Page 1 of 1

BRMIX error for large jobs

Posted: Mon Jan 06, 2014 1:41 am
by Rob
Hello,

I realize there are a number of threads about the error "BRMIX: very serious problems the old and the new charge density differ". I don't mean to needlessly add my own story, but my problem doesn't seem to be solved by any of the suggestions I've seen. In short, I see the BRMIX error once the system I'm computing reaches a certain level of complexity (more than about 15 atoms, or significant empty space as in the calculation of a surface or an isolated molecule).

I use VASP on 12-core Ubuntu workstations. I compiled the parallel version using gfortran, as in the following blog post: http://www.nsc.liu.se/~pla/blog/2013/11 ... on-ubuntu/. Simple calculations of bulk crystals work fine, but for larger systems, I see energies that are several orders of magnitude too negative, and eventually the BRMIX error. For example (note the NELMDL warning in the output below doesn't appear related to the BRMIX error):

running on 12 total cores
distrk: each k-point on 12 cores, 1 groups
distr: one band on 3 cores, 4 groups
using from now: INCAR
vasp.5.3.3 18Dez12 (build Nov 14 2013 23:22:48) complex
POSCAR found type information on POSCAR Pt
POSCAR found : 1 types and 14 ions
scaLAPACK will be used
LDA part: xc-table for Pade appr. of Perdew
POSCAR, INCAR and KPOINTS ok, starting setup
FFT: planning ...
WAVECAR not read
WARNING: random wavefunctions but no delay for mixing, default for NELMDL
entering main loop
N E dE d eps ncg rms rms(c)
DAV: 1 -0.472674876884E+11 -0.47267E+11 0.17746E+05 256 0.378E+04
DAV: 2 -0.351256385226E+04 0.47267E+11 -0.68224E+03 208 0.280E+02
DAV: 3 -0.255081764471E+04 0.96175E+03 -0.45684E+03 228 0.175E+02
DAV: 4 -0.107949211155E+05 -0.82441E+04 -0.73506E+04 216 0.201E+02
DAV: 5 -0.208555131626E+05 -0.10061E+05 -0.15271E+05 224 0.160E+02BRMIX: very serious problems
the old and the new charge density differ
old charge density: 140.00000 new 133.01969
0.107E+02
DAV: 6 -0.360794669596E+04 0.17248E+05 -0.71159E+03 284 0.258E+02BRMIX: very serious problems
the old and the new charge density differ
old charge density: 133.01969 new 130.92657
0.880E+01
DAV: 7 -0.252400659838E+04 0.10839E+04 -0.45792E+03 272 0.241E+02BRMIX: very serious problems
the old and the new charge density differ
old charge density: 130.92657 new 132.05849
0.779E+01
DAV: 8 -0.140082949723E+04 0.11232E+04 -0.26614E+03 272 0.309E+02BRMIX: very serious problems
the old and the new charge density differ
old charge density: 132.05849 new 125.25754
0.463E+01
DAV: 9 -0.302343008318E+04 -0.16226E+04 -0.38930E+03 272 0.269E+02BRMIX: very serious problems
the old and the new charge density differ
old charge density: 125.25754 new 123.47641
0.566E+01
DAV: 10 -0.601971966408E+04 -0.29963E+04 -0.36866E+03 272 0.319E+02BRMIX: very serious problems
the old and the new charge density differ
old charge density: 123.47641 new 145.19303
0.613E+01
DAV: 11 -0.974734670536E+03 0.50450E+04 0.44268E+02 276 0.449E+02BRMIX: very serious problems
the old and the new charge density differ
old charge density: 145.19303 new 148.13476
0.350E+01
DAV: 12 -0.115111193931E+04 -0.17638E+03 -0.94628E+02 272 0.398E+02BRMIX: very serious problems
the old and the new charge density differ
old charge density: 148.13476 new 146.20385
0.913E+01


And so on... I have tried basically all the suggestions that I've seen on these forums, none of which have shown any signs of improvement:

1) I have generated both serial and parallel executables, and I've tried different levels of optimization (OFLAG=-O3, -O2, -O1, and -O0). In all cases, calculations are fine for a small number of atoms, but fail when the structure is sufficiently complicated.

2) Aside from changes in my VASP compilation, I have tried the various combinations of changes that users have suggested in the input files. I've tried "ADDGRID = .TRUE.", "ISYM = 0", shifting an atom away from its high-symmetry position, "PREC = High", and varying the number of k-points.


Has anyone else observed the BRMIX error only when calculating larger systems? If anyone has an idea of why this error is occurring, or a suggestion of how I might get my workstations to handle more than the simplest systems, I would really appreciate it!

Best,
Rob

Re: BRMIX error for large jobs

Posted: Tue Sep 10, 2024 2:39 pm
by support_vasp

Hi,

We're sorry that we didn’t answer your question. This does not live up to the quality of support that we aim to provide. The team has since expanded. If we can still help with your problem, please ask again in a new post, linking to this one, and we will answer as quickly as possible.

Best wishes,

VASP