Computing roots of polynomials on vector processing machines (Q1089743): Difference between revisions

The algorithm of \textit{E. Durand} [Solutions numériques des équations algébriques, Tome I (1960; Zbl 0099.108), p. 279-281 ] and \textit{I. O. Kerner} [Numer. Math. 8, 290-294 (1966; Zbl 0202.436)] for the computation of roots of polynomials has not been of great use up to now. The reason is the existence of more efficient methods for computers of the SISD type (sequential). Actually vector processing machines such as CRAY-1 or CDC CYBER 205 must bring Durand's algorithm back to honour because of its possibility of extensive vectorization. However, as the method has its maximum efficiency for a polynomial with no multiple root a criterion using permutation-perturbation method of \textit{J. Vignes} [Math. Comput. Simulation 20, 227-249 (1978; Zbl 0437.65041)] is given to know whether the roots are all distinct or not. The optimum criterion for stopping the iterations is shown to be vectorizable and some useful properties of the method are given. Numerical examples are considered.