[mpich-discuss] MPI_Send and recv

[Sufeng Niu]

Hi, Pavan,

Thanks a lot, that is most confusing part for me in MPI. I would like to
clear this type of problems.

So the blocking begins to send message to itself, but there is no receive
posted. Thus, the MPI block it. Am I right here?

Thus, I need a nonblocking receive to wait the message here. Can I use
blocking receive?

Hi Jeff,

The weblink is https://computing.llnl.gov/tutorials/mpi/#Derived_Data_Types
the code is second example. Thank you

Thanks a lot!
Sufeng


On Tue, Jul 2, 2013 at 4:48 PM, Pavan Balaji <balaji at mcs.anl.gov> wrote:

>
> This is an incorrect program.  You need to post an Irecv before sending to
> yourself.
>
>  -- Pavan
>
>
> On 07/02/2013 04:39 PM, Sufeng Niu wrote:
>
>> Hi,
>>
>> I try to familiar with mpi self defined type api, and I directly run a
>> very simple code from LLNL webpage:
>>
>> //----------------------------**------------------------------**
>> ------------------------------**-------
>>
>> #include"mpi.h"
>> #include <stdio.h>
>> #define SIZE 4
>>
>> int main(argc,argv)
>> int argc;
>> char *argv[];  {
>> int numtasks, rank, source=0, dest, tag=1, i;
>> float a[SIZE][SIZE] =
>>    {1.0, 2.0, 3.0, 4.0,
>>     5.0, 6.0, 7.0, 8.0,
>>     9.0, 10.0, 11.0, 12.0,
>>    13.0, 14.0, 15.0, 16.0};
>> float b[SIZE];
>>
>> MPI_Status stat;
>> MPI_Datatype columntype;
>>
>> MPI_Init(&argc,&argv);
>> MPI_Comm_rank(MPI_COMM_WORLD, &rank);
>> MPI_Comm_size(MPI_COMM_WORLD, &numtasks);
>>
>> MPI_Type_vector(SIZE, 1, SIZE, MPI_FLOAT, &columntype);
>> MPI_Type_commit(&columntype);
>>
>> if (numtasks == SIZE) {
>>    if (rank == 0) {
>>       for (i=0; i<numtasks; i++)
>>         MPI_Send(&a[0][i], 1, columntype, i, tag, MPI_COMM_WORLD);
>>          }
>>
>>    MPI_Recv(b, SIZE, MPI_FLOAT, source, tag, MPI_COMM_WORLD, &stat);
>>    printf("rank= %d  b= %3.1f %3.1f %3.1f %3.1f\n",
>>          rank,b[0],b[1],b[2],b[3]);
>>    }
>> else
>>    printf("Must specify %d processors. Terminating.\n",SIZE);
>>
>> MPI_Type_free(&columntype);
>> MPI_Finalize();
>> }
>>
>> //----------------------------**------------------------------**
>> ------------------------------**-------
>>
>> However, it doesn't work. the program just suspend there. I figure out
>> the problem is at line MPI_Send, and MPI_Recv. the MPI_Send just send it
>> to itself. but I don't know why it doesn't work. Is it caused by dead
>> lock? why?
>>
>> Thank you very much!
>>
>> --
>> Best Regards,
>> Sufeng Niu
>> ECASP lab, ECE department, Illinois Institute of Technology
>> Tel: 312-731-7219
>>
>>
>> ______________________________**_________________
>> discuss mailing list     discuss at mpich.org
>> To manage subscription options or unsubscribe:
>> https://lists.mpich.org/**mailman/listinfo/discuss<https://lists.mpich.org/mailman/listinfo/discuss>
>>
>>
> --
> Pavan Balaji
> http://www.mcs.anl.gov/~balaji
>



-- 
Best Regards,
Sufeng Niu
ECASP lab, ECE department, Illinois Institute of Technology
Tel: 312-731-7219
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20130702/d7ff34e9/attachment.html>