mfem::GridFunction Class Reference

Class for grid function - Vector with associated FE space. More...

#include <gridfunc.hpp>

Inheritance diagram for mfem::GridFunction:

Collaboration diagram for mfem::GridFunction:

Public Types
enum	AvgType { ARITHMETIC , HARMONIC }

Public Member Functions
	GridFunction ()
	GridFunction (const GridFunction &orig)
	Copy constructor. The internal true-dof vector t_vec is not copied.
	GridFunction (FiniteElementSpace *f)
	Construct a GridFunction associated with the FiniteElementSpace *f.
	GridFunction (FiniteElementSpace *f, real_t *data)
	Construct a GridFunction using previously allocated array data.
	GridFunction (FiniteElementSpace *f, Vector &base, int base_offset=0)
	Construct a GridFunction using previously allocated Vector base starting at the given offset, base_offset.
	GridFunction (Mesh *m, std::istream &input)
	Construct a GridFunction on the given Mesh, using the data from input.
	GridFunction (Mesh *m, GridFunction *gf_array[], int num_pieces)
GridFunction &	operator= (const GridFunction &rhs)
	Copy assignment. Only the data of the base class Vector is copied.
void	MakeOwner (FiniteElementCollection *fec_)
	Make the GridFunction the owner of fec_owned and fes.
FiniteElementCollection *	OwnFEC ()
int	VectorDim () const
	Shortcut for calling FiniteElementSpace::GetVectorDim() on the underlying fes.
int	CurlDim () const
	Shortcut for calling FiniteElementSpace::GetCurlDim() on the underlying fes.
const Vector &	GetTrueVector () const
	Read only access to the (optional) internal true-dof Vector.
Vector &	GetTrueVector ()
	Read and write access to the (optional) internal true-dof Vector.
void	GetTrueDofs (Vector &tv) const
	Extract the true-dofs from the GridFunction.
void	SetTrueVector ()
	Shortcut for calling GetTrueDofs() with GetTrueVector() as argument.
virtual void	SetFromTrueDofs (const Vector &tv)
	Set the GridFunction from the given true-dof vector.
void	SetFromTrueVector ()
	Shortcut for calling SetFromTrueDofs() with GetTrueVector() as argument.
void	GetNodalValues (int i, Array< real_t > &nval, int vdim=1) const
	Returns the values in the vertices of i'th element for dimension vdim.
void	GetLaplacians (int i, const IntegrationRule &ir, Vector &laps, int vdim=1) const
void	GetLaplacians (int i, const IntegrationRule &ir, Vector &laps, DenseMatrix &tr, int vdim=1) const
void	GetHessians (int i, const IntegrationRule &ir, DenseMatrix &hess, int vdim=1) const
void	GetHessians (int i, const IntegrationRule &ir, DenseMatrix &hess, DenseMatrix &tr, int vdim=1) const
void	GetValuesFrom (const GridFunction &orig_func)
void	GetBdrValuesFrom (const GridFunction &orig_func)
void	GetVectorFieldValues (int i, const IntegrationRule &ir, DenseMatrix &vals, DenseMatrix &tr, int comp=0) const
void	ReorderByNodes ()
	For a vector grid function, makes sure that the ordering is byNODES.
void	GetNodalValues (Vector &nval, int vdim=1) const
	Return the values as a vector on mesh vertices for dimension vdim.
void	GetVectorFieldNodalValues (Vector &val, int comp) const
void	ProjectVectorFieldOn (GridFunction &vec_field, int comp=0)
void	GetDerivative (int comp, int der_comp, GridFunction &der) const
	Compute a certain derivative of a function's component. Derivatives of the function are computed at the DOF locations of der, and averaged over overlapping DOFs. Thus this function projects the derivative to the FiniteElementSpace of der.
real_t	GetDivergence (ElementTransformation &tr) const
void	GetCurl (ElementTransformation &tr, Vector &curl) const
void	GetGradient (ElementTransformation &tr, Vector &grad) const
	Gradient of a scalar function at a quadrature point.
void	GetGradients (ElementTransformation &tr, const IntegrationRule &ir, DenseMatrix &grad) const
	Extension of GetGradient(...) for a collection of IntegrationPoints.
void	GetGradients (const int elem, const IntegrationRule &ir, DenseMatrix &grad) const
	Extension of GetGradient(...) for a collection of IntegrationPoints.
void	GetVectorGradient (ElementTransformation &tr, DenseMatrix &grad) const
	Compute the vector gradient with respect to the physical element variable.
void	GetVectorGradientHat (ElementTransformation &T, DenseMatrix &gh) const
	Compute the vector gradient with respect to the reference element variable.
void	GetElementAverages (GridFunction &avgs) const
virtual void	GetElementDofValues (int el, Vector &dof_vals) const
void	ImposeBounds (int i, const Vector &weights, const Vector &lo_, const Vector &hi_)
void	ImposeBounds (int i, const Vector &weights, real_t min_=0.0, real_t max_=infinity())
void	RestrictConforming ()
void	ProjectGridFunction (const GridFunction &src)
	Project the src GridFunction to this GridFunction, both of which must be on the same mesh.
virtual void	ProjectCoefficient (Coefficient &coeff)
	Project coeff Coefficient to this GridFunction. The projection computation depends on the choice of the FiniteElementSpace fes. Note that this is usually interpolation at the degrees of freedom in each element (not L2 projection). For NURBS spaces these degrees of freedom are not available and L2 projection is resorted to as fallback.
void	ProjectCoefficient (Coefficient &coeff, Array< int > &dofs, int vd=0)
	Project coeff Coefficient to this GridFunction, using one element for each degree of freedom in dofs and nodal interpolation on that element.
void	ProjectCoefficient (VectorCoefficient &vcoeff)
	Project vcoeff VectorCoefficient to this GridFunction. The projection computation depends on the choice of the FiniteElementSpace fes. Note that this is usually interpolation at the degrees of freedom in each element (not L2 projection). For NURBS spaces these degrees of freedom are not available and L2 projection is resorted to as fallback.
void	ProjectCoefficient (VectorCoefficient &vcoeff, Array< int > &dofs)
	Project vcoeff VectorCoefficient to this GridFunction, using one element for each degree of freedom in dofs and nodal interpolation on that element.
void	ProjectCoefficient (VectorCoefficient &vcoeff, int attribute)
	Project vcoeff VectorCoefficient to this GridFunction, only projecting onto elements with the given attribute.
void	ProjectCoefficient (Coefficient *coeff[])
	Analogous to the version with argument vcoeff VectorCoefficient but using an array of scalar coefficients for each component.
virtual void	ProjectDiscCoefficient (VectorCoefficient &coeff)
	Project a discontinuous vector coefficient as a grid function on a continuous finite element space. The values in shared dofs are determined from the element with maximal attribute.
virtual void	ProjectDiscCoefficient (Coefficient &coeff, AvgType type)
	Projects a discontinuous coefficient so that the values in shared vdofs are computed by taking an average of the possible values.
virtual void	ProjectDiscCoefficient (VectorCoefficient &coeff, AvgType type)
	Projects a discontinuous vector coefficient so that the values in shared vdofs are computed by taking an average of the possible values.
std::unique_ptr< GridFunction >	ProlongateToMaxOrder () const
	Return a GridFunction with the values of this, prolongated to the maximum order of all elements in the mesh.
virtual void	CountElementsPerVDof (Array< int > &elem_per_vdof) const
	For each vdof, counts how many elements contain the vdof, as containment is determined by FiniteElementSpace::GetElementVDofs().
void	ProjectBdrCoefficient (Coefficient &coeff, const Array< int > &attr)
	Project a Coefficient on the GridFunction, modifying only DOFs on the boundary associated with the boundary attributes marked in the attr array.
virtual void	ProjectBdrCoefficient (VectorCoefficient &vcoeff, const Array< int > &attr)
	Project a VectorCoefficient on the GridFunction, modifying only DOFs on the boundary associated with the boundary attributes marked in the attr array.
virtual void	ProjectBdrCoefficient (Coefficient *coeff[], const Array< int > &attr)
	Project a set of Coefficients on the components of the GridFunction, modifying only DOFs on the boundary associated with the boundary attributed marked in the attr array.
void	ProjectBdrCoefficientNormal (VectorCoefficient &vcoeff, const Array< int > &bdr_attr)
virtual void	ProjectBdrCoefficientTangent (VectorCoefficient &vcoeff, const Array< int > &bdr_attr)
	Project the tangential components of the given VectorCoefficient on the boundary. Only boundary attributes that are marked in bdr_attr are projected. Assumes ND-type VectorFE GridFunction.
virtual real_t	ComputeL2Error (Coefficient *exsol[], const IntegrationRule *irs[]=NULL, const Array< int > *elems=NULL) const
	Returns ||exsol - u_h||_L2 for scalar or vector H1 or L2 elements.
virtual real_t	ComputeElementGradError (int ielem, VectorCoefficient *exgrad, const IntegrationRule *irs[]=NULL) const
	Returns ||grad u_ex - grad u_h||_L2 in element ielem for H1 or L2 elements.
virtual real_t	ComputeL2Error (Coefficient &exsol, const IntegrationRule *irs[]=NULL, const Array< int > *elems=NULL) const
	Returns ||u_ex - u_h||_L2 for H1 or L2 elements.
virtual real_t	ComputeL2Error (VectorCoefficient &exsol, const IntegrationRule *irs[]=NULL, const Array< int > *elems=NULL) const
	Returns ||u_ex - u_h||_L2 for vector fields.
virtual real_t	ComputeGradError (VectorCoefficient *exgrad, const IntegrationRule *irs[]=NULL) const
	Returns ||grad u_ex - grad u_h||_L2 for H1 or L2 elements.
virtual real_t	ComputeCurlError (VectorCoefficient *excurl, const IntegrationRule *irs[]=NULL) const
	Returns ||curl u_ex - curl u_h||_L2 for ND elements.
virtual real_t	ComputeDivError (Coefficient *exdiv, const IntegrationRule *irs[]=NULL) const
	Returns ||div u_ex - div u_h||_L2 for RT elements.
virtual real_t	ComputeDGFaceJumpError (Coefficient *exsol, Coefficient *ell_coeff, class JumpScaling jump_scaling, const IntegrationRule *irs[]=NULL) const
	Returns the Face Jumps error for L2 elements.
MFEM_DEPRECATED real_t	ComputeDGFaceJumpError (Coefficient *exsol, Coefficient *ell_coeff, real_t Nu, const IntegrationRule *irs[]=NULL) const
	Returns the Face Jumps error for L2 elements, with 1/h scaling.
virtual real_t	ComputeH1Error (Coefficient *exsol, VectorCoefficient *exgrad, Coefficient *ell_coef, real_t Nu, int norm_type) const
virtual real_t	ComputeH1Error (Coefficient *exsol, VectorCoefficient *exgrad, const IntegrationRule *irs[]=NULL) const
	Returns the error measured in H1-norm for H1 or L2 elements.
virtual real_t	ComputeHDivError (VectorCoefficient *exsol, Coefficient *exdiv, const IntegrationRule *irs[]=NULL) const
	Returns the error measured in H(div)-norm for RT elements.
virtual real_t	ComputeHCurlError (VectorCoefficient *exsol, VectorCoefficient *excurl, const IntegrationRule *irs[]=NULL) const
	Returns the error measured in H(curl)-norm for ND elements.
virtual real_t	ComputeMaxError (Coefficient &exsol, const IntegrationRule *irs[]=NULL) const
	Returns Max|u_ex - u_h| error for H1 or L2 elements.
virtual real_t	ComputeMaxError (Coefficient *exsol[], const IntegrationRule *irs[]=NULL) const
	Returns Max|u_ex - u_h| error for scalar or vector fields.
virtual real_t	ComputeMaxError (VectorCoefficient &exsol, const IntegrationRule *irs[]=NULL) const
	Returns Max|u_ex - u_h| error for vector fields.
virtual real_t	ComputeL1Error (Coefficient &exsol, const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_L1 for H1 or L2 elements.
virtual MFEM_DEPRECATED real_t	ComputeL1Error (Coefficient *exsol[], const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_L1 for H1 or L2 elements.
virtual real_t	ComputeW11Error (Coefficient *exsol, VectorCoefficient *exgrad, int norm_type, const Array< int > *elems=NULL, const IntegrationRule *irs[]=NULL) const
	Returns \(W^1_1\) norm (or portions thereof) for H1 or L2 elements.
virtual real_t	ComputeL1Error (VectorCoefficient &exsol, const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_L1 for vector fields.
virtual real_t	ComputeLpError (const real_t p, Coefficient &exsol, Coefficient *weight=NULL, const IntegrationRule *irs[]=NULL, const Array< int > *elems=NULL) const
	Returns ||u_ex - u_h||_Lp for H1 or L2 elements.
virtual void	ComputeElementLpErrors (const real_t p, Coefficient &exsol, Vector &error, Coefficient *weight=NULL, const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_Lp elementwise for H1 or L2 elements.
virtual void	ComputeElementL1Errors (Coefficient &exsol, Vector &error, const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_L1 elementwise for H1 or L2 elements.
virtual void	ComputeElementL2Errors (Coefficient &exsol, Vector &error, const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_L2 elementwise for H1 or L2 elements.
virtual void	ComputeElementMaxErrors (Coefficient &exsol, Vector &error, const IntegrationRule *irs[]=NULL) const
	Returns Max|u_ex - u_h| elementwise for H1 or L2 elements.
virtual real_t	ComputeLpError (const real_t p, VectorCoefficient &exsol, Coefficient *weight=NULL, VectorCoefficient *v_weight=NULL, const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_Lp for vector fields.
virtual void	ComputeElementLpErrors (const real_t p, VectorCoefficient &exsol, Vector &error, Coefficient *weight=NULL, VectorCoefficient *v_weight=NULL, const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_Lp elementwise for vector fields.
virtual void	ComputeElementL1Errors (VectorCoefficient &exsol, Vector &error, const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_L1 elementwise for vector fields.
virtual void	ComputeElementL2Errors (VectorCoefficient &exsol, Vector &error, const IntegrationRule *irs[]=NULL) const
	Returns ||u_ex - u_h||_L2 elementwise for vector fields.
virtual void	ComputeElementMaxErrors (VectorCoefficient &exsol, Vector &error, const IntegrationRule *irs[]=NULL) const
	Returns Max|u_ex - u_h| elementwise for vector fields.
virtual void	ComputeFlux (BilinearFormIntegrator &blfi, GridFunction &flux, bool wcoef=true, int subdomain=-1)
GridFunction &	operator= (real_t value)
	Redefine '=' for GridFunction = constant.
GridFunction &	operator= (const Vector &v)
	Copy the data from v.
virtual void	Update ()
	Transform by the Space UpdateMatrix (e.g., on Mesh change).
long	GetSequence () const
FiniteElementSpace *	FESpace ()
const FiniteElementSpace *	FESpace () const
virtual void	SetSpace (FiniteElementSpace *f)
	Associate a new FiniteElementSpace with the GridFunction.
virtual void	MakeRef (FiniteElementSpace *f, real_t *v)
	Make the GridFunction reference external data on a new FiniteElementSpace.
virtual void	MakeRef (FiniteElementSpace *f, Vector &v, int v_offset)
	Make the GridFunction reference external data on a new FiniteElementSpace.
void	MakeTRef (FiniteElementSpace *f, real_t *tv)
	Associate a new FiniteElementSpace and new true-dof data with the GridFunction.
void	MakeTRef (FiniteElementSpace *f, Vector &tv, int tv_offset)
	Associate a new FiniteElementSpace and new true-dof data with the GridFunction.
virtual void	Save (std::ostream &out) const
