Custom Query (116 matches)

At node/domain.cpp l693, a test read as

      if (ni.getValue() < 0 || ibegin.getValue() < 0 )
	// || (ibegin.getValue() + ni.getValue()) > ni_glo.getValue())
      
        ERROR("CDomain::checkLocalIDomain(void)",
              << "[ id = " << this->getId() << " , context = '" << CObjectFactory::GetCurrentContextId() << " ] "
              << "The local domain is wrongly defined,"
              << " check the attributes 'ni_glo' (" << ni_glo.getValue() << "), 'ni' (" << ni.getValue() << ") and 'ibegin' (" << ibegin.getValue() << ")");

The third condition does not make sense in the case of data_dim=1 and grid types curvilinear and regular , as each task may handle up to ni_glo*nj_glo grid cells.

It reads on line 792 as

      if (data_ni.isEmpty())
      {
        data_ni.setValue((data_dim == 1) ? (ni.getValue() * nj.getValue()) : ni.getValue());
      }

which is obviously wrong (set values should be reversed). This actually causes segmentation violation when nj is not set and data_dim==1; and this should lead to underestimated allocation when data_dim==2

We are experiencing a repetitive issue with XIOS 1.0 . It appeared using NEMO 3.6 stable and more than 2600 cores, and it seemed to be solved when using Intel 16 compiler and IMPI 5. However, after updating to NEMO 3.6 current stable, the problem appears when using 1920 or more cores. I don't really get how the NEMO revision change could affect to this, but there it is.

The problem is just in this line of client.cpp:

MPI_Send(buff,buffer.count(),MPI_CHAR,serverLeader,1,CXios::globalComm) ;

In the meanwhile the server.cpp is doing MPI_Iprobe continuosly in order to receive all the MPI_Send.

What we have observed is that using a high number of cores, around 80-100 of these cores get stucked at the MPI_Send, causing the run to hang and not complete. The fact that with a certain number of cores the issue appears 80% of the times but not always, made us think that could be related with the IMPI implementation.