The nonblocking calls have the same syntax as the blocking ones, with two exceptions:

1. The letter "I" (think of "initiate") appears in the name of the call, immediately following the first underscore: e.g., MPI_Irecv.
2. The final argument is a handle to an opaque (or hidden) request object that holds detailed information about the transaction. The request handle can be used for subsequent Wait and Test calls.

For example, a standard nonblocking send and its corresponding wait look like this:

C:

int MPI_Isend (const void *buf, int count, \
    MPI_Datatype datatype, int dest, int tag, \
    MPI_Comm comm, MPI_Request *request)

int MPI_Wait (MPI_Request *request, MPI_Status *status)

Fortran:

MPI_ISEND(buf, count, dtype, dest, tag, comm, request, ierror)

MPI_WAIT(request, status, ierror)

MPI_IRecv looks much the same:

int MPI_Irecv (void* buf, int count, MPI_Datatype datatype, \
    int source, int tag, MPI_Comm comm, MPI_Request *request)

MPI_IRECV(buf, count, dtype, source, tag, comm, request, &
    ierror)

The communication mode is still relevant for nonblocking calls and is still selected via the name of the send routine, though the precise behind-the-scenes behavior can differ to some degree from the equivalent blocking call. In any case, you can choose from among a variety of nonblocking send functions, including Synchronous (MPI_Issend), Ready (MPI_Irsend), and Buffered (MPI_Ibsend), as well as the standard send (MPI_Isend).