	Save the GridFunction to an output stream.
virtual void	Save (const char *fname, int precision=16) const
virtual void	Save (adios2stream &out, const std::string &variable_name, const adios2stream::data_type type=adios2stream::data_type::point_data) const
	Save the GridFunction to a binary output stream using adios2 bp format.
void	SaveVTK (std::ostream &out, const std::string &field_name, int ref)
	Write the GridFunction in VTK format. Note that Mesh::PrintVTK must be called first. The parameter ref > 0 must match the one used in Mesh::PrintVTK.
void	SaveSTL (std::ostream &out, int TimesToRefine=1)
	Write the GridFunction in STL format. Note that the mesh dimension must be 2 and that quad elements will be broken into two triangles.
virtual	~GridFunction ()
	Destroys grid function.
void	MakeRef (Vector &base, int offset, int size)
	Reset the Vector to be a reference to a sub-vector of base.
void	MakeRef (Vector &base, int offset)
	Reset the Vector to be a reference to a sub-vector of base without changing its current size.
Element index Get Value Methods
These methods take an element index and return the interpolated value of the field at a given reference point within the element. Warning These methods retrieve and use the ElementTransformation object from the mfem::Mesh. This can alter the state of the element transformation object and can also lead to unexpected results when the ElementTransformation object is already in use such as when these methods are called from within an integration loop. Consider using GetValue(ElementTransformation &T, ...) instead.
virtual real_t	GetValue (int i, const IntegrationPoint &ip, int vdim=1) const
virtual void	GetVectorValue (int i, const IntegrationPoint &ip, Vector &val) const
Element Index Get Values Methods
These are convenience methods for repeatedly calling GetValue for multiple points within a given element. The GetValues methods are optimized and should perform better than repeatedly calling GetValue. The GetVectorValues method simply calls GetVectorValue repeatedly. Warning These methods retrieve and use the ElementTransformation object from the mfem::Mesh. This can alter the state of the element transformation object and can also lead to unexpected results when the ElementTransformation object is already in use such as when these methods are called from within an integration loop. Consider using GetValues(ElementTransformation &T, ...) instead.
void	GetValues (int i, const IntegrationRule &ir, Vector &vals, int vdim=1) const
void	GetValues (int i, const IntegrationRule &ir, Vector &vals, DenseMatrix &tr, int vdim=1) const
void	GetVectorValues (int i, const IntegrationRule &ir, DenseMatrix &vals, DenseMatrix &tr) const
ElementTransformation Get Value Methods
These member functions are designed for use within GridFunctionCoefficient objects. These can be used with ElementTransformation objects coming from either Mesh::GetElementTransformation() or Mesh::GetBdrElementTransformation(). Note These methods do not reset the ElementTransformation object so they should be safe to use within integration loops or other contexts where the ElementTransformation is already in use.
virtual real_t	GetValue (ElementTransformation &T, const IntegrationPoint &ip, int comp=0, Vector *tr=NULL) const
virtual void	GetVectorValue (ElementTransformation &T, const IntegrationPoint &ip, Vector &val, Vector *tr=NULL) const
ElementTransformation Get Values Methods
These are convenience methods for repeatedly calling GetValue for multiple points within a given element. They work by calling either the ElementTransformation or FaceElementTransformations versions described above. Consequently, these methods should not be expected to run faster than calling the above methods in an external loop. Note These methods do not reset the ElementTransformation object so they should be safe to use within integration loops or other contexts where the ElementTransformation is already in use. These methods can also be used with FaceElementTransformations objects.
void	GetValues (ElementTransformation &T, const IntegrationRule &ir, Vector &vals, int comp=0, DenseMatrix *tr=NULL) const
void	GetVectorValues (ElementTransformation &T, const IntegrationRule &ir, DenseMatrix &vals, DenseMatrix *tr=NULL) const
Face Index Get Values Methods
These methods are designed to work with Discontinuous Galerkin basis functions. They compute field values on the interface between elements, or on boundary elements, by interpolating the field in a neighboring element. The side argument indices which neighboring element should be used: 0, 1, or 2 (automatically chosen). Warning These methods retrieve and use the FaceElementTransformations object from the mfem::Mesh. This can alter the state of the face element transformations object and can also lead to unexpected results when the FaceElementTransformations object is already in use such as when these methods are called from within an integration loop. Consider using GetValues(ElementTransformation &T, ...) instead.
int	GetFaceValues (int i, int side, const IntegrationRule &ir, Vector &vals, DenseMatrix &tr, int vdim=1) const
int	GetFaceVectorValues (int i, int side, const IntegrationRule &ir, DenseMatrix &vals, DenseMatrix &tr) const
Public Member Functions inherited from mfem::Vector
	Vector ()
	Vector (const Vector &)
	Copy constructor. Allocates a new data array and copies the data.
	Vector (Vector &&v)
	Move constructor. "Steals" data from its argument.
	Vector (int s)
	Creates vector of size s.
	Vector (real_t *data_, int size_)
	Creates a vector referencing an array of doubles, owned by someone else.
	Vector (Vector &base, int base_offset, int size_)
	Create a Vector referencing a sub-vector of the Vector base starting at the given offset, base_offset, and size size_.
	Vector (int size_, MemoryType mt)
	Create a Vector of size size_ using MemoryType mt.
	Vector (int size_, MemoryType h_mt, MemoryType d_mt)
	Create a Vector of size size_ using host MemoryType h_mt and device MemoryType d_mt.
template<typename CT , int N>
	Vector (const CT(&values)[N])
	Create a vector from a statically sized C-style array of convertible type.
template<typename CT , typename std::enable_if< std::is_convertible< CT, real_t >::value, bool >::type = true>
	Vector (std::initializer_list< CT > values)
	Create a vector using a braced initializer list.
virtual void	UseDevice (bool use_dev) const
	Enable execution of Vector operations using the mfem::Device.
virtual bool	UseDevice () const
	Return the device flag of the Memory object used by the Vector.
void	Load (std::istream **in, int np, int *dim)
	Reads a vector from multiple files.
void	Load (std::istream &in, int Size)
	Load a vector from an input stream.
void	Load (std::istream &in)
	Load a vector from an input stream, reading the size from the stream.
void	SetSize (int s)
	Resize the vector to size s.
void	SetSize (int s, MemoryType mt)
	Resize the vector to size s using MemoryType mt.
void	SetSize (int s, const Vector &v)
	Resize the vector to size s using the MemoryType of v.
void	SetData (real_t *d)
void	SetDataAndSize (real_t *d, int s)
	Set the Vector data and size.
void	NewDataAndSize (real_t *d, int s)
	Set the Vector data and size, deleting the old data, if owned.
void	NewMemoryAndSize (const Memory< real_t > &mem, int s, bool own_mem)
	Reset the Vector to use the given external Memory mem and size s.
void	MakeRef (Vector &base, int offset, int size)
	Reset the Vector to be a reference to a sub-vector of base.
void	MakeRef (Vector &base, int offset)
	Reset the Vector to be a reference to a sub-vector of base without changing its current size.
void	MakeDataOwner () const
	Set the Vector data (host pointer) ownership flag.
void	Destroy ()
	Destroy a vector.
void	DeleteDevice (bool copy_to_host=true)
	Delete the device pointer, if owned. If copy_to_host is true and the data is valid only on device, move it to host before deleting. Invalidates the device memory.
int	Size () const
	Returns the size of the vector.
int	Capacity () const
	Return the size of the currently allocated data array.
real_t *	GetData () const
	Return a pointer to the beginning of the Vector data.
MFEM_DEPRECATED	operator real_t * ()
	Conversion to double *. Deprecated.
MFEM_DEPRECATED	operator const real_t * () const
	Conversion to const double *. Deprecated.
real_t *	begin ()
	STL-like begin.
real_t *	end ()
	STL-like end.
const real_t *	begin () const
	STL-like begin (const version).
const real_t *	end () const
	STL-like end (const version).
Memory< real_t > &	GetMemory ()
	Return a reference to the Memory object used by the Vector.
const Memory< real_t > &	GetMemory () const
	Return a reference to the Memory object used by the Vector, const version.
void	SyncMemory (const Vector &v) const
	Update the memory location of the vector to match v.
void	SyncAliasMemory (const Vector &v) const
	Update the alias memory location of the vector to match v.
bool	OwnsData () const
	Read the Vector data (host pointer) ownership flag.
void	StealData (real_t **p)
	Changes the ownership of the data; after the call the Vector is empty.
real_t *	StealData ()
	Changes the ownership of the data; after the call the Vector is empty.
real_t &	Elem (int i)
	Access Vector entries. Index i = 0 .. size-1.
const real_t &	Elem (int i) const
	Read only access to Vector entries. Index i = 0 .. size-1.
real_t &	operator() (int i)
	Access Vector entries using () for 0-based indexing.
const real_t &	operator() (int i) const
	Read only access to Vector entries using () for 0-based indexing.
real_t &	operator[] (int i)
	Access Vector entries using [] for 0-based indexing.
const real_t &	operator[] (int i) const
	Read only access to Vector entries using [] for 0-based indexing.
real_t	operator* (const real_t *v) const
real_t	operator* (const Vector &v) const
	Return the inner-product.
Vector &	operator= (const real_t *v)
	Copy Size() entries from v.
Vector &	operator= (const Vector &v)
	Copy assignment.
Vector &	operator= (Vector &&v)
	Move assignment.
Vector &	operator= (real_t value)
	Redefine '=' for vector = constant.
Vector &	operator*= (real_t c)
Vector &	operator*= (const Vector &v)
	Component-wise scaling: (*this)(i) *= v(i)
Vector &	operator/= (real_t c)
Vector &	operator/= (const Vector &v)
	Component-wise division: (*this)(i) /= v(i)
Vector &	operator-= (real_t c)
Vector &	operator-= (const Vector &v)
Vector &	operator+= (real_t c)
Vector &	operator+= (const Vector &v)
Vector &	Add (const real_t a, const Vector &Va)
	(*this) += a * Va
Vector &	Set (const real_t a, const Vector &x)
	(*this) = a * x
void	SetVector (const Vector &v, int offset)
void	AddSubVector (const Vector &v, int offset)
void	Neg ()
	(*this) = -(*this)
void	Reciprocal ()
	(*this)(i) = 1.0 / (*this)(i)
void	Swap (Vector &other)
	Swap the contents of two Vectors.
void	cross3D (const Vector &vin, Vector &vout) const
void	median (const Vector &lo, const Vector &hi)
	v = median(v,lo,hi) entrywise. Implementation assumes lo <= hi.
void	GetSubVector (const Array< int > &dofs, Vector &elemvect) const
	Extract entries listed in dofs to the output Vector elemvect.
void	GetSubVector (const Array< int > &dofs, real_t *elem_data) const
	Extract entries listed in dofs to the output array elem_data.
void	SetSubVector (const Array< int > &dofs, const real_t value)
	Set the entries listed in dofs to the given value.
void	SetSubVector (const Array< int > &dofs, const Vector &elemvect)
	Set the entries listed in dofs to the values given in the elemvect Vector. Negative dof values cause the -dof-1 position in this Vector to receive the -val from elemvect.
void	SetSubVector (const Array< int > &dofs, real_t *elem_data)
	Set the entries listed in dofs to the values given the , elem_data array. Negative dof values cause the -dof-1 position in this Vector to receive the -val from elem_data.
void	AddElementVector (const Array< int > &dofs, const Vector &elemvect)
	Add elements of the elemvect Vector to the entries listed in dofs. Negative dof values cause the -dof-1 position in this Vector to add the -val from elemvect.
void	AddElementVector (const Array< int > &dofs, real_t *elem_data)
	Add elements of the elem_data array to the entries listed in dofs. Negative dof values cause the -dof-1 position in this Vector to add the -val from elem_data.
void	AddElementVector (const Array< int > &dofs, const real_t a, const Vector &elemvect)
	Add times the elements of the elemvect Vector to the entries listed in dofs. Negative dof values cause the -dof-1 position in this Vector to add the -a*val from elemvect.
void	SetSubVectorComplement (const Array< int > &dofs, const real_t val)
	Set all vector entries NOT in the dofs Array to the given val.
void	Print (std::ostream &out=mfem::out, int width=8) const
	Prints vector to stream out.
void	Print (adios2stream &out, const std::string &variable_name) const
void	Print_HYPRE (std::ostream &out) const
	Prints vector to stream out in HYPRE_Vector format.
void	PrintMathematica (std::ostream &out=mfem::out) const
	Prints vector as a List for importing into Mathematica.
void	PrintHash (std::ostream &out) const
	Print the Vector size and hash of its data.
void	Randomize (int seed=0)
	Set random values in the vector.
real_t	Norml2 () const
	Returns the l2 norm of the vector.
real_t	Normlinf () const
	Returns the l_infinity norm of the vector.
real_t	Norml1 () const
	Returns the l_1 norm of the vector.
real_t	Normlp (real_t p) const
	Returns the l_p norm of the vector.
real_t	Max () const
	Returns the maximal element of the vector.
real_t	Min () const
	Returns the minimal element of the vector.
real_t	Sum () const
	Return the sum of the vector entries.
real_t	DistanceSquaredTo (const real_t *p) const
	Compute the square of the Euclidean distance to another vector.
real_t	DistanceSquaredTo (const Vector &p) const
	Compute the square of the Euclidean distance to another vector.
real_t	DistanceTo (const real_t *p) const
	Compute the Euclidean distance to another vector.
real_t	DistanceTo (const Vector &p) const
	Compute the Euclidean distance to another vector.
int	CheckFinite () const
	Count the number of entries in the Vector for which isfinite is false, i.e. the entry is a NaN or +/-Inf.
virtual	~Vector ()
	Destroys vector.
virtual const real_t *	Read (bool on_dev=true) const
	Shortcut for mfem::Read(vec.GetMemory(), vec.Size(), on_dev).
virtual const real_t *	HostRead () const
	Shortcut for mfem::Read(vec.GetMemory(), vec.Size(), false).
virtual real_t *	Write (bool on_dev=true)
	Shortcut for mfem::Write(vec.GetMemory(), vec.Size(), on_dev).
virtual real_t *	HostWrite ()
	Shortcut for mfem::Write(vec.GetMemory(), vec.Size(), false).
virtual real_t *	ReadWrite (bool on_dev=true)
	Shortcut for mfem::ReadWrite(vec.GetMemory(), vec.Size(), on_dev).
virtual real_t *	HostReadWrite ()
	Shortcut for mfem::ReadWrite(vec.GetMemory(), vec.Size(), false).

