NeoPZ
|
Implements a matrix which computes the preconditioner developed by Dohrmann. Sub Structure. More...
#include <tpzdohrprecond.h>
Public Member Functions | |
TPZDohrPrecond (TPZDohrMatrix< TVar, TSubStruct > &origin, TPZAutoPointer< TPZDohrAssembly< TVar > > assemble) | |
Constructor with matrix. More... | |
TPZDohrPrecond (const TPZDohrPrecond ©) | |
Copy constructor. More... | |
TPZDohrPrecond () | |
Empty constructor for restoring. More... | |
~TPZDohrPrecond () | |
virtual TPZMatrix< TVar > * | Clone () const override |
std::list< TPZAutoPointer< TSubStruct > > & | Global () |
The matrix class is a placeholder for a list of substructures. More... | |
void | Initialize () |
Initialize the necessary datastructures. More... | |
void | SetNumThreads (int numthreads) |
virtual void | MultAdd (const TPZFMatrix< TVar > &x, const TPZFMatrix< TVar > &y, TPZFMatrix< TVar > &z, const TVar alpha, const TVar beta, const int opt) const override |
The only method any matrix class needs to implement. More... | |
virtual void | MultAddTBB (const TPZFMatrix< TVar > &x, const TPZFMatrix< TVar > &y, TPZFMatrix< TVar > &z, const TVar alpha, const TVar beta, const int opt) const |
void | SetSolver (TPZSolver< TVar > &solver) |
Specify the solution process for the coarse matrix. More... | |
void | ComputeV1 (const TPZFMatrix< TVar > &x, TPZFMatrix< TVar > &v1) const |
Compute the contribution of the coarse matrix. More... | |
void | ComputeV2 (const TPZFMatrix< TVar > &x, TPZFMatrix< TVar > &v2) const |
Compute the contribution of the sub domains. More... | |
int | ClassId () const override |
Routines to send and receive messages. More... | |
void | Read (TPZStream &buf, void *context) override |
Unpacks the object structure from a stream of bytes. More... | |
void | Write (TPZStream &buf, int withclassid) const override |
Packs the object structure in a stream of bytes. More... | |
Public Member Functions inherited from TPZMatrix< TVar > | |
TPZMatrix () | |
Simple constructor. More... | |
TPZMatrix (const TPZMatrix< TVar > &cp) | |
virtual | ~TPZMatrix () |
Simple destructor. More... | |
virtual int64_t | MemoryFootprint () const |
Returns the approximate size of the memory footprint (amount of memory required to store this object). More... | |
template<class TVar2 > | |
void | CopyFrom (TPZMatrix< TVar2 > ©) |
void | AutoFill (int64_t nrow, int64_t ncol, int symmetric) |
Fill matrix storage with randomic values. More... | |
virtual int | VerifySymmetry (REAL tol=1.e-13) const |
Checks if current matrix value is symmetric. More... | |
virtual int | Put (const int64_t row, const int64_t col, const TVar &value) |
Put values with bounds checking if DEBUG variable is defined. More... | |
virtual const TVar & | Get (const int64_t row, const int64_t col) const |
Get value with bound checking. More... | |
const TVar & | g (const int64_t row, const int64_t col) const |
Substitution for the () operator when const arguments are needed. More... | |
TVar & | operator() (const int64_t row, const int64_t col) |
The operators check on the bounds if the DEBUG variable is defined. More... | |
virtual TVar & | s (const int64_t row, const int64_t col) |
The operators check on the bounds if the DEBUG variable is defined. More... | |
TVar & | operator() (const int64_t row) |
The operators check on the bounds if the DEBUG variable is defined. More... | |
virtual int | PutVal (const int64_t, const int64_t, const TVar &val) |
Put values without bounds checking This method is faster than "Put" if DEBUG is defined. More... | |
virtual const TVar & | GetVal (const int64_t, const int64_t) const |
Get values without bounds checking This method is faster than "Get" if DEBUG is defined. More... | |
virtual void | AddKel (TPZFMatrix< TVar > &elmat, TPZVec< int64_t > &destinationindex) |
Add a contribution of a stiffness matrix. More... | |
virtual void | AddKel (TPZFMatrix< TVar > &elmat, TPZVec< int64_t > &sourceindex, TPZVec< int64_t > &destinationindex) |
Add a contribution of a stiffness matrix. More... | |
void | SetIsDecomposed (int val) |
Sets current matrix to decomposed state. More... | |
virtual bool | Compare (TPZSavable *copy, bool override=false) override |
Compare the object for identity with the object pointed to, eventually copy the object. More... | |
virtual bool | Compare (TPZSavable *copy, bool override=false) const override |
