论文部分内容阅读
航空、航天领域中有许多问题非常适合并且需要使用并行处理，尤其可以利用ＳＩＭＤ计算机，以提高处理速度，如流场计算，ＦＦＴ，矩阵运算等。但并行程序设计与串行程序设计相比较在实现上要更复杂。有许多影响并行程序效率的复杂因素需要通过实验进行研究，以不断优化算法。本文实现了一个在串行机上运行的适用于ＳＩＭＤ并行程序设计及性能评价的模拟环境—ＳＩＭＤＰ２（ＳＩＭＤＰｒｏｇｒａｍｍｉｎｇａｎｄＰｅｒｆｏｒｍａｎｃｅｅｖａｌｕａｔｉｎｇｅｎｖｉｒｏｎｍｅｎｔ）。该系统为研究者提供了一个廉价而灵活的ＳＩＭＤ计算机的并行算法研究及性能评价的实验环境。在该环境下，用户可分析、优化并行算法及评价所定义的ＳＩＭＤ系统结构
There are many problems in the field of aeronautics and astronautics, which are very suitable and require the use of parallel processing. In particular, the SIMD computer can be used to improve the processing speed, such as flow field calculation, FFT, matrix calculation and the like. However, the parallel programming and serial programming compared to the realization of more complex. There are many complications that affect the efficiency of parallel programs that need to be experimentally studied to optimize the algorithm. In this paper, SIMD Programming and Performance Evaluation Environment (SIMDP2), a simulation environment for SIMD parallel programming and performance evaluation, is implemented on a serial device. The system provides researchers with a cheap and flexible SIMD computer parallel algorithm research and performance evaluation of the experimental environment. In this environment, users can analyze and optimize parallel algorithms and evaluate the defined SIMD system architecture