A physicsists idea:
On the one hand
$$\int dV (\Delta f)^2 = \int dV \left(\sum_i \partial_i^2 f\right)^2=
\int dV \left(\sum_i \partial_i^2f\right)\left(\sum_j \partial_j^2f\right)\\=\sum_{i,j} \int dV (\partial_i^2f)(\partial_j^2 f)$$
On On the other hand
$$\int dV \sum_{i,j} \left(\partial_i \partial_j f \right)^2=\int dV \sum_{i,j} \left(\partial_i \partial_j f \right) \left(\partial_i \partial_j f \right)\\
=\sum_{i,j} \int dV \left(\partial_i \partial_j f \right) \left(\partial_i \partial_j f \right)$$
Let's look at the last expression.
We use partial integration to move a $\partial_i$ from the right parenthesis to the left. We will pick up a minus sign. Next, we use partial integration to move a $\partial_j$ from the left parenthesis to the right parenthesis. We will pick up another minus sign and have shown the identity.
EDIT:
I guess that the surface terms from the partial integration vanish due to the compact support of $f$. The appearance of the third derivative after the first partial integration should also not be a problem, since you specified that $f$ is a $C^3$ function.