基于MPI的静态并行算法库

Static Parallel Algorithm Lib Based on MPI

目前并行程序设计的状况是并行软件的发展落后于并行硬件,并行软件的缺乏是发展并行计算的主要障碍。基于这种状况,本文提出了构建并行算法库的思想,并给出了一个实例,说明如何将常用算法函数封装在一个高效优化的并行库中,从而可以被复用,从此免去并行程序员复杂、低效的劳动。

The current status of parallel programming is that parallel software development is behind the parallel hardware.The lack of parallel software is a major obstacle to the development of parallel computing.Based on this situation,constructing parallel algorithm library has been proposed to avoid the complicated and inefficient labor and give an example to show how to embed the frequently used algorithm function into a highly efficient parallel lib.