Description
The MakeSimplicial
method does not respect higher order meshes, this is particularly apparent when using gpu:
Git changeset ID: v0.13.0-485-ge44aa8a1-dirty
Running with 4 MPI processes
Detected 4 CUDA devices
Device configuration: cuda,cpu
Memory configuration: host-std,cuda
libCEED backend: /gpu/cuda/magma
Added 3171 elements in 2 iterations of local bisection for under-resolved interior boundaries
Added 698 duplicate vertices for interior boundaries in the mesh
Added 2094 duplicate boundary elements for interior boundaries in the mesh
Finished partitioning mesh into 4 subdomains
Characteristic length and time scales:
L₀ = 2.732e-03 m, t₀ = 9.113e-03 ns
Mesh curvature order: 2
Mesh bounding box:
(Xmin, Ymin, Zmin) = (-1.366e-03, -1.366e-03, -1.000e-03) m
(Xmax, Ymax, Zmax) = (+1.366e-03, +1.366e-03, +1.000e-03) m
Parallel Mesh Stats:
minimum average maximum total
vertices 2394 2594 2768 10376
edges 16569 17150 17640 68602
faces 28028 28404 28766 113619
elements 13802 13848 13934 55392
neighbors 3 3 3
minimum maximum
h 0 0.21655
kappa -nan -nan
This will affect all mesh cracking capability too for higher order meshes