Using Ultrasoft pseudopotentials

[Nikita]

Hello!
I wonder if the new version of EPW ( version 5) could work with ultrasoft pseudopotential ? It looks like a number of mostly verified pseudopotentials are ultrasoft https://www.materialscloud.org/discover ... ficiency#0.
If no, am I right that http://www.quantum-simulation.org/potentials/sg15_oncv/ could be a suitable library for using?

Respectfully,
Nikita

[sponce]

Dear Nikita,

Indeed EPW does not yet support US. However this is in our todo list.

For norm-conserving, you can use the SG15 indeed.
I also recommand the Pseudo-Dojo ones http://www.pseudo-dojo.org/

Best,
Samuel

[Nikita]

Thank you for your answer, Samuel!

By the way, I have another question. It looks like using ibrav=0 during pw.x runs is not suitable for EPW. Is that right? Cause I have an error reading my pw.x output and don`t get what else could be the reason of error.

Respectfully,
Nikita

[sponce]

Hello,

Ibrav == 0 has a special treatment but should work in EPW.
What is the error message exactly ?

Best,
Samuel

[Nikita]

Well, then if ibrav=0 is not a problem, I don`t get what is it.

This is the input I have: https://www.dropbox.com/s/z33g1g45rvaxt ... r.zip?dl=0
And this is the error:
task # 11
from pw_readschemafile : error # 1
xml data file not found

Respectfully,
Nikita

[sponce]

Hello,

It seems EPW is unable to read some data from the nscf run.
See the reading PW routine PW/src/pw_restart_new.f90 around line 700+

The files might not be in the correct directory and not have the same filename etc

Best,
Samuel

[Nikita]

Yes, you were right - that works for me. However, I specified in EPW input the flag oudir before and got that error. Anyway now it works.

Thank you!
Respectfully,
Nikita