Problem with IBZKPT and HSE?
Posted: Thu Sep 05, 2013 3:18 pm
I am having a strange problem:
(Using vasp.5.3.3)
When running an HSE calculation, starting from GGA wave-functions.
If I run using the automatic k-point generation scheme, everything is fine.
However, if I take the IBZKPT file generated from the GGA run, and use it to start the HSE calculation, with the ONLY difference being that I copy the IBZKPT onto the KPOINTS file, I get a warning about the k-point mesh being non-commensurate with the symmetry of the lattice. (If I use automatic generation, the IBZKPT file generated is exactly the same between the GGA and HSE calculations.)
I am already using a gamma centred mesh. It does not make sense that the output of the k-point generation routine is not usable as input.
Turning ISYM = 0 does nothing. Any suggestions?
This is reproducible on all of the systems I have tried - including bulk Si.
I include some simple input files below that will produce the above behaviour.
**************************
HSE INCAR:
**************************
SYSTEM = bulk Si Calculation
GGA = PE
ISTART = 1
NBANDS = 24
ISMEAR = -5
#comment these lines for the GGA INCAR
LHFCALC = .TRUE.
AEXX = 0.25
AGGAX = 0.75
HFSCREEN = 0.2
PRECFOCK = N
**************************
KPOINTS:
**************************
Automatic mesh
0 ! number of k-points = 0 ->automatic generation scheme
Gamma ! fully automatic: MP-set
6 6 6 ! 8x8x8
**************************
POSCAR
**************************
Silicon Test
1.00000
0.000000 2.715500 2.715500
2.715500 0.000000 2.715500
2.715500 2.715500 0.000000
Si
2
Selective Dynamics
Direct
0.12500 0.12500 0.12500 F F F
0.87500 0.87500 0.87500 F F F
(Using vasp.5.3.3)
When running an HSE calculation, starting from GGA wave-functions.
If I run using the automatic k-point generation scheme, everything is fine.
However, if I take the IBZKPT file generated from the GGA run, and use it to start the HSE calculation, with the ONLY difference being that I copy the IBZKPT onto the KPOINTS file, I get a warning about the k-point mesh being non-commensurate with the symmetry of the lattice. (If I use automatic generation, the IBZKPT file generated is exactly the same between the GGA and HSE calculations.)
I am already using a gamma centred mesh. It does not make sense that the output of the k-point generation routine is not usable as input.
Turning ISYM = 0 does nothing. Any suggestions?
This is reproducible on all of the systems I have tried - including bulk Si.
I include some simple input files below that will produce the above behaviour.
**************************
HSE INCAR:
**************************
SYSTEM = bulk Si Calculation
GGA = PE
ISTART = 1
NBANDS = 24
ISMEAR = -5
#comment these lines for the GGA INCAR
LHFCALC = .TRUE.
AEXX = 0.25
AGGAX = 0.75
HFSCREEN = 0.2
PRECFOCK = N
**************************
KPOINTS:
**************************
Automatic mesh
0 ! number of k-points = 0 ->automatic generation scheme
Gamma ! fully automatic: MP-set
6 6 6 ! 8x8x8
**************************
POSCAR
**************************
Silicon Test
1.00000
0.000000 2.715500 2.715500
2.715500 0.000000 2.715500
2.715500 2.715500 0.000000
Si
2
Selective Dynamics
Direct
0.12500 0.12500 0.12500 F F F
0.87500 0.87500 0.87500 F F F