[Wien] more accurate total energy

[Peter Blaha]

> At the moment I'm doing calculations for which I need very accurate energies. So I want to add 2 more digits to the total energy printed in the outputfiles.

The total energy is calculated in mixer, where it reads the scf file and
adds up all partial energies: The relevant lines are:

:DEN  : DENSITY INTEGRALS =            -757.695273   (Ry)     written in lapw0

:SUM  : SUM OF EIGENVALUES =            -12.512277            written in lapw2
                                                  (and read+written in sumpara)
:1S 01: 1S                   -356.368828 Ry
:2S 01: 2S                    -38.029220 Ry
:2PP01: 2P*                   -31.500320 Ry
:2P 01: 2P                    -31.076109 Ry
and similar                                                   written in lcore



> Therefore I have changed the format for total energy from F20.6 to F20.8 in the following files:
>
> SRC_mixer/mixer.f
> SRC_mixer/scfana.f
> SRC_mini/wrtscf.f
> SRC_elast/anaelast.f
> SRC_lcore/hfsd.f
> SRC_sumpara/sumpara.f
> SRC_sumpara/scfsum.f
> SRC_telnes/hfsd.f
> SRC_lapw0/energy.F
> SRC_lapw0/lapw0.F
> SRC_lapw2/l2main.F

Sounds reasonable, but it is hard to tell if you have got all of them
(or if all are necessary (telnes ?)

> and the format for eigenvalues, since they are used to calculate the total energy, from F13.7 to F13.9 in the following files:
>
> SRC_lapwso/hmsec.f
> SRC_lapwso/kptout.f
> SRC_spaghetti/get_ei.f
> SRC_lapw1/prtkpt.F

I don't think this was necessary. The eigenvalues are read internally from
case.energy or case.vector, and have a large precision there.

                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------