From: Ian Scott <ian.m.scott@st...>  20020326 18:26:28

> I have some more questions to vnl_sparse_matrix_linear_system. > I have an impression that this class is only an adaptor to > vnl_linear_system. So the question is: > >  does vnl_sparse_matrix_linear_system adress the sparseness? > I need to > solve a sparse matrix like 50,000 x 50,000. Will it handle it? You'll need to try  it depends on the amount of memory, your matrix density, and the particular implementation details. > >  What is the safe way to get x from Ax=b and/or how do I get > inverse of > sparse A in order to use mult from vnl_sparse_matrix to get > x=A^{1}*b. I'd suggest using the VXL documentation search engine to answer questions like this. http://www.isbe.man.ac.uk/search_vxl.html http://www.isbe.man.ac.uk/cgibin/htsearch?config=htdigvxl;words=sparse%20v nl; seems to suggest a few possibilities, e.g. vnl_lsqr Ian. 