Protected Member Functions
void	SaveSTLTri (std::ostream &out, real_t p1[], real_t p2[], real_t p3[])
void	ProjectDeltaCoefficient (DeltaCoefficient &delta_coeff, real_t &integral)
void	SumFluxAndCount (BilinearFormIntegrator &blfi, GridFunction &flux, Array< int > &counts, bool wcoef, int subdomain)
void	ProjectDiscCoefficient (VectorCoefficient &coeff, Array< int > &dof_attr)
void	LegacyNCReorder ()
	Loading helper.
void	Destroy ()
void	AccumulateAndCountZones (Coefficient &coeff, AvgType type, Array< int > &zones_per_vdof)
	Accumulates (depending on type) the values of coeff at all shared vdofs and counts in how many zones each vdof appears.
void	AccumulateAndCountZones (VectorCoefficient &vcoeff, AvgType type, Array< int > &zones_per_vdof)
	Accumulates (depending on type) the values of vcoeff at all shared vdofs and counts in how many zones each vdof appears.
void	AccumulateAndCountDerivativeValues (int comp, int der_comp, GridFunction &der, Array< int > &zones_per_dof) const
	Used for the serial and parallel implementations of the GetDerivative() method; see its documentation.
void	AccumulateAndCountBdrValues (Coefficient *coeff[], VectorCoefficient *vcoeff, const Array< int > &attr, Array< int > &values_counter)
void	AccumulateAndCountBdrTangentValues (VectorCoefficient &vcoeff, const Array< int > &bdr_attr, Array< int > &values_counter)
void	ComputeMeans (AvgType type, Array< int > &zones_per_vdof)
void	UpdatePRef ()
	P-refinement version of Update().

Protected Attributes
FiniteElementSpace *	fes
	FE space on which the grid function lives. Owned if fec_owned is not NULL.
FiniteElementCollection *	fec_owned
	Used when the grid function is read from a file. It can also be set explicitly, see MakeOwner().
long	fes_sequence
Vector	t_vec
Protected Attributes inherited from mfem::Vector
Memory< real_t >	data
int	size

Detailed Description

Class for grid function - Vector with associated FE space.

Definition at line 30 of file gridfunc.hpp.

Member Enumeration Documentation

AvgType

enum mfem::GridFunction::AvgType

Enumerator
ARITHMETIC
HARMONIC

Definition at line 427 of file gridfunc.hpp.

Constructor & Destructor Documentation

GridFunction() [1/7]

mfem::GridFunction::GridFunction ( )

inline

Definition at line 75 of file gridfunc.hpp.

GridFunction() [2/7]

mfem::GridFunction::GridFunction ( const GridFunction & orig )

inline

Copy constructor. The internal true-dof vector t_vec is not copied.

Definition at line 78 of file gridfunc.hpp.

GridFunction() [3/7]

mfem::GridFunction::GridFunction ( FiniteElementSpace * f )

inline

Construct a GridFunction associated with the FiniteElementSpace *f.