Compare the object for identity with the object pointed to, eventually copy the object. More... | |
virtual void | GetSub (const TPZVec< int64_t > &indices, TPZFMatrix< TVar > &block) const |
Extract the block indicated by the indices from the matrix. More... | |
bool | CompareValues (TPZMatrix< TVar > &M, TVar tol) |
Compare values of this to B, with a precision tolerance tol. More... | |
template<> | |
void | Print (const char *name, std::ostream &out, const MatrixOutputFormat form) const |
template<> | |
void | Print (const char *name, std::ostream &out, const MatrixOutputFormat form) const |
template<> | |
void | Print (const char *name, std::ostream &out, const MatrixOutputFormat form) const |
template<> | |
void | SolveCG (int64_t &numiterations, TPZSolver< std::complex< float > > &preconditioner, const TPZFMatrix< std::complex< float > > &F, TPZFMatrix< std::complex< float > > &result, TPZFMatrix< std::complex< float > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveCG (int64_t &numiterations, TPZSolver< std::complex< double > > &preconditioner, const TPZFMatrix< std::complex< double > > &F, TPZFMatrix< std::complex< double > > &result, TPZFMatrix< std::complex< double > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveCG (int64_t &numiterations, TPZSolver< std::complex< long double > > &preconditioner, const TPZFMatrix< std::complex< long double > > &F, TPZFMatrix< std::complex< long double > > &result, TPZFMatrix< std::complex< long double > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveGMRES (int64_t &numiterations, TPZSolver< std::complex< float > > &preconditioner, TPZFMatrix< std::complex< float > > &H, int &numvectors, const TPZFMatrix< std::complex< float > > &F, TPZFMatrix< std::complex< float > > &result, TPZFMatrix< std::complex< float > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveGMRES (int64_t &numiterations, TPZSolver< std::complex< double > > &preconditioner, TPZFMatrix< std::complex< double > > &H, int &numvectors, const TPZFMatrix< std::complex< double > > &F, TPZFMatrix< std::complex< double > > &result, TPZFMatrix< std::complex< double > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveGMRES (int64_t &numiterations, TPZSolver< std::complex< long double > > &preconditioner, TPZFMatrix< std::complex< long double > > &H, int &numvectors, const TPZFMatrix< std::complex< long double > > &F, TPZFMatrix< std::complex< long double > > &result, TPZFMatrix< std::complex< long double > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveBICG (int64_t &numiterations, TPZSolver< std::complex< float > > &preconditioner, const TPZFMatrix< std::complex< float > > &F, TPZFMatrix< std::complex< float > > &result, REAL &tol) |
template<> | |
void | SolveBICG (int64_t &numiterations, TPZSolver< std::complex< double > > &preconditioner, const TPZFMatrix< std::complex< double > > &F, TPZFMatrix< std::complex< double > > &result, REAL &tol) |
template<> | |
void | SolveBICG (int64_t &numiterations, TPZSolver< std::complex< long double > > &preconditioner, const TPZFMatrix< std::complex< long double > > &F, TPZFMatrix< std::complex< long double > > &result, REAL &tol) |
template<> | |
void | SolveBICGStab (int64_t &numiterations, TPZSolver< std::complex< float > > &preconditioner, const TPZFMatrix< std::complex< float > > &F, TPZFMatrix< std::complex< float > > &result, TPZFMatrix< std::complex< float > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveBICGStab (int64_t &numiterations, TPZSolver< std::complex< double > > &preconditioner, const TPZFMatrix< std::complex< double > > &F, TPZFMatrix< std::complex< double > > &result, TPZFMatrix< std::complex< double > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveBICGStab (int64_t &numiterations, TPZSolver< std::complex< long double > > &preconditioner, const TPZFMatrix< std::complex< long double > > &F, TPZFMatrix< std::complex< long double > > &result, TPZFMatrix< std::complex< long double > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveIR (int64_t &numiterations, TPZSolver< std::complex< float > > &preconditioner, const TPZFMatrix< std::complex< float > > &F, TPZFMatrix< std::complex< float > > &result, TPZFMatrix< std::complex< float > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveIR (int64_t &numiterations, TPZSolver< std::complex< double > > &preconditioner, const TPZFMatrix< std::complex< double > > &F, TPZFMatrix< std::complex< double > > &result, TPZFMatrix< std::complex< double > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
