deal.II version GIT relicensing-2076-g6b43d56a25 2024-11-02 12:30:00+00:00
|
#include <deal.II/lac/block_sparse_matrix.h>
Static Public Member Functions | |
static ::ExceptionBase & | ExcBlockDimensionMismatch () |
static ::ExceptionBase & | ExcIncompatibleRowNumbers (int arg1, int arg2, int arg3, int arg4) |
static ::ExceptionBase & | ExcIncompatibleColNumbers (int arg1, int arg2, int arg3, int arg4) |
static ::ExceptionBase & | ExcInUse (int arg1, std::string arg2, std::string arg3) |
static ::ExceptionBase & | ExcNoSubscriber (std::string arg1, std::string arg2) |
Protected Member Functions | |
void | collect_sizes () |
void | vmult_block_block (BlockVectorType &dst, const BlockVectorType &src) const |
void | vmult_block_nonblock (BlockVectorType &dst, const VectorType &src) const |
void | vmult_nonblock_block (VectorType &dst, const BlockVectorType &src) const |
void | vmult_nonblock_nonblock (VectorType &dst, const VectorType &src) const |
void | Tvmult_block_block (BlockVectorType &dst, const BlockVectorType &src) const |
void | Tvmult_block_nonblock (BlockVectorType &dst, const VectorType &src) const |
void | Tvmult_nonblock_block (VectorType &dst, const BlockVectorType &src) const |
void | Tvmult_nonblock_nonblock (VectorType &dst, const VectorType &src) const |
void | prepare_add_operation () |
void | prepare_set_operation () |
Protected Attributes | |
BlockIndices | row_block_indices |
BlockIndices | column_block_indices |
Table< 2, ObserverPointer< BlockType, BlockMatrixBase< SparseMatrix< number > > > > | sub_objects |
Private Types | |
using | map_value_type = decltype(counter_map)::value_type |
using | map_iterator = decltype(counter_map)::iterator |
Private Member Functions | |
void | check_no_subscribers () const noexcept |
Private Attributes | |
ObserverPointer< const BlockSparsityPattern, BlockSparseMatrix< number > > | sparsity_pattern |
TemporaryData | temporary_data |
std::atomic< unsigned int > | counter |
std::map< std::string, unsigned int > | counter_map |
std::vector< std::atomic< bool > * > | validity_pointers |
const std::type_info * | object_info |
Static Private Attributes | |
static std::mutex | mutex |
Blocked sparse matrix based on the SparseMatrix class. This class implements the functions that are specific to the SparseMatrix base objects for a blocked sparse matrix, and leaves the actual work relaying most of the calls to the individual blocks to the functions implemented in the base class. See there also for a description of when this class is useful.
Definition at line 49 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::BaseClass = BlockMatrixBase<SparseMatrix<number> > |
Typedef the base class for simpler access to its own alias.
Definition at line 55 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::BlockType = typename BaseClass::BlockType |
Typedef the type of the underlying matrix.
Definition at line 60 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::value_type = typename BaseClass::value_type |
Import the alias from the base class.
Definition at line 65 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::pointer = typename BaseClass::pointer |
Definition at line 66 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::const_pointer = typename BaseClass::const_pointer |
Definition at line 67 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::reference = typename BaseClass::reference |
Definition at line 68 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::const_reference = typename BaseClass::const_reference |
Definition at line 69 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::size_type = typename BaseClass::size_type |
Definition at line 70 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::iterator = typename BaseClass::iterator |
Definition at line 71 of file block_sparse_matrix.h.
using BlockSparseMatrix< number >::const_iterator = typename BaseClass::const_iterator |
Definition at line 72 of file block_sparse_matrix.h.
|
inherited |
Definition at line 362 of file block_matrix_base.h.
|
privateinherited |
The data type used in counter_map.
Definition at line 238 of file enable_observer_pointer.h.
|
privateinherited |
The iterator type used in counter_map.
