Hi,

  I had the following setup:

 Rank 0 process on node 1 wants to send an array of particular size to Rank 1 process on same node.

1. What are the optimisations that can be done/invoked while running mpirun  to perform this memory to memory transfer efficiently?

2. Is there any performance gain  if 2 processes that are exchanging data arrays are kept on the same node rather than on different nodes connected by infiniband?

 Awaiting a reply,

-Chev