DUNE PDELab (git)
Class to document the stationary local operator interface. More...
#include <dune/pdelab/localoperator/interface.hh>
Methods for selective assembly | |
template<typename EG > | |
bool | skip_entity (const EG &eg) const |
whether to assembly methods associated with a given entity More... | |
template<typename IG > | |
bool | skip_intersection (const IG &ig) const |
whether to assembly methods associated with a given intersection More... | |
Methods for the sparsity pattern | |
template<typename LFSU , typename LFSV , typename LocalPattern > | |
void | pattern_volume (const LFSU &lfsu, const LFSV &lfsv, LocalPattern &pattern) const |
get an element's contribution to the sparsity pattern More... | |
template<typename LFSU , typename LFSV , typename LocalPattern > | |
void | pattern_volume_post_skeleton (const LFSU &lfsu, const LFSV &lfsv, LocalPattern &pattern) const |
get an element's contribution to the sparsity pattern after the intersections have been handled More... | |
template<typename LFSU , typename LFSV , typename LocalPattern > | |
void | pattern_skeleton (const LFSU &lfsu_s, const LFSV &lfsv_s, const LFSU &lfsu_n, const LFSV &lfsv_n, LocalPattern &pattern_sn, LocalPattern &pattern_ns) const |
get an internal intersection's contribution to the sparsity pattern More... | |
template<typename LFSU , typename LFSV , typename LocalPattern > | |
void | pattern_boundary (const LFSU &lfsu_s, const LFSV &lfsv_s, LocalPattern &pattern_ss) const |
get a boundary intersection's contribution to the sparsity pattern More... | |
Methods for the residual – non-constant parts | |
template<typename EG , typename LFSU , typename X , typename LFSV , typename R > | |
void | alpha_volume (const EG &eg, const LFSU &lfsu, const X &x, const LFSV &lfsv, R &r) const |
get an element's contribution to alpha More... | |
template<typename EG , typename LFSU , typename X , typename LFSV , typename R > | |
void | alpha_volume_post_skeleton (const EG &eg, const LFSU &lfsu, const X &x, const LFSV &lfsv, R &r) const |
get an element's contribution to alpha after the intersections have been handled More... | |
template<typename IG , typename LFSU , typename X , typename LFSV , typename R > | |
void | alpha_skeleton (const IG &ig, const LFSU &lfsu_s, const X &x_s, const LFSV &lfsv_s, const LFSU &lfsu_n, const X &x_n, const LFSV &lfsv_n, R &r_s, R &r_n) const |
get an internal intersections's contribution to alpha More... | |
template<typename IG , typename LFSU , typename X , typename LFSV , typename R > | |
void | alpha_boundary (const IG &ig, const LFSU &lfsu_s, const X &x_s, const LFSV &lfsv_s, R &r_s) const |
get a boundary intersections's contribution to alpha More... | |
Methods for the residual – constant parts | |
template<typename EG , typename LFSV , typename R > | |
void | lambda_volume (const EG &eg, const LFSV &lfsv, R &r) const |
get an element's contribution to lambda More... | |
template<typename EG , typename LFSV , typename R > | |
void | lambda_volume_post_skeleton (const EG &eg, const LFSV &lfsv, R &r) const |
get an element's contribution to lambda after the intersections have been handled More... | |
template<typename IG , typename LFSV , typename R > | |
void | lambda_skeleton (const IG &ig, const LFSV &lfsv_s, const LFSV &lfsv_n, R &r_s, R &r_n) const |
get an internal intersections's contribution to lambda More... | |
template<typename IG , typename LFSV , typename R > | |
void | lambda_boundary (const IG &ig, const LFSV &lfsv_s, R &r_s) const |
get a boundary intersections's contribution to lambda More... | |
Methods for the application of the jacobian for linear problems | |
template<typename EG , typename LFSU , typename X , typename LFSV , typename Y > | |
void | jacobian_apply_volume (const EG &eg, const LFSU &lfsu, const X &z, const LFSV &lfsv, Y &y) const |
Applies an element's jacobian to a vector for a linear problem. More... | |
template<typename EG , typename LFSU , typename X , typename LFSV , typename Y > | |
void | jacobian_apply_volume_post_skeleton (const EG &eg, const LFSU &lfsu, const X &z, const LFSV &lfsv, Y &y) const |
template<typename IG , typename LFSU , typename X , typename LFSV , typename Y > | |
void | jacobian_apply_skeleton (const IG &ig, const LFSU &lfsu_s, const X &z_s, const LFSV &lfsv_s, const LFSU &lfsu_n, const X &z_n, const LFSV &lfsv_n, Y &y_s, Y &y_n) const |
template<typename IG , typename LFSU , typename X , typename LFSV , typename Y > | |
void | jacobian_apply_boundary (const IG &ig, const LFSU &lfsu_s, const X &z_s, const LFSV &lfsv_s, Y &y_s) const |