Definition at line 243 of file enable_observer_pointer.h.
|
default |
Constructor; initializes the matrix to be empty, without any structure, i.e. the matrix is not usable at all. This constructor is therefore only useful for matrices which are members of a class. All other matrices should be created at a point in the data flow where all necessary information is available.
You have to initialize the matrix before usage with reinit(BlockSparsityPattern). The number of blocks per row and column are then determined by that function.
BlockSparseMatrix< number >::BlockSparseMatrix | ( | const BlockSparsityPattern & | sparsity | ) |
Constructor. Takes the given matrix sparsity structure to represent the sparsity pattern of this matrix. You can change the sparsity pattern later on by calling the reinit() function.
This constructor initializes all sub-matrices with the sub-sparsity pattern within the argument.
You have to make sure that the lifetime of the sparsity structure is at least as long as that of this matrix or as long as reinit() is not called with a new sparsity structure.
|
overridevirtual |
Destructor.
BlockSparseMatrix & BlockSparseMatrix< number >::operator= | ( | const BlockSparseMatrix< number > & | ) |
Pseudo copy operator only copying empty objects. The sizes of the block matrices need to be the same.
|
inline |
This operator assigns a scalar to a matrix. Since this does usually not make much sense (should we set all matrix entries to this value? Only the nonzero entries of the sparsity pattern?), this operation is only allowed if the actual value to be assigned is zero. This operator only exists to allow for the obvious notation matrix=0
, which sets all elements of the matrix to zero, but keep the sparsity pattern previously used.
Definition at line 382 of file block_sparse_matrix.h.
void BlockSparseMatrix< number >::clear | ( | ) |
Release all memory and return to a state just like after having called the default constructor. It also forgets the sparsity pattern it was previously tied to.
This calls SparseMatrix::clear on all sub-matrices and then resets this object to have no blocks at all.
|
virtual |
Reinitialize the sparse matrix with the given sparsity pattern. The latter tells the matrix how many nonzero elements there need to be reserved.
Basically, this function only calls SparseMatrix::reinit() of the sub-matrices with the block sparsity patterns of the parameter.
You have to make sure that the lifetime of the sparsity structure is at least as long as that of this matrix or as long as reinit(const SparsityPattern &) is not called with a new sparsity structure.
The elements of the matrix are set to zero by this function.
bool BlockSparseMatrix< number >::empty | ( | ) | const |
Return whether the object is empty. It is empty if either both dimensions are zero or no BlockSparsityPattern is associated.
size_type BlockSparseMatrix< number >::get_row_length | ( | const size_type | row | ) | const |
Return the number of entries in a specific row.
size_type BlockSparseMatrix< number >::n_nonzero_elements | ( | ) | const |
Return the number of nonzero elements of this matrix. Actually, it returns the number of entries in the sparsity pattern; if any of the entries should happen to be zero, it is counted anyway.
size_type BlockSparseMatrix< number >::n_actually_nonzero_elements | ( | const double | threshold = 0.0 | ) | const |
Return the number of actually nonzero elements. Just counts the number of actually nonzero elements (with absolute value larger than threshold) of all the blocks.
const BlockSparsityPattern & BlockSparseMatrix< number >::get_sparsity_pattern | ( | ) | const |
Return a (constant) reference to the underlying sparsity pattern of this matrix.
Though the return value is declared const
, you should be aware that it may change if you call any nonconstant function of objects which operate on it.
std::size_t BlockSparseMatrix< number >::memory_consumption | ( | ) | const |
Determine an estimate for the memory consumption (in bytes) of this object.
|
inline |
Matrix-vector multiplication: let \(dst = M*src\) with \(M\) being this matrix.
Definition at line 398 of file block_sparse_matrix.h.
|
inline |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block column.
Definition at line 409 of file block_sparse_matrix.h.
|
inline |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block row.
Definition at line 420 of file block_sparse_matrix.h.
|
inline |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block.
Definition at line 431 of file block_sparse_matrix.h.
|
inline |
Matrix-vector multiplication: let \(dst = M^T*src\) with \(M\) being this matrix. This function does the same as vmult() but takes the transposed matrix.
