18. Let \( \mathbf{A},\mathbf{B},\mathbf{C} \), and \( \mathbf{D} \) be matrices. The computation of \( ((\mathbf{AB})\mathbf{C})\mathbf{D} \) requires the same number of floating-point operations as the computation of \( (\mathbf{AB})(\mathbf{CD}) \)