apply a boundary intersections's jacobian for a linear problem. More... | |
Methods for the application of the jacobian for nonlinear problems | |
template<typename EG , typename LFSU , typename X , typename Z , typename LFSV , typename Y > | |
void | jacobian_apply_volume (const EG &eg, const LFSU &lfsu, const X &x, const Z &z, const LFSV &lfsv, Y &y) const |
Applies an element's jacobian to a vector for a nonlinear problem. More... | |
template<typename EG , typename LFSU , typename X , typename Z , typename LFSV , typename Y > | |
void | jacobian_apply_volume_post_skeleton (const EG &eg, const LFSU &lfsu, const X &x, const Z &z, const LFSV &lfsv, Y &y) const |
template<typename IG , typename LFSU , typename X , typename Z , typename LFSV , typename Y > | |
void | jacobian_apply_skeleton (const IG &ig, const LFSU &lfsu_s, const X &x_s, const Z &z_s, const LFSV &lfsv_s, const LFSU &lfsu_n, const X &x_n, const Z &z_n, const LFSV &lfsv_n, Y &y_s, Y &y_n) const |
template<typename IG , typename LFSU , typename X , typename Z , typename LFSV , typename Y > | |
void | jacobian_apply_boundary (const IG &ig, const LFSU &lfsu_s, const X &x_s, const Z &z_s, const LFSV &lfsv_s, Y &y_s) const |
apply a boundary intersections's jacobian for a nonlinear problem. More... | |
Methods to extract the jacobian | |
template<typename EG , typename LFSU , typename X , typename LFSV , typename LocalMatrix > | |
void | jacobian_volume (const EG &eg, const LFSU &lfsu, const X &x, const LFSV &lfsv, LocalMatrix &mat) const |
get an element's jacobian More... | |
template<typename EG , typename LFSU , typename X , typename LFSV , typename LocalMatrix > | |
void | jacobian_volume_post_skeleton (const EG &eg, const LFSU &lfsu, const X &x, const LFSV &lfsv, LocalMatrix &mat) const |
get an element's jacobian after the intersections have been handled More... | |
template<typename IG , typename LFSU , typename X , typename LFSV , typename LocalMatrix > | |
void | jacobian_skeleton (const IG &ig, const LFSU &lfsu_s, const X &x_s, const LFSV &lfsv_s, const LFSU &lfsu_n, const X &x_n, const LFSV &lfsv_n, LocalMatrix &mat_ss, LocalMatrix &mat_sn, LocalMatrix &mat_ns, LocalMatrix &mat_nn) const |
apply an internal intersections's jacobians More... | |
template<typename IG , typename LFSU , typename X , typename LFSV , typename LocalMatrix > | |
void | jacobian_boundary (const IG &ig, const LFSU &lfsu_s, const X &x_s, const LFSV &lfsv_s, LocalMatrix &mat_ss) const |
get a boundary intersections's jacobian More... | |
Flags selective assembly | |
enum | |
Whether to do selective assembly on the elements, i.e. whether or not skip_entity() should be called. | |
enum | |
Whether to do selective assembly on the intersections, i.e. whether or not skip_intersection() should be called. | |
Flags for the sparsity pattern | |
enum | |
Whether to assemble the pattern on the elements, i.e. whether or not pattern_volume() should be called. | |
enum | |
Whether to assemble the pattern on the elements after the skeleton has been handled, i.e. whether or not pattern_volume_post_skeleton() should be called. | |
enum | |
Whether to assemble the pattern on the interior intersections, i.e. whether or not pattern_skeleton() should be called. | |
enum | |
Whether to assemble the pattern on the boundary intersections, i.e. whether or not pattern_boundary() should be called. | |
Flags for the non-constant part of the residual and the jacobian | |
enum | |
Whether to call the local operator's alpha_volume(), jacobian_apply_volume() and jacobian_volume(). | |
enum | |
Whether to call the local operator's alpha_volume_post_skeleton(), jacobian_apply_volume_post_skeleton() and jacobian_volume_post_skeleton(). | |
enum | |
Whether to call the local operator's alpha_skeleton(), jacobian_apply_skeleton() and jacobian_skeleton(). | |
enum | |
Whether to call the local operator's alpha_boundary(), jacobian_apply_boundary() and jacobian_boundary(). | |
Flags for the constant part of the residual | |
enum | |
Whether to call the local operator's lambda_volume(). | |
enum | |
Whether to call the local operator's lambda_volume_post_skeleton(). | |
enum | |
Whether to call the local operator's lambda_skeleton(). | |
enum | |
Whether to call the local operator's lambda_boundary(). | |
Special flags | |
enum | |
Whether to visit the skeleton methods from both sides. | |
enum | |
Wheter the local operator describes a linear problem. | |
Detailed Description
Class to document the stationary local operator interface.
