Runge-Kutta-Nyström-type parallel block predictor-corrector methods (Q1282327)

The authors investigate a particular class of explicit Runge-Kutta-Nyström-type block predictor-corrector methods for use on parallel computers. The approach consists of applying the predictor-corrector method not only at step points, but also at off-step points (block points). By using direct Runge-Kutta-Nyström correctors, the authors obtain block parallel iterated Runge-Kutta-Nyström (BPIRKN) methods possessing both faster convergence and smaller truncation error. The algorithm described in the paper requires two sequential derivative evaluations per step for any order of accuracy. Comparisons with the highly efficient code ODEX2 show many advantages of BPIRKN methods.