When using the diagonalization method for semiconductors, should the SMEAR parameter be included?

[yis...@163.com]

Dear CP2K developers,

Hello everyone! I am a new CP2K user and have recently been using the diagonalization algorithm to optimize a nickel hydroxide (Ni(OH)₂) unit cell. During my calculations, I noticed that adding the SMEAR parameter accelerates SCF convergence. I am using the FERMI_DIRAC method with an electronic temperature of ELECTRONIC_TEMPERATURE = 300 K.

Since nickel hydroxide is a semiconductor, I would like to ask for your advice: Is it recommended to include the SMEAR parameter in structural optimizations for semiconductor systems? Could using this parameter affect the accuracy of the optimization results?

I look forward to your guidance. Thank you very much!