Definition at line 442 of file block_sparse_matrix.h.
|
inline |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block row.
Definition at line 453 of file block_sparse_matrix.h.
|
inline |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block column.
Definition at line 464 of file block_sparse_matrix.h.
|
inline |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block.
Definition at line 475 of file block_sparse_matrix.h.
|
inline |
Apply the Jacobi preconditioner, which multiplies every element of the src
vector by the inverse of the respective diagonal element and multiplies the result with the relaxation parameter omega
.
All diagonal blocks must be square matrices for this operation.
Definition at line 486 of file block_sparse_matrix.h.
|
inline |
Apply the Jacobi preconditioner to a simple vector.
The matrix must be a single square block for this.
Definition at line 507 of file block_sparse_matrix.h.
void BlockSparseMatrix< number >::print_formatted | ( | std::ostream & | out, |
const unsigned int | precision = 3 , |
||
const bool | scientific = true , |
||
const unsigned int | width = 0 , |
||
const char * | zero_string = " " , |
||
const double | denominator = 1. , |
||
const char * | separator = " " |
||
) | const |
Print the matrix in the usual format, i.e., as a matrix and not as a list of nonzero elements. For better readability, elements not in the matrix are displayed as empty space, while matrix elements which are explicitly set to zero are displayed as such.
The parameters allow for a flexible setting of the output format: precision
and scientific
are used to determine the number format, where scientific = false
means fixed point notation. A zero entry for width
makes the function compute a width, but it may be changed to a positive value, if output is crude.
Additionally, a character for an empty value may be specified in zero_string
, and a character to separate row entries can be set in separator
.
Finally, the whole matrix can be multiplied with a common denominator
to produce more readable output, even integers.
|
inherited |
Copy the matrix given as argument into the current object.
Copying matrices is an expensive operation that we do not want to happen by accident through compiler generated code for operator=
. (This would happen, for example, if one accidentally declared a function argument of the current type by value rather than by reference.) The functionality of copying matrices is implemented in this member function instead. All copy operations of objects of this type therefore require an explicit function call.
The source matrix may be a matrix of arbitrary type, as long as its data type is convertible to the data type of this matrix.
The function returns a reference to this
.
|
inherited |
Access the block with the given coordinates.
|
inherited |
Access the block with the given coordinates. Version for constant objects.
|
inherited |
Return the dimension of the codomain (or range) space. Note that the matrix is of dimension \(m \times n\).
|
inherited |
Return the dimension of the domain space. Note that the matrix is of dimension \(m \times n\).
|
inherited |
Return the number of blocks in a column. Returns zero if no sparsity pattern is presently associated to this matrix.
|
inherited |
Return the number of blocks in a row. Returns zero if no sparsity pattern is presently associated to this matrix.
|
inherited |
Set the element (i,j)
to value
. Throws an error if the entry does not exist or if value
is not a finite number. Still, it is allowed to store zero values in non-existent fields.
|
inherited |
Set all elements given in a FullMatrix into the sparse matrix locations given by indices
. In other words, this function writes the elements in full_matrix
into the calling matrix, using the local-to-global indexing specified by indices
for both the rows and the columns of the matrix. This function assumes a quadratic sparse matrix and a quadratic full_matrix, the usual situation in FE calculations.
The optional parameter elide_zero_values
can be used to specify whether zero values should be set anyway or they should be filtered away (and not change the previous content in the respective element if it exists). The default value is false
, i.e., even zero values are treated.
|
inherited |
Same function as before, but now including the possibility to use rectangular full_matrices and different local-to-global indexing on rows and columns, respectively.
|
inherited |
Set several elements in the specified row of the matrix with column indices as given by col_indices
to the respective value.
The optional parameter elide_zero_values
can be used to specify whether zero values should be set anyway or they should be filtered away (and not change the previous content in the respective element if it exists). The default value is false
, i.e., even zero values are treated.
|
inherited |
Set several elements to values given by values
in a given row in columns given by col_indices into the sparse matrix.
