QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: nisreen on November 2, 2020, 09:00

Title: SCF convergence
Post by: nisreen on November 2, 2020, 09:00

Dear team,

I'm working on a large system (interface) with more than 1000 atom, however the SCF iteration is not converging at all, I've tried to increase the K-points and changing the basis functions but same issue, can I adopt the no SCF calculations? is it accurate enough? if not, what can I do to solve the SCF convergence issue?

any suggestions will be much appreciated

thanks

Title: Re: SCF convergence
Post by: mlee on November 3, 2020, 08:12

This is not easy to guide it.
But if you have a large or complex configuration, you can try to converge step by step as a rough and constraint condition at first, then tighter condition using the pre-optimization configuration.