This class is for documentation purposes only. Each method given here is controlled by a flag. If the corresponding flag for a method is false, that method is never called and it is permissible for the method to be missing entirely from the local operator. The flags are those from LocalOperatorDefaultFlags.
There are five categories of methods, denoted by the first part of their name:
pattern_*
(): Methods for the sparsity pattern, controlled by thedoPattern*
flags,alpha_*
(): Methods for the non-constant parts of the residual, controlled by thedoAlpha*
flags,lambda_*
(): Methods for the constant parts of the residual, controlled by thedoLambda*
flags,jacobian_apply_*
(): Methods for the application of the jacobian, controlled by thedoAlpha*
flags, and finallyjacobian_*
(): Methods to extract the jacobian, controlled by thedoAlpha*
flags.
There are four classes of methods, denoted by the last part of their name:
*_volume
(): methods called on the entities before iterating over the intersections, controlled by thedo*Volume
flags,*_volume_post_skeleton
(): methods called on the entities after iterating over the intersections, controlled by thedo*VolumePostSkeleton
flags,*_skeleton
(): methods called on the interior intersections, controlled by thedo*Skeleton
flags, and finally*_boundary
(): methods called on the boundary intersections, controlled by thedo*Boundary
flags.
Not all combinations of categories and methods do actually exist.
To assemble the global sparsity pattern, residual or jacobian, the GridOperator iterates over the elements of the grid. For each element, it will call the appropriate *_volume
() method. Then it will iterate through the elements intersections and call the appropriate *_skeleton
() or *_boundary
() methods on the intersection. Finally it will call the appropriate *_volume_post_skeleton() method.
The special flag doSkeletonTwoSided controls whether each interior intersection is visited once or twice. If it is true, each intersection may be given to *_skeleton
() twice – the second time with the meaning of inside and outside exchanged. Note the "may": In the parallel case only interior entities are visited, so intersections at a processor boundary will only be visited once per processor in any case.
If doSkeletonTwoSided is false (the default), each intersection will only be visited once – and the orientation in which it is visited is left unspecified, except that the inside entity will always be an interior entity. Note that it will be visited once per process, so intersecions at processor boundaries are still visited twice when all processes are considered.
The alpha
and lambda
categories are a bit special in that the GridOperator uses them together – each time a method on the local operator should be called, it will first call the alpha_*
() method and then call the lambda_*
() method.
If the controlling flag for a method is false, the call to the method is omitted in such a way that the method does not even have to be present on the local operator. If both the do*Skeleton
and do*Boundary
flags are false, the iteration through the intersections is skipped.
Member Function Documentation
◆ alpha_boundary()
|
inline |
get a boundary intersections's contribution to alpha
- Parameters
-
ig IntersectionGeometry describing the intersection. lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. x_s Local position in the trial GridFunctionSpace in the inside entity. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. r_s Local part of the residual in the inside entity.
- Note
- It is permissible to include contributions of the residual which are independent of
x_s
here (they have to be omitted from lambda_boundary() in that case, of course). This is the difference to jacobian_apply_boundary(). -
The method should not clear
r_s
; it should just add its entries to it.
This method is controlled by the flag doAlphaBoundary. For a given element, it's calls happen intermingled with the calls to alpha_skeleton(), but after the call to alpha_volume() and before the call to alpha_volume_post_skeleton().
