TDHF: unexpected output parameters in OUTCAR

[ashwin]

I am running the TDHF example for Silicon using the INCAR from https://cms.mpi.univie.ac.at/wiki/index ... lculations and the usual POSCAR, POTCAR, and KPOINTS from http://www.vasp.at/vasp-workshop/examples/Si_BSE.tgz.

While my calculations seem to produce reasonable results, there is something a bit unexpected in the OUTCAR. Specifically, I use:

ALGO = TDHF; LHFCALC = .TRUE. ; HFSCREEN = 0.2; AEXX=0.25

in the INCAR. The OUTCAR says "parameters for screened Coulomb W: AEXX= 0.2500000 HFSCREEN= 0.2000000" as expected. Rather unexpectedly though, it also says:

Exchange correlation treatment:
GGA = -- GGA type
...
LHFCALC = T Hartree Fock is set to
...
AEXX = 1.0000 exact exchange contribution
HFSCREEN= 0.0000 screening length (either q_TF or 0.3 A-1)
HFSCREENC= 0.2000 screening length for correlation (either q_TF or 0.3 A-1)
...

Why is AEXX = 1.0 and HFSCREEN = 0.0 in the exchange-correlation block in conflict with the INCAR and the "parameters for screened Coulomb" written out later on?

[admin]

The values printed in the list of parameters of the exchange correlation treatment
at the beginning of OUTCAR are not correct. Correct values are printed
in the line "parameters for screened Coulomb W:"
Switch to the newest version of the vasp code. This prints the parameters correctly.

[ashwin]

Thanks, Admin. Unfortunately, upgrading to the newest version does not resolve this problem. I am using vasp.5.4.1 05Feb16 and I have applied all of the released patches.