Parallel Algorithm of Restarting Block Davidson MethodBased on PC Network Parallel Environment
Wang Shun-xu
Abstract
Wang Shun-xu
Abstract
A parallel restarting block Davidson method is presented in this paper for calculating extreme eigenpairs of large sparse matrix A based on PC network parallel environment. The individual processors run under the control of a program base on the partial orthogonal base of projection subspace V. If the dimension of the subspace V is greater than m, the subspace V is constructed with new Ritz vector, and the iteration is restarting. MPI is installed in the operation system of windows 2000 on microcomputers to construct a PC network parallel computing environment. The results of numerical experiments show that the algorithm is highly effective.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
A parallel restarting block Davidson method is presented in this paper for calculating extreme eigenpairs of large sparse matrix A based on PC network parallel environment. The individual processors run under the control of a program base on the partial orthogonal base of projection subspace V. If the dimension of the subspace V is greater than m, the subspace V is constructed with new Ritz vector, and the iteration is restarting. MPI is installed in the operation system of windows 2000 on microcomputers to construct a PC network parallel computing environment. The results of numerical experiments show that the algorithm is highly effective.
Key concepts: Subspace topology, Block (permutation group theory), Construct (python library), Computer science, Algorithm, Dimension (graph theory), Parallel computing, Base (topology)