I'm pretty sure the intent was that MPI_Recv should NOT return MPI_ERR_PENDING as there is no request on which the error can be pending, but I don't know if much thought was given to allowing MPI_Recv to ignore acknowledge ranks.
Dave



From:        Wesley Bland <wbland@icl.utk.edu>
To:        "MPI 3.0 Fault Tolerance and Dynamic Process Control working Group" <mpi3-ft@lists.mpi-forum.org>,
Date:        03/15/2013 12:45 PM
Subject:        Re: [Mpi3-ft] MPI_Recv + MPI_Comm_failure_ack
Sent by:        mpi3-ft-bounces@lists.mpi-forum.org

---

I think you are correct in your evaluation, though I also think that wasn't our intent. I think the intent (unless I'm forgetting a discussion) was to allow MPI_ERR_PENDING to be returned by MPI_RECV and let MPI_COMM_FAILURE_ACK cover both cases. Can anyone else confirm that this was the goal.

If that's the case, it's something we'll need to fix in the text.

Thanks,
Wesley


On Fri, Mar 15, 2013 at 12:32 PM, David Solt <dsolt@us.ibm.com> wrote:
Based on the proposal:

MPI_Comm_failure_ack(blah, blah)

This local operation gives the users a way to acknowledge all locally noticed failures on
comm. After the call, unmatched MPI_ANY_SOURCE receptions that would have raised an
error code MPI_ERR_PENDING due to process failure (see Section 17.2.2) proceed without
further reporting of errors due to those acknowledged failures.

I think this clearly indicates that MPI_Recv is uninfluenced by calls to MPI_Comm_failure_ack.  Therefore, there is no way to call MPI_Recv(MPI_ANY_SOURCE) and ignore failures reported by MPI_Comm_failure_ack.  

I believe the following code will NOT work (i.e. after the first failure, the MPI_Recv will continuously fail):


MPI_Comm_size(intercomm, &size);
while (failures < size) {
        err = MPI_Recv(blah, blah, MPI_ANY_SOURCE, intercomm, &status);
        if (err == MPI_PROC_FAILED) {
                MPI_Comm_failure_ack(intercomm);
                MPI_Comm_failure_get_acked(intercomm, &group);
                MPI_Group_size(group, &failures);
        } else {
                /* process received data */
        }
}

and has to be written as:

MPI_Comm_size(intercomm, &size);
while (failures < size) {

        if (request == MPI_REQUEST_NULL) {
                err = MPI_Irecv(blah, blah, MPI_ANY_SOURCE, intercomm, &request);
        }
        err = MPI_Wait(&request, &status);

        if (err == MPI_ERR_PENDING) {
                MPI_Comm_failure_ack(intercomm);
                MPI_Comm_failure_get_acked(intercomm, &group);
                MPI_Group_size(group, &failures);
        } else {
                /* process received data */
        }
}

Am I correct in my thinking?
If so, was there a reason why MPI_Recv could not also "obey" MPI_Comm_failure_ack calls?

Thanks,
Dave
_______________________________________________
mpi3-ft mailing list
mpi3-ft@lists.mpi-forum.org
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft
_______________________________________________
mpi3-ft mailing list
mpi3-ft@lists.mpi-forum.org
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft

_______________________________________________
mpi3-ft mailing list
mpi3-ft@lists.mpi-forum.org
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft