problem with PAW+GW cal. of SiO2 (nonsymmorphic operations)

Moderators: gmatteo, bruneval, rigna

Post Reply
User avatar
gmatteo
Posts: 269
Joined: Sun Aug 16, 2009 5:40 pm

problem with PAW+GW cal. of SiO2 (nonsymmorphic operations)

Post by gmatteo » Thu Mar 04, 2010 8:38 pm

Dear all,

I was trying to do the GW correction calculation for SiO2 using PAW psps, The
following error message
was met with during the 3rd dataset to calculate the screening:
-------------------------------------------------------------------------------------------------
screening.F90:648:ERROR
Cryst%nsym/=Dtset%nsym, check pawinit and symrhoij
-P-0000
-P-0000 leave_new : decision taken to exit ...
----------------------------------------------------------------------------------------------------
after I change the parameter "symmorphi " from 0 to 1, it worked.
but in tutorial_GW1,
http://www.abinit.org/documentation/hel ... n_gw1.html
it's advised to set this input variable as 0,
I am not sure how much difference would come from the change of this variable?
If anyone is familiar with GW, advices and comments will be welcomed.

Thanks,
The following is my input file ,and the PAW psps I used were downloaded from the
ABINIT PAW table website .
______________________________________________________________________
#crystalline SiO2
#Calculation of the GW corrections
#
# option for output
prtden 1
ndtset 4
gwpara 2 # GW parallelisation on bands
nsym 0
# dataset1: usual self-consistent ground-state calculation
nband1 32 #number of (occ. and emp.) bands to be computed
prtden1 1
# dataset2: calculation of KSS file
iscf2 -2
tolwfr2 1.0d-2
getden2 -1
nband2 60 # number of bands to be computed
kssform2 3
nbandkss2 100 # number of bands in KSS files (-1 maximum)
# dataset3: calculation of the screening (epsilon^-1 matrix)
optdriver3 3 # Screening cal.
gwcalctyp3 2 # GW calculation using numerical integration
symchi3 1 # Symmetryze \chi_0, summation over IBZ
getkss3 -1 # Obtain KSS file from previous dataset
nband3 60 # Bands to be used in the screening cal.
ecutwfn3 2.5 #
ecuteps3 2.5 #
nfreqim3 10
nfreqre3 10
freqremax3 2.0
#ppmfrq3 0.5 # Plasmon-pole frequency
fftgw3 11 # FFT for GW
awtr3 1 # Adler-Wiser expression assuming Time-reversal
# dataset4: calculation of the self-energy matrix elements (GW corrections)
optdriver4 4 # Self-energy calculation
gwcalctyp4 2 #
symsigma4 1 #
getkss4 -2 # Obtain KSS file from dataset 1
getscr4 -1 # Obtain SCR file form previous dataset
nband4 60 # Bands to be used in the Self-Energy cal.
ecutwfn4 2.5 #
ecutsigx4 2.5 #
nkptgw4 1 # number of k-points where to cal. GW correction
kptgw4 0.00 0.00 0.00
bdgw4 12 18 # calculate GW corrections for bands from 12~18

#################### COMMON PART ##################################
#Definition of the unit cell
acell 3*5.41 angstrom
rprim 0.5 0.5 0.0
0.0 0.5 0.5
0.5 0.0 0.5
xred 0.0 0.0 0.0
0.25 0.25 0.25
0.75 0.75 0.75
natom 3
ntypat 2
typat 1 2 2
znucl 14 8

# Definition of the k-point grid
kptopt 1
ngkpt 4 4 4
nshiftk 4
shiftk 0.5 0.5 0.5
0.5 0.0 0.0
0.0 0.5 0.0
0.0 0.0 0.5
istwfk *1
iscf 17 # PAW SCF convergence method
#toldfe 1.0d-10
tolvrs 1.0d-8
# energy cutoff [Ha]:
ecut 10
pawecutdg 20.0
#ecutsm 0.5 # ecut smearing useful when relaxation
occopt 4 #cold smearing
tsmear 0.005
# option for the choice of calculations
diemac 12 # a rough guess of the dielectric constant
ixc 7
nstep 200
# Use only symmorphic operations
symmorphi 0



--

State Key Laboratory for Surface Physics and Chemistry
P.O. Box 718-36,Mianyang 621907
Sichuan, P.R. China
Email:shipeng_2000@163.com

User avatar
gmatteo
Posts: 269
Joined: Sun Aug 16, 2009 5:40 pm

Re: problem with PAW+GW cal. of SiO2 (nonsymmorphic operations)

Post by gmatteo » Thu Mar 04, 2010 8:45 pm

The GW documentation is not up-to-date.
Now the GW code can treat nonsymmorphic operations both in screening and in sigma.
Nonsymmorphic operations can always be removed by using symmorphi=0 but
one has to be consistent through all the steps of the calculation.
If the KSS file has been generated with symmorphi=0, then also the screening
calculation as well as the sigma part have to be performed with the same value.

Best Regards
Matteo Giantomassi

sappler
Posts: 1
Joined: Thu Mar 04, 2010 11:53 pm

Re: problem with PAW+GW cal. of SiO2 (nonsymmorphic operations)

Post by sappler » Fri Mar 05, 2010 7:58 am

Dear Matteo Giantomassi,
Thanks for your quick reply,
and here's another question: is there a good way to get GW correction for common k-points?
for example: If you want to calculate the band structure along high symmetry lines, such as L-W-gamma, it's difficult to generate the k-mesh which includes the special k-points you are intrested in, during the KSS file generation procedure. If not included, then the following route to calculate screening would complain the kptgw is not one point in the k-mesh generated before.
I think this is a general problem when doing GW correction calculations.

Any comments and advices are welcomed, thanks,

Shipeng

Post Reply