Noisy DOS

[Dhirendra]

I am getting a very noisy DOS on a supercell with >56 atoms (Si and Ni). Any suggestions on how to obtain neat DOS plots?

[Jess Wellendorff]

In general I would recommend to use a fine k-point sampling and the default Tetrahedron method for DOS calculations for bulk configurations. If you double the k-point sampling and the DOS spectrum does not change, perhaps the noise is not really noise after all.

[Dhirendra]

Thanks, may be an nscf calculations with dense grid will help.