We need to be able to see you entire send and receive calls, but at a first pass, it's probably that he sizes of your send and receive don't match. Send us a minimum working example (not your entire code). Wesley

On Nov 29, 2013, at 1:56 AM, Soheil Hooshdaran <[email protected]> wrote:

Hello. After executing MPI_Send() for the first time in the root process, my program blocks.

int the root process, I have: ... while(i++<iWorldSize) { cout << "executing send for the " << i-1 << "th time" << endl; MPI_Send(&A[i-1], n*m/iWorldSize, MPI_INT, j, itag, MPI_COMM_WORLD); cout << "executed send for the " << i << "th time" << endl; }//end while ...

and in the slave processes I have

... MPI_Status status;

int *buffer = new int[m*(n/iWorldSize)];

cout << "\n\n\n**********Executing receive*******" << endl; MPI_Recv(&buffer, m*(n/iWorld ...

where did I go wrong?

Thank you _______________________________________________ discuss mailing list [email protected] To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss