Dear all,
I found something rather curious and I was wondering if this is systematic
or a glitch. When working with hybrids (HSE06) PAW pseudos work much better
than NC despite everything else (equilibrium structure, bands, phonons)
being virtually identical. Is there a reason for PAW showing much
This is quite surprising: there has been much more development work on
hybrid functionals with NC PP than with PAW. If you have simple cases of NC
PP calculatons failing for no apparent good reason, please report them
(preferrably on gitlab.com)
Paolo
On Mon, Sep 17, 2018 at 12:03 AM, Christoph W