Defines accuracy for truncation of low rank blocks.
#include <TTruncAcc.hh>
Inheritance diagram for TTruncAcc:
Public Member Functions | |
| TTruncAcc () | |
| TTruncAcc (const size_t k, const double absolute_eps=CFG::Arith::abs_eps) | |
| TTruncAcc (const double relative_eps, const double absolute_eps=CFG::Arith::abs_eps) | |
| TTruncAcc (const TTruncAcc &ta) | |
| template<typename T > | |
| size_t | trunc_rank (const BLAS::Vector< T > &sv) const |
| virtual const TTruncAcc & | acc (const TBlockCluster *) const |
| return accuracy description for individual subblock | |
| virtual const TTruncAcc & | acc (const TMatrix *) const |
| return accuracy description for individual submatrix | |
| const TTruncAcc & | operator() (const TBlockCluster *bc) const |
| abbreviation via () operator | |
| const TTruncAcc & | operator() (const TMatrix *M) const |
| abbreviation via () operator | |
| size_t | rank () const |
| return fixed rank | |
| size_t | max_rank () const |
| return maximal rank | |
| double | rel_eps () const |
| return relative accuracy | |
| double | abs_eps () const |
| return absolute accuracy | |
| bool | is_fixed_rank () const |
| return true if accuracy is fixed rank | |
| bool | is_exact () const |
| return true if accuracy is "exact" | |
| void | set_max_rank (const size_t k) |
| set maximal rank in truncation | |
| TTruncAcc & | operator= (const TTruncAcc &ta) |
| copy operator | |
| virtual String | to_string () const |
| return string representation | |
Constructor & Destructor Documentation
|
inline |
construct accuracy object for exact truncation
|
inline |
construct accuracy object for fixed rank truncation
|
inline |
construct accuracy object for fixed accuracy truncation
Member Function Documentation
| size_t trunc_rank | ( | const BLAS::Vector< T > & | sv | ) | const |
return truncation rank based on internal strategy and given singular values sv
