> In ALL the cases, with different imput values, and different buffer
> sizes (big enough), the performance of both approacges is EXACTLY the
> same. In other words, I couldn´t speedup my application even though
> there is enough calculous time to overlap with communication.
I just want to state that I have also found that MPI_Send/MPI_Recv and
MPI_Isend/MPI_Irecv have very similar performance when used in a 4th
Order Finite Difference code. However, it is a little easier to code
the Isend/Irecv communication than the Send/Recv communication because
you don't have to worry about deadlocks....The communication pattern on
the other hand makes a significant difference.
That's my 2 cents.
-J
|
