基于CRS提出了一种适合于分布式并行环境改进的平方共轭残差方法——ICRS.通过算法重构,ICRS方法将CRS方法所需要的2个全局同步化点降低到了1个,没有数据相关性.理论分析和实验表明ICRS方法比CRS方法具有更好的并行性和可扩展性.
Based on the CRS method, an ICRS method is designed for distributed parallel environments. The improved method reduces two global synchronization points to one by changing the computation sequence in the CRS method and all inner products per iteration are independent. Theoretical analysis and numerical experiments show that the ICRS method has better parallelism and scalability than the CRS method.