On 2021-10-13 02:35, Drew Parsons wrote:
..
Debugging a bit further (with MPI_IN_PLACE removed), I can identify
that the bug is in dolfinx not openmpi (unless there are two bugs
here).
Comparing detailed debug output from 2 threads, I find one thread
skips the facet loop in compute_nonlocal_dual_graph() in dolfinx'
mesh/graphbuild.cpp, while the other thread crashes at
buffer[pos[dest] + max_num_vertices_per_facet] += cell_offset;
because pos[dest] is 0, but max_num_vertices_per_facet=-1.
...
I'll test that again restoring MPI_IN_PLACE to confirm dolfinx
max_num_vertices_per_facet=-1 is the true problem here.
Yeah, with MPI_IN_PLACE in place, compute_nonlocal_dual_graph() still
gets pos[dest] is 0, but max_num_vertices_per_facet=-1, causing the
segfault in dolfinx.
Apologies for the distraction, I'll reassign this bug to dolfinx.
Drew