The optional parameter elide_zero_values
can be used to specify whether zero values should be inserted anyway or they should be filtered away. The default value is false
, i.e., even zero values are inserted/replaced.
|
inherited |
Add value
to the element (i,j). Throws an error if the entry does not exist or if value
is not a finite number. Still, it is allowed to store zero values in non-existent fields.
|
inherited |
Add all elements given in a FullMatrix<double> into sparse matrix locations given by indices
. In other words, this function adds the elements in full_matrix
to the respective entries in calling matrix, using the local-to-global indexing specified by indices
for both the rows and the columns of the matrix. This function assumes a quadratic sparse matrix and a quadratic full_matrix, the usual situation in FE calculations.
The optional parameter elide_zero_values
can be used to specify whether zero values should be added anyway or these should be filtered away and only non-zero data is added. The default value is true
, i.e., zero values won't be added into the matrix.
|
inherited |
Same function as before, but now including the possibility to use rectangular full_matrices and different local-to-global indexing on rows and columns, respectively.
|
inherited |
Set several elements in the specified row of the matrix with column indices as given by col_indices
to the respective value.
The optional parameter elide_zero_values
can be used to specify whether zero values should be added anyway or these should be filtered away and only non-zero data is added. The default value is true
, i.e., zero values won't be added into the matrix.
|
inherited |
Add an array of values given by values
in the given global matrix row at columns specified by col_indices in the sparse matrix.
The optional parameter elide_zero_values
can be used to specify whether zero values should be added anyway or these should be filtered away and only non-zero data is added. The default value is true
, i.e., zero values won't be added into the matrix.
|
inherited |
Add matrix
scaled by factor
to this matrix, i.e. the matrix factor*matrix
is added to this
. If the sparsity pattern of the calling matrix does not contain all the elements in the sparsity pattern of the input matrix, this function will throw an exception.
Depending on MatrixType, however, additional restrictions might arise. Some sparse matrix formats require matrix
to be based on the same sparsity pattern as the calling matrix.
|
inherited |
Return the value of the entry (i,j). This may be an expensive operation and you should always take care where to call this function. In order to avoid abuse, this function throws an exception if the wanted element does not exist in the matrix.
|
inherited |
This function is mostly like operator()() in that it returns the value of the matrix entry (i,j)
. The only difference is that if this entry does not exist in the sparsity pattern, then instead of raising an exception, zero is returned. While this may be convenient in some cases, note that it is simple to write algorithms that are slow compared to an optimal solution, since the sparsity of the matrix is not used.
|
inherited |
Return the main diagonal element in the ith row. This function throws an error if the matrix is not quadratic and also if the diagonal blocks of the matrix are not quadratic.
This function is considerably faster than the operator()(), since for quadratic matrices, the diagonal entry may be the first to be stored in each row and access therefore does not involve searching for the right column number.
|
inherited |
Call the compress() function on all the subblocks of the matrix.
See Compressing distributed objects for more information.
|
inherited |
Multiply the entire matrix by a fixed factor.
|
inherited |
Divide the entire matrix by a fixed factor.
|
inherited |
Adding Matrix-vector multiplication. Add \(M*src\) on \(dst\) with \(M\) being this matrix.
|
inherited |
Adding Matrix-vector multiplication. Add MTsrc to dst with M being this matrix. This function does the same as vmult_add() but takes the transposed matrix.
|
inherited |
Return the norm of the vector v with respect to the norm induced by this matrix, i.e. vTMv). This is useful, e.g. in the finite element context, where the LT-norm of a function equals the matrix norm with respect to the mass matrix of the vector representing the nodal values of the finite element function. Note that even though the function's name might suggest something different, for historic reasons not the norm but its square is returned, as defined above by the scalar product.