void | SolveIR (int64_t &numiterations, TPZSolver< std::complex< long double > > &preconditioner, const TPZFMatrix< std::complex< long double > > &F, TPZFMatrix< std::complex< long double > > &result, TPZFMatrix< std::complex< long double > > *residual, REAL &tol, const int FromCurrent) |
template<> | |
bool | SolveEigenvaluesJacobi (int64_t &numiterations, REAL &tol, TPZVec< std::complex< float > > *Sort) |
template<> | |
bool | SolveEigenvaluesJacobi (int64_t &numiterations, REAL &tol, TPZVec< std::complex< double > > *Sort) |
template<> | |
bool | SolveEigenvaluesJacobi (int64_t &numiterations, REAL &tol, TPZVec< std::complex< long double > > *Sort) |
virtual void | Multiply (const TPZFMatrix< TVar > &A, TPZFMatrix< TVar > &res, int opt=0) const |
It mutiplies itself by TPZMatrix<TVar>A putting the result in res. More... | |
virtual void | Add (const TPZMatrix< TVar > &A, TPZMatrix< TVar > &res) const |
It adds itself to TPZMatrix<TVar>A putting the result in res. More... | |
virtual void | Residual (const TPZFMatrix< TVar > &x, const TPZFMatrix< TVar > &rhs, TPZFMatrix< TVar > &res) |
Computes res = rhs - this * x. More... | |
virtual void | Substract (const TPZMatrix< TVar > &A, TPZMatrix< TVar > &result) const |
It substracts A from storing the result in result. More... | |
virtual void | Identity () |
Converts the matrix in an identity matrix. More... | |
virtual void | Transpose (TPZMatrix< TVar > *const T) const |
It makes *T the transpose of current matrix. More... | |
int | Inverse (TPZFMatrix< TVar > &Inv, DecomposeType dec) |
It makes Inv =[this]. IMPORTANT OBSERVATION –> The original matrix (calling object) no is more equal. It containts the some decomposition (LU or Cholesky or ...) More... | |
TVar | MatrixNorm (int p, int64_t numiter=2000000, REAL tol=1.e-10) const |
Computes the matrix norm of this. More... | |
TVar | ConditionNumber (int p, int64_t numiter=2000000, REAL tol=1.e-10) |
Computes the matrix condition number of this. More... | |
virtual int | PutSub (const int64_t sRow, const int64_t sCol, const TPZFMatrix< TVar > &Source) |
It puts submatrix Source on actual matrix structure. More... | |
virtual int | GetSub (const int64_t sRow, const int64_t sCol, const int64_t rowSize, const int64_t colSize, TPZFMatrix< TVar > &Target) const |
Gets submatrix storing it on Target. More... | |
virtual int | AddSub (const int64_t sRow, const int64_t sCol, const TPZFMatrix< TVar > &Source) |
It adds Source matrix on current matrix from position (sRow, sCol) More... | |
virtual int | InsertSub (const int64_t sRow, const int64_t sCol, const int64_t rowSize, const int64_t colSize, const int64_t pRow, const int64_t pCol, TPZMatrix< TVar > *Target) const |
Inserts a submatrix from current object on matrix *Target with no redimentioning. More... | |
virtual int | AddSub (const int64_t sRow, const int64_t sCol, const int64_t rowSize, const int64_t colSize, const int64_t pRow, const int64_t pCol, TPZMatrix< TVar > *pA) const |
Adds a submatrix from current object in *Target. More... | |
virtual void | UpdateFrom (TPZAutoPointer< TPZMatrix< TVar > >) |
Updates the values of the matrix based on the values of the matrix. More... | |
virtual int | IsSimetric () const |
Checks if the current matrix is symmetric. More... | |
int | IsSquare () const |
Checks if current matrix is square. More... | |
virtual void | Simetrize () |
Simetrizes copies upper plan to the lower plan, making its data simetric. More... | |
virtual int | IsDefPositive () const |
Checks if current matrix is definite positive. More... | |
int | IsDecomposed () const |
Checks if current matrix is already decomposed. More... | |
virtual int | Decompose_LU (std::list< int64_t > &singular) |
Decomposes the current matrix using LU decomposition. More... | |
virtual int | Decompose_LU () |
virtual int | Decompose_Cholesky () |
Decomposes the current matrix using Cholesky method. The current matrix has to be symmetric. More... | |
virtual int | Decompose_Cholesky (std::list< int64_t > &singular) |
Decomposes the current matrix using Cholesky method. More... | |
virtual int | Decompose_LDLt (std::list< int64_t > &singular) |
Decomposes the current matrix using LDLt. The current matrix has to be symmetric. "L" is lower triangular with 1.0 in its diagonal and "D" is a Diagonal matrix. More... | |
