Base class for building matrices implementing basic management and parallel construction.
#include <TMatBuilder.hh>
Public Member Functions |
| TMatBuilder (const bool coarsening=CFG::Arith::coarsen_build) |
void | set_coarsening (const bool b) |
| (de-) activate coarsening with standard accuracy
|
void | set_coarsening (const TTruncAcc &acc) |
| activate coarsening with accuracy acc
|
virtual TMatrix * | build (const uint nthreads, const TBlockClusterTree *bct, const TTruncAcc &acc, TProgressBar *progress=NULL) const |
| build the H-matrix with block-wise accuracy defined by acc
|
virtual TMatrix * | build (const uint nthreads, const TBlockCluster *bc, const TTruncAcc &acc, TProgressBar *progress=NULL) const |
virtual TMatrix * | build (const uint nthreads, const TBlockClusterTree *cluster, const matform_t matformat, const TTruncAcc &acc, TProgressBar *progress=NULL) const |
| same as
|
virtual TMatrix * | build (const uint nthreads, const TBlockCluster *cluster, const matform_t matformat, const TTruncAcc &acc, TProgressBar *progress=NULL) const |
| same as
|
Constructor & Destructor Documentation
TMatBuilder |
( |
const bool |
coarsening = CFG::Arith::coarsen_build | ) |
|
construct matrix construction object with standard coarsening, e.g. same precision as in block construction
Member Function Documentation
same as standard
- See Also
- build, but build block matrices for given block cluster without permutations etc.
- See Also
- build but with user defined matrix format matformat
- See Also
- build but with user defined matrix format matformat
Member Data Documentation
if true, the accuracy during coarsening is identical to the accuracy during block construction (standard accuracy)