An implementation of vortex methods for modeling 2D incompressible flows using the CUDA technology

Исследована возможность ускорения вычислений в методе вихревых элементов - бессеточном лагранжевом методе вычислительной гидродинамики - за счет использования графических ускорителей. Реализован алгоритм, основанный на авторской модификации метода вихревых элементов, позволяющий проводить все необходимые вычисления непосредственно на графическом ускорителе с использованием технологии CUDA. Скорость решения типичной задачи на одном ускорителе GeForce GTX 970 или Tesla C2050 оказывается соизмеримой со скоростью решения аналогичной задачи на кластере, содержащем 30-40 ядер, при использовании технологии MPI. Полученные результаты говорят о высокой эффективности применения графических ускорителей при решении задач гидродинамики вихревыми методами. The possibility of computation speedup in the vortex element method (a meshfree Lagrangian method of computational fluid dynamics) using the graphics accelerators is studied. An algorithm based on the authors' modification of the vortex element method is implemented; this algorithm allows one to perform all the necessary computations directly on a GPU using the CUDA technology. The speed of solving a typical problem on a single GeForce GTX 970 or Tesla C2050 accelerator is comparable with the speed of solving a similar problem on a cluster containing 30-40 cores with the use of the MPI technology. The numerical results obtained confirm a high efficiency of using graphics accelerators when solving the problems of hydrodynamics with vortex methods.

Interpretive MPI for Parallel Computing

Message Passing Interface (MPI) is a standardized library specification designed for message-passing parallel programming on large-scale distributed systems. A number of MPI libraries have been implemented to allow users to develop portable programs using the scientific programming languages, Fortran, C and C++. Ch is an embeddable C/C++ interpreter that provides an interpretive environment for C/C++ based scripts and programs. Combining Ch with any MPI C/C++ library provides the functionality for rapid development of MPI C/C++ programs without compilation. In this article, the method of interfacing Ch scripts with MPI C implementations is introduced by using the MPICH2 C library as an example. The MPICH2-based Ch MPI package provides users with the ability to interpretively run MPI C program based on the MPICH2 C library. Running MPI programs through the MPICH2-based Ch MPI package across heterogeneous platforms consisting of Linux and Windows machines is illustrated. Comparisons for the bandwidth, latency, and parallel computation speedup between C MPI, Ch MPI, and MPI for Python in an Ethernet-based environment comprising identical Linux machines are presented. A Web-based example is given to demonstrate the use of Ch and MPICH2 in C based CGI scripting to facilitate the development of Web-based applications for parallel computing.