Correct Usage of UPF Files from pseudopotentiallibrary.org

[Gopal Iyer]

Dear All,

I am trying to use the .upf pseudopotential files (boron, in this case) from the Pseudopotential Library for the Quantum ESPRESSO (6.4) PwSCF portions of my Nexus scripts. Here are some things I tried and the corresponding error messages I received:

• Using the .upf file as-is: file ./B.upf not readable
• Renaming the .upf file to .ncpp: Wrong nlc or nnl
• I also tried running QE with the .xml files, but it crashed, as expected, because QE doesn't support XML PPs

I feel like I am missing something trivial here. Any advice on how to fix these issues would be appreciated. Thanks!

Best,

Gopal

[Paul R. C. Kent]

This sounds like a file accessibility problem.

Have you tried using the files outside of Nexus? You could also look at the QE inputs created by Nexus and verify that the PATHs are OK.

Perhaps also check the UPF files by eye for any obviously bad header lines, mysterious characters, truncation etc.

I suggest to solve the usage problem with QE directly and then tackle Nexus if necessary.

[Gopal Iyer]

Hi Paul,

I tried running it separately with QE but the problem persists. To double-check, I also tried it with other externally available NCPPs with the same directory structure and it worked fine.

Also, I noticed that the UPF/NCPP file has no linebreaks anywhere, at least in the form that I was able to download it. Is this supposed to be the case? I couldn't spot any other glaring issues with it. I have attached it here for reference. Thanks!

[Guangming Wang]

Hi Gopal,

I tried a test calculation of B atom calculation(single atom in a box with 100 Ry cutoff). The result looks OK to me. I got the PBE energy -5.21629459 Ry. (1 Ha = 2 Ry)

The exact energy from our previous paper is −2.61531(13) Ha. This is the paper link. https://pubs.acs.org/doi/pdf/10.1021/acs.jctc.9b00962

I am attaching my run below. I ran the calculation on CORI NERSC machine with QE 6.4.1 version.

Best,

Guangming

[Gopal Iyer]

Hi Guangming,

I see that the UPF file you are using is formatted quite differently from the one available on pseudopotentiallibrary.org (yours has markup language–like tags that UPF files are supposed to have). Could you please tell me how you generated that file? Ideally, I would like to reproduce the procedure for other elements. Did you perhaps use ppconvert or opium? Thanks!

Best,

Gopal

[Guangming Wang]

That's the same file as shown on the https://pseudopotentiallibrary.org actually.

I would recommend to use "wget" to obtain the file. It might bring up some format problems when directly copying the texts, etc.

For example, here, you can do:

wget https://pseudopotentiallibrary.org/recipes/B/ccECP/B.upf

It should give you the correct format UPF file.

[Gopal Iyer]

Okay, that explains it. Thank you very much!

[Guangming Wang]

You are very welcome!