[Wien] New exchange-correlation potential

[Stefaan Cottenier]

> 1. In GaAs.inm_vresp, YES has been replaced by NO;
> 2. I have done the calculation as following steps:
>     For the first scf run, I have set NR2V and indxc=5 in GaAs.in0;
>     Change NR2V to R2V and run one more scf cycle;
>     save lda;
>     Set indxc=28 in GaAs.in0;
>     cp GaAs.in0 GaAs.in0_grr and set index=50 in GaAs.in0_grr
>     run_lapw -p -cc 0.00001

Did you change the mixing method to PRATT in the run_lapw step? And did 
you check afterwards whether the scf-cycle had properly converged? If 
you didn't change it to PRATT, it will have probably diverged, and your 
band gap will be a 'random' result. Just do 'grep :DIS case.scf' to verify.

Stefaan