摘要
使用基于MPI并行编程方法对IFortran77串行计算程序进行并行处理,主要工作在循环级的可并行性研究和并行实现。文中给出并行代码在国家高性能计算中心(合肥)的曙光-1000机上运行的并行加速比以及相应的系统并行效率。
In this article, we make parallelizing processing to a Fortran77 serial routine by MPI-based parallel-programming, and put emphasis upon parallelability research and implement of cycling-layer in source programme.The parallelized code was debugged and executed on Dawn-1000 computer in National High Performance Computing Center at Hefei. The actual paralleling speed ratio and the relative paralleling efficiency are given. ;;;
出处
《计算机工程》
CAS
CSCD
北大核心
2001年第7期46-47,87,共3页
Computer Engineering
基金
国家高性能计算基金资助项目(99207)
