Abstract: | An adaptive algorithm for solving large nonsymmetric linear systems is presented in this paper. The new algorithm combines polynomial preconditioning technique with the CGNR method. Residual polynomial is used in the preconditioning to estimate the eigenvalues of the s.p.d. matrix A TA, and the residual polynomial is generated from several steps of CGNR by recurrence. The algorithm is adaptive during its implementation. The robustness is maintained, and the iteration convergence is speeded up. Two numerical test results are also reported. |