|
| int | get_num_iterations () |
| |
| real_t | get_final_residual_norm () |
| |
| bool | has_converged () |
| |
| void | on_iteration_complete (const gko::LinOp *op, const gko::size_type &iteration, const gko::LinOp *residual, const gko::LinOp *solution, const gko::LinOp *residual_norm) const override |
| |
| void | on_iteration_complete (const gko::LinOp *op, const gko::size_type &iteration, const gko::LinOp *residual, const gko::LinOp *solution, const gko::LinOp *residual_norm, const gko::LinOp *implicit_sq_residual_norm) const override |
| |
| void | on_iteration_complete (const gko::LinOp *op, const gko::LinOp *rhs, const gko::LinOp *solution, const gko::size_type &iteration, const gko::LinOp *residual, const gko::LinOp *residual_norm, const gko::LinOp *implicit_sq_residual_norm, const gko::array< gko::stopping_status > *status, bool stopped) const override |
| |
| | ConvergenceLogger () |
| |
Custom gko::log::Logger base class for logging the final residual norm and iteration count. This is necessary because some solvers (e.g., IR) will not have the final residual available if they stop due to reaching maximum iterations prior to convergence, and the standard Convergence logger will not respect the use of derived types (VectorWrapper/ParallelVectorWrapper) when computing the final residual.
This base class should not be used directly; solvers should use EnableConvergenceLogger, which has vector type as a template parameter, instead.
Definition at line 493 of file ginkgo.hpp.