1: \begin{abstract}

2: Most computational studies in chemistry and materials science are

3: based on the use of density functional theory. Although the exact

4: density functional is unknown, several density functional approximations

5: (DFAs) offer a good balance of affordable computational cost and semi-quantitative

6: accuracy for applications. The development of DFAs still continues

7: on many fronts, and several new DFAs aiming for improved accuracy

8: are published every year. However, the numerical behavior of these

9: DFAs is an often overlooked problem. In this work, we look at all

10: 592 DFAs for three-dimensional systems available in Libxc 5.2.2 and

11: examine the convergence of the density functional total energy based

12: on tabulated atomic Hartree--Fock wave functions. We show that several

13: recent DFAs, including the celebrated SCAN family of functionals,

14: show impractically slow convergence with typically used numerical

15: quadrature schemes, making these functionals unsuitable both for routine

16: applications or high-precision studies, as thousands of radial quadrature

17: points may be required to achieve sub-$\mu E_{h}$ accurate total

18: energies for these functionals, while standard quadrature grids like

19: the SG-3 grid only contain $\mathcal{O}(100)$ radial quadrature points.

20: These results are both a warning to users to always check the sufficiency

21: of the quadrature grid when adopting novel functionals, as well as

22: a guideline to the theory community to develop better behaved density

23: functionals.

24: \end{abstract}

25: