Re: MTL and C++ compilers - some thoughts
Gunter Winkler <[email protected]>
| Newsgroups | gmane.comp.lib.mtl.devel |
|---|---|
| Message-ID | <[email protected]> |
On Monday, 2. September 2002 14:27, you wrote: > specialized version of the critical functions is not a nice solution. > So i suggest to think about re-design of the iterator classes for > sparse matrices (in compressed2d.h compressed_iter.h). > > What do the list members thing ?! Good idea. I'd like to have a way to say: 'I have at most 10 nonzeros per row and at least 1' to avoid lots of inserts(..). This would be much more convenient than building the data/index array by yourself. We would need two procedures to allocate storage and compact the matrix. btw: I currently do FEM für (2D) convection-diffusion problems. My program spends most of its time in find(..), insert(..) and vec_ref(..) of compressed2D :-( regards Gunter Winkler _______________________________________________ This list is archived at http://www.osl.iu.edu/MailArchives/mtl-devel/