GW for large supercells

[chandrima]

Hi
I have been trying to run a GW calculation for a 40 atom supercell of SrTiO3 with an O vacancy. Although I run a ground state calculation with LOPTICS = .TRUE. in the INCAR file, I keep getting the error message :
The derivative of the wavefunctions with respect to k (WAVEDER) |
| can not be found. You must redo the groundstate calculations |
| using LOPTICS=.TRUE. in order to write the WAVEDER file.

when I do the GW calculation. Also the WAVEDER file is present after I do the ground state calculation. I've attached my INCAR file below :

***********************************************


ISTART = 1
ICHARG = 0
NPAR = 160
# ALGO = very_fast

NELMIN = 4
NELM = 1000
EDIFF = 1E-06 stopping-criterion for ELM
ENCUT = 680

EDIFFG = -4E-02
NSW = 50 number of steps for IOM
IBRION = 1 ionic relax: 0-MD 1-quasi-New 2-CG
POTIM = 0.5

NEDOS = 1000
NBAND = 800
ISMEAR = 0
SIGMA = 0.05
LOPTICS =.TRUE.
LREAL = .FALSE.

LORBIT=11
ISPIN =1
ALGO = GW0
NOMEGA = 50

[admin]

this error message does not only show up if WAVEDER does not exist at all or is empty, but also if some crucial parameters differ between the run from which WAVEDER was written and the one by which it is read (nr of k-points, total number of bands,...). please check whether thismight have happened.