Obviously, the matrix needs to be square for this operation.
|
inherited |
Return the frobenius norm of the matrix, i.e. the square root of the sum of squares of all entries in the matrix.
|
inherited |
Compute the matrix scalar product \(\left(u,Mv\right)\).
|
inherited |
Compute the residual r=b-Ax. Write the residual into dst
.
|
inherited |
Print the matrix to the given stream, using the format (line,col) value
, i.e. one nonzero entry of the matrix per line. The optional flag outputs the sparsity pattern in a different style according to the underlying sparse matrix type.
|
inherited |
Iterator starting at the first entry.
|
inherited |
Iterator starting at the first entry of row r
.
|
inherited |
Iterator starting at the first entry.
|
inherited |
Iterator starting at the first entry of row r
.
|
inherited |
Final iterator.
|
inherited |
Final iterator of row r
.
|
inherited |
Final iterator.
|
inherited |
Final iterator of row r
.
|
inherited |
Return a reference to the underlying BlockIndices data of the rows.
|
inherited |
Return a reference to the underlying BlockIndices data of the columns.
|
staticinherited |
Exception
|
staticinherited |
Exception
|
protectedinherited |
This function collects the sizes of the sub-objects and stores them in internal arrays, in order to be able to relay global indices into the matrix to indices into the subobjects. You must call this function each time after you have changed the size of the sub-objects.
Derived classes should call this function whenever the size of the sub-objects has changed and the X_block_indices
arrays need to be updated.
Note that this function is not public since not all derived classes need to export its interface. For example, for the usual deal.II SparseMatrix class, the sizes are implicitly determined whenever reinit() is called, and individual blocks cannot be resized. For that class, this function therefore does not have to be public. On the other hand, for the PETSc classes, there is no associated sparsity pattern object that determines the block sizes, and for these the function needs to be publicly available. These classes therefore export this function.
|
protectedinherited |
Matrix-vector multiplication: let \(dst = M*src\) with \(M\) being this matrix.
Due to problems with deriving template arguments between the block and non-block versions of the vmult/Tvmult functions, the actual functions are implemented in derived classes, with implementations forwarding the calls to the implementations provided here under a unique name for which template arguments can be derived by the compiler.
|
protectedinherited |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block column.
Due to problems with deriving template arguments between the block and non-block versions of the vmult/Tvmult functions, the actual functions are implemented in derived classes, with implementations forwarding the calls to the implementations provided here under a unique name for which template arguments can be derived by the compiler.
|
protectedinherited |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block row.
Due to problems with deriving template arguments between the block and non-block versions of the vmult/Tvmult functions, the actual functions are implemented in derived classes, with implementations forwarding the calls to the implementations provided here under a unique name for which template arguments can be derived by the compiler.
|
protectedinherited |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block.
Due to problems with deriving template arguments between the block and non-block versions of the vmult/Tvmult functions, the actual functions are implemented in derived classes, with implementations forwarding the calls to the implementations provided here under a unique name for which template arguments can be derived by the compiler.
|
protectedinherited |
Matrix-vector multiplication: let \(dst = M^T*src\) with \(M\) being this matrix. This function does the same as vmult() but takes the transposed matrix.
Due to problems with deriving template arguments between the block and non-block versions of the vmult/Tvmult functions, the actual functions are implemented in derived classes, with implementations forwarding the calls to the implementations provided here under a unique name for which template arguments can be derived by the compiler.
|
protectedinherited |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block row.
Due to problems with deriving template arguments between the block and non-block versions of the vmult/Tvmult functions, the actual functions are implemented in derived classes, with implementations forwarding the calls to the implementations provided here under a unique name for which template arguments can be derived by the compiler.
|
protectedinherited |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block column.
Due to problems with deriving template arguments between the block and non-block versions of the vmult/Tvmult functions, the actual functions are implemented in derived classes, with implementations forwarding the calls to the implementations provided here under a unique name for which template arguments can be derived by the compiler.
|
protectedinherited |
Matrix-vector multiplication. Just like the previous function, but only applicable if the matrix has only one block.
