[QE-users] workaround for projwfc with SG15 ONCV

[Zeeshan Ahmad]

Hi,

I ran a calculation using SG15 ONCV pseudopotentials (for one of the elements) which don't have the PSWFC section for projwfc.x. My system contains Pb and I atoms, out of the pseudopotential for I was taken from https://github.com/pipidog/ONCVPSP <https://github.com/pipidog/ONCVPSP> (FR SG15, contains the PSWFC section) but the Pb psuedopotential was Pb_ONCV_PBE_FR-1.0.upf and does not contain PSWFC section.

To get projwfc running (without rerunning the expensive hybrid scf calculation), I regenerated the FR pseudopotential for Pb with oncvpsp-3.3.1 version and with the same mesh size (There is still difference in total_psenergy in PP_HEADER). Using this pseudopotential, I added the PSWFC section in the used Pb pseudopotential located in outdir/prefix.save/. However, when I run projwfc.x, I get the following error: 

Error in routine fill_nlmchi (1):
     wrong # of atomic wfcs

On checking, the value of nwfc is 256 and the value of natomwfc is 128, equal to the value from beginning of output file:

"    Starting wfcs are  112 randomized atomic wfcs +  108 random wfcs”

So, it seems there is no possibility of hacking my way into running projwfc.x and I would have to rerun the scf calculation with the new pseuodopotential?

Also, when I run projwfc.x without adding PSWFC section to Pb pseudopotential located in outdir/prefix.save/, the projwfc runs without giving an error but it assumes all atoms are I. Is this a bug?


Thanks,
Zeeshan

--
Zeeshan Ahmad
Postdoctoral Researcher
Pritzker School of Molecular Engineering
University of Chicago

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20210219/f32388b4/attachment.html>