This paper presents a divide-and-conquer algorithm for solving eigenvalue problem of nonsymmetric matrices. The new algorithm bases on Languerre iteration. Theoretical analysis and Numerical results show that our algorithm is faster, and able to obtain more different eigenvalues than J. J. Dengarra’s algorithm presented in [1]. Above all, our afeorithm is well suitable to parallel implementation. Numerical results of parallel computing are also presented in this paper. The parallel efficiency is encouraging.
展开▼
