摘要
在三类不同体系结构的并行多处理机系统上，通过求解大型稀疏线性代数方程组的高效并行迭代算法的设计与实现，指出了高效并行算法的设计必须与某类高性能并行计算机相结合，才能使得同一解算问题在不同类型的高性能并行机上获得最佳性能价格比。
By the design and implementation of high efficent parallel algorithms for the solutions of very larger sparse linear algebraic equations on three types of parallel computers with different architectures, this paper points out that only if the parallel algorithms for solutions of concrete applied problems are designed according to the architectural characteristics of parallel computers, optimal performance can be obtained.
出处
《高技术通讯》
EI
CAS
CSCD
1998年第8期15-18,共4页
Chinese High Technology Letters
基金
863计划资助项目
关键词
并行计算机
并行算法
并行效率
设计
Parallel computer, Parallel algorithms, Parallel efficiency
