On the numerical treatment of a multi-dimensional parabolic boundary-value control problem