On Thu, 27 May 2010, Andrea Hawkins wrote:

> I'm wondering though, does the inflow_ID and ouflow_ID correspond to
> the same numbers as the edges of an element are numbered? i.e. 0 for
> the bottom, 1 for the right side, 2 for the top, and 3 for the left
> side?

The PeriodicBoundary ids are the same as the BoundaryInfo IDs.  If you
generated your domain with build_square, build_cube, etc. then I
believe our convention is to use the same ordering as for a single
element's sides.

> I've tried the code above and get the following error:
>
> Assertion `!_b' failed.
> [0] /h1/ahawkins/LIBRARIES/LIBMESH/libmesh-r3801/include/parallel/threads.h,
> line 64, compiled Apr 29 2010 at 12:08:34
> terminate called after throwing an instance of 'libMesh::LogicError'
>  what():  Error in libMesh internal logic
>
> where the traceback is:
> Stack frames: 10
> 0: print_trace(std::ostream&)
> 1: _ZN7Threads11BoolAcquireC9ERb
> 2: Threads::BoolAcquire::BoolAcquire(bool&)
> 3: void Threads::parallel_reduce<StoredRange<MeshBase::const_node_iterator,
> Node const*>, (anonymous
> namespace)::FindBBox>(StoredRange<MeshBase::const_node_iterator, Node
> const*> const&, (anonymous namespace)::FindBBox&)
> 4: MeshTools::bounding_box(MeshBase const&)
> 5: PointLocatorTree::init(Trees::BuildType)
> 6: _ZN16PointLocatorTreeC9ERK8MeshBasePK16PointLocatorBase
> 7: PointLocatorTree::PointLocatorTree(MeshBase const&, PointLocatorBase 
> const*)
> 8: PointLocatorBase::build(MeshEnums::PointLocatorType, MeshBase
> const&, PointLocatorBase const*)
> 9: MeshBase::point_locator() const

Damn; this is a bug whose correct fix I worked out but never found
time to implement or test.  Would you mind testing it for me?  The fix
should be in the SVN head (r3822) now.
---
Roy

------------------------------------------------------------------------------

_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to