/*******************************************************************
* *
* File : cvdense.h *
* Programmers : Scott D. Cohen, Alan C. Hindmarsh, and *
* Radu Serban @ LLNL *
* Version of : 26 June 2002 *
*-----------------------------------------------------------------*
* Copyright (c) 2002, The Regents of the University of California *
* Produced at the Lawrence Livermore National Laboratory *
* All rights reserved *
* For details, see sundials/cvode/LICENSE *
*-----------------------------------------------------------------*
* This is the header file for the CVODE dense linear solver, *
* CVDENSE. *
* *
* Note: The type integertype must be large enough to store the *
* value of the linear system size N. *
* *
*******************************************************************/
#ifdef __cplusplus /* wrapper to enable C++ usage */
extern "C" {
#endif
#ifndef _cvdense_h
#define _cvdense_h
#include <stdio.h>
#include "cvode.h"
#include "sundialstypes.h"
#include "dense.h"
#include "nvector.h"
/******************************************************************
* *
* CVDENSE solver statistics indices *
*----------------------------------------------------------------*
* The following enumeration gives a symbolic name to each *
* CVDENSE statistic. The symbolic names are used as indices into *
* the iopt and ropt arrays passed to CVodeMalloc. *
* The CVDENSE statistics are: *
* *
* iopt[DENSE_NJE] : number of Jacobian evaluations, i.e. of *
* calls made to the dense Jacobian routine *
* (default or user-supplied). *
* *
* iopt[DENSE_LRW] : size (in realtype words) of real workspace *
* matrices and vectors used by this solver. *
* *
* iopt[DENSE_LIW] : size (in integertype words) of integer *
* workspace vectors used by this solver. *
* *
******************************************************************/
enum { DENSE_NJE=CVODE_IOPT_SIZE, DENSE_LRW, DENSE_LIW };
/******************************************************************
* *
* CVDENSE solver constants *
*----------------------------------------------------------------*
* CVD_MSBJ : maximum number of steps between dense Jacobian *
* evaluations *
* *
* CVD_DGMAX : maximum change in gamma between dense Jacobian *
* evaluations *
* *
******************************************************************/
#define CVD_MSBJ 50
#define CVD_DGMAX RCONST(0.2)
/******************************************************************
* *
* Type : CVDenseJacFn *
*----------------------------------------------------------------*
* A dense Jacobian approximation function Jac must have the *
* prototype given below. Its parameters are: *
* *
* N is the length of all vector arguments. *
* *
* J is the dense matrix (of type DenseMat) that will be loaded *
* by a CVDenseJacFn with an approximation to the Jacobian matrix *
* J = (df_i/dy_j) at the point (t,y). *
* J is preset to zero, so only the nonzero elements need to be *
* loaded. Two efficient ways to load J are: *
* *
* (1) (with macros - no explicit data structure references) *
* for (j=0; j < N; j++) { *
* col_j = DENSE_COL(J,j); *
* for (i=0; i < N; i++) { *
* generate J_ij = the (i,j)th Jacobian element *
* col_j[i] = J_ij; *
* } *
* } *
* *
* (2) (without macros - explicit data structure references) *
* for (j=0; j < N; j++) { *
* col_j = (J->data)[j]; *
* for (i=0; i < N; i++) { *
* generate J_ij = the (i,j)th Jacobian element *
* col_j[i] = J_ij; *
* } *
* } *
* *
* The DENSE_ELEM(A,i,j) macro is appropriate for use in small *
* problems in which efficiency of access is NOT a major concern. *
* *
* f is the right hand side function for the ODE problem. *
* *
* f_data is a pointer to user data to be passed to f, the same *
* as the F_data parameter passed to CVodeMalloc. *
* *
* t is the current value of the independent variable. *
* *
* y is the current value of the dependent variable vector, *
* namely the predicted value of y(t). *
* *
* fy is the vector f(t,y). *
* *
* ewt is the error weight vector. *
* *
* h is a tentative step size in t. *
* *
* uround is the machine unit roundoff. *
* *
* jac_data is a pointer to user data - the same as the jac_data *
* parameter passed to CVDense. *
* *
* nfePtr is a pointer to the memory location containing the *
* CVODE problem data nfe = number of calls to f. The Jacobian *
* routine should update this counter by adding on the number *
* of f calls made in order to approximate the Jacobian, if any. *
* For example, if the routine calls f a total of N times, then *
* the update is *nfePtr += N. *
* *
* vtemp1, vtemp2, and vtemp3 are pointers to memory allocated *
* for vectors of length N which can be used by a CVDenseJacFn *
* as temporary storage or work space. *
* *
******************************************************************/
typedef void (*CVDenseJacFn)(integertype N, DenseMat J, RhsFn f, void *f_data,
realtype t, N_Vector y, N_Vector fy, N_Vector ewt,
realtype h, realtype uround, void *jac_data,
long int *nfePtr, N_Vector vtemp1,
N_Vector vtemp2, N_Vector vtemp3);
/******************************************************************
* *
* Function : CVDense *
*----------------------------------------------------------------*
* A call to the CVDense function links the main CVODE integrator *
* with the CVDENSE linear solver. *
* *
* cvode_mem is the pointer to CVODE memory returned by *
* CVodeMalloc. *
* *
* djac is the dense Jacobian approximation routine to be used. *
* A user-supplied djac routine must be of type *
* CVDenseJacFn. Pass NULL for djac to use the default *
* difference quotient routine CVDenseDQJac supplied *
* with this solver. *
* *
* jac_data is a pointer to user data which is passed to the *
* djac routine every time it is called. *
* *
* The return values of CVDense are: *
* SUCCESS = 0 if successful *
* LMEM_FAIL = -1 if there was a memory allocation failure *
* *
* NOTE: The dense linear solver assumes a serial implementation *
* of the NVECTOR package. Therefore, CVDense will first *
* test for a compatible N_Vector internal representation *
* by checking (1) the machine environment ID tag and *
* (2) that the functions N_VMake, N_VDispose, N_VGetData, *
* and N_VSetData are implemented. *
* *
******************************************************************/
int CVDense(void *cvode_mem, CVDenseJacFn djac, void *jac_data);
/******************************************************************
* *
* Function : CVReInitDense *
*----------------------------------------------------------------*
* A call to the CVReInitDense function resets the link between *
* the main CVODE integrator and the CVDENSE linear solver. *
* After solving one problem using CVDENSE, call CVReInit and then*
* CVReInitDense to solve another problem of the same size, if *
* there is a change in the CVDense parameters djac or jac_data. *
* If there is no change in parameters, it is not necessary to *
* call either CVReInitDense or CVDense for the new problem. *
* *
* All arguments to CVReInitDense have the same names and meanings*
* as those of CVDense. The cvode_mem argument must be identical *
* to its value in the previous CVDense call. *
* *
* The return values of CVReInitDense are: *
* SUCCESS = 0 if successful, or *
* LMEM_FAIL = -1 if the cvode_mem argument is NULL *
* *
* NOTE: CVReInitDense performs the same compatibility tests as *
* CVDense. *
* *
******************************************************************/
int CVReInitDense(void *cvode_mem, CVDenseJacFn djac, void *jac_data);
#endif
#ifdef __cplusplus
}
#endif