Functions
void	intDriver (const int NUM, const double t, const double t_end, const double pr_global, double y_global)
	Integration driver for the CPU integrators. More...

int	integrate (const double t_start, const double t_end, const double pr, double *y)
	A header definition of the integrate method, that must be implemented by various solvers. More...

void	init_solver_log ()
	Initializes solver specific items for logging. More...

void	solver_log ()
	Executes solver specific logging tasks. More...

void	initialize_solver (int num_threads)
	Initializes the solver. More...

void	cleanup_solver (int num_threads)
	Cleans up the created solvers. More...

const char *	solver_name ()
	Returns a descriptive solver name. More...

void	accelerInt_initialize (int num_threads)
	Initializes the solver. More...

void	accelerInt_integrate (const int NUM, const double t_start, const double t_end, const double stepsize, double __restrict__ y_host, const double __restrict__ var_host)
	integrate NUM odes from time `t` to time `t_end`, using stepsizes of `t_step` More...

void	accelerInt_cleanup (int num_threads)
	Cleans up the solver. More...

void	write_log (int NUM, double t, const double y_host, FILE pFile)
	Writes state vectors to file. More...

int	main (int argc, char *argv[])

void	check_error (int, int)
	Checks the return code of the given thread (IVP) for an error, and exits if found. More...

Detailed Description

Include common code.

Function Documentation

◆ accelerInt_cleanup()

void generic::accelerInt_cleanup ( int num_threads )

Cleans up the solver.

Parameters

[in] num_threads The number of OpenMP threads to use

Definition at line 63 of file solver_interface.c.

◆ accelerInt_initialize()

void generic::accelerInt_initialize ( int num_threads )

Initializes the solver.

Parameters

[in] num_threads The number of OpenMP threads to use

Definition at line 23 of file solver_interface.c.

◆ accelerInt_integrate()

void generic::accelerInt_integrate	(	const int	NUM,
		const double	t_start,
		const double	t_end,
		const double	stepsize,
		double *__restrict__	y_host,
		const double *__restrict__	var_host
	)

integrate NUM odes from time t to time t_end, using stepsizes of t_step

Parameters

[in] NUM The number of ODEs to integrate. This should be the size of the leading dimension of y_host and var_host.

See also: accelerint_indx

Parameters

[in]	t_start	The system time
[in]	t_end	The end time
[in]	stepsize	The integration step size. If `stepsize` < 0, the step size will be set to `t_end - t`
[in,out]	y_host	The state vectors to integrate.
[in]	var_host	The parameters to use in dydt() and eval_jacob()
[in]	NUM	The number of ODEs to integrate. This should be the size of the leading dimension of `y_host` and `var_host`.

See also: accelerint_indx

Parameters

[in]	t	The system time
[in]	t_end	The end time
[in]	stepsize	The integration step size. If `stepsize` < 0, the step size will be set to `t_end - t`
[in,out]	y_host	The state vectors to integrate.
[in]	var_host	The parameters to use in dydt() and eval_jacob()

Definition at line 39 of file solver_interface.c.

◆ check_error()

void generic::check_error	(	int	tid,
		int	code
	)

Checks the return code of the given thread (IVP) for an error, and exits if found.

Parameters

tid	The thread (IVP) index
code	The return code of the thread

See also: Return codes of Integrators

Parameters

tid	The thread (IVP) index
code	The return code of the thread

See also: Return codes of EXP4 integrator

Definition at line 13 of file radau2a_props.c.

◆ cleanup_solver()

void generic::cleanup_solver ( int num_threads )

Cleans up the created solvers.

Parameters

num_threads The number of OpenMP threads used

Frees and cleans up allocated CVODE memory.

Parameters

num_threads The number of OpenMP threads used

Frees and cleans up allocated RK78 memory.

Definition at line 165 of file cvodes_init.c.

◆ init_solver_log()

void generic::init_solver_log ( )

Initializes solver specific items for logging.

Initializes stepsize logging for stiffness measurement

Initializes solver specific items for logging.

See also: solver_options.cuh

Initializes the Krylov subspace logging files (if LOG_OUTPUT is defined)

See also: solver_options.cuh

Definition at line 190 of file cvodes_init.c.

◆ initialize_solver()

void generic::initialize_solver ( int num_threads )

Initializes the solver.

Parameters

num_threads The number of OpenMP threads to use

Various definitions in the generated options, e.g. FINITE_DIFFERENCE, CV_MAX_ERRTEST_FAILS, etc. affect the options set for the CVODEs solver.

See also: solver_options.h

The RHS function dydt_cvodes() will be used in the solver.