◆ alpha_skeleton()
|
inline |
get an internal intersections's contribution to alpha
- Parameters
-
ig IntersectionGeometry describing the intersection. lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. x_s Local position in the trial GridFunctionSpace in the inside entity. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. lfsu_n LocalFunctionSpace of the trial GridFunctionSpace in the outside entity. x_n Local position in the trial GridFunctionSpace in the outside entity. lfsv_n LocalFunctionSpace of the test GridFunctionSpace in the outside entity. r_s Local part of the residual in the inside entity. r_n Local part of the residual in the outside entity.
- Note
- It is permissible to include contributions of the residual which are independent of
x_s
andx_n
here (they have to be omitted from lambda_skeleton() in that case, of course). This is the difference to jacobian_apply_skeleton(). -
The method should not clear
r_s
andr_n
; it should just add its entries to them.
This method is controlled by the flag doAlphaSkeleton. For a given element, it's calls happen intermingled with the calls to alpha_boundary(), but after the call to alpha_volume() and before the call to alpha_volume_post_skeleton().
◆ alpha_volume()
|
inline |
get an element's contribution to alpha
- Parameters
-
eg ElementGeometry describing the entity. lfsu LocalFunctionSpace of the trial GridFunctionSpace. x Local position in the trial GridFunctionSpace. lfsv LocalFunctionSpace of the test GridFunctionSpace. r Local part of the residual.
- Note
- It is permissible to include contributions of the residual which are independent of
x
here (they have to be omitted from lambda_volume() in that case, of course). This is the difference to jacobian_apply_volume(). -
x
andr
are of type std::vector. -
The method should not clear
r
; it should just add its entries to it.
This method is controlled by the flag doAlphaVolume. For a given element, it is called before the alpha_skeleton() and/or alpha_boundary() methods are called (if they are called at all).
◆ alpha_volume_post_skeleton()
|
inline |
get an element's contribution to alpha after the intersections have been handled
- Parameters
-
eg ElementGeometry describing the entity. lfsu LocalFunctionSpace of the trial GridFunctionSpace. x Local position in the trial GridFunctionSpace. lfsv LocalFunctionSpace of the test GridFunctionSpace. r Local part of the residual.
- Note
- It is permissible to include contributions of the residual which are independent of
x
here (they have to be omitted from lambda_volume_post_skeleton() in that case, of course). This is the difference to jacobian_apply_volume_post_skeleton(). -
The method should not clear
r
; it should just add its entries to it.
This method is controlled by the flag doAlphaVolumePostSkeleton. For a given element, it is called after the alpha_skeleton() and/or alpha_boundary() methods are called (if they are called at all).
◆ jacobian_apply_boundary() [1/2]
|
inline |
apply a boundary intersections's jacobian for a nonlinear problem.
This method performs the following update:
\[ y = y + (\nabla_u \alpha_{\text{bnd}})\Big|_{x_s} z, \]
where \(u\) is the solution and the Jacobian is evaluated at \(u = x_s\).
- Parameters
-
ig IntersectionGeometry describing the intersection. lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. x_s Local position in the trial space for the inside entity at which to evaluate the Jacobian. z_s Local position in the trial space for the inside entity to which to apply the Jacobian. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. y_s Local part of the residual in the inside entity.
- Note
- This is different from alpha_boundary(), since the result will be linear in
x
, whereas alpha_boundary() may include contributions to the the residual which are constant inx
. -
The method should not clear
y_s
; it should just add its entries to it. -
This method assumes that the residual is not linear and the Jacobian is not constant; the Jacobian should be evaluated for the vector
x
and applied to the vectorz
.
This method is controlled by the flag doAlphaBoundary. For a given element, it's calls happen intermingled with the calls to jacobian_apply_skeleton(), but after the call to jacobian_apply_volume() and before the call to jacobian_apply_volume_post_skeleton().
◆ jacobian_apply_boundary() [2/2]
|
inline |
apply a boundary intersections's jacobian for a linear problem.
This method performs the following update:
\[ y = y + (\nabla_u \alpha_{\text{bnd}}) z, \]
where \(u\) is the solution and the residual is considered linear in \(u\).
- Parameters
-
ig IntersectionGeometry describing the intersection. lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. z_s Local position in the trial space for the inside entity to which to apply the Jacobian. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. y_s Local part of the residual in the inside entity.
