VASP 5.2 band structure calculation error!!

[wcwang]

Hi, All,
I’m testing VASP 5.2 GWA to do the band structure. There is a fatal error about WAVEDER file. To make it clear, I list what I did below.
Step.1 Do ground state DFT calculation
Step.2 Do G0W0 calculation
Step.3 Do self consistent GW correction to get WAVECAR and CHARGE files for the sequent band structure calculation.
Step.4 Do one kpoint band structure test.
From step 1 to step 3 which follow the manual, there is no problem at all. However, I got one error in the last step, it shows,
“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.�
In fact, I already copied WAVEDER file from step. 1 to the current calculation, why does not it work here. The step 4’s INCAR and KPOINTS are listed below.
INCAR:
SYSTEM = Si
NBANDS = 96
ISMEAR = 0; SIGMA = 0.05
ALGO=GW0
NOMEGA=50
NELM=4
KPOINTS:
auto
1
r
0 0 0 1
Any suggestion is appreciated, thanks.

[panda]

Did you redo the groundstate calculations using LOPTICS=.TRUE. in order to write the WAVEDER file? Because I do not see it specified in your INCAR! Following the advice of the error message is always the first step... Simply copying the WAVEDER from Step 1, if in Step 1 LOPTICS = .TRUE. is not specified will not work. You will have to start over at step 1 with LOPTICS = .TRUE. and then your calculation should work.

[wcwang]

Thanks for the information. As a matter of fact, Kpoint sampling matters a lot. If I set G-center sheme, there is no prolem at all. However, if I specify some kpoint, say 0 0 0, there shows the erros as I first posted.

[panda]

Hmm, in that case I have no idea. When you specify gamma is this in auto K-point generation scheme?