EigAdept-The Expert Eigensolver Toolbox
EigAdept-The Expert Eigensolver Toolbox
复制标题
EigAdept-专家特征求解器工具箱
DOI：
--
复制
发表时间：
2006
期刊：
影响因子：
--
通讯作者：
Yeliang Zhang
中科院分区：
文献类型：
--
作者：
Xiaoye S. Li;Osni A. Marques;Yeliang Zhang
特征值和特征向量的计算在大范围的计算机模拟中是一个重要且通常耗时的阶段。最近的努力，在发展的特征求解器库给用户强大的算法，而不需要用户花很多时间在编程。然而，鉴于领域科学家可用的各种数值算法，为日益复杂和大型的应用程序选择“最佳”算法是一项艰巨的任务。在本文中，我们讨论了一种方法和一个软件工具箱，其目的是引导用户通过迷宫的各种特征值求解器与不同的配置，并在确定最好的特征值求解器的基础上的应用类型和矩阵属性。
The computation of eigenvalues and eigenvectors is an important and often time-consuming phase in a large range of computer simulations. Recent efforts in the development of eigensolver libraries have given users robust algorithms without the need for users to spend much time in programming. Yet, given the variety of numerical algorithms that are available to domain scientists, choosing the “best” algorithm for increasingly sophisticated and larger applications is a daunting task. In this paper, we discuss a methodology and a software toolbox that aims at guiding the user through the maze of various eigensolvers with different configurations, and at determining the best eigensolver based on the application type and the matrix properties.