- Note
- This is different from alpha_boundary(), since the result will be linear in
x
, whereas alpha_boundary() may include contributions to the the residual which are constant inx
. -
The method should not clear
y_s
; it should just add its entries to it. -
This method assumes that the residual is linear and the Jacobian is constant; as such, the method isn't passed a linearization point at which to evaluate the Jacobian, but only the vector
z_s
to which the Jacobian should be applied.
This method is controlled by the flag doAlphaBoundary. For a given element, it's calls happen intermingled with the calls to jacobian_apply_skeleton(), but after the call to jacobian_apply_volume() and before the call to jacobian_apply_volume_post_skeleton().
◆ jacobian_apply_skeleton() [1/2]
|
inline |
Applies an internal intersections's jacobians to the appropriate vectors for a nonlinear problem. This method performs the following updates:
\begin{align*} y_s & = y_s + (\nabla_{u_s} \alpha_{\text{skel},s})\Big|_{(x_s,x_n)} z_s + (\nabla_{u_n} \alpha_{\text{skel},s})\Big|_{(x_s,x_n)} z_n, \\ y_n & = y_n + (\nabla_{u_s} \alpha_{\text{skel},n})\Big|_{(x_s,x_n)} z_s + (\nabla_{u_n} \alpha_{\text{skel},n})\Big|_{(x_s,x_n)} z_n, \end{align*}
where \(u_s\) and \(u_n\) are the solution on the inside and the outside of the intersection and the Jacobian is evaluated at \(u = (x_s,x_n)\)$.
- Parameters
-
ig IntersectionGeometry describing the intersection. lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. x_s Local position in the trial space for the inside entity at which to evaluate the Jacobian. z_s Local position in the trial space for the inside entity to which to apply the Jacobian. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. lfsu_n LocalFunctionSpace of the trial GridFunctionSpace in the outside entity. x_n Local position in the trial space for the outside entity at which to evaluate the Jacobian. z_n Local position in the trial space for the outside entity to which to apply the Jacobian. lfsv_n LocalFunctionSpace of the test GridFunctionSpace in the outside entity. y_s Where to store the inside entity's result. y_n Where to store the outside entity's result.
- Note
- This is different from alpha_skeleton(), since the result will be linear in
z_s
andz_n
, whereas alpha_skeleton() may include contributions to the the residual which are constant inz_s
andz_n
. -
The method should not clear
y_s
andy_n
; it should just add its entries to them. -
This method assumes that the residual is not linear and the Jacobian is not constant; the Jacobian should be evaluated for the vector
x
and applied to the vectorz
.
This method is controlled by the flag doAlphaSkeleton. For a given element, it's calls happen intermingled with the calls to jacobian_apply_boundary(), but after the call to jacobian_apply_volume() and before the call to jacobian_apply_volume_post_skeleton().
◆ jacobian_apply_skeleton() [2/2]
|
inline |
Applies an internal intersections's jacobians to the appropriate vectors for a linear problem. This method performs the following updates:
\begin{align*} y_s & = y_s + (\nabla_{u_s} \alpha_{\text{skel},s}) z_s + (\nabla_{u_n} \alpha_{\text{skel},s}) z_n, \\ y_n & = y_n + (\nabla_{u_s} \alpha_{\text{skel},n}) z_s + (\nabla_{u_n} \alpha_{\text{skel},n}) z_n, \end{align*}
where \(u_s\) and \(u_n\) are the solution on the inside and outside of the intersection and the residual is considered linear in \(u_{\{s,n\}}\).
- Parameters
-
ig IntersectionGeometry describing the intersection. lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. z_s Local position in the trial space for the inside entity to which to apply the Jacobian. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. lfsu_n LocalFunctionSpace of the trial GridFunctionSpace in the outside entity. z_n Local position in the trial space for the outside entity to which to apply the Jacobian. lfsv_n LocalFunctionSpace of the test GridFunctionSpace in the outside entity. y_s Where to store the inside entity's result. y_n Where to store the outside entity's result.
- Note
- This is different from alpha_skeleton(), since the result will be linear in
z_s
andz_n
, whereas alpha_skeleton() may include contributions to the the residual which are constant inz_s
andz_n
. -
The method should not clear
y_s
andy_n
; it should just add its entries to them. -
This method assumes that the residual is linear and the Jacobian is constant; as such, the method isn't passed linearization points at which to evaluate the Jacobian, but only the vectors
z_s
andz_n
to which the Jacobian should be applied.
