[mpich-discuss] Error in MPI call.

[Zhou, Hui]

HI hritikesh semwal,

The error says a connection error: Connection reset by peer. Are you running 8 processes on a single node or 8 nodes?

--
Hui Zhou


From: hritikesh semwal via discuss <discuss at mpich.org>
Reply-To: "discuss at mpich.org" <discuss at mpich.org>
Date: Tuesday, April 28, 2020 at 8:45 PM
To: "discuss at mpich.org" <discuss at mpich.org>
Cc: hritikesh semwal <hritikesh.semwal at gmail.com>
Subject: [mpich-discuss] Error in MPI call.

Hello all,

Greetings of the day.

I am working on the development of a parallel CFD solver in which I am using distributed graph communicator with MPI neighbourhood collective operation MPI_Neighbor_alltoallw. While running on 2 and 4 processors, the code is working fine but when I tried it on 8 processors it is showing following error,

Fatal error in PMPI_Neighbor_alltoallw: Unknown error class, error stack:
PMPI_Neighbor_alltoallw(299)..............:  MPI_Neighbor_alltoallw(sendbuf=0x2fa8c20, sendcounts=0x1cd2300, sdispls=0x1cd2270, sendtypes=0x1cd2180, recvbuf=0x2fac1d0, recvcounts=0x7f167615d010, rdispls=0x7f1667ace010, recvtypes=0x7f1660786010, comm=comm=0x84000002)
PMPI_Neighbor_alltoallw(285)..............:
MPIR_Neighbor_alltoallw_impl(174).........:
MPIR_Neighbor_alltoallw_intra_auto(93)....:
MPIR_Neighbor_alltoallw_nb(33)............:
MPIR_Wait_impl(77)........................:
MPIDI_CH3i_Progress_wait(242).............: an error occurred while handling an event returned by MPIDI_CH3I_Sock_Wait()
MPIDI_CH3I_Progress_handle_sock_event(479):
MPIDI_CH3I_Socki_handle_read(4013)........: connection failure (set=0,sock=1,errno=104:Connection reset by peer)

Can anyone tell me what is the solution to this problem? This would be a great help.

Thank you.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20200429/39a0880c/attachment-0001.html>