Due to problems with deriving template arguments between the block and non-block versions of the vmult/Tvmult functions, the actual functions are implemented in derived classes, with implementations forwarding the calls to the implementations provided here under a unique name for which template arguments can be derived by the compiler.
|
protectedinherited |
Some matrix types, in particular PETSc, need to synchronize set and add operations. This has to be done for all matrices in the BlockMatrix. This routine prepares adding of elements by notifying all blocks. Called by all internal routines before adding elements.
|
protectedinherited |
Notifies all blocks to let them prepare for setting elements, see prepare_add_operation().
|
inherited |
Subscribes a user of the object by storing the pointer validity
. The subscriber may be identified by text supplied as identifier
.
Definition at line 131 of file enable_observer_pointer.cc.
|
inherited |
Unsubscribes a user from the object.
identifier
and the validity
pointer must be the same as the one supplied to subscribe(). Definition at line 151 of file enable_observer_pointer.cc.
|
inlineinherited |
Return the present number of subscriptions to this object. This allows to use this class for reference counted lifetime determination where the last one to unsubscribe also deletes the object.
Definition at line 322 of file enable_observer_pointer.h.
|
inlineinherited |
List the subscribers to the input stream
.
Definition at line 339 of file enable_observer_pointer.h.
|
inherited |
List the subscribers to deallog
.
Definition at line 199 of file enable_observer_pointer.cc.
Read or write the data of this object to or from a stream for the purpose of serialization using the BOOST serialization library.
This function does not actually serialize any of the member variables of this class. The reason is that what this class stores is only who subscribes to this object, but who does so at the time of storing the contents of this object does not necessarily have anything to do with who subscribes to the object when it is restored. Consequently, we do not want to overwrite the subscribers at the time of restoring, and then there is no reason to write the subscribers out in the first place.
Definition at line 331 of file enable_observer_pointer.h.
|
privatenoexceptinherited |
Check that there are no objects subscribing to this object. If this check passes then it is safe to destroy the current object. It this check fails then this function will either abort or print an error message to deallog (by using the AssertNothrow mechanism), but will not throw an exception.
Definition at line 53 of file enable_observer_pointer.cc.
|
private |
Pointer to the block sparsity pattern used for this matrix. In order to guarantee that it is not deleted while still in use, we subscribe to it using the ObserverPointer class.
Definition at line 370 of file block_sparse_matrix.h.
|
protectedinherited |
Index arrays for rows and columns.
Definition at line 846 of file block_matrix_base.h.
|
protectedinherited |
Definition at line 847 of file block_matrix_base.h.
|
protectedinherited |
Array of sub-matrices.
Definition at line 852 of file block_matrix_base.h.
|
privateinherited |
A set of scratch arrays that can be used by the add() and set() functions that take pointers to data to pre-sort indices before use. Access from multiple threads is synchronized via the mutex variable that is part of the structure.
Definition at line 1064 of file block_matrix_base.h.
|
mutableprivateinherited |
Store the number of objects which subscribed to this object. Initially, this number is zero, and upon destruction it shall be zero again (i.e. all objects which subscribed should have unsubscribed again).
The creator (and owner) of an object is counted in the map below if HE manages to supply identification.
We use the mutable
keyword in order to allow subscription to constant objects also.
This counter may be read from and written to concurrently in multithreaded code: hence we use the std::atomic
class template.
Definition at line 227 of file enable_observer_pointer.h.
|
mutableprivateinherited |
In this map, we count subscriptions for each different identification string supplied to subscribe().
Definition at line 233 of file enable_observer_pointer.h.
|
mutableprivateinherited |
In this vector, we store pointers to the validity bool in the ObserverPointer objects that subscribe to this class.
Definition at line 249 of file enable_observer_pointer.h.
|
mutableprivateinherited |
Pointer to the typeinfo object of this object, from which we can later deduce the class name. Since this information on the derived class is neither available in the destructor, nor in the constructor, we obtain it in between and store it here.
Definition at line 257 of file enable_observer_pointer.h.
|
staticprivateinherited |
A mutex used to ensure data consistency when accessing the mutable
members of this class. This lock is used in the subscribe() and unsubscribe() functions, as well as in list_subscribers()
.
Definition at line 280 of file enable_observer_pointer.h.