Definition at line 83 of file gridfunc.hpp.

GridFunction() [4/7]

mfem::GridFunction::GridFunction ( FiniteElementSpace * f, real_t * data )

inline

Construct a GridFunction using previously allocated array data.

The GridFunction does not assume ownership of data which is assumed to be of size at least f->GetVSize(). Similar to the Vector constructor for externally allocated array, the pointer data can be NULL. The data array can be replaced later using the method SetData().

Definition at line 92 of file gridfunc.hpp.

GridFunction() [5/7]

mfem::GridFunction::GridFunction ( FiniteElementSpace * f, Vector & base, int base_offset = 0 )

inline

Construct a GridFunction using previously allocated Vector base starting at the given offset, base_offset.

Definition at line 98 of file gridfunc.hpp.

GridFunction() [6/7]

mfem::GridFunction::GridFunction	(	Mesh *	m,
		std::istream &	input )

Construct a GridFunction on the given Mesh, using the data from input.

The content of input should be in the format created by the method Save(). The reconstructed FiniteElementSpace and FiniteElementCollection are owned by the GridFunction.

Definition at line 38 of file gridfunc.cpp.

GridFunction() [7/7]

mfem::GridFunction::GridFunction	(	Mesh *	m,
		GridFunction *	gf_array[],
		int	num_pieces )

Definition at line 79 of file gridfunc.cpp.

~GridFunction()

virtual mfem::GridFunction::~GridFunction ( )

inlinevirtual

Destroys grid function.

Definition at line 1552 of file gridfunc.hpp.

Member Function Documentation

AccumulateAndCountBdrTangentValues()

void mfem::GridFunction::AccumulateAndCountBdrTangentValues ( VectorCoefficient & vcoeff, const Array< int > & bdr_attr, Array< int > & values_counter )

protected

Definition at line 2205 of file gridfunc.cpp.

AccumulateAndCountBdrValues()

void mfem::GridFunction::AccumulateAndCountBdrValues ( Coefficient * coeff[], VectorCoefficient * vcoeff, const Array< int > & attr, Array< int > & values_counter )

protected

Definition at line 2053 of file gridfunc.cpp.

AccumulateAndCountDerivativeValues()

void mfem::GridFunction::AccumulateAndCountDerivativeValues ( int comp, int der_comp, GridFunction & der, Array< int > & zones_per_dof ) const

protected

Used for the serial and parallel implementations of the GetDerivative() method; see its documentation.

Definition at line 1316 of file gridfunc.cpp.

AccumulateAndCountZones() [1/2]

void mfem::GridFunction::AccumulateAndCountZones ( Coefficient & coeff, AvgType type, Array< int > & zones_per_vdof )

protected

Accumulates (depending on type) the values of coeff at all shared vdofs and counts in how many zones each vdof appears.

Definition at line 1957 of file gridfunc.cpp.

AccumulateAndCountZones() [2/2]

void mfem::GridFunction::AccumulateAndCountZones ( VectorCoefficient & vcoeff, AvgType type, Array< int > & zones_per_vdof )

protected

Accumulates (depending on type) the values of vcoeff at all shared vdofs and counts in how many zones each vdof appears.

Definition at line 1998 of file gridfunc.cpp.

ComputeCurlError()

real_t mfem::GridFunction::ComputeCurlError ( VectorCoefficient * excurl, const IntegrationRule * irs[] = NULL ) const

virtual

Returns ||curl u_ex - curl u_h||_L2 for ND elements.

Parameters

[in]	excurl	Pointer to a VectorCoefficient object reproducing the expected curl of the vector field, curl u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 2968 of file gridfunc.cpp.

ComputeDGFaceJumpError() [1/2]

real_t mfem::GridFunction::ComputeDGFaceJumpError ( Coefficient * exsol, Coefficient * ell_coeff, class JumpScaling jump_scaling, const IntegrationRule * irs[] = NULL ) const

virtual

Returns the Face Jumps error for L2 elements.

Computes:

\[\sqrt{\sum_{f\in faces}\int_f js(f) ell(f) (2 u_{ex} - u_1 - u_2)^2}\]

Where js[f] is the jump_scaling evaluated on the face f and ell is the average of ell_coef evaluated in the two elements sharing the face f.

Parameters

[in]	exsol	Pointer to a Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in]	ell_coeff	Pointer to a Coefficient object used to compute the averaged value ell in the above integral.
[in]	jump_scaling	Can be configured to provide scaling by nu, nu/h, or nu*p^2/h
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of faces.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 3051 of file gridfunc.cpp.

ComputeDGFaceJumpError() [2/2]

real_t mfem::GridFunction::ComputeDGFaceJumpError	(	Coefficient *	exsol,
		Coefficient *	ell_coeff,
		real_t	Nu,
		const IntegrationRule *	irs[] = NULL ) const

Returns the Face Jumps error for L2 elements, with 1/h scaling.

Note

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Deprecated

See ComputeDGFaceJumpError(Coefficient *exsol, Coefficient *ell_coeff, class JumpScaling jump_scaling, const IntegrationRule *irs[]) const for the preferred implementation.

Definition at line 3168 of file gridfunc.cpp.

ComputeDivError()

real_t mfem::GridFunction::ComputeDivError ( Coefficient * exdiv, const IntegrationRule * irs[] = NULL ) const

virtual

Returns ||div u_ex - div u_h||_L2 for RT elements.

Parameters

[in]	exdiv	Pointer to a Coefficient object reproducing the expected divergence of the vector field, div u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 3012 of file gridfunc.cpp.

ComputeElementGradError()

real_t mfem::GridFunction::ComputeElementGradError ( int ielem, VectorCoefficient * exgrad, const IntegrationRule * irs[] = NULL ) const

virtual

Returns ||grad u_ex - grad u_h||_L2 in element ielem for H1 or L2 elements.

Parameters

[in]	ielem	Index of the element in which to compute the L2 error.
[in]	exgrad	Pointer to a VectorCoefficient object reproducing the expected gradient of the scalar field, grad u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Definition at line 2887 of file gridfunc.cpp.

ComputeElementL1Errors() [1/2]

virtual void mfem::GridFunction::ComputeElementL1Errors ( Coefficient & exsol, Vector & error, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns ||u_ex - u_h||_L1 elementwise for H1 or L2 elements.

Compute the \(L^1\) error in each element of the mesh and store the results in the Vector error. The result should be of length number of elements, for example an L2 GridFunction of order zero using map type VALUE.

Parameters

[in]	exsol	Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in,out]	error	Vector to contain the element-wise \(L^1\) errors
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Uses ComputeElementLpError internally. See the ComputeElementLpError documentation for generalizations of this error computation.

Definition at line 1183 of file gridfunc.hpp.

ComputeElementL1Errors() [2/2]

virtual void mfem::GridFunction::ComputeElementL1Errors ( VectorCoefficient & exsol, Vector & error, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns ||u_ex - u_h||_L1 elementwise for vector fields.

Compute the \(L^1\) error in each element of the mesh and store the results in the Vector error. The result should be of length number of elements, for example an L2 GridFunction of order zero using map type VALUE.

Computes:

\[\int_{elem} |scalar\_error|\]

Where

\[scalar\_error = \sqrt{(u_{ex} - u_h) \cdot (u_{ex} - u_h)}\]

Parameters

[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in,out]	error	Vector to contain the element-wise \(L^1\) errors
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Uses ComputeElementLpError internally. See the ComputeElementLpError documentation for generalizations of this error computation.

Definition at line 1380 of file gridfunc.hpp.

ComputeElementL2Errors() [1/2]

virtual void mfem::GridFunction::ComputeElementL2Errors ( Coefficient & exsol, Vector & error, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns ||u_ex - u_h||_L2 elementwise for H1 or L2 elements.

Compute the \(L^2\) error in each element of the mesh and store the results in the Vector error. The result should be of length number of elements, for example an L2 GridFunction of order zero using map type VALUE.

Computes:

\[(\int_{elem} |u_{ex} - u_h|^2)^{1/2}\]

Parameters

[in]	exsol	Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in,out]	error	Vector to contain the element-wise \(L^2\) errors
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Uses ComputeElementLpError internally. See the ComputeElementLpError documentation for generalizations of this error computation.

Definition at line 1221 of file gridfunc.hpp.

ComputeElementL2Errors() [2/2]

virtual void mfem::GridFunction::ComputeElementL2Errors ( VectorCoefficient & exsol, Vector & error, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns ||u_ex - u_h||_L2 elementwise for vector fields.

Compute the \(L^2\) error in each element of the mesh and store the results in the Vector error. The result should be of length number of elements, for example an L2 GridFunction of order zero using map type VALUE.

Computes:

\[(\int_{elem} |scalar\_error|^2)^{1/2}\]

Where

\[scalar\_error = \sqrt{(u_{ex} - u_h) \cdot (u_{ex} - u_h)}\]

Parameters

[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in,out]	error	Vector to contain the element-wise \(L^2\) errors
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Uses ComputeElementLpError internally. See the ComputeElementLpError documentation for generalizations of this error computation.

Definition at line 1421 of file gridfunc.hpp.

ComputeElementLpErrors() [1/2]

void mfem::GridFunction::ComputeElementLpErrors ( const real_t p, Coefficient & exsol, Vector & error, Coefficient * weight = NULL, const IntegrationRule * irs[] = NULL ) const

virtual

Returns ||u_ex - u_h||_Lp elementwise for H1 or L2 elements.

Compute the Lp error in each element of the mesh and store the results in the Vector error. The result should be of length number of elements, for example an L2 GridFunction of order zero using map type VALUE.

Computes:

\[(\int_{elem} w \, |u_{ex} - u_h|^p)^{1/p}\]

Parameters

[in]	p	Real value indicating the exponent of the \(L^p\) norm. To avoid domain errors p should have a positive value, either finite or infinite.
[in]	exsol	Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in,out]	error	Vector to contain the element-wise \(L^p\) errors
[in]	weight	Optional pointer to a Coefficient object reproducing a weighting function, w.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Definition at line 3452 of file gridfunc.cpp.

ComputeElementLpErrors() [2/2]

void mfem::GridFunction::ComputeElementLpErrors ( const real_t p, VectorCoefficient & exsol, Vector & error, Coefficient * weight = NULL, VectorCoefficient * v_weight = NULL, const IntegrationRule * irs[] = NULL ) const

virtual

Returns ||u_ex - u_h||_Lp elementwise for vector fields.

Compute the \(L^p\) error in each element of the mesh and store the results in the Vector error. The result should be of length number of elements, for example an L2 GridFunction of order zero using map type VALUE.

Computes:

\[(\int_{elem} w \, |scalar\_error|^p)^{1/p}\]

Where

\[scalar\_error = |v\_weight \cdot (u_{ex} - u_h)|\]

or

\[scalar\_error = \sqrt{(u_{ex} - u_h) \cdot (u_{ex} - u_h)}\]

Parameters

[in]	p	Real value indicating the exponent of the \(L^p\) norm. To avoid domain errors p should have a positive value, either finite or infinite.
[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in,out]	error	Vector to contain the element-wise \(L^p\) errors
[in]	weight	Optional pointer to a Coefficient object reproducing a weighting function, w.
[in]	v_weight	Optional pointer to a VectorCoefficient object reproducing a weighting vector as shown above.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Definition at line 3600 of file gridfunc.cpp.

