QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: Inaoton on May 19, 2017, 11:54

Title: Using Initial State
Post by: Inaoton on May 19, 2017, 11:54

Sir
        I run a script for a crystal subjected to a pressure of 10Gpa. And now i want to run for a pressure of 20Gpa.(Here i want to calculate enthalpy)
        In that case, if i use the output .nc file of the previous calculation (10Gpa) as the initial state, and give a pressure 20Gpa
        1. Will it give the same result ? (Comparing with the result run without no initial state)
        2. Will it converge quickly?
        3.Can  the same method  be used for calculating any other properties at different pressures?(like elastic properties, optical properties)
        Please guide

Title: Re: Using Initial State
Post by: Jess Wellendorff on May 22, 2017, 12:39

I am guessing you run OptimizeGeometry with various pressures. You can then uase the initial_state argument to the ATK calculator to start one calculation from the self-consistent ground state from another calculation.
1+2: Yes, if you start the calculation at 20 GPa from the configuration obtained at 10 GPA, without changing the calculator, you should obtain the same result, and often (but not always) in fewer SCF iterations.
3: Yes, but bear in mind that this works best if you do not make drastic changes to the configuration (like changing elements) or to the calculator (like changing k-point sampling).

Title: Re: Using Initial State
Post by: lknife on May 23, 2017, 22:13

I have a question: if you optimized the structure under a pressure of 10GPa and got a configuration, after that, you re-optimize the previously obtained structure under a pressure of 20 GPa, will the final configuration be the same as the configuration being optimized under 20 GPa or 20+10=30 GPa?

Title: Re: Using Initial State
Post by: Anders Blom on May 23, 2017, 22:24

20 of course