This method is controlled by the flag doAlphaSkeleton. For a given element, it's calls happen intermingled with the calls to jacobian_apply_boundary(), but after the call to jacobian_apply_volume() and before the call to jacobian_apply_volume_post_skeleton().
◆ jacobian_apply_volume() [1/2]
|
inline |
Applies an element's jacobian to a vector for a nonlinear problem.
This method performs the following update:
\[ y = y + (\nabla_u \alpha_{\text{vol}})\Big|_x z, \]
where \(u\) is the solution and the Jacobian is evaluated at \(u=x\).
- Parameters
-
eg ElementGeometry describing the entity. lfsu LocalFunctionSpace of the trial GridFunctionSpace. x Local position in the trial space at which to evaluate the Jacobian. z Local position in the trial space to which to apply the Jacobian. lfsv LocalFunctionSpace of the test GridFunctionSpace. y Where to store the result.
- Note
- This is different from alpha_volume(), since the result will be linear in
z
, whereas alpha_volume() may include contributions to the the residual which are constant inz
. -
The method should not clear
y
; it should just add its entries to it. -
This method assumes that the residual is not linear and the Jacobian is not constant; the Jacobian should be evaluated for the vector
x
and applied to the vectorz
. - Your linear operator should export the flag isLinear=false in this case, see LocalOperatorDefaultFlags.
This method is controlled by the flag doAlphaVolume. For a given element, it is called before the jacobian_apply_skeleton() and/or jacobian_apply_boundary() methods are called (if they are called at all).
◆ jacobian_apply_volume() [2/2]
|
inline |
Applies an element's jacobian to a vector for a linear problem.
This method performs the following update:
\[ y = y + (\nabla_u \alpha_{\text{vol}}) z, \]
where \(u\) is the solution and the residual is considered linear in \(u\).
- Parameters
-
eg ElementGeometry describing the entity. lfsu LocalFunctionSpace of the trial GridFunctionSpace. z Local position in the trial space to which to apply the Jacobian. lfsv LocalFunctionSpace of the test GridFunctionSpace. y Where to store the result.
- Note
- This is different from alpha_volume(), since the result will be linear in
z
, whereas alpha_volume() may include contributions to the the residual which are constant inz
. -
The method should not clear
y
; it should just add its entries to it. -
This method assumes that the residual is linear and the Jacobian is constant; as such, the method isn't passed a linearization point at which to evaluate the Jacobian, but only the vector
z
to which the Jacobian should be applied. - Your operator should export the flag isLinear=true in this case. This is done by default.
This method is controlled by the flag doAlphaVolume. For a given element, it is called before the jacobian_apply_skeleton() and/or jacobian_apply_boundary() methods are called (if they are called at all).
◆ jacobian_apply_volume_post_skeleton() [1/2]
|
inline |
Applies an element's jacobian to a vector for a nonlinear problem after the intersections have been handled. This method performs the following update:
\[ y = y + (\nabla_u \alpha_{\text{vol\_post\_skel}})\Big|_x z, \]
where \(u\) is the solution and the Jacobian is evaluated at \(u=x\).
- Parameters
-
eg ElementGeometry describing the entity. lfsu LocalFunctionSpace of the trial GridFunctionSpace. x Local position in the trial space at which to evaluate the Jacobian. z Local position in the trial space to which to apply the Jacobian. lfsv LocalFunctionSpace of the test GridFunctionSpace. y Where to store the result.
- Note
- This is different from alpha_volume_post_skeleton(), since the result will be linear in
z
, whereas alpha_volume_post_skeleton() may include contributions to the the residual which are constant inz
. -
The method should not clear
y
; it should just add its entries to it. -
This method assumes that the residual is not linear and the Jacobian is not constant; the Jacobian should be evaluated for the vector
x
and applied to the vectorz
.
This method is controlled by the flag doAlphaVolumePostSkeleton. For a given element, it is called after the jacobian_apply_skeleton() and/or jacobian_apply_boundary() methods are called (if they are called at all).
◆ jacobian_apply_volume_post_skeleton() [2/2]
|
inline |
Applies an element's jacobian to a vector for a linear problem after the intersections have been handled. This method performs the following update:
\[ y = y + (\nabla_u \alpha_{\text{vol\_post\_skel}}) z, \]
where \(u\) is the solution and the residual is considered linear in \(u\).
- Parameters
-
eg ElementGeometry describing the entity. lfsu LocalFunctionSpace of the trial GridFunctionSpace. z Local position in the trial space to which to apply the Jacobian. lfsv LocalFunctionSpace of the test GridFunctionSpace. y Where to store the result.
