Add procs_io, list of processors belonging to different compute nodes
This is useful to create (write) files on the different MPI compute nodes. For example for 10 cores allocated like: 3 on node1, 5 on node2 and 2 on node 3, procs_io is (0, 3, 8). See for example modifications in NEB/src/path_gen_inputs.f90. Files are written on each MPI node.