Massively Parallel Particle Simulations on Graphics Processing Units with CUDA

Technische Universit¨at Braunschweig, Institute of Dynamics and Vibrations, Schleinitzstrase 20, 38106 BraunschweigParticle methods are a powerful tool to model dynamic systems. Thereby, the system is discretized by a large number of parti-cles, which are interacting via local, predefined particle-particle interaction laws. The resulting computational effort includesneighborhood search, computation of interaction forces and state update via time integration. Particle methods are used in alot of different fields of applications like computer science, physics and engineering sciences.As the analyzed systems’ number of particles constantly grow, performance enhancement has become an important part ofpresent algorithm development. Besides the well-established approach of algorithm parallelization on multi-core CPUs orCPU clusters, modern graphics processing units (GPUs) present a different and trend-setting possibility for massive paral-lelization even on desktop computers. Among the top four supercomputers of the world, three are already using NVIDIAGPUs.In late 2006, NVIDIA introduced the first GPUs optimized for general purpose calculations. This was followed by the intro-duction of a new computing architecture differing from the standard graphics user-interface like OpenGL. This architectureis called Compute Unified Device Architecture (CUDA). It enables the user to program the GPU using standard C commandswith few additional runtime functions. The differences in architecture between CPU and GPU result in a completely differentalgorithm implementation.So, a performance evaluation of different types of particle systems implemented on a GPU using CUDA and on a standardCPU is presented.