Explore the performance of OpenMP and CUDA in sparse matrix-vector multiplication and their implications for hybrid computing environments.