Registered Member
|
Hi everybody,
I'm well aware that there are many topics on this, but I still can't find help for my task, so.. What I'm trying to do is to compute L*M^(-1)*L, where L and M are two Sparse N*N matrices. N is very big (many thousands) and I can't compute explicitly the inverse of M so I should try another method, i tried to get a decomposition of M through a solver and then solve it for L, like this: SimplicialCholesky<SparseMatrix<double> > chol(M); A=chol.solve(L); A=L*A; but after the second instruction it seems that A is an empty matrix (or at least is what I thought after seeing that is .outerSize() value is 0), while I'm totally sure that neither M or L are empty. What am I doing wrong? Thanks |
Moderator
|
hm, this should do the job. you can everything went well by checking chol.info()==Success before solving. Compiling without -DNDEBUG might help too to detect errors in your code.
|
Registered users: Bing [Bot], Google [Bot], Yahoo [Bot]