[mpich-discuss] Is this possible for mpich2?

Pavan Balaji balaji at mcs.anl.gov
Wed Jan 12 08:19:20 CST 2011


[please keep mpich-discuss cc'ed].

MPI_Isend, MPI_Irecv, MPI_Iprobe.

  -- Pavan

On 01/12/2011 08:18 AM, Eugene N wrote:
> Thank you, dear Pavan
>
> And if i need both the clients and server to continue their computations,
> and not to wait, what are the non-blocking alternatives?
>
> Eugene
>
> 2011/1/12 Pavan Balaji <balaji at mcs.anl.gov <mailto:balaji at mcs.anl.gov>>
>
>
>     You should be able to do this with a combination of MPI_Probe and
>     MPI_Recv on the server side, and simple MPI_Send on the client side.
>
>     while (true) {
>             MPI_Probe(...);
>             /* Check for size of incoming data and malloc */
>             MPI_Recv();
>     }
>
>       -- Pavan
>
>
>     On 01/12/2011 06:58 AM, Eugene N wrote:
>
>         Greetings,
>
>         I was wondering, whether mpich2 can be used in the following
>         scenario:
>
>         Client-Server application, where N clients send data to server, at
>         random moments
>         in time, and server must collect them all, not caring much from
>         whence
>         the data came.
>
>         sort of like this:
>
>         SERVER: while(true) { recieve something from anybody}
>
>         CLIENT1: if (i feel like it) {send something to server }
>         ...
>         CLIENTN: if (i feel like it) {send something to server }
>
>
>         The problem is this - i don't really know what sort of
>         interaction  this
>         is in professional terms (is it synchronous, asynchronous, another
>         kind?) and therefore what function i must look up in my
>         reference manual
>         pages.
>
>         Also, there is another snag, the size of the data, being sent, is
>         unknown (a random sized block of data)?
>
>         I'd love to hear some opinions and advice on the matter
>
>
>         Thanks in advance
>
>         Eugene
>
>
>
>
>
>
>
>
>
>         _______________________________________________
>         mpich-discuss mailing list
>         mpich-discuss at mcs.anl.gov <mailto:mpich-discuss at mcs.anl.gov>
>         https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>
>
>     --
>     Pavan Balaji
>     http://www.mcs.anl.gov/~balaji <http://www.mcs.anl.gov/%7Ebalaji>
>
>

-- 
Pavan Balaji
http://www.mcs.anl.gov/~balaji


More information about the mpich-discuss mailing list