MPI Communication Benchmarks Results for MPI_Gather()
2 total tasks: 2 tasks per node (Linux), 16 tasks per node (IBM)
4 total tasks: 2 tasks per node (Linux), 16 tasks per node (IBM)
8 total tasks: 2 tasks per node (Linux), 16 tasks per node (IBM)
16 total tasks: 2 tasks per node (Linux), 16 tasks per node (IBM)
32 total tasks: 2 tasks per node (Linux), 16 tasks per node (IBM)
64 total tasks: 2 tasks per node (Linux), 16 tasks per node (IBM)
128 total tasks: 2 tasks per node (Linux), 16 tasks per node (IBM)
256 total tasks: 2 tasks per node (Linux), 16 tasks per node (IBM)