- Note
- This is different from alpha_volume_post_skeleton(), since the result will be linear in
z
, whereas alpha_volume_post_skeleton() may include contributions to the the residual which are constant inz
. -
The method should not clear
y
; it should just add its entries to it. -
This method assumes that the residual is linear and the Jacobian is constant; as such, the method isn't passed a linearization point at which to evaluate the Jacobian, but only the vector
z
to which the Jacobian should be applied.
This method is controlled by the flag doAlphaVolumePostSkeleton. For a given element, it is called after the jacobian_apply_skeleton() and/or jacobian_apply_boundary() methods are called (if they are called at all).
◆ jacobian_boundary()
|
inline |
get a boundary intersections's jacobian
- Parameters
-
ig IntersectionGeometry describing the intersection. lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. x_s Local position in the trial GridFunctionSpace in the inside entity. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. mat_ss Where to store the contribution to the inside entity's jacobian.
- Note
- The method should not clear
mat_ss
; it should just add its entries to it.
This method is controlled by the flag doAlphaBoundary. For a given element, it's calls happen intermingled with the calls to jacobian_skeleton(), but after the call to jacobian_volume() and before the call to jacobian_volume_post_skeleton().
◆ jacobian_skeleton()
|
inline |
apply an internal intersections's jacobians
- Parameters
-
ig IntersectionGeometry describing the intersection. lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. x_s Local position in the trial GridFunctionSpace in the inside entity. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. lfsu_n LocalFunctionSpace of the trial GridFunctionSpace in the outside entity. x_n Local position in the trial GridFunctionSpace in the outside entity. lfsv_n LocalFunctionSpace of the test GridFunctionSpace in the outside entity. mat_ss Where to store the contribution to the inside entity's jacobian. mat_sn Where to store the contribution to the interaction jacobian between the inside and the outside entity. mat_ns Where to store the contribution to the interaction jacobian between the outside and the inside entity. mat_nn Where to store the contribution to the outside entity's jacobian.
- Note
- The method should not clear
mat_ss
,mat_sn
,mat_ns
,mat_nn
; it should just add its entries to them.
This method is controlled by the flag doAlphaSkeleton. For a given element, it's calls happen intermingled with the calls to jacobian_boundary(), but after the call to jacobian_volume() and before the call to jacobian_volume_post_skeleton().
◆ jacobian_volume()
|
inline |
get an element's jacobian
- Parameters
-
eg ElementGeometry describing the entity. lfsu LocalFunctionSpace of the trial GridFunctionSpace. x Local position in the trial GridFunctionSpace. lfsv LocalFunctionSpace of the test GridFunctionSpace. mat Where to store the contribution to the jacobian.
- Note
- The method should not clear
mat
; it should just add its entries to it.
This method is controlled by the flag doAlphaVolume. For a given element, it is called before the jacobian_skeleton() and/or jacobian_boundary() methods are called (if they are called at all).
◆ jacobian_volume_post_skeleton()
|
inline |
get an element's jacobian after the intersections have been handled
- Parameters
-
eg ElementGeometry describing the entity. lfsu LocalFunctionSpace of the trial GridFunctionSpace. x Local position in the trial GridFunctionSpace. lfsv LocalFunctionSpace of the test GridFunctionSpace. mat Where to store the contribution to the jacobian.
- Note
- The method should not clear
mat
; it should just add its entries to it.
This method is controlled by the flag doAlphaVolumePostSkeleton. For a given element, it is called after the jacobian_skeleton() and/or jacobian_boundary() methods are called (if they are called at all).
◆ lambda_boundary()
|
inline |
get a boundary intersections's contribution to lambda
- Parameters
-
ig IntersectionGeometry describing the intersection. inside entity. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. r_s Local part of the residual in the inside entity.
- Note
- The method should not clear
r_s
; it should just add its entries to it.
This method is controlled by the flag doLambdaBoundary. For a given element, it's calls happen intermingled with the calls to lambda_skeleton(), but after the call to lambda_volume() and before the call to lambda_volume_post_skeleton().
◆ lambda_skeleton()
|
inline |
get an internal intersections's contribution to lambda
- Parameters
-
ig IntersectionGeometry describing the intersection. inside entity. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. lfsv_n LocalFunctionSpace of the test GridFunctionSpace in the outside entity. r_s Local part of the residual in the inside entity. r_n Local part of the residual in the outside entity.
