Hello,

I'm trying to solve Ax=b where A is sparse and symmetric quasi-definite. SparseLU is too slow for my use case (1312x1312 with 5564 non-zero elements). Then I tried SimplicialLDLT since LDL factorization is guaranteed to exist for a symmetric quasi-definite matrix. However, the doc says that only SPD matrix is supported so I'm not sure if this is allowed.

When using SimplicialLDLT on my problem, I observed something weird. When the upper triangle part is used, the solver finds the right solution; when the lower triangle part is used, the numerical factorization fails. How is this happening?

Also, what is the solver recommended for a sparse symmetric quasi-definite system if SimplicialLDLT is not the right option? Thanks!