Finally I get the following error while running step-40 with 24
processes(single node) or 48 processes(two nodes).
|
--------------------------------------------------------
Anerror occurred inline <250>of file
</mnt/lustre/serc/secbteja/bhanu3/dealii-master/source/lac/petsc_parallel_vector.cc>infunction
voiddealii::PETScWrappers::MPI::Vector::create_vector(unsignedint,unsignedint)
Theviolated condition was:
size()==n
Additionalinformation:
Dimension4294967295notequal to 0.
Stacktrace:
-----------
#0 /mnt/lustre/serc/secbteja/bhanu3/dealiiinst/lib/libdeal_II.g.so.9.1.0-pre:
dealii::PETScWrappers::MPI::Vector::create_vector(unsigned int, unsigned int)
#1 /mnt/lustre/serc/secbteja/bhanu3/dealiiinst/lib/libdeal_II.g.so.9.1.0-pre:
dealii::PETScWrappers::MPI::Vector::Vector()
#2 ./step-40: Step40::LaplaceProblem<2>::LaplaceProblem()
#3 ./step-40: main
--------------------------------------------------------
|
Incidentally 4294967295 = 2^32 - 1.
Please suggest.
Bhanu,
it's hard to tell whether this is specific to Cray, is a compiler bug, a bug
in deal.II, or anything else. Does it work with 1 processor? With 2
processors? With 4? Etc? The program does run on my machine with 24
processors, so the question is what exactly is different on your system.
By the way: The error happens in this function
```
void
Vector::create_vector(const size_type n, const size_type local_size)
{
(void)n;
Assert(local_size <= n, ExcIndexRange(local_size, 0, n));
ghosted = false;
const PetscErrorCode ierr =
VecCreateMPI(communicator, local_size, PETSC_DETERMINE, &vector);
AssertThrow(ierr == 0, ExcPETScError(ierr));
Assert(size() == n, ExcDimensionMismatch(size(), n)); // line 250
}
```
So the error suggests that after the function creates the vector, size() of
the vector returns (unsigned int)-1. It's not clear to me how this can happen.
If you run this in a debugger, can you check what the arguments 'n' and
'local_size' to this function are?
Best
W.
--
------------------------------------------------------------------------
Wolfgang Bangerth email: [email protected]
www: http://www.math.colostate.edu/~bangerth/
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
