DENFER

Message

rudra-b · #1 Post by **rudra-b** » Sat Jun 09, 2007 5:22 pm

in a gamma point calculation i am getting an warnning like:
vasp.4.6.21 13Feb03 gamma-only
POSCAR found : 2 types and 8 ions
LDA part: xc-table for Ceperly-Alder, Vosko type interpolation para-ferro
POSCAR, INCAR and KPOINTS ok, starting setup
WARNING: wrap around errors must be expected
FFT: planning ... 3
reading WAVECAR
initial charge from wavefunction
entering main loop
N E dE d eps ncg rms rms(c)
DAV: 1 -0.676711387020E+04 -0.67671E+04 -0.12361E+05 250 0.108E+03
WARNING: DENFER: can't reach specified precision
Number of Electrons is NELECT = 53.0000000001705
DAV: 2 -0.700301389455E+04 -0.23590E+03 -0.23309E+03 312 0.359E+02
WARNING: DENFER: can't reach specified precision
Number of Electrons is NELECT = 52.9999999997584
DAV: 3 -0.706351552510E+04 -0.60502E+02 -0.57868E+02 312 0.206E+02
WARNING: DENFER: can't reach specified precision
Number of Electrons is NELECT = 53.0000000001137
WARNING: DENFER: can't reach specified precision
Number of Electrons is NELECT = 52.9999999998721
DAV: 4 -0.708285803918E+04 -0.19343E+02 -0.18959E+02 312 0.109E+02
WARNING: DENFER: can't reach specified precision
Number of Electrons is NELECT = 52.9999999998863
DAV: 5 -0.708759138287E+04 -0.47333E+01 -0.47145E+01 312 0.519E+01B RMIX: very serious problems
the old and the new charge density differ
old charge density: 52.50000 new 52.99994
0.179E+02
WARNING: DENFER: can't reach specified precision
Number of Electrons is NELECT = 53.0000000001137
DAV: 6 -0.412291051987E+04 0.29647E+04 -0.19215E+03 178 0.205E+02 0.152E+02
DAV: 7 -0.297687435536E+04 0.11460E+04 -0.29234E+03 264 0.259E+02 0.151E+02

can anyone plz. suggest the source of the problem?

#2 Post by **admin** » Fri Jun 15, 2007 12:12 pm

the charge density cannot be integrated up to the required numberof electrons, (using the gamma point only, Fermi spearing and a broadening at the fermi edge as given by SIGMA (please check OUTCAR for the actual SIGMA)
please check:
1) is it justified to use gamma only (in case you have a bulk cell, 1 k-point is scertainly not enough for a cell of 8 atoms)
2) for a cluster or a molecule, the default SIGMA is not appropriate (line spectra, no broadening)
3) alternatively, you may use gaussian integration (ISMEAR=0)