ComputeElementMaxErrors() [1/2]

virtual void mfem::GridFunction::ComputeElementMaxErrors ( Coefficient & exsol, Vector & error, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns Max|u_ex - u_h| elementwise for H1 or L2 elements.

Compute the \(L^\infty\) error in each element of the mesh and store the results in the Vector error. The result should be of length number of elements, for example an L2 GridFunction of order zero using map type VALUE.

Parameters

[in]	exsol	Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in,out]	error	Vector to contain the element-wise \(L^\infty\) errors
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Uses ComputeElementLpError internally. See the ComputeElementLpError documentation for generalizations of this error computation.

Definition at line 1250 of file gridfunc.hpp.

ComputeElementMaxErrors() [2/2]

virtual void mfem::GridFunction::ComputeElementMaxErrors ( VectorCoefficient & exsol, Vector & error, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns Max|u_ex - u_h| elementwise for vector fields.

Compute the \(L^\infty\) error in each element of the mesh and store the results in the Vector error. The result should be of length number of elements, for example an L2 GridFunction of order zero using map type VALUE.

Computes:

\[max_{elem} |scalar\_error|\]

Where

\[scalar\_error = \sqrt{(u_{ex} - u_h) \cdot (u_{ex} - u_h)}\]

Parameters

[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in,out]	error	Vector to contain the element-wise \(L^\infty\) errors
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Uses ComputeElementLpError internally. See the ComputeElementLpError documentation for generalizations of this error computation.

Computes the maximum magnitude of the difference vector not the component-wise maximum difference of the vector fields.

Definition at line 1459 of file gridfunc.hpp.

ComputeFlux()

void mfem::GridFunction::ComputeFlux ( BilinearFormIntegrator & blfi, GridFunction & flux, bool wcoef = true, int subdomain = -1 )

virtual

Reimplemented in mfem::ParGridFunction.

Definition at line 338 of file gridfunc.cpp.

ComputeGradError()

real_t mfem::GridFunction::ComputeGradError ( VectorCoefficient * exgrad, const IntegrationRule * irs[] = NULL ) const

virtual

Returns ||grad u_ex - grad u_h||_L2 for H1 or L2 elements.

Parameters

[in]	exgrad	Pointer to a VectorCoefficient object reproducing the expected gradient of the scalar field, grad u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

This function only computes the error of the gradient in the interior of the elements. In the context of discontinuous Galerkin (DG) methods it may also be desirable to compute the error in the jumps across element interfaces using ComputeDGFaceJumpError().

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 2925 of file gridfunc.cpp.

ComputeH1Error() [1/2]

real_t mfem::GridFunction::ComputeH1Error ( Coefficient * exsol, VectorCoefficient * exgrad, Coefficient * ell_coef, real_t Nu, int norm_type ) const

virtual

This method is kept for backward compatibility.

Returns either the H1-seminorm, or the DG face jumps error, or both depending on norm_type = 1, 2, 3. Additional arguments for the DG face jumps norm: ell_coeff: mesh-depended coefficient (weight) Nu: scalar constant weight

Reimplemented in mfem::ParGridFunction.

Definition at line 3177 of file gridfunc.cpp.

ComputeH1Error() [2/2]

real_t mfem::GridFunction::ComputeH1Error ( Coefficient * exsol, VectorCoefficient * exgrad, const IntegrationRule * irs[] = NULL ) const

virtual

Returns the error measured in H1-norm for H1 or L2 elements.

Computes the norm using the \(L^2\) norms of the function and its gradient

\[\sqrt{norm\_u^2 + norm\_du^2}\]

Where

\[norm\_u = \|u_{ex} - u_h\|_{L^2}\]

and

\[norm\_du = \|du_{ex} - \nabla u_h\|_{L^2}\]

Parameters

[in]	exsol	Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in]	exgrad	VectorCoefficient object reproducing the anticipated values of the gradient of the scalar field, du_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

For L2 elements this returns what could be called a "broken" H1-norm.

Reimplemented in mfem::ParGridFunction.

Definition at line 3194 of file gridfunc.cpp.

ComputeHCurlError()

real_t mfem::GridFunction::ComputeHCurlError ( VectorCoefficient * exsol, VectorCoefficient * excurl, const IntegrationRule * irs[] = NULL ) const

virtual

Returns the error measured in H(curl)-norm for ND elements.

Computes the norm using the \(L^2\) norms of the function and its curl

\[\sqrt{norm\_u^2 + norm\_du^2}\]

Where

\[norm\_u = \|u_{ex} - u_h\|_{L^2}\]

and

\[norm\_du = \|du_{ex} - \nabla\times u_h\|_{L^2}\]

Parameters

[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in]	excurl	VectorCoefficient object reproducing the anticipated values of the curl of the vector field, du_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 3212 of file gridfunc.cpp.

ComputeHDivError()

real_t mfem::GridFunction::ComputeHDivError ( VectorCoefficient * exsol, Coefficient * exdiv, const IntegrationRule * irs[] = NULL ) const

virtual

Returns the error measured in H(div)-norm for RT elements.

Computes the norm using the \(L^2\) norms of the function and its divergence

\[\sqrt{norm\_u^2 + norm\_du^2}\]

Where

\[norm\_u = \|u_{ex} - u_h\|_{L^2}\]

and

\[norm\_du = \|du_{ex} - \nabla\cdot u_h\|_{L^2}\]

Parameters

[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in]	exdiv	VectorCoefficient object reproducing the anticipated values of the divergence of the vector field, du_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 3203 of file gridfunc.cpp.

ComputeL1Error() [1/3]

virtual real_t mfem::GridFunction::ComputeL1Error ( Coefficient & exsol, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns ||u_ex - u_h||_L1 for H1 or L2 elements.

Computes:

\[\sum_{elems} \int_{elem} |u_{ex} - u_h|\]

Parameters

[in]	exsol	Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Uses ComputeLpError internally. See the ComputeLpError documentation for generalizations of this error computation.

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Reimplemented in mfem::ParGridFunction.

Definition at line 962 of file gridfunc.hpp.

ComputeL1Error() [2/3]

virtual MFEM_DEPRECATED real_t mfem::GridFunction::ComputeL1Error ( Coefficient * exsol[], const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns ||u_ex - u_h||_L1 for H1 or L2 elements.

Computes:

\[\sum_{elems} \int_{elem} |u_{ex} - u_h|\]

Parameters

[in]	exsol	Pointer to an array of Coefficient objects reproducing the anticipated values of the scalar field, u_ex. Only the first entry of this array will be accessed.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Uses ComputeW11Error internally. See the ComputeW11Error documentation for generalizations of this error computation.

Warning

While this function is nominally equivalent to ComputeLpError, with appropriate arguments, the returned errors may differ noticeably because ComputeLpError uses a higher order integration rule by default.

Deprecated

See ComputeL1Error(Coefficient &exsol, const IntegrationRule *irs[]) const for the preferred implementation.

Reimplemented in mfem::ParGridFunction.

Definition at line 1002 of file gridfunc.hpp.

ComputeL1Error() [3/3]

virtual real_t mfem::GridFunction::ComputeL1Error ( VectorCoefficient & exsol, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns ||u_ex - u_h||_L1 for vector fields.

Computes:

\[\sum_{elems} \int_{elem} |scalar\_error|\]

Where

\[scalar\_error = \sqrt{(u_{ex} - u_h) \cdot (u_{ex} - u_h)}\]

Parameters

[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Uses ComputeLpError internally. See the ComputeLpError documentation for generalizations of this error computation.

Reimplemented in mfem::ParGridFunction.

Definition at line 1075 of file gridfunc.hpp.

ComputeL2Error() [1/3]

virtual real_t mfem::GridFunction::ComputeL2Error ( Coefficient & exsol, const IntegrationRule * irs[] = NULL, const Array< int > * elems = NULL ) const

inlinevirtual

Returns ||u_ex - u_h||_L2 for H1 or L2 elements.

Parameters

[in]	exsol	Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.
[in]	elems	Optional pointer to a marker array, with a length equal to the number of local elements, indicating which elements to integrate over. Only those elements corresponding to non-zero entries in elems will contribute to the computed L2 error.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 588 of file gridfunc.hpp.

ComputeL2Error() [2/3]

real_t mfem::GridFunction::ComputeL2Error ( Coefficient * exsol[], const IntegrationRule * irs[] = NULL, const Array< int > * elems = NULL ) const

virtual

Returns ||exsol - u_h||_L2 for scalar or vector H1 or L2 elements.

Parameters

[in]	exsol	Pointer to an array of scalar Coefficient objects, one for each component of the vector field. The length of the array should be at least equal to FiniteElementSpace::GetVDim().
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.
[in]	elems	Optional pointer to a marker array, with a length equal to the number of local elements, indicating which elements to integrate over. Only those elements corresponding to non-zero entries in elems will contribute to the computed L2 error.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 2786 of file gridfunc.cpp.

ComputeL2Error() [3/3]

real_t mfem::GridFunction::ComputeL2Error ( VectorCoefficient & exsol, const IntegrationRule * irs[] = NULL, const Array< int > * elems = NULL ) const

virtual

Returns ||u_ex - u_h||_L2 for vector fields.

Parameters

[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.
[in]	elems	Optional pointer to a marker array, with a length equal to the number of local elements, indicating which elements to integrate over. Only those elements corresponding to non-zero entries in elems will contribute to the computed L2 error.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 2844 of file gridfunc.cpp.

ComputeLpError() [1/2]

real_t mfem::GridFunction::ComputeLpError ( const real_t p, Coefficient & exsol, Coefficient * weight = NULL, const IntegrationRule * irs[] = NULL, const Array< int > * elems = NULL ) const

virtual

Returns ||u_ex - u_h||_Lp for H1 or L2 elements.

Computes:

\[(\sum_{elems} \int_{elem} w \, |u_{ex} - u_h|^p)^{1/p}\]

Parameters

[in]	p	Real value indicating the exponent of the \(L^p\) norm. To avoid domain errors p should have a positive value, either finite or infinite.
[in]	exsol	Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in]	weight	Optional pointer to a Coefficient object reproducing a weighting function, w.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.
[in]	elems	Optional pointer to a marker array, with a length equal to the number of local elements, indicating which elements to integrate over. Only those elements corresponding to non-zero entries in elems will contribute to the computed L2 error.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 3387 of file gridfunc.cpp.

ComputeLpError() [2/2]

real_t mfem::GridFunction::ComputeLpError ( const real_t p, VectorCoefficient & exsol, Coefficient * weight = NULL, VectorCoefficient * v_weight = NULL, const IntegrationRule * irs[] = NULL ) const

virtual

Returns ||u_ex - u_h||_Lp for vector fields.

