I’ve noticed several cases that used ismear=-5 for relaxation task, like mp-30 (Cu)
But why? As far as I know this tetrahedron method is accurate for the final single point calculation after the relaxation. Any rationale for not using ismear 1 or 2 for metal ?
Hi @JM_Choi, you’re right that ISMEAR = -5 shouldn’t be used for relaxations. This is a mistake in some older data in MP.
In many cases, the calculation used for the relaxed structure has already been corrected. In your example, mp-30, the calculation used to relax the structure uses a reasonable ISMEAR.
You can see all individual calculations that build a material entry on its calculation details page.