External optional components of Abinit (BigDFT, Wannier90...) and codes using/providing data from/to Abinit (AtomPAW, ONCVPSP, EXC, DP, Yambo...)
Moderators: ebousquet, bxu
-
jennydaman
- Posts: 3
- Joined: Tue Dec 13, 2016 10:08 pm
Post
by jennydaman » Tue Dec 13, 2016 10:16 pm
Hi, I'm performing convergence studies on monolayer silicon (silicene) with abinit 7.8.2 . I'm using the file Si.GGA_PBE-JTH.xml from the JTH PAW atomic datasets table as my pseudopotential, but no matter the configuration, I always get this warning 300+ times:
Code: Select all
pawio_print_ij: WARNING -
The matrix seems to have high value(s) !
( 1 components have a value greater than 50.0).
It can cause instabilities during SCF convergence.
Action: you should check your atomic dataset (psp file)
and look for "high" projector functions...
Also, I get this once towards the beginning of the file:
Code: Select all
pawpsp_calc: WARNING -
Max. radius for Vloc was too large (> 20.00 a.u.) !
Numeric noise was possible.
Mesh size for Vloc has been set to 1772.
Compensation charge density is not taken into account in XC energy/potential
beta integral value: -0.150771E+06
pspatm: atomic psp has been read and splines computed
Is there any way to fix this, or is this even a significant issue? My only other work around is to use the old 2009 pseudopotentials...
Thank you for your time.
-
ebousquet
- Posts: 469
- Joined: Tue Apr 19, 2011 11:13 am
- Location: University of Liege, Belgium
Post
by ebousquet » Thu Dec 15, 2016 7:47 pm
Dear jennydaman,
If you don't encounter instabilities in the SCF convergence, then I would say this is fine. The new JTH might be less good on that aspect than the old one, but if the old one works better you can stay with it (your remark will be use to improve the next PAW table). If it is too troublesome, you can also test the dataset given by GBRV:
http://www.physics.rutgers.edu/gbrv/.
Best wishes,
Eric