Hello

I would like to know how to compute the determinant of a sparse symmetric matrix in Eigen? Or any simple way to evaluate if it is singular or not?


thanks

If your matrix is symmetric positive definite, then Eigen::SimplicialLDLt will do (from the SparseCholesky module). You need to use the default branch. Otherwise, you will have to rely on an external library, either SuperLU or UmfPack via our SuperLU or UmfPackLU wrapper classes (SuperLUSupport and UmfPackSupport modules).

I am solving A*x = B where x is sparse symmetric. I am using simplicialCholesky, however, I need to check whether A is positive definite or not. Seems that there is no way to work around.

ggael wrote:If your matrix is symmetric positive definite, then Eigen::SimplicialLDLt will do (from the SparseCholesky module). You need to use the default branch. Otherwise, you will have to rely on an external library, either SuperLU or UmfPack via our SuperLU or UmfPackLU wrapper classes (SuperLUSupport and UmfPackSupport modules).

What about calling the info() method after the numerical factorization?