ABINIT Discussion Forums

Dear all
I calculate the electron-phonon interaction with follow input file:
#########################################################################
ndtset 6

#
# DATASET 1 : make ground state wavefunctions and density
#
tolwfr1 1.0d-15
#nline1 8 # This is to expedite the convergence of higher-lying bands
rfphon1 0 # for DS1 do _not_ do perturbation
nqpt1 0 # for DS1 do _not_ do perturbation
prtwf1 1 # need GS wavefunctions for further runs
getwfk1 0

qpt2 0.00000000E+00 0.00000000E+00 0.00000000E+00
qpt3 0.00000000E+00 0.00000000E+00 0.00000000E+00
qpt4 5.00000000E-01 0.00000000E+00 0.00000000E+00
qpt5 5.00000000E-01 5.00000000E-01 0.00000000E+00
qpt6 5.00000000E-01 5.00000000E-01 5.00000000E-01

#Set 2 : Response function calculation of d/dk wave function

iscf2 -3 # Need this non-self-consistent option for d/dk
kptopt2 2 # Modify default to use time-reversal symmetry
rfphon2 0 # Cancel default
rfelfd2 2 # Calculate d/dk wave function only
tolwfr2 1.0d-15 # Use wave function residual criterion instead
prtwf2 1

#Set 3 : Response function calculation of Q=0 phonons and electric field pert.

getddk3 2 # d/dk wave functions from last dataset
kptopt3 2 # Modify default to use time-reversal symmetry
rfelfd3 3 # Electric-field perturbation response only

#
# general data for all phonon calculations:
#
rfatpol 1 12
rfdir 1 1 1
rfphon 1
prtwf 0 # for response function runs, do not keep first order wavefunctions
tolvrs 1.0e-8
getwfk 1
nqpt 1
prepgkk 1 # force all perturbations to be calculated for q-points considered
prtgkk 1 # print out GKK files containing electron-phonon coupling

#
# Common data for GS and perturbation runs
#

#
# the kpoint grid is minimalistic to keep the calculation
# manageable.
#
ngkpt 4 4 4
kptopt 3 # keep all k-points. Also for ground state, because of eventual transport calculations
#
# use a centered grid for the kpoints: obligatory for el-ph for the moment
#
nshiftk 1
shiftk 0.0 0.0 0.0

#
# as is the kinetic energy cutoff
#
ecut 40.0

acell 3*1.00
rprim
-4.89864463178838E+00 4.89864463178838E+00 1.08828974363585E+01
4.89864463178838E+00 -4.89864463178839E+00 1.08828974363585E+01
4.89864463178838E+00 4.89864463178838E+00 -1.08828974363585E+01

nband 48

#
# include metallic occupation function with a small smearing
#
occopt 1
#tsmear 0.001
natom 12
typat 1 1 1 1 1 1 1 1 2 2 3 3
xred 1.0872288846E+00 4.8421399949E-01 9.0813587221E-01
5.7607812728E-01 1.1790930124E+00 9.1864127789E-02
1.2342139995E+00 -1.7392187272E-01 -1.0301488514E-01
9.2909301242E-01 2.3372288846E+00 1.1030148851E+00
-3.3722888463E-01 -2.3421399949E-01 -4.0813587221E-01
1.1739218727E+00 1.0709069876E+00 1.4081358722E+00
-4.8421399949E-01 4.2392187272E-01 6.0301488514E-01
8.2090698758E-01 -8.7228884628E-02 3.9698511486E-01
7.5000000000E-01 2.5000000000E-01 5.0000000000E-01
1.0000000000E+00 -5.5509608618E-17 3.4028916580E-21
1.2500000000E+00 7.5000000000E-01 5.0000000000E-01
1.5000000000E+00 5.0000000000E-01 0.0000000000E+00

nstep 200
ntypat 3
znucl 8 23 83

#########################################################################

As the totorial said, the q grid must be sub grid of k mesh. I use 2*2*2 q grid and 4*4*4 k grid. But, when I use mrgddb to merge DDBs after the calculation, an error occurs as follow:

.......
read 1 blocks from the input DDB

read the input derivative database number 2
compare the current and input DDB information
chki8 : ERROR -
Comparing integers for variable nkpt.
Value from input DDB is 36 and
from transfer DDB is 64.
Action : check your DDBs.

leave_new : decision taken to exit ...
.......

I then found that the nkpt for 6 datatsets are 64, 36, 36, 64, 64 and 64 respectively.
Can anyone tell me how to solve this problem?

Thanks a lot
Yours
Wjj

Hi

It's strange that you have more kpoint in the first dataset (GS calculation) than in the d/dk computation and phonon/electric field perturbations.

You should add in your first dataset kptopt1= 2. You'll then have the same kpoint grid in the first three datasets.

I don't know if it will solve your problem, but you can begin with this.

Boris

Using kptopt 3 for all dtset would be the simplest.

M.