UT theoretical chemistry code forum

Hi, all,

I'm new to bader analysis. I want to do a test to check if the increased FFT grid resolution will importantly influence the bader charge of my system. My question is how to increade the FFT grid resolution in the code? From the source code, it seems the FFT grids numbers are read from the CHGCAR file. So if I want to do a convergence bader analysis test using 80^3, 100^3, 200^3, 300^3 and 500^3 grid numbers, I must do 5 times VASP calculations using NGX, Y, Z=80, 100, 200, 300, 500, respectively?

Thanks!

Dongshan

Yes, that's right.

Thank Graeme's reply.

More detailed, Should I start the vasp calculation from scratch? May I use the optimized (relaxed) structure from the previous vasp calculation which used the default FFT grid number and just do one-step ionic step calculation using the new FFT grid number? Since my system is quite huge, if all of vasp calculations at different FFT grid numbers need to re-do from scratch, it will be a very tough project.

Dongshan

You can do a single-energy point calculation. For large grids, the issue is usually memory instead of cpu.