Tensor product refinement#3
Conversation
alexfikl
force-pushed
the
tensor-product-refinement
branch
from
5211cd5 to
25104db
Compare
inducer
left a comment
inducer
left a comment
There was a problem hiding this comment.
Thanks for working on this. A few comments below. I think I'd like to merge this after inducer#70, just to keep that from becoming even larger than it already is.
| # {{{ shape-dependent tesselation helpers | ||
|
|
||
| @singledispatch | ||
| def get_child_basis_vertex_indices(shape: mp.Shape, child): |
There was a problem hiding this comment.
get_child_basis_vertex_indices doesn't make a great deal of sense to me as a name. What is it supposed to do? (Document, maybe rename?)
There was a problem hiding this comment.
Removed this and its horrible name completely as it was just used in map_unit_nodes_to_children.
| assert len(child) == shape.nvertices | ||
| return child[:shape.dim + 1] |
There was a problem hiding this comment.
Does simplex-as-default make sense? (If I'm understanding this right?)
There was a problem hiding this comment.
Removed this as it was only used in the map_unit_nodes_to_children function.
| # | ||
| # * lines and squares don't require any reordering | ||
|
|
||
| return [child[i] for i in [0, 1, 2, 4][:shape.dim + 1]] |
There was a problem hiding this comment.
This sounds like it's limited to up to 3D. Add handling for larger-D cases.
There was a problem hiding this comment.
Should be slightly more general after c7d09c9.
| from meshmode.mesh import _ModepyElementGroup | ||
| if not isinstance(group, _ModepyElementGroup): | ||
| raise TypeError(f"groups of type '{type(group.mesh_el_group).__name__}'" | ||
| " are not supported.") |
| from meshmode.mesh import _ModepyElementGroup | ||
| if not isinstance(group, _ModepyElementGroup): | ||
| raise TypeError(f"groups of type '{type(group.mesh_el_group).__name__}'" | ||
| " are not supported.") |
| if isinstance(meg, _ModepyElementGroup): | ||
| from meshmode.mesh.refinement.tesselate import map_unit_nodes_to_children | ||
| mapped_unit_nodes = map_unit_nodes_to_children( | ||
| meg._modepy_shape, fine_unit_nodes, record.tesselation) | ||
| else: |
I'm definitely fine with merging that first! |
|
@inducer Hm.. firedrake seems to be looking for an older EDIT: Guess it's incoming with OP2/PyOP2#605 |
|
Yikes. Thanks for telling me about that. I had no idea they were using that. I guess bringing it back now won't help either. |
Tensor Product Additions
Co-authored-by: Andreas Klöckner <inform@tiker.net>
alexfikl
force-pushed
the
tensor-product-refinement
branch
from
4537029 to
cfb3f70
Compare
Increase terminate timeout in NodalDGContext See merge request inducer/meshmode!106
Slightly loosen test_boundary_interpolation tolerance See merge request inducer/meshmode!107
Co-authored-by: Andreas Klöckner <inform@tiker.net>
Co-authored-by: Andreas Klöckner <inform@tiker.net>
Co-authored-by: Andreas Klöckner <inform@tiker.net>
Co-authored-by: Andreas Klöckner <inform@tiker.net>
Co-authored-by: Andreas Klöckner <inform@tiker.net>
alexfikl
force-pushed
the
tensor-product-refinement
branch
from
df892c4 to
da75962
Compare
|
Continued in inducer#97. |
2 participants
See inducer#70