[Wien] Some (bug ?) in SRC_lapwdm/l2main.frc in determining NEMIN, NEMAX

[Peter Blaha]

Do you have eigenvalues lower than EMIN ? Did you use an EMIN larger than
EF ?
I've never used this in lapwdm, because usually you do NOT want to cut
away the low energy states, but they should be all used. (This is
different to lapw2, where you may want to cut-off semicre states before
plotting).

Anyway, compare the diagonal elements of the density matrices with the
QTLs of lapw2. Are they the same ? If yes, your fix could be ok...???

>   It seems that in rare cases (NEMAX+1) would be smaller  than NEMIN,
> which leads  to feed negative 4th argument to zgemm/dgemm, program
> loader refuses to go further and  lapwdm crashes.
>
> According to the following lines of code, I really don't see any logic
> flow to guarantee that (NEMAX+1)>=NEMIN:
>
>  IF(E(NUM).LT.EMIN) NEMIN=NEMIN+1
> IF(abs(WEIGHT(NUM)).gt.1d-5) NEMAX=NEMAX+1
>
>
> So I changed them to
>      IF(abs(WEIGHT(NUM)).gt.1d-5) then
>         NEMAX=NEMAX+1
>         IF(E(NUM).LT.EMIN) NEMIN=NEMIN+1
>       ENDIF
>
> And it works and the program does not crash. However, I am not very sure
> whether the correction changes the physics significantly. Or should
> we just ignore those parts feeding negative 4th argument to zgemm  ? Or
> better treatment ?
> These lines of codes are the same in wien2k and wien2003.
>
> Sincerely,
>
> Tianjiao
>
>
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>


                                      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/
--------------------------------------------------------------------------