virtual int | Decompose_LDLt () |
Decomposes the current matrix using LDLt. More... | |
virtual int | Substitution (TPZFMatrix< TVar > *B) const |
Computes Forward and Backward substitution for a "LU" decomposed matrix. More... | |
virtual int | Subst_Forward (TPZFMatrix< TVar > *b) const |
Computes B = Y, where A*Y = B, A is lower triangular. More... | |
virtual int | Subst_Backward (TPZFMatrix< TVar > *b) const |
Computes B = Y, where A*Y = B, A is upper triangular. More... | |
virtual int | Subst_LForward (TPZFMatrix< TVar > *b) const |
Computes B = Y, where A*Y = B, A is lower triangular with A(i,i)=1. More... | |
virtual int | Subst_LBackward (TPZFMatrix< TVar > *b) const |
Computes B = Y, where A*Y = B, A is upper triangular with A(i,i)=1. More... | |
virtual int | Subst_Diag (TPZFMatrix< TVar > *b) const |
Computes B = Y, where A*Y = B, A is diagonal matrix. More... | |
virtual void | Input (std::istream &in=std::cin) |
Input operation. More... | |
virtual void | Print (std::ostream &out) const |
virtual void | Print (const char *name, std::ostream &out=std::cout, const MatrixOutputFormat form=EFormatted) const |
It prints the matrix data in a MatrixFormat Rows X Cols. More... | |
int64_t | Rows () const |
Returns number of rows. More... | |
int64_t | Cols () const |
Returns number of cols. More... | |
virtual int64_t | Dim () const |
Returns the dimension of the matrix if the matrix is square. More... | |
virtual int | Resize (const int64_t newRows, const int64_t newCols) |
Redimensions a matriz keeping the previous values. More... | |
virtual int | Redim (const int64_t newRows, const int64_t newCols) |
Redimensions the matrix reinitializing it with zero. More... | |
virtual int | Zero () |
Zeroes the matrix. More... | |
virtual void | SolveJacobi (int64_t &numiterations, const TPZFMatrix< TVar > &F, TPZFMatrix< TVar > &result, TPZFMatrix< TVar > *residual, TPZFMatrix< TVar > &scratch, REAL &tol, const int FromCurrent=0) |
Solves the linear system using Jacobi method. . More... | |
virtual void | SolveSOR (int64_t &numiterations, const TPZFMatrix< TVar > &F, TPZFMatrix< TVar > &result, TPZFMatrix< TVar > *residual, TPZFMatrix< TVar > &scratch, const REAL overrelax, REAL &tol, const int FromCurrent=0, const int direction=1) |
Solves the linear system using Successive Over Relaxation method (Gauss Seidel). . More... | |
virtual void | SolveSSOR (int64_t &numiterations, const TPZFMatrix< TVar > &F, TPZFMatrix< TVar > &result, TPZFMatrix< TVar > *residual, TPZFMatrix< TVar > &scratch, const REAL overrelax, REAL &tol, const int FromCurrent=0) |
Solves the linear system using Symmetric Successive Over Relaxation method (Gauss Seidel). . More... | |
virtual void | SolveCG (int64_t &numiterations, TPZSolver< TVar > &preconditioner, const TPZFMatrix< TVar > &F, TPZFMatrix< TVar > &result, TPZFMatrix< TVar > *residual, REAL &tol, const int FromCurrent=0) |
Solves the linear system using Conjugate Gradient method. . More... | |
virtual void | SolveBICG (int64_t &numiterations, TPZSolver< TVar > &preconditioner, const TPZFMatrix< TVar > &F, TPZFMatrix< TVar > &result, REAL &tol) |
Solves the linear system using Bi-Conjugate Gradient method. . More... | |
virtual void | SolveBICGStab (int64_t &numiterations, TPZSolver< TVar > &preconditioner, const TPZFMatrix< TVar > &F, TPZFMatrix< TVar > &result, TPZFMatrix< TVar > *residual, REAL &tol, const int FromCurrent=0) |
Solves the linear system using Bi-Conjugate Gradient stabilized method. . More... | |
virtual void | SolveGMRES (int64_t &numiterations, TPZSolver< TVar > &preconditioner, TPZFMatrix< TVar > &H, int &numvectors, const TPZFMatrix< TVar > &F, TPZFMatrix< TVar > &result, TPZFMatrix< TVar > *residual, REAL &tol, const int FromCurrent) |
Solves the linear system using Generalized Minimal Residual (GMRES) method. . More... | |
virtual void | SolveIR (int64_t &numiterations, TPZSolver< TVar > &preconditioner, const TPZFMatrix< TVar > &F, TPZFMatrix< TVar > &result, TPZFMatrix< TVar > *residual, REAL &tol, const int FromCurrent=0) |
Solves the linear system using IR method. . More... | |
virtual bool | SolveEigenvaluesJacobi (int64_t &numiterations, REAL &tol, TPZVec< TVar > *Sort=0) |
Transforms this matrix in a diagonal matrix, where the diagonal values are its eigenvalues. This method is efficient only for small matrices. More... | |
