Structured mesh repartitioning is an important class of mesh partitioning, which has useful applications in many areas of scientific computing. The project targets to mesh repartitioning problems in real large scale parallel numerical simulation applications, bases on the features of structured mesh in data structure and computing algorithms, researches structured mesh repartitioning algorithms in algorithms designing, algorithms analyzing, software developing, verification researching, etc. Compared with general ones, the repartitioning algorithm of this project will have better performance and profounder theoretic analyzing results. The research achievements will support two application programs to fulfill large scale numerical simulations. The project will provide the theoretic foundation of structured mesh parallel repartitioning algorithms within this class, while it will have great significance for structured mesh related parallel computing.
结构网格重剖分是一类重要的网格剖分,在科学计算的多个领域都有着重要的应用。本项目面向大规模并行数值模拟应用中的网格重剖分问题,围绕结构网格在数据结构和计算方法的特点,从结构网格并行重剖分算法设计、结构网格并行重剖分算法分析、结构网格并行重剖分软件研制、结构网格并行重剖分验证研究等方面,开展结构网格重剖分算法的研究。与通用网格重剖分算法相比,本项目研究的网格重剖分算法将具有更好的性能,并且有着更深刻的理论分析结果。研究成果将支持两个应用程序实现大规模数值模拟。项目将为该类结构网格并行重剖分提供理论基础,对结构网格相关的并行计算有重要意义。
{{i.achievement_title}}
数据更新时间:2023-05-31
武功山山地草甸主要群落类型高光谱特征
基于边信息的高光谱图像恢复模型
基于空间特征重标定网络的遥感图像场景分类
A Fast Algorithm for Computing Dominance Classes
GSI同化GPS可降水量对四川盆地不同类型降水模拟影响
面向全空间有限元数值模拟的四面体网格并行生成算法研究
三维动态混合网格自适应移动与重剖分方法研究
面向GPU的Delaunay三角剖分网格细化算法研究
基于网格的大规模并行关系查询算法的研究