Class SUNNonLinSolWrapper

Inheritance Relationships

Derived Types

Class Documentation

class SUNNonLinSolWrapper

A RAII wrapper for SUNNonLinearSolver structs which solve the nonlinear system F (y) = 0 or G(y) = y.

Subclassed by amici::SUNNonLinSolFixedPoint, amici::SUNNonLinSolNewton

Public Functions

explicit SUNNonLinSolWrapper(SUNNonlinearSolver sol)

SUNNonLinSolWrapper from existing SUNNonlinearSolver.

Parameters:

sol –

virtual ~SUNNonLinSolWrapper()
SUNNonLinSolWrapper(SUNNonLinSolWrapper const &other) = delete

Copy constructor.

Parameters:

other –

SUNNonLinSolWrapper(SUNNonLinSolWrapper &&other) noexcept

Move constructor.

Parameters:

other –

SUNNonLinSolWrapper &operator=(SUNNonLinSolWrapper const &other) = delete

Copy assignment.

Parameters:

other –

Returns:

SUNNonLinSolWrapper &operator=(SUNNonLinSolWrapper &&other) noexcept

Move assignment.

Parameters:

other –

Returns:

SUNNonlinearSolver get() const

Get the wrapped SUNNonlinearSolver.

Returns:

SUNNonlinearSolver

SUNNonlinearSolver_Type get_type() const

Get type ID of the solver.

Returns:

int setup(N_Vector y, void *mem)

Setup solver.

Parameters:

  • y – the initial iteration passed to the nonlinear solver.

  • mem – the sundials integrator memory structure.

Returns:

int solve(N_Vector y0, N_Vector y, N_Vector w, realtype tol, bool callLSetup, void *mem)

solve the nonlinear system F (y) = 0 or G(y) = y.

Parameters:

  • y0 – the initial iterate for the nonlinear solve. This must remain unchanged throughout the solution process.

  • y – the solution to the nonlinear system

  • w – the solution error weight vector used for computing weighted error norms.

  • tol – the requested solution tolerance in the weighted root-mean- squared norm.

  • callLSetup – a flag indicating that the integrator recommends for the linear solver setup function to be called.

  • mem – the sundials integrator memory structure.

Returns:

int set_sys_fn(SUNNonlinSolSysFn SysFn)

Set function to evaluate the nonlinear residual function F(y) = 0 or the fixed point function G(y) = y.

Parameters:

SysFn –

Returns:

int set_l_setup_fn(SUNNonlinSolLSetupFn SetupFn)

Set linear solver setup function.

Parameters:

SetupFn –

Returns:

int set_l_solve_fn(SUNNonlinSolLSolveFn SolveFn)

Set linear solver solve function.

Parameters:

SolveFn –

Returns:

int set_conv_test_fn(SUNNonlinSolConvTestFn CTestFn, void *ctest_data)

Set function to test for convergence.

Parameters:

  • CTestFn –

  • ctest_data –

Returns:

int set_max_iters(int maxiters)

Set maximum number of non-linear iterations.

Parameters:

maxiters –

Returns:

long int get_num_iters() const

get_num_iters

Returns:

int get_cur_iter() const

get_cur_iter

Returns:

long int get_num_conv_fails() const

get_num_conv_fails

Returns:

Protected Functions

void initialize()

initialize

Protected Attributes

SUNNonlinearSolver solver = nullptr

the wrapper solver