virtual bool | SolveEigensystemJacobi (int64_t &numiterations, REAL &tol, TPZVec< TVar > &Eigenvalues, TPZFMatrix< TVar > &Eigenvectors) const |
Compute Eigenvalues and Eigenvectors of this matrix. This method is efficient only for small matrices. More... | |
virtual int | SolveDirect (TPZFMatrix< TVar > &F, const DecomposeType dt, std::list< int64_t > &singular) |
Solves the linear system using Direct methods. More... | |
virtual int | SolveDirect (TPZFMatrix< TVar > &F, const DecomposeType dt) |
Solves the linear system using Direct methods. More... | |
virtual int | Decompose (const DecomposeType dt, std::list< int64_t > &singular) |
decompose the system of equations acording to the decomposition scheme More... | |
int | Solve_LU (TPZFMatrix< TVar > *B, std::list< int64_t > &singular) |
Solves the linear system using LU method . More... | |
int | Solve_LU (TPZFMatrix< TVar > *B) |
Solves the linear system using LU method . More... | |
virtual int | Solve_Cholesky (TPZFMatrix< TVar > *B) |
Solves the linear system using Cholesky method . More... | |
int | Solve_Cholesky (TPZFMatrix< TVar > *B, std::list< int64_t > &singular) |
Solves the linear system using Cholesky method . More... | |
int | Solve_LDLt (TPZFMatrix< TVar > *B, std::list< int64_t > &singular) |
Solves the linear system using LDLt method . More... | |
int | Solve_LDLt (TPZFMatrix< TVar > *B) |
Solves the linear system using LDLt method . More... | |
Public Member Functions inherited from TPZSavable | |
TPZSavable () | |
virtual | ~TPZSavable () |
virtual std::list< std::map< std::string, uint64_t > > | VersionHistory () const |
virtual std::pair< std::string, uint64_t > | Version () const |
Public Member Functions inherited from TPZRegisterClassId | |
template<typename T > | |
TPZRegisterClassId (int(T::*)() const) | |
TPZRegisterClassId ()=default | |
Private Attributes | |
std::list< TPZAutoPointer< TSubStruct > > | fGlobal |
The matrix class is a placeholder for a list of substructures. More... | |
TPZStepSolver< TVar > * | fCoarse |
The global matrix associated with the coarse degrees of freedom. More... | |
int64_t | fNumCoarse |
Size of the coarse system. More... | |
int | fNumThreads |
Number of threads used during preconditioning. More... | |
TPZAutoPointer< TPZDohrAssembly< TVar > > | fAssemble |
Additional Inherited Members | |
Static Public Member Functions inherited from TPZMatrix< TVar > | |
static int | Error (const char *msg, const char *msg2=0) |
Returns error messages. More... | |
static TVar | ReturnNearestValue (TVar val, TPZVec< TVar > &Vec, TVar tol) |
Retorna o valor mais proximo a "val" (exceto valores no intervalo -tol <= val <= +tol) contido no vetor Vec. More... | |
Static Public Member Functions inherited from TPZSavable | |
static std::set< TPZRestoreClassBase * > & | RestoreClassSet () |
This static function guarantees that the gMap object is available when needed. More... | |
static std::map< int, TPZRestore_t > & | ClassIdMap () |
This static function guarantees that the gMap object is available when needed. More... | |
static std::pair< std::string, uint64_t > | NeoPZVersion () |
static void | Register (TPZRestoreClassBase *restore) |
static void | RegisterClassId (int classid, TPZRestore_t fun) |
static TPZSavable * | CreateInstance (const int &classId) |
Protected Member Functions inherited from TPZMatrix< TVar > | |
void | PrepareZ (const TPZFMatrix< TVar > &y, TPZFMatrix< TVar > &z, const TVar beta, const int opt) const |
Is an auxiliar method used by MultiplyAdd. More... | |
TPZMatrix (const int64_t row, const int64_t col) | |
Constructor. More... | |
virtual int | Clear () |
It clears data structure. More... | |
Static Protected Member Functions inherited from TPZMatrix< TVar > | |
static void | Swap (int64_t *a, int64_t *b) |
Swaps contents of a in b and b in a. More... | |
Protected Attributes inherited from TPZMatrix< TVar > | |
int64_t | fRow |
Number of rows in matrix. More... | |
int64_t | fCol |
Number of cols in matrix. More... | |
char | fDecomposed |
Decomposition type used to decompose the current matrix. More... | |
char | fDefPositive |
Definite Posistiveness of current matrix. More... | |
Static Protected Attributes inherited from TPZMatrix< TVar > | |
static TVar | gZero |
Initializing value to static variable. More... | |
Implements a matrix which computes the preconditioner developed by Dohrmann. Sub Structure.
