obtained. This case can be regarded as a paradigm of extensions of parallel execution model.