This paper discusses a new parallelization approach of the dynamic relaxation method, which is programmed with the NVIDIA CUDA API and executed on the graphics cards (GPU) of a computer. The main advantage of a GPU card ...