Definition at line 32 of file tpzdohrprecond.h.
TPZDohrPrecond< TVar, TSubStruct >::TPZDohrPrecond | ( | TPZDohrMatrix< TVar, TSubStruct > & | origin, |
TPZAutoPointer< TPZDohrAssembly< TVar > > | assemble | ||
) |
Constructor with matrix.
Definition at line 37 of file tpzdohrprecond.cpp.
References TPZDohrPrecond< TVar, TSubStruct >::fNumThreads, and TPZDohrMatrix< TVar, TSubStruct >::NumThreads().
TPZDohrPrecond< TVar, TSubStruct >::TPZDohrPrecond | ( | const TPZDohrPrecond< TVar, TSubStruct > & | copy | ) |
Copy constructor.
Definition at line 44 of file tpzdohrprecond.cpp.
References TPZDohrPrecond< TVar, TSubStruct >::fCoarse.
TPZDohrPrecond< TVar, TSubStruct >::TPZDohrPrecond | ( | ) |
Empty constructor for restoring.
Definition at line 54 of file tpzdohrprecond.cpp.
Referenced by TPZDohrPrecond< TVar, TSubStruct >::Clone().
TPZDohrPrecond< TVar, TSubStruct >::~TPZDohrPrecond | ( | ) |
Definition at line 59 of file tpzdohrprecond.cpp.
References TPZDohrPrecond< TVar, TSubStruct >::fCoarse.
|
inlineoverridevirtual |
Implements TPZMatrix< TVar >.
Definition at line 60 of file tpzdohrprecond.h.
References TPZDohrPrecond< TVar, TSubStruct >::TPZDohrPrecond().
void TPZDohrPrecond< TVar, TSubStruct >::ComputeV1 | ( | const TPZFMatrix< TVar > & | x, |
TPZFMatrix< TVar > & | v1 | ||
) | const |
Compute the contribution of the coarse matrix.
Definition at line 384 of file tpzdohrprecond.cpp.
References TPZMatrix< TVar >::Cols(), ELDLt, TPZDohrPrecond< TVar, TSubStruct >::fAssemble, TPZDohrPrecond< TVar, TSubStruct >::fCoarse, TPZDohrPrecond< TVar, TSubStruct >::fGlobal, TPZDohrPrecond< TVar, TSubStruct >::fNumCoarse, LOGPZ_DEBUG, TPZMatrix< TVar >::Print(), and TPZFMatrix< TVar >::Zero().
Referenced by TPZDohrPrecond< TVar, TSubStruct >::MultAdd(), and TPZDohrPrecond< TVar, TSubStruct >::SetNumThreads().
void TPZDohrPrecond< TVar, TSubStruct >::ComputeV2 | ( | const TPZFMatrix< TVar > & | x, |
TPZFMatrix< TVar > & | v2 | ||
) | const |
Compute the contribution of the sub domains.
Definition at line 450 of file tpzdohrprecond.cpp.
References TPZDohrPrecond< TVar, TSubStruct >::fAssemble, TPZDohrPrecond< TVar, TSubStruct >::fGlobal, LOGPZ_DEBUG, and TPZMatrix< TVar >::Print().
