MFEM
v4.5.2
Finite element discretization library
|
Go to the source code of this file.
Namespaces | |
mfem | |
Functions | |
template<typename real_t > | |
void | mfem::CalcShapeMatrix (const FiniteElement &fe, const IntegrationRule &ir, real_t *B, const Array< int > *dof_map=NULL) |
Store mass-like matrix B for each integration point on the reference element. For tensor product evaluation, this is only called on the 1D reference element, and higher dimensions are put together from that. The element mass matrix can be written \( M_E = B^T D_E B \) where the B built here is the B, and is unchanging across the mesh. The diagonal matrix \( D_E \) then contains all the element-specific geometry and physics data. More... | |
template<typename real_t > | |
void | mfem::CalcGradTensor (const FiniteElement &fe, const IntegrationRule &ir, real_t *G, const Array< int > *dof_map=NULL) |
store gradient matrix G for each integration point on the reference element. For tensor product evaluation, this is only called on the 1D reference element, and higher dimensions are put together from that. The element stiffness matrix can be written \[ S_E = \sum_{k=1}^{nq} G_{k,i}^T (D_E^G)_{k,k} G_{k,j} \] where \( nq \) is the number of quadrature points, \( D_E^G \) contains all the information about the element geometry and coefficients (Jacobians etc.), and \( G \) is the matrix built in this routine, which is the same for all elements in a mesh. More... | |
template<typename real_t > | |
void | mfem::CalcShapes (const FiniteElement &fe, const IntegrationRule &ir, real_t *B, real_t *G, const Array< int > *dof_map) |