Hi Stefano,

> I'm observing a very strange behaviour using
> PetscWrappers::MPI:sparse_matrix. Namely, if I run two times using nprocs =1
> I obtain (as expected) the same matrix_output (except for matrix names in
> matlab_output but it is right). But when using np > 1, matrix values are
> different: not only between nprocs = 1 and nprocs > 1 runs, but also between
> runs with same nprocs value greater than one: it seems it prints out the
> matrix before the PETSc assembly process has terminated (and not all cached
> values as been transferred into PETSc Mat) . Is it normal?

Do you print out the data after the call to compress()? Also note that
the matrix depends on the number of processes because the degrees of
freedom are ordered in a different way. The matrix should be the same
with the same number of processes of course.

Is this also happening in step-18 for you? Anything interesting
happening in debug mode or running with valgrind? Can you construct a
minimal example that shows this?

-- 
Timo Heister
http://www.math.tamu.edu/~heister/
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to