When given a vector weight, compute the pointwise (scalar) error as the dot product of the vector error with the vector weight. Otherwise, the scalar error is the l_2 norm of the vector error.

Computes:

\[(\sum_{elems} \int_{elem} w \, |scalar\_error|^p)^{1/p}\]

Where

\[scalar\_error = |v\_weight \cdot (u_{ex} - u_h)|\]

or

\[scalar\_error = \sqrt{(u_{ex} - u_h) \cdot (u_{ex} - u_h)}\]

Parameters

[in]	p	Real value indicating the exponent of the \(L^p\) norm. To avoid domain errors p should have a positive value, either finite or infinite.
[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in]	weight	Optional pointer to a Coefficient object reproducing a weighting function, w.
[in]	v_weight	Optional pointer to a VectorCoefficient object reproducing a weighting vector as shown above.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Reimplemented in mfem::ParGridFunction.

Definition at line 3511 of file gridfunc.cpp.

ComputeMaxError() [1/3]

virtual real_t mfem::GridFunction::ComputeMaxError ( Coefficient & exsol, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns Max|u_ex - u_h| error for H1 or L2 elements.

Compute the \(L^\infty\) error across the entire domain.

Parameters

[in]	exsol	Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

Uses ComputeLpError internally. See the ComputeLpError documentation for generalizations of this error computation.

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Reimplemented in mfem::ParGridFunction.

Definition at line 869 of file gridfunc.hpp.

ComputeMaxError() [2/3]

real_t mfem::GridFunction::ComputeMaxError ( Coefficient * exsol[], const IntegrationRule * irs[] = NULL ) const

virtual

Returns Max|u_ex - u_h| error for scalar or vector fields.

Compute the \(L^\infty\) error across the entire domain.

Computes:

\[max_{elems} (max_{elem} |scalar\_error|)\]

Where

\[scalar\_error = max_{d=0\ldots vdim}|u_{ex}[d] - u_h[d]|\]

Parameters

[in]	exsol	Pointer to an array of scalar Coefficient objects, one for each component of the vector field. The length of the array should be at least equal to FiniteElementSpace::GetVDim().
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

This implementation of the max error of a vector field computes the max norm over vector components rather than the magnitude of the vector.

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Reimplemented in mfem::ParGridFunction.

Definition at line 3221 of file gridfunc.cpp.

ComputeMaxError() [3/3]

virtual real_t mfem::GridFunction::ComputeMaxError ( VectorCoefficient & exsol, const IntegrationRule * irs[] = NULL ) const

inlinevirtual

Returns Max|u_ex - u_h| error for vector fields.

Compute the \(L^\infty\) error across the entire domain.

Computes:

\[max_{elems} (max_{elem} |scalar\_error|)\]

Where

\[scalar\_error = \sqrt{(u_{ex} - u_h) \cdot (u_{ex} - u_h)}\]

Parameters

[in]	exsol	VectorCoefficient object reproducing the anticipated values of the vector field, u_ex.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

Uses ComputeLpError internally. See the ComputeLpError documentation for generalizations of this error computation.

Computes the maximum magnitude of the difference vector not the component-wise maximum difference of the vector fields.

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Reimplemented in mfem::ParGridFunction.

Definition at line 931 of file gridfunc.hpp.

ComputeMeans()

void mfem::GridFunction::ComputeMeans ( AvgType type, Array< int > & zones_per_vdof )

protected

Definition at line 2268 of file gridfunc.cpp.

ComputeW11Error()

real_t mfem::GridFunction::ComputeW11Error ( Coefficient * exsol, VectorCoefficient * exgrad, int norm_type, const Array< int > * elems = NULL, const IntegrationRule * irs[] = NULL ) const

virtual

Returns \(W^1_1\) norm (or portions thereof) for H1 or L2 elements.

Computes for norm_type == 1 the \(L^1\) norm of \(u\):

\[(\sum_{elems} \int_{elem} |u_{ex} - u_h|\]

Computes for norm_type == 2 the \(L^1\) semi-norm of \(\nabla u\):

\[(\sum_{elems} \int_{elem} |du_{ex} - \nabla u_h|\]

Computes for norm_type == 3 the \(W^1_1\) norm of \(u\):

\[(\sum_{elems} \int_{elem} |u_{ex} - u_h| + |du_{ex} - \nabla u_h|\]

Parameters

[in]	exsol	Pointer to Coefficient object reproducing the anticipated values of the scalar field, u_ex.
[in]	exgrad	Pointer to VectorCoefficient object reproducing the anticipated values of the gradient of the scalar field, du_ex.
[in]	norm_type	Integer value of 1, 2, or 3 indicating the type of norm to compute (see above).
[in]	elems	Optional pointer to a marker array, with a length equal to the number of local elements, indicating which elements to integrate over. Only those elements corresponding to non-zero entries in elems will contribute to the computed \(W^1_1\) error.
[in]	irs	Optional pointer to an array of custom integration rules e.g. higher order than the default rules. If present the array will be indexed by Geometry::Type.

Note

If an array of integration rules is provided through irs, be sure to include valid rules for each element type that may occur in the list of elements.

Quadratures with negative weights (as in some simplex integration rules in MFEM) can produce negative integrals even with non-negative integrands. To avoid returning negative errors this function uses the absolute values of the element-wise integrals. This may lead to results which are not entirely consistent with such integration rules.

Definition at line 3277 of file gridfunc.cpp.

CountElementsPerVDof()

void mfem::GridFunction::CountElementsPerVDof ( Array< int > & elem_per_vdof ) const

virtual

For each vdof, counts how many elements contain the vdof, as containment is determined by FiniteElementSpace::GetElementVDofs().

Reimplemented in mfem::ParGridFunction.

Definition at line 1940 of file gridfunc.cpp.

CurlDim()

int mfem::GridFunction::CurlDim

(

)

const

Shortcut for calling FiniteElementSpace::GetCurlDim() on the underlying fes.

Definition at line 358 of file gridfunc.cpp.

Destroy()

void mfem::GridFunction::Destroy ( )

protected

Definition at line 157 of file gridfunc.cpp.

FESpace() [1/2]

FiniteElementSpace * mfem::GridFunction::FESpace ( )

inline

Definition at line 1484 of file gridfunc.hpp.

FESpace() [2/2]

const FiniteElementSpace * mfem::GridFunction::FESpace ( ) const

inline

Definition at line 1485 of file gridfunc.hpp.

GetBdrValuesFrom()

void mfem::GridFunction::GetBdrValuesFrom

(

const GridFunction &

orig_func

)

Definition at line 1147 of file gridfunc.cpp.

GetCurl()

void mfem::GridFunction::GetCurl	(	ElementTransformation &	tr,
		Vector &	curl ) const

Definition at line 1483 of file gridfunc.cpp.

GetDerivative()

void mfem::GridFunction::GetDerivative	(	int	comp,
		int	der_comp,
		GridFunction &	der ) const

Compute a certain derivative of a function's component. Derivatives of the function are computed at the DOF locations of der, and averaged over overlapping DOFs. Thus this function projects the derivative to the FiniteElementSpace of der.

Parameters

[in]	comp	Index of the function's component to be differentiated. The index is 1-based, i.e., use 1 for scalar functions.
[in]	der_comp	Use 0/1/2 for derivatives in x/y/z directions.
[out]	der	The resulting derivative (scalar function). The FiniteElementSpace of this function must be set before the call.

Definition at line 1369 of file gridfunc.cpp.

GetDivergence()

real_t mfem::GridFunction::GetDivergence

(

ElementTransformation &

tr

)

const

Definition at line 1397 of file gridfunc.cpp.

GetElementAverages()

void mfem::GridFunction::GetElementAverages

(

GridFunction &

avgs

)

const

Compute \( (\int_{\Omega} (*this) \psi_i)/(\int_{\Omega} \psi_i) \), where \( \psi_i \) are the basis functions for the FE space of avgs. Both FE spaces should be scalar and on the same mesh.

Definition at line 1722 of file gridfunc.cpp.

GetElementDofValues()

void mfem::GridFunction::GetElementDofValues ( int el, Vector & dof_vals ) const

virtual

Sets the output vector dof_vals to the values of the degrees of freedom of element el.

Reimplemented in mfem::ParGridFunction.

Definition at line 1762 of file gridfunc.cpp.

GetFaceValues()

int mfem::GridFunction::GetFaceValues	(	int	i,
		int	side,
		const IntegrationRule &	ir,
		Vector &	vals,
		DenseMatrix &	tr,
		int	vdim = 1 ) const

Compute a collection of scalar values from within the face indicated by the index i.

Definition at line 661 of file gridfunc.cpp.

GetFaceVectorValues()

int mfem::GridFunction::GetFaceVectorValues	(	int	i,
		int	side,
		const IntegrationRule &	ir,
		DenseMatrix &	vals,
		DenseMatrix &	tr ) const

Compute a collection of vector values from within the face indicated by the index i.

Definition at line 1057 of file gridfunc.cpp.

GetGradient()

void mfem::GridFunction::GetGradient	(	ElementTransformation &	tr,
		Vector &	grad ) const

Gradient of a scalar function at a quadrature point.

Note

It is assumed that the IntegrationPoint of interest has been specified by ElementTransformation::SetIntPoint() before calling GetGradient().

Can be used from a ParGridFunction when tr is an ElementTransformation of a face-neighbor element and face-neighbor data has been exchanged.

Definition at line 1575 of file gridfunc.cpp.

GetGradients() [1/2]

void mfem::GridFunction::GetGradients ( const int elem, const IntegrationRule & ir, DenseMatrix & grad ) const

inline

Extension of GetGradient(...) for a collection of IntegrationPoints.

Definition at line 349 of file gridfunc.hpp.

GetGradients() [2/2]

void mfem::GridFunction::GetGradients	(	ElementTransformation &	tr,
		const IntegrationRule &	ir,
		DenseMatrix &	grad ) const

Extension of GetGradient(...) for a collection of IntegrationPoints.

Definition at line 1640 of file gridfunc.cpp.

GetHessians() [1/2]

void mfem::GridFunction::GetHessians	(	int	i,
		const IntegrationRule &	ir,
		DenseMatrix &	hess,
		DenseMatrix &	tr,
		int	vdim = 1 ) const

Definition at line 648 of file gridfunc.cpp.

GetHessians() [2/2]

void mfem::GridFunction::GetHessians	(	int	i,
		const IntegrationRule &	ir,
		DenseMatrix &	hess,
		int	vdim = 1 ) const

Definition at line 605 of file gridfunc.cpp.

GetLaplacians() [1/2]

void mfem::GridFunction::GetLaplacians	(	int	i,
		const IntegrationRule &	ir,
		Vector &	laps,
		DenseMatrix &	tr,
		int	vdim = 1 ) const

Definition at line 593 of file gridfunc.cpp.

GetLaplacians() [2/2]

void mfem::GridFunction::GetLaplacians	(	int	i,
		const IntegrationRule &	ir,
		Vector &	laps,
		int	vdim = 1 ) const

Definition at line 566 of file gridfunc.cpp.

