Cantera  2.4.0
Public Member Functions | Protected Attributes | List of all members
MultiJac Class Reference

Class MultiJac evaluates the Jacobian of a system of equations defined by a residual function supplied by an instance of class OneDim. More...

#include <MultiJac.h>

Inheritance diagram for MultiJac:
[legend]
Collaboration diagram for MultiJac:
[legend]

Public Member Functions

 MultiJac (OneDim &r)
 
void eval (doublereal *x0, doublereal *resid0, double rdt)
 Evaluate the Jacobian at x0. More...
 
doublereal elapsedTime () const
 Elapsed CPU time spent computing the Jacobian. More...
 
int nEvals () const
 Number of Jacobian evaluations. More...
 
int age () const
 Number of times 'incrementAge' has been called since the last evaluation. More...
 
void incrementAge ()
 Increment the Jacobian age. More...
 
void updateTransient (doublereal rdt, integer *mask)
 
void setAge (int age)
 Set the Jacobian age. More...
 
vector_inttransientMask ()
 
void incrementDiagonal (int j, doublereal d)
 
- Public Member Functions inherited from BandMatrix
 BandMatrix ()
 Base Constructor. More...
 
 BandMatrix (size_t n, size_t kl, size_t ku, doublereal v=0.0)
 Creates a banded matrix and sets all elements to zero. More...
 
 BandMatrix (const BandMatrix &y)
 
BandMatrixoperator= (const BandMatrix &y)
 
void resize (size_t n, size_t kl, size_t ku, doublereal v=0.0)
 Resize the matrix problem. More...
 
void bfill (doublereal v=0.0)
 Fill or zero the matrix. More...
 
doublereal & operator() (size_t i, size_t j)
 Index into the (i,j) element. More...
 
doublereal operator() (size_t i, size_t j) const
 Constant Index into the (i,j) element. More...
 
doublereal & value (size_t i, size_t j)
 Return a changeable reference to element (i,j). More...
 
doublereal value (size_t i, size_t j) const
 Return the value of element (i,j). More...
 
size_t index (size_t i, size_t j) const
 Returns the location in the internal 1D array corresponding to the (i,j) element in the banded array. More...
 
doublereal _value (size_t i, size_t j) const
 Return the value of the (i,j) element for (i,j) within the bandwidth. More...
 
virtual size_t nRows () const
 Return the number of rows in the matrix. More...
 
size_t nColumns () const
 Number of columns. More...
 
size_t nSubDiagonals () const
 Number of subdiagonals. More...
 
size_t nSuperDiagonals () const
 Number of superdiagonals. More...
 
size_t ldim () const
 Return the number of rows of storage needed for the band storage. More...
 
virtual void mult (const doublereal *b, doublereal *prod) const
 Multiply A*b and write result to prod. More...
 
virtual void leftMult (const doublereal *const b, doublereal *const prod) const
 Multiply b*A and write result to prod. More...
 
int factor ()
 Perform an LU decomposition, the LAPACK routine DGBTRF is used. More...
 
int solve (const doublereal *const b, doublereal *const x)
 Solve the matrix problem Ax = b. More...
 
int solve (doublereal *b, size_t nrhs=1, size_t ldb=0)
 Solve the matrix problem Ax = b. More...
 
virtual vector_fp::iterator begin ()
 Returns an iterator for the start of the band storage data. More...
 
vector_fp::iterator end ()
 Returns an iterator for the end of the band storage data. More...
 
vector_fp::const_iterator begin () const
 Returns a const iterator for the start of the band storage data. More...
 
vector_fp::const_iterator end () const
 Returns a const iterator for the end of the band storage data. More...
 
virtual void zero ()
 Zero the matrix elements. More...
 
virtual doublereal rcond (doublereal a1norm)
 Returns an estimate of the inverse of the condition number for the matrix. More...
 
virtual int factorAlgorithm () const
 Returns the factor algorithm used. More...
 
virtual doublereal oneNorm () const
 Returns the one norm of the matrix. More...
 
virtual doublereal * ptrColumn (size_t j)
 Return a pointer to the top of column j. More...
 
virtual doublereal *const * colPts ()
 Return a vector of const pointers to the columns. More...
 
virtual size_t checkRows (doublereal &valueSmall) const
 Check to see if we have any zero rows in the Jacobian. More...
 
virtual size_t checkColumns (doublereal &valueSmall) const
 Check to see if we have any zero columns in the Jacobian. More...
 
int info () const
 LAPACK "info" flag after last factor/solve operation. More...
 
