Parallelization of Toeplitz Solvers

摘要: We shall describe parallel algorithms for solving Toeplitz and block systems, based on the Levinson type methods implemented in serial Package (by Arushanian et al. [1]). A matrix A, with blocks i , j has structure if satisfy = − all i, such that 0 ≤ n 1 each is a general matrix. The implementation performance of these MIMD shared memory machines will be discussed. It shown contribution inner product calculations to time complexity negligible typical applications.