摘要
近年来,随着日常生活等实际应用领域中大规模稀疏矩阵求解问题的推动,代数多重网格(AMG)算法及其并行化的研究成为了数值计算领域的热点。本文在原始AMG算法和MPRS算法的基础上,对现有的并行AMG算法提出了一种优化的动态阈值算法(DVRS)。在Visual Studio 2008环境下,数值计算实验结果表明,新算法适用于更广泛的领域,与原有的并行AMG算法相比,改善了AMG并行计算的可扩展性。
In recent years, the need for solving large-scale sparse matrix in the practical application domain of life has sparked great interest in the research of algebraic multigrid algorithm(AMG) and its parallelization. Based on initial AMG method and MPRS algorithm, the authors propose an optimized DVRS algorithm for existing parallel AMG algorithm. Under the environment of Visual Studio 2008, the experiments and experimental results show that the new method is suitable for more widespread domain and improves the algorithmic scalability of the parallel AMG computing remarkably.
出处
《电子技术(上海)》
2011年第1期18-21,共4页
Electronic Technology