The paper deals with the application of parallel computation methods to the numerical solution of the nonlinear boundary value problem for the degenerate two-dimensional differential heat conduction equation. The nonlinearity of the problem stems from the power dependence of the thermal conductivity coefficient on temperature. The solution algorithm is based on the boundary element method with the application of the dual reciprocity method enabling all the computations to be brought to the boundary of the problem solution domain. A program has been developed from the presented computational algorithm. To accelerate the computation as much as possible, we use parallel programming processes and graphics processors. The program is written in the С++ programming language with the use of the OpenMP and OpenCL open standards. An example is considered to illustrate the work of the algorithm and the program; the calculation accuracy and the calculation speed are analyzed.
Keywords: parallel computation, OpenMP, OpenCL, nonlinear heat conduction problem, boundary element method, analytical integration, radial basis functions.DOI: 10.17804/2410-9908.2016.6.080-091
References
1.Aguilar-Leal O., Fuentes-Aguilar R.Q., Chairez I., Garcia-Gonzalez A., Huegel J.C. Distributed parameter system identification using finite element differential neural networks. Applied Soft Computing, 2016, vol. 43, pp. 633-642. DOI: 10.1016/j.asoc.2016 Petaccia G., Leporati F., Torti E. OpenMP and CUDA simulations of Sella Zerbino Dam break on unstructured grids. Computational Geosciences, 2016, vol. 20, no. 10, pp. 1123-1132. DOI: 10.1007. Sundarapandian M., Kalpathi R., Siochi R.A.C., Kadam A.S. Lung diaphragm tracking in CBCT images using spatio-temporal MRF.