Jaroslav Hajek wrote: > On Wed, Mar 10, 2010 at 9:52 PM, David Bateman <dbate...@dbateman.org> wrote: > >> Yeah, things like UMFPACK internally block the sparse matrices and treat the >> sub-blocks with blas (ie dense kernels), so its a well known trick. >> > > Internally? Or can UMFPACK be forced to solve a block sparse matrix > directly (i.e. using block structure you specify)? That would make it > worth considering a real-life block sparse extension. > Internally with a CSC matrix passed to the routines.. If you were happy to pass your block matrix in CSC form and play with some of the UMFPACK Control parameters (UMFACK_BLOCK_SIZE comes to mind) it might be possible to do something reasonable with block matrices, though I'm not really sure of the details..
D. ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Octave-dev mailing list Octave-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/octave-dev
