摘要
有效地进行任务划分、控制并行粒度,才能充分利用并行计算机的资源,通过对复杂连续系统仿真程序特点的分析,提出了以状态方程为核心、以右函数段的计算为主要对象的自动任务划分算法,使用结果表明具有很好的并行效果。
The parallel computer's resource can be used efficiently by proper task partitioning and parallel granularity controlling. By analyzing the characteristic of simulation program, this paper points out a task partitioning algorithm which take state equations as core and right-hand function as object. The algorithm has been used in the automatic parallel tools which we have developed and the application results have showed good effect on parallel simulation.
出处
《计算机工程与设计》
CSCD
北大核心
2000年第2期13-17,共5页
Computer Engineering and Design