num prob error

[ben_howchen1]

Hi all,

during the electronic convergence steps i get an error that reads like so


'RMM: 24 -0.284489649610E+03 -0.37419E-05 -0.16277E-05436119 0.156E-04 0.718E-03
RMM: 25 -0.284488938092E+03 0.71152E-03 -0.43201E-01319317 0.106E-02 0.934E-03
RMM: 26 -0.284489645676E+03 -0.70758E-03 -0.21207E-03365826 0.103E-03 0.510E-03
RMM: 27 -0.284489644941E+03 0.73486E-06 -0.53108E-03251050 0.703E-04 0.366E-03
RMM: 28 -0.284489259837E+03 0.38510E-03 -0.33236E-01306906 0.760E-03 0.360E-03
num prob RMM: 29 -0.284489626105E+03 -0.36627E-03 -0.11239E-02277313 0.123E-03 0.278E-03
num prob num prob RMM: 30 -0.284489649502E+03 -0.23397E-04 -0.56723E-05256700 0.171E-04 0.256E-03
num prob num prob num prob num prob num prob RMM: 31 -0.284488623998E+03 0.10255E-02 -0.48868E-01264230 0.596E-03 0.489E-03
num prob num prob num prob num prob num prob num prob num prob RMM: 32 -0.284489644279E+03 -0.10203E-02 -0.29015E-03256917 0.740E-04 0.245E-03
num prob num prob num prob num prob num prob num prob num prob num prob num prob RMM: 33 -0.284489649575E+03 -0.52958E-05 -0.22839E-05239618 0.100E-04 0.242E-03
num prob num prob num prob num prob num prob num prob num prob num prob num prob num prob num prob num prob RMM: 34 -0.284489649599E+03 -0.23676E-07 -0.13530E-07239456 0.355E-05
5 F= -.28448965E+03 E0= -.28448032E+03 d E =-.186741E-01'

Whats odd is in terms of the energetics it is of a sensible value and converges correctly. You can only see something is off when looking at the DOS and it has a range of -60 to +200 with density all over the positive which doesnt make much sense

INCAR:
ALGO = NORMAL
PREC = Accurate
IBRION = -1
ISMEAR = 0
SIGMA = 0.02
NEDOS = 1500
NBANDS =1000


I have tried adjusting the electronic minimisation algorithm, switching a gamma centred k-grid, turning off symmetry as recommended by similar posts and the problem persists.
Thanks in advance


Any ideas?

[henrique_miranda]

This error is specific to the RMM-DIIS algorithm.
You can find more information on potential problems and how to fix them here: https://www.vasp.at/wiki/index.php/IALGO#RMM-DIIS

You say that you already tried different minimization algorithms (ALGO tag) so I will assume that your problem is not related to that.
Here are a couple of suggestions:
1. Are you sure that NBANDS is enough and not too much for the system you are studying? Try running the calculation without the NBANDS tag in the INCAR in which case VASP will choose the number of bands automatically.
2. You are using gaussian smearing for the DOS calculation, are you sure this is correct? This might lead to problems in case you have a metallic system. Try running with ISMEAR=-1 which should work for both metals and semiconductors. If the calculation works then you can check if the system is metallic or not and then choose the most adequate value for ISMEAR.

[ben_howchen1]

Thanks for the advice,

Adjusting the smearing method to -1 did solve the issue

Cheers!