- Public Member Functions inherited from GeneralMatrix
 GeneralMatrix ()
 Base Constructor. More...
 
virtual int factorQR ()
 Factors the A matrix using the QR algorithm, overwriting A. More...
 
virtual doublereal rcondQR ()
 Returns an estimate of the inverse of the condition number for the matrix. More...
 
virtual void useFactorAlgorithm (int fAlgorithm)
 Change the way the matrix is factored. More...
 
virtual void clearFactorFlag ()
 clear the factored flag More...
 
virtual bool factored () const
 true if the current factorization is up to date with the matrix More...
 

Protected Attributes

OneDimm_resid
 Residual evaluator for this Jacobian. More...
 
vector_fp m_r1
 
doublereal m_rtol
 
doublereal m_atol
 
doublereal m_elapsed
 
vector_fp m_ssdiag
 
vector_int m_mask
 
int m_nevals
 
int m_age
 
size_t m_size
 
size_t m_points
 
- Protected Attributes inherited from BandMatrix
vector_fp data
 Matrix data. More...
 
vector_fp ludata
 Factorized data. More...
 
size_t m_n
 Number of rows and columns of the matrix. More...
 
size_t m_kl
 Number of subdiagonals of the matrix. More...
 
size_t m_ku
 Number of super diagonals of the matrix. More...
 
doublereal m_zero
 value of zero More...
 
std::unique_ptr< PivData > m_ipiv
 Pivot vector. More...
 
std::vector< doublereal * > m_colPtrs
 Vector of column pointers. More...
 
std::vector< double * > m_lu_col_ptrs
 
vector_int iwork_
 Extra work array needed - size = n. More...
 
vector_fp work_
 Extra dp work array needed - size = 3n. More...
 
int m_info
 
- Protected Attributes inherited from GeneralMatrix
int m_factored
 Indicates whether the matrix is factored. More...
 

Detailed Description

Class MultiJac evaluates the Jacobian of a system of equations defined by a residual function supplied by an instance of class OneDim.

The residual function may consist of several linked 1D domains, with different variables in each domain.

Definition at line 22 of file MultiJac.h.

Member Function Documentation

◆ eval()

void eval ( doublereal *  x0,
doublereal *  resid0,
double  rdt 
)

Evaluate the Jacobian at x0.

The unperturbed residual function is resid0, which must be supplied on input. The third parameter 'rdt' is the reciprocal of the time step. If zero, the steady-state Jacobian is evaluated.

Definition at line 43 of file MultiJac.cpp.

References BandMatrix::bfill(), OneDim::eval(), OneDim::loc(), MultiJac::m_resid, Cantera::npos, OneDim::nVars(), and BandMatrix::value().

Referenced by MultiNewton::solve().

◆ elapsedTime()

doublereal elapsedTime ( ) const
inline

Elapsed CPU time spent computing the Jacobian.

Definition at line 36 of file MultiJac.h.

◆ nEvals()

int nEvals ( ) const
inline

Number of Jacobian evaluations.

Definition at line 41 of file MultiJac.h.

Referenced by MultiNewton::solve().

◆ age()

int age ( ) const
inline

Number of times 'incrementAge' has been called since the last evaluation.

Definition at line 46 of file MultiJac.h.

Referenced by MultiJac::setAge(), and MultiNewton::solve().

◆ incrementAge()

void incrementAge ( )
inline

Increment the Jacobian age.

Definition at line 51 of file MultiJac.h.

Referenced by MultiNewton::solve().

◆ setAge()

void setAge ( int  age)
inline

Set the Jacobian age.

Definition at line 58 of file MultiJac.h.

References MultiJac::age().

Referenced by Domain1D::needJacUpdate(), and MultiNewton::solve().

Member Data Documentation

◆ m_resid

OneDim* m_resid
protected

Residual evaluator for this Jacobian.

This is a pointer to the residual evaluator. This object isn't owned by this Jacobian object.

Definition at line 74 of file MultiJac.h.

Referenced by MultiJac::eval().


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