Referenced by TPZDohrPrecond< TVar, TSubStruct >::MultAdd(), and TPZDohrPrecond< TVar, TSubStruct >::SetNumThreads().
|
inline |
The matrix class is a placeholder for a list of substructures.
Definition at line 63 of file tpzdohrprecond.h.
References TPZDohrPrecond< TVar, TSubStruct >::fGlobal, and TPZDohrPrecond< TVar, TSubStruct >::Initialize().
Referenced by TPZDohrStructMatrix::Assemble().
void TPZDohrPrecond< TVar, TSubStruct >::Initialize | ( | ) |
Initialize the necessary datastructures.
It will compute the coarse matrix, coarse residual and any other necessary data structures
Definition at line 329 of file tpzdohrprecond.cpp.
References TPZDohrPrecond< TVar, TSubStruct >::fAssemble, TPZDohrPrecond< TVar, TSubStruct >::fCoarse, TPZDohrPrecond< TVar, TSubStruct >::fGlobal, TPZDohrPrecond< TVar, TSubStruct >::fNumCoarse, nsub, and VisualMatrix().
Referenced by TPZDohrStructMatrix::Assemble(), TPZDohrStructMatrix::AssembleTBB(), and TPZDohrPrecond< TVar, TSubStruct >::Global().
|
overridevirtual |
The only method any matrix class needs to implement.
x | Is x on the above operation. It must be a vector! |
y | Is y on the above operation |
z | Is z on the above operation |
alpha | Is alpha on the above operation |
beta | Is beta on the above operation |
opt | Indicates if is Transpose or not In this case the variable x represents the rhs and z the result of the preconditioning When used as a preconditioner y will be zero In fact, it will compute It computes but z and x can not overlap in memory. |
Reimplemented from TPZMatrix< TVar >.
Definition at line 194 of file tpzdohrprecond.cpp.
References TPZDohrAssembleList< TVar >::AddItem(), TPZMatrix< TVar >::Cols(), TPZDohrPrecond< TVar, TSubStruct >::ComputeV1(), TPZDohrPrecond< TVar, TSubStruct >::ComputeV2(), TPZMatrix< TVar >::Error(), TPZDohrPrecond< TVar, TSubStruct >::fAssemble, TPZDohrPrecond< TVar, TSubStruct >::fGlobal, TPZDohrPrecond< TVar, TSubStruct >::fNumThreads, TPZTimeTemp::fPreCond, LOGPZ_DEBUG, TPZDohrPrecond< TVar, TSubStruct >::MultAddTBB(), TPZMatrix< TVar >::PrepareZ(), TPZMatrix< TVar >::Print(), TPZStack< T, NumExtAlloc >::Push(), PZ_PTHREAD_CREATE, PZ_PTHREAD_JOIN, TPZfTime::ReturnTimeDouble(), TPZMatrix< TVar >::Rows(), and tempo.
Referenced by TPZDohrPrecond< TVar, TSubStruct >::SetNumThreads().
|
virtual |
Copy of the MultAdd using TBB
Definition at line 129 of file tpzdohrprecond.cpp.
References TPZMatrix< TVar >::Cols(), TPZMatrix< TVar >::Error(), TPZDohrPrecond< TVar, TSubStruct >::fAssemble, TPZDohrPrecond< TVar, TSubStruct >::fGlobal, TPZMatrix< TVar >::PrepareZ(), PZ_PTHREAD_CREATE, PZ_PTHREAD_JOIN, pzenviroment, and TPZMatrix< TVar >::Rows().
Referenced by TPZDohrPrecond< TVar, TSubStruct >::MultAdd(), and TPZDohrPrecond< TVar, TSubStruct >::SetNumThreads().
|
overridevirtual |
Unpacks the object structure from a stream of bytes.
buf | The buffer containing the object in a packed form |
context |
Reimplemented from TPZMatrix< TVar >.
Definition at line 509 of file tpzdohrprecond.cpp.
References TPZDohrPrecond< TVar, TSubStruct >::fAssemble, TPZDohrMatrix< TVar, TSubStruct >::fAssembly, TPZDohrPrecond< TVar, TSubStruct >::fCoarse, TPZDohrPrecond< TVar, TSubStruct >::fGlobal, TPZDohrPrecond< TVar, TSubStruct >::fNumCoarse, TPZDohrPrecond< TVar, TSubStruct >::fNumThreads, TPZPersistenceManager::GetInstance(), TPZStream::Read(), TPZMatrix< TVar >::Read(), and TPZDohrMatrix< TVar, TSubStruct >::SubStructures().
