The ABINIT Users Mailing List ( CLOSED )

["zhang yigang" <zhangyg@mail.igcas.ac.cn>]

Dear Abinit users:

 

I encounter a problem in the GW calculation of Abinit, and wish someone would like to kindly help me out.

 

Because the calculation of the Screening takes a lot of time ( and memory, too), I split it into many proportions on different computers for different q-points, and then merge them using mrgscr.  When I do that, the program mrgscr complains:

......

r-rotation tables set up

**error surot 10: g-shell not closed

......

 

I am using Abinit version 5.2.3. The input file for the fist part of calculation, i.e., GS and KSS is

#Dataset 1 : usual self-consistent ground-state calculation
#Definition of the k-point grid
kptopt1 1                      #automatic generation
kptrlatt1  6 6 3 2 -2 3 8 4 0  #kptrlen of 41, has 43 k-points
nshiftk1 1
shiftk1 0.0 0.0 0.0
istwfk1 43*1
#Definition of the SCF procedure
toldfe1 1.0d-10                 #tolerance on total energy
prtden1 1
nband1 150

 

#Dateset 2 : Calculation of the KSS file
#Definition of the k-points
kptopt2 1                      #automatic generation
kptrlatt2  6 6 3 2 -2 3 8 4 0  #kptrlen of 41, has 43 k-points
nshiftk2 1
shiftk2 0.0 0.0 0.0
istwfk2 43*1                   #option needed for special k-points like Gamma
#Definition of the SCF procedure
iscf2 -2                       #non-self-consistent calculation
getden2 -1                     #read previous density file
tolwfr2 1.0d-14                #still get it converged
nband2 150
#Definition of parameters for the calculation of kss file
nbandkss2 150                  #number of bands in the KSS file

 

The input file for the second part of the calculation, i.e., SCR is

 #Dataset 3 : Calculation of the screening (epsilon^-1 matrix)
optdriver3 3      #screening calculation
getkss3 2         #obtain kss
nband3 150        #bands to be used in the screening calculation
ecutwfn3 10.0     #cutoff energy of the planewave set to represent the wavefunction
ecuteps3 10.0     #cutoff energy of the planewave to represent the dielectric matrix
nqptdm3 2
qptdm3 0.000010 0.000020  0.000030
       0.000000 0.000000 -0.333333

 

with the common part of the input file as:

#definition of the primitive cell & structure
acell  8.0071777375E+00  8.0071777375E+00  1.1599268320E+01
rprim  2.9111461655E-01 -9.5668818328E-01  0.0000000000E+00
       2.9111461655E-01  9.5668818328E-01  0.0000000000E+00
       0.0000000000E+00  0.0000000000E+00  1.0000000000E+00
ntypat 3
znucl 12 14 8
natom 10
typat 2*1 2*2 6*3
xred
-2.5326455804E-01 2.5326455804E-01  2.5000000000E-01
2.5326455804E-01 -2.5326455804E-01  7.5000000000E-01
0.0000000000E+00  0.0000000000E+00  0.0000000000E+00
0.0000000000E+00  0.0000000000E+00  5.0000000000E-01
-9.3051419496E-01 9.3051419496E-01  2.5000000000E-01
9.3051419496E-01 -9.3051419496E-01  7.5000000000E-01
-6.3507149034E-01 6.3507149034E-01  4.4300772496E-01
6.3507149034E-01 -6.3507149034E-01  9.4300772496E-01
-6.3507149034E-01 6.3507149034E-01  5.6992275041E-02
6.3507149034E-01 -6.3507149034E-01 -4.4300772496E-01

 

#definition of planewave basis set
ecut 50 #default unit for energy is Hartree

 

#Use only symmorphic operations
symmorphi 0

 

In the LOG file for calculating the screening, there does have a warning like:

setshells: WARNING-

The value ecutwfn=  0.100000E+02 given in the input file leads to
  the same values for nshwfn and npwwfn as ecutwfn=  0.972991E+01
  This value will be adopted for the calculation.

 

A similar warning also exists for ecuteps, but I am not sure if they are related to the above mrgscr error.

 

 

I am looking forward to your kind help.

 

with best regrads!

yigang zhang