Search

versions. The first uses a single CPU thread. The second employs OpenACC to use the system GPU. The third divides the computational domain into a 2D grid processed by separate cores, with MPI handling