If FINITE_DIFFERENCE is not defined, the jacobian function in eval_jacob_cvodes will be used in the CVODE solver. Else, the CVODE finite difference based on the RHS function will be used.

Parameters

num_threads	The number of OpenMP threads to use
num_threads	Unused

Definition at line 49 of file cvodes_init.c.

◆ intDriver()

void generic::intDriver	(	const int	NUM,
		const double	t,
		const double	t_end,
		const double *	pr_global,
		double *	y_global
	)

Integration driver for the CPU integrators.

Parameters

[in]	NUM	The (non-padded) number of IVPs to integrate
[in]	t	The current system time
[in]	t_end	The IVP integration end time
[in]	pr_global	The system constant variable (pressures / densities)
[in,out]	y_global	The system state vectors at time t. Returns system state vectors at time t_end
[in]	NUM	The (non-padded) number of IVPs to integrate
[in]	t	The current system time
[in]	t_end	The IVP integration end time
[in]	pr_global	The system constant variable (pressures / densities)
[in,out]	y_global	The system state vectors at time t. Returns system state vectors at time t_end

This is generic driver for CPU integrators

Definition at line 28 of file solver_generic.c.

◆ integrate()

int generic::integrate	(	const double	t_start,
		const double	t_end,
		const double	pr,
		double *	y
	)

A header definition of the integrate method, that must be implemented by various solvers.

Parameters

[in]	t_start	the starting IVP integration time
[in]	t_end	the IVP integration endtime
[in]	pr	the IVP constant variable (presssure/density)
[in,out]	y	The IVP state vector at time t_start. At end of this function call, the system state at time t_end is stored here

A header definition of the integrate method, that must be implemented by various solvers.

Parameters

[in]	t_start	The starting time
[in]	t_end	The end integration time
[in]	pr	The system constant variable (pressure/density)
[in,out]	y	The system state vector at time `t_start`. Overwritten with the system state at time `t_end`

Returns: Return code,

See also: Return codes of Radau-IIa integrator

A header definition of the integrate method, that must be implemented by various solvers.

Parameters

t_start	The initial integration time
t_end	The final integration timestep
pr	User data passed to the RHS function dydt() - commonly used for the Pressure term
y	The state vector

Returns: The result of this integration step

See also: Return codes of EXP4 integrator

A header definition of the integrate method, that must be implemented by various solvers.

Parameters

t_start	The initial integration time
t_end	The final integration timestep
pr	User data passed to the RHS function dydt() - commonly used for the Pressure term
y	The state vector

Returns: The result of this integration step

See also: Return codes of EXPRB43 integrator

Definition at line 520 of file radau2a.c.

◆ main()

int generic::main	(	int	argc,
		char *	argv[]
	)

Main function

Parameters

[in]	argc	command line argument count
[in]	argv	command line argument vector

This allows running the integrators from the command line. The syntax is as follows:
./solver-name [num_threads] [num_IVPs]

num_threads [Optional, Default:1]
- The number OpenMP threads to utilize
- The number of threads cannot be greater than recognized by OpenMP via omp_get_max_threads()
num_IVPs [Optional, Default:1]
- The number of initial value problems to solve.
- This must be less than the number of conditions in the data file if SAME_IC is not defined.
- If SAME_IC is defined, then the initial conditions in the mechanism files will be used.

Definition at line 87 of file solver_main.c.

◆ solver_log()

void generic::solver_log ( )

Executes solver specific logging tasks.

Logs errors, step-sizes, and krylov subspace size (if LOG_OUTPUT is defined)

See also: solver_options.cuh

Definition at line 192 of file cvodes_init.c.

◆ solver_name()

const char* generic::solver_name ( )

Returns a descriptive solver name.

Definition at line 181 of file cvodes_init.c.

◆ write_log()

void generic::write_log	(	int	NUM,
		double	t,
		const double *	y_host,
		FILE *	pFile
	)

Writes state vectors to file.

Parameters

[in]	NUM	the number of state vectors to write
[in]	t	the current system time
[in]	y_host	the current state vectors
[in]	pFile	the opened binary file object

The resulting file is updated as: system time
temperature, mass fractions (State #1)
temperature, mass fractions (State #2)...

Definition at line 47 of file solver_main.c.

Functions

Detailed Description

Function Documentation

◆ accelerInt_cleanup()

◆ accelerInt_initialize()

◆ accelerInt_integrate()

◆ check_error()

◆ cleanup_solver()

◆ init_solver_log()

◆ initialize_solver()

◆ intDriver()

◆ integrate()

◆ main()

◆ solver_log()

◆ solver_name()

◆ write_log()