Author : Pooja S. Patil 1
Date of Publication :21st August 2017
Abstract: Numerical methods for solving fractional differential equations are computationally heavy due to the need of floatingpoint operations, the non-local nature of the fractional differential operators and more importantly, the data flow inside the entire memory system of a computer. Hence such systems can be implemented on Graphics Processing Unit (GPU) which has the parallel computing power for quicker simulation. A GPU has a number of threads where each thread can execute different program. MATLAB and Parallel Computing toolboxe can be used to access the computational power of GPU and MATLAB code can be implemented on the GPU. This helps us to achieve significant & faster computation than a normal CPU system. In this paper an attempt is made to implement numerical method for simple fractional ordinary differential equation (FODE) on a Dual Core CPU and NVIDIA GPU. This paper presents the relative performance of a GPU v/s CPU for fractional Euler’s method to solve FODE. From the results presented, it is observed that GPU provides two times speed up for fractional Euler’s method.
Reference :
-
[1] Diethelm, Kai.“The analysis of fractional differential equations: An application-oriented exposition using differential operators of Caputo type.”Springer, Verlag Berlin Heidelberg, 2010.
[2] Machado, J. Tenreiro, Virginia Kiryakova, and Francesco Mainardi. “Recent history of fractional calculus.”Communications in Nonlinear Science and Numerical Simulation Elsevier 16, no. 3 (2011): 1140- 1153.
[3] Weilbeer, Marc.“Efficient numerical methods for fractional differential equations and their analytical background.”Papierflieger, 2005.
[4] ur Rehman, Mujeeb, and Rahmat Ali Khan. “The Legendre wavelet method for solving fractional differential equations.”Communications in Nonlinear Science and Numerical Simulation Elsevier 16, no. 11 (2011): 4163-4173.
[5] Suh, Jung W., and Youngmin Kim.“Accelerating MATLAB with GPU computing: A primer with examples.”Newnes, 2013.
[6] Higham, Desmond J., Xuerong Mao, and Andrew M. Stuart. “Strong convergence of Euler-type methods for nonlinear stochastic differential equations.”SIAM Journal on Numerical Analysis 40, no. 3 (2002): 1041- 1063.
[7] Lambert, John Denholm. “Numerical methods for ordinary differential systems: the initial value problem.”John Wiley & Sons, Inc.New York, NY, USA, 1991.
[8] Li, Changpin, and Fanhai Zeng. “Numerical methods for fractional calculus.”Vol. 24. CRC Press, 2015.
[9] Lubich, Ch. “A stability analysis of convolution quadraturea for Abel- Volterra integral equations.”IMA journal of numerical analysis 6, no. 1 (1986): 87-101. Oxford University Press
[10] Flynn, Michael J. “Some computer organizations and their effectiveness.” IEEE transactions on computers 100, no. 9 : 948-960, IEEE Computer Society Washington, DC, USA, 1972.
[11] Baida Zhang, Shuai Xu, Feng Zhang, Yuan Bi and Linqi Huang, “Accelerating MatLab code using GPU: A review of tools and strategies”, Artificial Intelligence, Management Science and Electronic Commerce (AIMSEC), 2nd International Conference on, pp. 1875-1878, IEEE, 2011
[12] Reese, Jill, and Sarah Zaranek. “Gpu programming in matlab.” MathWorks News & Notes. Natick, MA: The MathWorks Inc (2012): 22-5.
[13] Altman, Yair M.“Accelerating MATLAB Performance: 1001 tips to speed up MATLAB programs.”CRC Press, 2014.
[14] Lippert, Anthony. “Nvidia gpu architecture for general purpose computing. ”NVIDIA presentation, April (2009).
[15] Luebke, David, and M. Harris. “General-purpose computation on graphics hardware.”In Workshop, SIGGRAPH. 2004.