Referenced by TPZDohrPrecond< TVar, TSubStruct >::SetNumThreads().
|
inline |
Specify the number of threads for preconditioning
Definition at line 75 of file tpzdohrprecond.h.
References TPZDohrPrecond< TVar, TSubStruct >::ClassId(), TPZDohrPrecond< TVar, TSubStruct >::ComputeV1(), TPZDohrPrecond< TVar, TSubStruct >::ComputeV2(), TPZDohrPrecond< TVar, TSubStruct >::MultAdd(), TPZDohrPrecond< TVar, TSubStruct >::MultAddTBB(), substruct_tst14.test::numthreads, TPZDohrPrecond< TVar, TSubStruct >::Read(), TPZDohrPrecond< TVar, TSubStruct >::SetSolver(), and TPZDohrPrecond< TVar, TSubStruct >::Write().
void TPZDohrPrecond< TVar, TSubStruct >::SetSolver | ( | TPZSolver< TVar > & | solver | ) |
Specify the solution process for the coarse matrix.
Referenced by TPZDohrPrecond< TVar, TSubStruct >::SetNumThreads().
|
overridevirtual |
Packs the object structure in a stream of bytes.
buf | Buffer which will receive the bytes |
withclassid |
Reimplemented from TPZMatrix< TVar >.
Definition at line 525 of file tpzdohrprecond.cpp.
References TPZDohrPrecond< TVar, TSubStruct >::fCoarse, TPZDohrPrecond< TVar, TSubStruct >::fNumCoarse, TPZDohrPrecond< TVar, TSubStruct >::fNumThreads, TPZStream::Write(), TPZMatrix< TVar >::Write(), and TPZPersistenceManager::WritePointer().
Referenced by TPZDohrPrecond< TVar, TSubStruct >::SetNumThreads().
|
private |
Definition at line 45 of file tpzdohrprecond.h.
Referenced by TPZDohrPrecond< TVar, TSubStruct >::ComputeV1(), TPZDohrPrecond< TVar, TSubStruct >::ComputeV2(), TPZDohrPrecond< TVar, TSubStruct >::Initialize(), TPZDohrPrecond< TVar, TSubStruct >::MultAdd(), TPZDohrPrecond< TVar, TSubStruct >::MultAddTBB(), and TPZDohrPrecond< TVar, TSubStruct >::Read().
|
private |
The global matrix associated with the coarse degrees of freedom.
Definition at line 37 of file tpzdohrprecond.h.
Referenced by TPZDohrPrecond< TVar, TSubStruct >::ComputeV1(), TPZDohrPrecond< TVar, TSubStruct >::Initialize(), TPZDohrPrecond< TVar, TSubStruct >::Read(), TPZDohrPrecond< TVar, TSubStruct >::TPZDohrPrecond(), TPZDohrPrecond< TVar, TSubStruct >::Write(), and TPZDohrPrecond< TVar, TSubStruct >::~TPZDohrPrecond().
|
private |
The matrix class is a placeholder for a list of substructures.
Definition at line 35 of file tpzdohrprecond.h.
Referenced by TPZDohrPrecond< TVar, TSubStruct >::ComputeV1(), TPZDohrPrecond< TVar, TSubStruct >::ComputeV2(), TPZDohrPrecond< TVar, TSubStruct >::Global(), TPZDohrPrecond< TVar, TSubStruct >::Initialize(), TPZDohrPrecond< TVar, TSubStruct >::MultAdd(), TPZDohrPrecond< TVar, TSubStruct >::MultAddTBB(), and TPZDohrPrecond< TVar, TSubStruct >::Read().
|
private |
Size of the coarse system.
Definition at line 40 of file tpzdohrprecond.h.
Referenced by TPZDohrPrecond< TVar, TSubStruct >::ComputeV1(), TPZDohrPrecond< TVar, TSubStruct >::Initialize(), TPZDohrPrecond< TVar, TSubStruct >::Read(), and TPZDohrPrecond< TVar, TSubStruct >::Write().
|
private |
Number of threads used during preconditioning.
Definition at line 43 of file tpzdohrprecond.h.
Referenced by TPZDohrPrecond< TVar, TSubStruct >::MultAdd(), TPZDohrPrecond< TVar, TSubStruct >::Read(), TPZDohrPrecond< TVar, TSubStruct >::TPZDohrPrecond(), and TPZDohrPrecond< TVar, TSubStruct >::Write().