Development of advanced computational fluid dynamics tools and their application to simulation of internal turbulent flows

¹ Faculty of Power Engineering, Baltic State Technical University 1, 1-ya Krasnoarmeyskaya Str., St. Petersburg 190005, Russia
² Faculty of Mathematics and Mechanics, St. Petersburg State University Universitetsky Prosp., Old Petergof, St. Petersburg 198504, Russia
³ Faculty of Science, Engineering and Computing, Kingston University Friars Av., Roehampton Vale, London SW15 3DW, United Kingdom

Abstract

Modern graphics processing units (GPU) provide architectures and new programming models that enable to harness their large processing power and to design computational fluid dynamics (CFD) simulations at both high performance and low cost. Possibilities of the use of GPUs for the simulation of internal fluid flows are discussed. The finite volume method is applied to solve three-dimensional (3D) unsteady compressible Euler and Navier-Stokes equations on unstructured meshes. Compute Inified Device Architecture (CUDA) technology is used for programming implementation of parallel computational algorithms. Solution of some fluid dynamics problems on GPUs is presented and approaches to optimization of the CFD code related to the use of different types of memory are discussed. Speedup of solution on GPUs with respect to the solution on central processor unit (CPU) is compared with the use of different meshes and different methods of distribution of input data into blocks. Performance measurements show that numerical schemes developed achieve 20 to 50 speedup on GPU hardware compared to CPU reference implementation. The results obtained provide promising perspective for designing a GPU-based software framework for applications in CFD.