Unable to allocate memory for merge block during coprocessing

ParaView Support
1

I am getting a std::bad_alloc when using the merge block feature in a Catalyst script.

ERROR: In ParaView/VTK/Common/Core/vtkGenericDataArray.txx, line 452
vtkIdTypeArray (0x2aab39a989f0): Unable to allocate 154940 elements of size 8 bytes.

Without the merge block operation (keeping everything downstream the same) it works fine, but I would like to reduce my file output (hence the merge).

Is there something I can do as a use to increase memory (the machine is no where near limited by RAM for this case) or is this perhaps an issue with mergeBlock, etc.??

Thanks for the help!

3

This actually is occurring regularly with my implementation, not just with merge blocks. I am linking with a fortran code. I do not think I am creating excessive data. Below is my adapter file that updates the data structures. I am using intel 17.0 compilers on RHEL 6 cluster.

extern "C" void updatefield_(int* nblki,
                         double* data,
                         char** name)
{
vtkCPInputDataDescription* idd = vtkCPPythonAdaptorAPI::GetCoProcessorData()- 
>GetInputDescriptionByName("input");
vtkMultiBlockDataSet* mbds = vtkMultiBlockDataSet::SafeDownCast(idd->GetGrid());
vtkMultiPieceDataSet* mpds = vtkMultiPieceDataSet::SafeDownCast(mbds->GetBlock(0));
vtkStructuredGrid* grid = vtkStructuredGrid::SafeDownCast(mpds->GetPiece((*nblki-1)));
int ncls = grid->GetNumberOfCells();
vtkDoubleArray* field = vtkDoubleArray::SafeDownCast(grid->GetCellData()->GetArray(*name));
//Check if field alrady exists
if (field == nullptr)
{
  vtkNew<vtkDoubleArray> newfield;
  newfield->SetNumberOfComponents(1);
  newfield->SetName(*name);
  newfield->SetNumberOfTuples(grid->GetNumberOfCells());
  for(int i=0; i<ncls; i++)
    {
      newfield->SetValue(i,*(data+i));
    }
  grid->GetCellData()->AddArray(newfield);
  }
//Update field
else
{
  for(int i=0; i<ncls; i++)
    {
      field->SetValue(i,*(data+i));
    }
}
}
4

See this post for solution