[Wien] Segmentation fault in f7splt (lapw2)

Pavel Ondračka pavel.ondracka at email.cz
Wed Oct 9 11:25:18 CEST 2013


Dear prof. Blaha,

I'm still not sure if we have an understanding. This is probably my
fault for not being clear enough.

In your first mail you said, that F7SPLT is being called _only_ when
ISPLIT is set to 15. This however is not what the corresponding code
says:

csplit.f:76 !   IF(L.EQ.3 .and. isplit(jatom).eq.15)     &
csplit.f:77     IF(L.EQ.3)                               &
csplit.f:78          CALL F7SPLT
(ALM(1,NUM),BLM(1,NUM),cLM(1,NUM,ipip),MULT(JATOM), &
csplit.f:79          PEI(L),num,coord,dmat_f)

When I was talking about code being commented out, I wasn't referring to
the F7SPLT call at line (csplit.f:78) but rather to the line that
contains check for ISPLIT = 15 (csplit.f:76). 

So basically in stock WIEN2k_13.1 F7SPLT is being called every time
regardless of ISPLIT value and from your responses I have a feeling that
this isn't intentional.

Best regards
Pavel Ondračka

Peter Blaha píše v Út 08. 10. 2013 v 16:02 +0200:
> At present it could be commented out.
> However, it is planned to calculate the density matrix (for -orb) in 
> lapw2 instead of an extra lapwdm call and at that stage one would need 
> this subroutine.

> Am 08.10.2013 11:45, schrieb Pavel Ondračka:
> > Peter Blaha píše v Út 08. 10. 2013 v 10:45 +0200:
> >> You are right.
> >> Anyway, a quick fix should be to add a line
> >>
> >>         ipip=max(ilo(3),1)
> >>
> >> However, in comparison to p3splt and d5splt.f I can see that there are
> >> cross-terms with local orbitals missing in most of these equations, so
> >> I'd very much doubt that these numbers are correct when you have f-los.
> >>
> >> In any case, the fsplit is not really used for anything (except when you
> >> use ISPLIT=15, which should not be done and the qtl program is a much
> >> better alternative.
> >
> > So I'm still not quite sure if I get it.
> > The ipip=max(ilo(3),1) ipip fix is quite straightforward, however I
> > still don't quite understand why the F7SPLT is being called every time?
> >
> > So should line
> > csplit.f:76 !   IF(L.EQ.3 .and. isplit(jatom).eq.15)     &
> > contain the exclamation mark? According to your argumentation I judge
> > that this is another bug? Because right now the "isplit(jatom).eq.15"
> > check is commented out and F7SPLT call is guarded just by "IF(L.EQ.3)"
> > check (csplit.f:77), hence F7SPLT is called independently of ISPLIT
> > value.
> >
> >
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> > SEARCH the MAILING-LIST at:  http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
> >
> 




More information about the Wien mailing list