Random symmetrical matrices for generalized eigenvalue problem - How to solve

Question

Assume that we have $A$ and $B$ and they both are real symmetrical random matrices, e.g they have both negative and positive real values.

Now I want to solve $$AV = BVD$$ where $V$ are real eigenvectors and $D$ are real eigenvalues.

This is equivalent to the MATLAB command:

[V, D] = eig(A, B)

Is that possible to solve? What decomposition should I use here? I have looked up some tools from NetLib about solving generalized eigenvalue problem. But they only show that $B$ must have real positive symmetrical values. https://www.netlib.org/lapack/lug/node54.html

So I assume that NetLib have no solution for my math problem.