Hi guys, I have a question about Sparse Matrix SimplicialLLT method
I'm using it to solve a large scale liner equations Ax = B, and some values of inv(A)
my code is as follows

Code: Select all

        SimplicialLLT  < SparseMatrix < double > > llt;
   llt.compute(A);
   X= llt.solve(B);
        L = llt.matrixL()


Actually I got the right X solution
But the very strange thing is L matrix is right only when the matrix A is small
When A is over like 2000*2000, the L matrix is totally wrong.
I compared L*L.transpose() with A in MATLAB, it's different, but the X is still right.
Is there something I missed? Hope for your help~