> Is that wrong?
No, this looks alright.
What I would do is print, right before you call the solver, the sizes of all
blocks of rVec as well as the locally available elements. I.e., do something
like this:
std::cout << get_this_mpi_process(mpi_communicator) << " "
<< rVec.size() << " "
<< rVec.n_blocks() << " "
<< rVec.block(0).size() << " "
<< rVec.block(1).size() << " "
<< rVec.block(0).local_range().first() << " "
<< rVec.block(0).local_range().second() << " "
<< rVec.block(1).local_range().first() << " "
<< rVec.block(1).local_range().second() << " "
<< std::endl
If one of these numbers looks suspicious, you'll need to investigate what is
going on.
Best
W.
--
-------------------------------------------------------------------------
Wolfgang Bangerth email: [EMAIL PROTECTED]
www: http://www.math.tamu.edu/~bangerth/
_______________________________________________
