Dear all,
I was trying to calculate niobium (bcc structure) phonon dispersion using response function method.
Here I followed tutorial http://www.abinit.org/doc/helpfiles/for-v8.2/tutorial/lesson_rf2.html
So I used trf2_2.in to generate my qpt for trf2_1.in.
When I use ngkpt 4 4 4, everything is right and I can get my phonon dispersion.
But if I used ngkpt 5 5 5, anaddb said body centered lattice must use even ngqpt.
--- !ERROR
message: |
For body-centered lattices, the numbers ngqpt(1:3)
must be even, while they are :
ngqpt(1)*nshiftk = 5
ngqpt(2)*nshiftk = 5
ngqpt(3)*nshiftk = 5
Action : modify ngqpt(1:3) in the input file.
src_file: smpbz.F90
src_line: 330
Could you tell me why it must be even? I thought I can use any grid I want.
Thanks,
Xing
Phonon Calculation [SOLVED]
Moderators: mverstra, joaocarloscabreu
Re: Phonon Calculation [SOLVED]
Hi Xing,
you can use a _centered_ default grid (shiftk 0 0 0 nshiftk 1) of any density, including 5 5 5.
The _optimal_ grid for BCC systems uses 2 shifts and has to be even (see paragraph under shiftk in the input variables web site) - this gives you the best quality integral for a given number of irreducible k-points.
Further, as your q-points have to be differences of k-point grids, using a prime number like 5 or 7 is very constraining. You will have to do ngqpt 5 5 5 as well...
cheers
you can use a _centered_ default grid (shiftk 0 0 0 nshiftk 1) of any density, including 5 5 5.
The _optimal_ grid for BCC systems uses 2 shifts and has to be even (see paragraph under shiftk in the input variables web site) - this gives you the best quality integral for a given number of irreducible k-points.
Further, as your q-points have to be differences of k-point grids, using a prime number like 5 or 7 is very constraining. You will have to do ngqpt 5 5 5 as well...
cheers
Matthieu Verstraete
University of Liege, Belgium
University of Liege, Belgium