GetNodalValues() [1/2]

void mfem::GridFunction::GetNodalValues	(	int	i,
		Array< real_t > &	nval,
		int	vdim = 1 ) const

Returns the values in the vertices of i'th element for dimension vdim.

Definition at line 392 of file gridfunc.cpp.

GetNodalValues() [2/2]

void mfem::GridFunction::GetNodalValues	(	Vector &	nval,
		int	vdim = 1 ) const

Return the values as a vector on mesh vertices for dimension vdim.

Definition at line 1913 of file gridfunc.cpp.

GetSequence()

long mfem::GridFunction::GetSequence ( ) const

inline

Return update counter, similar to Mesh::GetSequence(). Used to check if it is up to date with the space.

Definition at line 1482 of file gridfunc.hpp.

GetTrueDofs()

void mfem::GridFunction::GetTrueDofs

(

Vector &

tv

)

const

Extract the true-dofs from the GridFunction.

Definition at line 363 of file gridfunc.cpp.

GetTrueVector() [1/2]

Vector & mfem::GridFunction::GetTrueVector ( )

inline

Read and write access to the (optional) internal true-dof Vector.

Note that t_vec is set if it is not allocated or set already.

Definition at line 140 of file gridfunc.hpp.

GetTrueVector() [2/2]

const Vector & mfem::GridFunction::GetTrueVector ( ) const

inline

Read only access to the (optional) internal true-dof Vector.

Definition at line 133 of file gridfunc.hpp.

GetValue() [1/2]

real_t mfem::GridFunction::GetValue ( ElementTransformation & T, const IntegrationPoint & ip, int comp = 0, Vector * tr = NULL ) const

virtual

Return a scalar value from within the element indicated by the ElementTransformation Object.

Reimplemented in mfem::ParGridFunction.

Definition at line 720 of file gridfunc.cpp.

GetValue() [2/2]

real_t mfem::GridFunction::GetValue ( int i, const IntegrationPoint & ip, int vdim = 1 ) const

virtual

Return a scalar value from within the given element.

Reimplemented in mfem::ParGridFunction.

Definition at line 446 of file gridfunc.cpp.

GetValues() [1/3]

void mfem::GridFunction::GetValues	(	ElementTransformation &	T,
		const IntegrationRule &	ir,
		Vector &	vals,
		int	comp = 0,
		DenseMatrix *	tr = NULL ) const

Compute a collection of scalar values from within the element indicated by the ElementTransformation object.

Definition at line 835 of file gridfunc.cpp.

GetValues() [2/3]

void mfem::GridFunction::GetValues	(	int	i,
		const IntegrationRule &	ir,
		Vector &	vals,
		DenseMatrix &	tr,
		int	vdim = 1 ) const

Compute a collection of vector values from within the element indicated by the index i.

Definition at line 555 of file gridfunc.cpp.

GetValues() [3/3]

void mfem::GridFunction::GetValues	(	int	i,
		const IntegrationRule &	ir,
		Vector &	vals,
		int	vdim = 1 ) const

Compute a collection of scalar values from within the element indicated by the index i.

Definition at line 518 of file gridfunc.cpp.

GetValuesFrom()

void mfem::GridFunction::GetValuesFrom

(

const GridFunction &

orig_func

)

Definition at line 1101 of file gridfunc.cpp.

GetVectorFieldNodalValues()

void mfem::GridFunction::GetVectorFieldNodalValues	(	Vector &	val,
		int	comp ) const

Definition at line 1249 of file gridfunc.cpp.

GetVectorFieldValues()

void mfem::GridFunction::GetVectorFieldValues	(	int	i,
		const IntegrationRule &	ir,
		DenseMatrix &	vals,
		DenseMatrix &	tr,
		int	comp = 0 ) const

Definition at line 1185 of file gridfunc.cpp.

GetVectorGradient()

void mfem::GridFunction::GetVectorGradient	(	ElementTransformation &	tr,
		DenseMatrix &	grad ) const

Compute the vector gradient with respect to the physical element variable.

Definition at line 1664 of file gridfunc.cpp.

GetVectorGradientHat()

void mfem::GridFunction::GetVectorGradientHat	(	ElementTransformation &	T,
		DenseMatrix &	gh ) const

Compute the vector gradient with respect to the reference element variable.

Definition at line 1381 of file gridfunc.cpp.

GetVectorValue() [1/2]

void mfem::GridFunction::GetVectorValue ( ElementTransformation & T, const IntegrationPoint & ip, Vector & val, Vector * tr = NULL ) const

virtual

Return a vector value from within the element indicated by the ElementTransformation Object.

Reimplemented in mfem::ParGridFunction.

Definition at line 855 of file gridfunc.cpp.

GetVectorValue() [2/2]

void mfem::GridFunction::GetVectorValue ( int i, const IntegrationPoint & ip, Vector & val ) const

virtual

Return a vector value from within the given element.

Reimplemented in mfem::ParGridFunction.

Definition at line 473 of file gridfunc.cpp.

GetVectorValues() [1/2]

void mfem::GridFunction::GetVectorValues	(	ElementTransformation &	T,
		const IntegrationRule &	ir,
		DenseMatrix &	vals,
		DenseMatrix *	tr = NULL ) const

Compute a collection of vector values from within the element indicated by the ElementTransformation object.

Definition at line 995 of file gridfunc.cpp.

GetVectorValues() [2/2]

void mfem::GridFunction::GetVectorValues	(	int	i,
		const IntegrationRule &	ir,
		DenseMatrix &	vals,
		DenseMatrix &	tr ) const

Definition at line 711 of file gridfunc.cpp.

ImposeBounds() [1/2]

void mfem::GridFunction::ImposeBounds	(	int	i,
		const Vector &	weights,
		const Vector &	lo_,
		const Vector &	hi_ )

Impose the given bounds on the function's DOFs while preserving its local integral (described in terms of the given weights) on the i'th element through SLBPQ optimization. Intended to be used for discontinuous FE functions.

Definition at line 1825 of file gridfunc.cpp.

ImposeBounds() [2/2]

void mfem::GridFunction::ImposeBounds	(	int	i,
		const Vector &	weights,
		real_t	min_ = 0.0,
		real_t	max_ = infinity() )

Definition at line 1861 of file gridfunc.cpp.

LegacyNCReorder()

void mfem::GridFunction::LegacyNCReorder ( )

protected

Loading helper.

Definition at line 3939 of file gridfunc.cpp.

MakeOwner()

void mfem::GridFunction::MakeOwner ( FiniteElementCollection * fec_ )

inline

Make the GridFunction the owner of fec_owned and fes.

If the new FiniteElementCollection, fec_, is NULL, ownership of fec_owned and fes is taken away.

Definition at line 122 of file gridfunc.hpp.

MakeRef() [1/4]

void mfem::GridFunction::MakeRef ( FiniteElementSpace * f, real_t * v )

virtual

Make the GridFunction reference external data on a new FiniteElementSpace.

This method changes the FiniteElementSpace associated with the GridFunction and sets the pointer v as external data in the GridFunction.

Reimplemented in mfem::ParGridFunction.

Definition at line 233 of file gridfunc.cpp.

MakeRef() [2/4]

void mfem::GridFunction::MakeRef ( FiniteElementSpace * f, Vector & v, int v_offset )

virtual

Make the GridFunction reference external data on a new FiniteElementSpace.

This method changes the FiniteElementSpace associated with the GridFunction and sets the data of the Vector v (plus the v_offset) as external data in the GridFunction.

Note

This version of the method will also perform bounds checks when the build option MFEM_DEBUG is enabled.

Reimplemented in mfem::ParGridFunction.

Definition at line 241 of file gridfunc.cpp.

MakeRef() [3/4]

void mfem::Vector::MakeRef ( Vector & base, int offset )

inline

Reset the Vector to be a reference to a sub-vector of base without changing its current size.

Definition at line 211 of file vector.hpp.

MakeRef() [4/4]

void mfem::Vector::MakeRef ( Vector & base, int offset, int size )

inline

Reset the Vector to be a reference to a sub-vector of base.

Definition at line 207 of file vector.hpp.

MakeTRef() [1/2]

void mfem::GridFunction::MakeTRef	(	FiniteElementSpace *	f,
		real_t *	tv )

Associate a new FiniteElementSpace and new true-dof data with the GridFunction.

• If the prolongation matrix of f is trivial (i.e. its method FiniteElementSpace::GetProlongationMatrix() returns NULL), then the method MakeRef() is called with the same arguments.
• Otherwise, the method SetSpace() is called with argument f.
• The internal true-dof vector is set to reference tv.

Definition at line 251 of file gridfunc.cpp.

MakeTRef() [2/2]

void mfem::GridFunction::MakeTRef	(	FiniteElementSpace *	f,
		Vector &	tv,
		int	tv_offset )

Associate a new FiniteElementSpace and new true-dof data with the GridFunction.

• If the prolongation matrix of f is trivial (i.e. its method FiniteElementSpace::GetProlongationMatrix() returns NULL), this method calls MakeRef() with the same arguments.
• Otherwise, this method calls SetSpace() with argument f.
• The internal true-dof vector is set to reference the sub-vector of tv starting at the offset tv_offset.

Definition at line 265 of file gridfunc.cpp.

operator=() [1/3]

GridFunction & mfem::GridFunction::operator= ( const GridFunction & rhs )

inline

Copy assignment. Only the data of the base class Vector is copied.

It is assumed that this object and rhs use FiniteElementSpaces that have the same size.

Note

Defining this method overwrites the implicitly defined copy assignment operator.

Definition at line 116 of file gridfunc.hpp.

operator=() [2/3]

GridFunction & mfem::GridFunction::operator=

(

const Vector &

v

)

Copy the data from v.

The size of v must be equal to the size of the associated FiniteElementSpace fes.

Definition at line 3692 of file gridfunc.cpp.

operator=() [3/3]

GridFunction & mfem::GridFunction::operator=

(

real_t

value

)

Redefine '=' for GridFunction = constant.

Definition at line 3686 of file gridfunc.cpp.

OwnFEC()

FiniteElementCollection * mfem::GridFunction::OwnFEC ( )

inline

Definition at line 124 of file gridfunc.hpp.

ProjectBdrCoefficient() [1/3]

void mfem::GridFunction::ProjectBdrCoefficient ( Coefficient & coeff, const Array< int > & attr )

inline

Project a Coefficient on the GridFunction, modifying only DOFs on the boundary associated with the boundary attributes marked in the attr array.

Definition at line 481 of file gridfunc.hpp.

ProjectBdrCoefficient() [2/3]

void mfem::GridFunction::ProjectBdrCoefficient ( Coefficient * coeff[], const Array< int > & attr )

virtual

Project a set of Coefficients on the components of the GridFunction, modifying only DOFs on the boundary associated with the boundary attributed marked in the attr array.

If a Coefficient pointer in the array coeff is NULL, that component will not be touched.

Reimplemented in mfem::ParGridFunction.

Definition at line 2664 of file gridfunc.cpp.

