GAINARU ANA

Using hardware gather-scatter capabilities to optimize MPI all-to-all