Hardware-targeted semi-implicit extrapolation ODE solvers

One of the most valuable problems in control systems design is fast and accurate numerical simulation. The developer of a particular simulation system should choose the most efficient numerical integration method, taking into account features of the hardware platform. Today personal computers became a powerful tool for engineering and scientific applications, but have some disadvantages comparing to modern supercomputers: lack of parallelism, scalar processors and special problem requirements to effective program execution on GPU are the first problems that the simulation software faces. In this paper, we propose an efficient semi-implicit extrapolation D-method that shows the best performance on the modern personal computer than the other studied integration methods, and is suitable for embedded platforms as well. A brief description of the method is given, and experimental results are presented.

[1]  Roberto Barrio,et al.  Breaking the limits: The Taylor series method , 2011, Appl. Math. Comput..

[2]  Eberhard Schäfer,et al.  A new approach to explain the “high irradiance responses” of photomorphogenesis on the basis of phytochrome , 1975 .

[3]  E. Hairer,et al.  Solving Ordinary Differential Equations II: Stiff and Differential-Algebraic Problems , 2010 .

[4]  Edmond Chow,et al.  Fine-Grained Parallel Incomplete LU Factorization , 2015, SIAM J. Sci. Comput..

[5]  Alan H. Karp,et al.  A variable order Runge-Kutta method for initial value problems with rapidly varying right-hand sides , 1990, TOMS.

[6]  P. Deuflhard.  Order and stepsize control in extrapolation methods , 1983 .

[7]  Jeffrey T. Draper,et al.  Floating-point division and square root using a Taylor-series expansion algorithm , 2007, 2007 50th Midwest Symposium on Circuits and Systems.

[8]  J. Sprott,et al.  Some simple chaotic flows. , 1994, Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics.

[9]  D. N. Butusov,et al.  Simulation of dynamical systems based on parallel numerical integration methods , 2015, 2015 IEEE NW Russia Young Researchers in Electrical and Electronic Engineering Conference (EIConRusNW).