- Note
- The method should not clear
r_s
andr_n
; it should just add its entries to them.
This method is controlled by the flag doLambdaSkeleton. For a given element, it's calls happen intermingled with the calls to lambda_boundary(), but after the call to lambda_volume() and before the call to lambda_volume_post_skeleton().
◆ lambda_volume()
|
inline |
get an element's contribution to lambda
- Parameters
-
eg ElementGeometry describing the entity. lfsv LocalFunctionSpace of the test GridFunctionSpace. r Local part of the residual.
- Note
- The method should not clear
r
; it should just add its entries to it.
This method is controlled by the flag doLambdaVolume. For a given element, it is called before the lambda_skeleton() and/or lambda_boundary() methods are called (if they are called at all).
◆ lambda_volume_post_skeleton()
|
inline |
get an element's contribution to lambda after the intersections have been handled
- Parameters
-
eg ElementGeometry describing the entity. lfsv LocalFunctionSpace of the test GridFunctionSpace. r Local part of the residual.
- Note
- The method should not clear
r
; it should just add its entries to it.
This method is controlled by the flag doLambdaVolumePostSkeleton. For a given element, it is called after the lambda_skeleton() and/or lambda_boundary() methods are called (if they are called at all).
◆ pattern_boundary()
|
inline |
get a boundary intersection's contribution to the sparsity pattern
- Parameters
-
lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. pattern_ss Local sparsity pattern for the inside entity.
- Note
- The method should not clear the pattern; it should just add its entries to it.
This method is controlled by the flag doPatternBoundary. For a given element, it's calls happen intermingled with the calls to pattern_skeleton(), but after the call to pattern_volume() and before the call to pattern_volume_post_skeleton().
◆ pattern_skeleton()
|
inline |
get an internal intersection's contribution to the sparsity pattern
- Parameters
-
lfsu_s LocalFunctionSpace of the trial GridFunctionSpace in the inside entity. lfsv_s LocalFunctionSpace of the test GridFunctionSpace in the inside entity. lfsu_n LocalFunctionSpace of the trial GridFunctionSpace in the outside entity. lfsv_n LocalFunctionSpace of the test GridFunctionSpace in the outside entity. pattern_sn Local sparsity pattern. pattern_ns Local sparsity pattern.
- Note
- The method should not clear the patterns; it should just add its entries to them.
This method is controlled by the flag doPatternSkeleton. For a given element, it's calls happen intermingled with the calls to pattern_boundary(), but after the call to pattern_volume() and before the call to pattern_volume_post_skeleton().
◆ pattern_volume()
|
inline |
get an element's contribution to the sparsity pattern
- Parameters
-
lfsu LocalFunctionSpace of the trial GridFunctionSpace. lfsv LocalFunctionSpace of the test GridFunctionSpace. pattern Local sparsity pattern.
- Note
- The method should not clear the pattern; it should just add its entries to it.
This method is controlled by the flag doPatternVolume. For a given element, it is called before the pattern_skeleton() and/or pattern_boundary() methods are called (if they are called at all).
◆ pattern_volume_post_skeleton()
|
inline |
get an element's contribution to the sparsity pattern after the intersections have been handled
- Parameters
-
lfsu LocalFunctionSpace of the trial GridFunctionSpace. lfsv LocalFunctionSpace of the test GridFunctionSpace. pattern Local sparsity pattern.
- Note
- The method should not clear the pattern; it should just add its entries to it.
This method is controlled by the flag doPatternVolume. For a given element, it is called before the pattern_skeleton() and/or pattern_boundary() methods are called (if they are called at all).
◆ skip_entity()
|
inline |
whether to assembly methods associated with a given entity
- Parameters
-
eg ElementGeometry describing the entity.
This method is controlled by the flag doSkipEntity. For a given element, it is called before the any other local method. If return value is true, all volume, skeleton, and boundary methods are skipped.
◆ skip_intersection()
|
inline |
whether to assembly methods associated with a given intersection
- Parameters
-
ig IntersectionGeometry describing the intersection.
This method is controlled by the flag doSkipIntersection. For a given intersection, it is called after local volume methods and before any skeleton and boundary methods. If return value is true, all skeleton, and boundary methods are skipped.
The documentation for this class was generated from the following file:
- dune/pdelab/localoperator/interface.hh