Hi -

I am trying to use a SparseMatrix<complex<float> > matrix for a matrix vector multiply where the matrix is symmetric but not Hermitian. Code:

SparseMatrix<complex<float> > Aij;

// some code to initialize Aij which is symmetric (only fill upper) ...

// now do y = A*x
VectorXcf y = Aij.selfadjointView<Eigen::Upper>()*x;

The problem is the matrix Aij is symmetric but not equal to its conjugate transpose (i.e. not Hermitian). The above code treats the matrix as if it were Hermitian and so gets the wrong result. I could not find an operator that gives a view of the matrix as symmetric but not Hermitian. Is there one? I am using Eigen 3.

Help greatly appreciated!

Hi, I can only tell you that's in our todo list...