ProjectBdrCoefficient() [3/3]

void mfem::GridFunction::ProjectBdrCoefficient ( VectorCoefficient & vcoeff, const Array< int > & attr )

virtual

Project a VectorCoefficient on the GridFunction, modifying only DOFs on the boundary associated with the boundary attributes marked in the attr array.

Reimplemented in mfem::ParGridFunction.

Definition at line 2646 of file gridfunc.cpp.

ProjectBdrCoefficientNormal()

void mfem::GridFunction::ProjectBdrCoefficientNormal	(	VectorCoefficient &	vcoeff,
		const Array< int > &	bdr_attr )

Project the normal component of the given VectorCoefficient on the boundary. Only boundary attributes that are marked in 'bdr_attr' are projected. Assumes RT-type VectorFE GridFunction.

Definition at line 2694 of file gridfunc.cpp.

ProjectBdrCoefficientTangent()

void mfem::GridFunction::ProjectBdrCoefficientTangent ( VectorCoefficient & vcoeff, const Array< int > & bdr_attr )

virtual

Project the tangential components of the given VectorCoefficient on the boundary. Only boundary attributes that are marked in bdr_attr are projected. Assumes ND-type VectorFE GridFunction.

Reimplemented in mfem::ParGridFunction.

Definition at line 2769 of file gridfunc.cpp.

ProjectCoefficient() [1/6]

void mfem::GridFunction::ProjectCoefficient ( Coefficient & coeff )

virtual

Project coeff Coefficient to this GridFunction. The projection computation depends on the choice of the FiniteElementSpace fes. Note that this is usually interpolation at the degrees of freedom in each element (not L2 projection). For NURBS spaces these degrees of freedom are not available and L2 projection is resorted to as fallback.

Reimplemented in mfem::ParGridFunction.

Definition at line 2360 of file gridfunc.cpp.

ProjectCoefficient() [2/6]

void mfem::GridFunction::ProjectCoefficient	(	Coefficient &	coeff,
		Array< int > &	dofs,
		int	vd = 0 )

Project coeff Coefficient to this GridFunction, using one element for each degree of freedom in dofs and nodal interpolation on that element.

Definition at line 2421 of file gridfunc.cpp.

ProjectCoefficient() [3/6]

void mfem::GridFunction::ProjectCoefficient

(

Coefficient *

coeff[]

)

Analogous to the version with argument vcoeff VectorCoefficient but using an array of scalar coefficients for each component.

Definition at line 2551 of file gridfunc.cpp.

ProjectCoefficient() [4/6]

void mfem::GridFunction::ProjectCoefficient

(

VectorCoefficient &

vcoeff

)

Project vcoeff VectorCoefficient to this GridFunction. The projection computation depends on the choice of the FiniteElementSpace fes. Note that this is usually interpolation at the degrees of freedom in each element (not L2 projection). For NURBS spaces these degrees of freedom are not available and L2 projection is resorted to as fallback.

Definition at line 2445 of file gridfunc.cpp.

ProjectCoefficient() [5/6]

void mfem::GridFunction::ProjectCoefficient	(	VectorCoefficient &	vcoeff,
		Array< int > &	dofs )

Project vcoeff VectorCoefficient to this GridFunction, using one element for each degree of freedom in dofs and nodal interpolation on that element.

Definition at line 2495 of file gridfunc.cpp.

ProjectCoefficient() [6/6]

void mfem::GridFunction::ProjectCoefficient	(	VectorCoefficient &	vcoeff,
		int	attribute )

Project vcoeff VectorCoefficient to this GridFunction, only projecting onto elements with the given attribute.

Definition at line 2525 of file gridfunc.cpp.

ProjectDeltaCoefficient()

void mfem::GridFunction::ProjectDeltaCoefficient ( DeltaCoefficient & delta_coeff, real_t & integral )

protected

Definition at line 2293 of file gridfunc.cpp.

ProjectDiscCoefficient() [1/4]

void mfem::GridFunction::ProjectDiscCoefficient ( Coefficient & coeff, AvgType type )

virtual

Projects a discontinuous coefficient so that the values in shared vdofs are computed by taking an average of the possible values.

Reimplemented in mfem::ParGridFunction.

Definition at line 2626 of file gridfunc.cpp.

ProjectDiscCoefficient() [2/4]

void mfem::GridFunction::ProjectDiscCoefficient ( VectorCoefficient & coeff )

virtual

Project a discontinuous vector coefficient as a grid function on a continuous finite element space. The values in shared dofs are determined from the element with maximal attribute.

Reimplemented in mfem::ParGridFunction.

Definition at line 2620 of file gridfunc.cpp.

ProjectDiscCoefficient() [3/4]

void mfem::GridFunction::ProjectDiscCoefficient ( VectorCoefficient & coeff, Array< int > & dof_attr )

protected

Project a discontinuous vector coefficient in a continuous space and return in dof_attr the maximal attribute of the elements containing each degree of freedom.

Definition at line 2588 of file gridfunc.cpp.

ProjectDiscCoefficient() [4/4]

void mfem::GridFunction::ProjectDiscCoefficient ( VectorCoefficient & coeff, AvgType type )

virtual

Projects a discontinuous vector coefficient so that the values in shared vdofs are computed by taking an average of the possible values.

Reimplemented in mfem::ParGridFunction.

Definition at line 2637 of file gridfunc.cpp.

ProjectGridFunction()

void mfem::GridFunction::ProjectGridFunction

(

const GridFunction &

src

)

Project the src GridFunction to this GridFunction, both of which must be on the same mesh.

The current implementation assumes that all elements use the same projection matrix.

Definition at line 1773 of file gridfunc.cpp.

ProjectVectorFieldOn()

void mfem::GridFunction::ProjectVectorFieldOn	(	GridFunction &	vec_field,
		int	comp = 0 )

Definition at line 1280 of file gridfunc.cpp.

ProlongateToMaxOrder()

std::unique_ptr< GridFunction > mfem::GridFunction::ProlongateToMaxOrder

(

)

const

Return a GridFunction with the values of this, prolongated to the maximum order of all elements in the mesh.

Definition at line 4001 of file gridfunc.cpp.

ReorderByNodes()

void mfem::GridFunction::ReorderByNodes

(

)

For a vector grid function, makes sure that the ordering is byNODES.

Definition at line 1222 of file gridfunc.cpp.

RestrictConforming()

void mfem::GridFunction::RestrictConforming

(

)

On a non-conforming mesh, make sure the function lies in the conforming space by multiplying with R and then with P, the conforming restriction and prolongation matrices of the space, respectively.

Definition at line 1900 of file gridfunc.cpp.

Save() [1/3]

void mfem::GridFunction::Save ( adios2stream & out, const std::string & variable_name, const adios2stream::data_type type = adios2stream::data_type::point_data ) const

virtual

Save the GridFunction to a binary output stream using adios2 bp format.

Reimplemented in mfem::ParGridFunction.

Definition at line 3732 of file gridfunc.cpp.

Save() [2/3]

void mfem::GridFunction::Save ( const char * fname, int precision = 16 ) const

virtual

Save the GridFunction to a file. The given precision will be used for ASCII output.

Reimplemented in mfem::ParGridFunction.

Definition at line 3724 of file gridfunc.cpp.

Save() [3/3]

void mfem::GridFunction::Save ( std::ostream & out ) const

virtual

Save the GridFunction to an output stream.

Reimplemented in mfem::ParGridFunction.

Definition at line 3699 of file gridfunc.cpp.

SaveSTL()

void mfem::GridFunction::SaveSTL	(	std::ostream &	out,
		int	TimesToRefine = 1 )

Write the GridFunction in STL format. Note that the mesh dimension must be 2 and that quad elements will be broken into two triangles.

Definition at line 3839 of file gridfunc.cpp.

SaveSTLTri()

void mfem::GridFunction::SaveSTLTri ( std::ostream & out, real_t p1[], real_t p2[], real_t p3[] )

protected

Definition at line 3819 of file gridfunc.cpp.

SaveVTK()

void mfem::GridFunction::SaveVTK	(	std::ostream &	out,
		const std::string &	field_name,
		int	ref )

Write the GridFunction in VTK format. Note that Mesh::PrintVTK must be called first. The parameter ref > 0 must match the one used in Mesh::PrintVTK.

Definition at line 3740 of file gridfunc.cpp.

SetFromTrueDofs()

void mfem::GridFunction::SetFromTrueDofs ( const Vector & tv )

virtual

Set the GridFunction from the given true-dof vector.

Reimplemented in mfem::ParGridFunction.

Definition at line 378 of file gridfunc.cpp.

SetFromTrueVector()

void mfem::GridFunction::SetFromTrueVector ( )

inline

Shortcut for calling SetFromTrueDofs() with GetTrueVector() as argument.

Definition at line 153 of file gridfunc.hpp.

SetSpace()

void mfem::GridFunction::SetSpace ( FiniteElementSpace * f )

virtual

Associate a new FiniteElementSpace with the GridFunction.

The GridFunction is resized using the SetSize() method.

Reimplemented in mfem::ParGridFunction.

Definition at line 225 of file gridfunc.cpp.

SetTrueVector()

void mfem::GridFunction::SetTrueVector ( )

inline

Shortcut for calling GetTrueDofs() with GetTrueVector() as argument.

Definition at line 147 of file gridfunc.hpp.

SumFluxAndCount()

void mfem::GridFunction::SumFluxAndCount ( BilinearFormIntegrator & blfi, GridFunction & flux, Array< int > & counts, bool wcoef, int subdomain )

protected

Definition at line 281 of file gridfunc.cpp.

Update()

void mfem::GridFunction::Update ( )

virtual

Transform by the Space UpdateMatrix (e.g., on Mesh change).

Reimplemented in mfem::ParGridFunction.

Definition at line 167 of file gridfunc.cpp.

UpdatePRef()

void mfem::GridFunction::UpdatePRef ( )

protected

P-refinement version of Update().

Definition at line 205 of file gridfunc.cpp.

VectorDim()

int mfem::GridFunction::VectorDim

(

)

const

Shortcut for calling FiniteElementSpace::GetVectorDim() on the underlying fes.

Definition at line 353 of file gridfunc.cpp.

Member Data Documentation

fec_owned

FiniteElementCollection* mfem::GridFunction::fec_owned

protected

Used when the grid function is read from a file. It can also be set explicitly, see MakeOwner().

If not NULL, this pointer is owned by the GridFunction.

Definition at line 40 of file gridfunc.hpp.

fes

FiniteElementSpace* mfem::GridFunction::fes

protected

FE space on which the grid function lives. Owned if fec_owned is not NULL.

Definition at line 34 of file gridfunc.hpp.

fes_sequence

long mfem::GridFunction::fes_sequence

protected

Definition at line 42 of file gridfunc.hpp.

t_vec

Vector mfem::GridFunction::t_vec

protected

Optional, internal true-dof vector: if the FiniteElementSpace fes has a non-trivial (i.e. not NULL) prolongation operator, this Vector may hold associated true-dof values - either owned or external.

Definition at line 47 of file gridfunc.hpp.

---

The documentation for this class was generated from the following files:

• fem/gridfunc.hpp
• fem/gridfunc.cpp