On 12/07/2017 11:55 PM, Phani Motamarri wrote:
This works fine in serial but when running in parallel on 8 procs, the above
function gives a seg fault at line 854 of dof_tools.cc
// dof numbers on the centers of the lines bounding this
// face
for (unsigned int line=0; line<4; ++line)
for (unsigned int dof=0; dof!=fe.dofs_per_vertex; ++dof)
*selected_dofs[face->line(line)->child(0)->vertex_dof_index(1,dof)] = true;*
The line where the code seg faults has been marked bold.
My hanging node mesh generation process starts with a base uniform mesh
containing 64 elements in a cube of domain 40 units and I kept refining just
the layer of elements around the center of cube till the mesh size becomes
less than 0.05. Final mesh contains around 512 elements.
This may be a bug, but it's hard to tell without a testcase that allows
someone to look into it. Can you come up with a small program that
demonstrates the issue? It doesn't have to do anything useful -- just build a
DoFHandler in parallel and call this function.
Best
W.
--
------------------------------------------------------------------------
Wolfgang Bangerth email: bange...@colostate.edu
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 dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
