My Community

Hi,

I encounter warning: " PSMAXN for non-local potential too small". I saw from previous discussions that ENCUT is too large compared to ENCUT and has to be decreased or harder potentials need to be used, and the order of atoms in POTCAR file should be ordered descending in ENMAX and LREAL should be set to False.

I tried all of the above suggestions (sv_GW potentials in POTCAR), but I still get the error for ENCUT = 1000, I use R2SCAN functional with VdW correction, the k-grid is 24 4 6 Gamma-point. My system is compose of Ta Ni Se and the hardest potentials I can find are sv_GWs. Do you have any suggestion how to mitigate this warning?

You can find the INCAR and POTCAR files in the attachment.

Regards,
Burak

Hi Burak,

Thansk for reaching out. I will try to help you resolve the issue.

First I would like to reproduce the error on my end, so I would ask you to upload all the input and output files according to the forum posting guidelines.

Kind regards

Hi Manuel,

thanks, please find the files in the attachment

Regards,

I was able to reproduce your case and tried different configurations. The only two ways to consistently get rid of the warning message is to reduce ENCUT or to switch to the norm-conserving pseudopotentials, so

• Ni_sv_GW_nc

• Se_sv_GW_nc

• Ta_sv_GW_nc

These POTCAR files are the "hardest" we have in our dataset and I was able to run your example at ENCUT=1000 without the warning.

Hi Manuel,

thanks a lot. Where can I find these potentials? Which version of VASP these come with?

I also looked at
https://www.vasp.at/wiki/index.php/Avai ... potentials
but could not find mention of these potentials. Maybe I am missing a point.

Regards,
Burak

I'm sorry for the confusion. The nc potentials are used internally by VASP developers for testing but are not publicly available. I was not aware of this fact when I posted my response. For now, the only way I could find to get rid of the warning message is to lower ENCUT.

Kind regards,

Hi Manuel,

thanks, do you know in a foreseeable future, these potentials will be available to us?

Regards,
Burak

It is something that is being discussed. However, I cannot tell you when, if at all, these potentials would be released publicly. Not all of the norm-conserving potentials have been tested with the same rigor as the currently available potentials. At the very least we would have to test the entire database thoroughly so that the quality of the potentials lives up to our standards.