ViennaCL - The Vienna Computing Library
1.5.0
|
A tag for the conjugate gradient Used for supplying solver parameters and for dispatching the solve() function. More...
#include <mixed_precision_cg.hpp>
Public Member Functions | |
mixed_precision_cg_tag (double tol=1e-8, unsigned int max_iterations=300, float inner_tol=1e-2f) | |
The constructor. | |
double | tolerance () const |
Returns the relative tolerance. | |
float | inner_tolerance () const |
Returns the relative tolerance. | |
unsigned int | max_iterations () const |
Returns the maximum number of iterations. | |
unsigned int | iters () const |
Return the number of solver iterations: | |
void | iters (unsigned int i) const |
double | error () const |
Returns the estimated relative error at the end of the solver run. | |
void | error (double e) const |
Sets the estimated relative error at the end of the solver run. |
A tag for the conjugate gradient Used for supplying solver parameters and for dispatching the solve() function.
mixed_precision_cg_tag | ( | double | tol = 1e-8 , |
unsigned int | max_iterations = 300 , |
||
float | inner_tol = 1e-2f |
||
) | [inline] |
The constructor.
tol | Relative tolerance for the residual (solver quits if ||r|| < tol * ||r_initial||) |
max_iterations | The maximum number of iterations |
inner_tol | Inner tolerance for the low-precision iterations |
double error | ( | ) | const [inline] |
Returns the estimated relative error at the end of the solver run.
void error | ( | double | e | ) | const [inline] |
Sets the estimated relative error at the end of the solver run.
float inner_tolerance | ( | ) | const [inline] |
Returns the relative tolerance.
unsigned int iters | ( | ) | const [inline] |
Return the number of solver iterations:
void iters | ( | unsigned int | i | ) | const [inline] |
unsigned int max_iterations | ( | ) | const [inline] |
Returns the maximum number of iterations.
double tolerance | ( | ) | const [inline] |
Returns the relative tolerance.