[Wien] FSM calculation problem
Peter Blaha
pblaha at theochem.tuwien.ac.at
Tue Jun 7 12:03:45 CEST 2005
I've checked the fsm calculations on your computer.
runfsm -m 0.1
Then I checked case.in2up and case.in2dn. It shows:
TOT (TOT,FOR,QTL,EFG,FERMI)
-9.0 16,1 0.50 0.05 EMIN, NE, ESEPERMIN, ESEPER0
So the script has changed NE from 16.0 to 16,1 !!!
This is of course WRONG ! since fortran needs 16.1 instead of 16,1
This line is produced by awk :
awk -vm=$m '/NE/{$0=" "$1" "$2+m" "$3" "$4" "$5" "$6" "$7" "$8" "$9" "};//' $file.in2_fsm>$file.in2up
and here is the problem. In some languages (I know it only from German,
but it seems it is the same in Spanish)
a floating point number is printed with a comma, while in english it is
printed with the correct decimal point.
For a SUSE system start as superuser,
yast2
system --> edit /etc/sysconfig editor
and choose
system --> environment --> language
and set
RC_LC_NUMERIC to en_US
PS: If anybody knows an awk option to set this properly without playing with
system variables, please let me know.
Regards
> I have downloaded and installed the latest version of tsch available for the
> Ubuntu 5.04 distribution, to no avail. I am not at all an expert in Linux,
> however, and I could not find anything related to a wait command problem with
> SUSE 9.2 or other distributions in the tsch mailing list you pointed out.
> I have also run the commands by hand, as you suggested and followed the
> previous instructions for inspection of scfXX files you gave. They seemed to
> me to be in order, that is, no error messages and the .error files are empty,
> but when I ended up with mixer, which I believe is the last step in this (as
> in any) sp_lapw calculation, the case.scf file appeared empty!! (Could it be
> that it only fills up when running a whole iterated set of calculation
> cycles?).
> The sum of the electrons in case.scfup and dn files is 16, which is the
> correct value in case.in2. But there are not any in2up/dn files this time!! Is
> this normal?.
> I have not run any fsm calculation this time, however, because I don't know
> how to or even if it is possible to run only a cycle of this kind of script.
> Sorry if my doubts are troublesome or if I'm not clear in explaining.
> Thanks for your efforts in helping.
> Roberto.
>
> At 07:39 06/06/2005, you wrote:
> > Hi,
> > As I mentioned already to your colleage, I do not think it is a problem
> > with WIEN2k.
> >
> > Maybe you are right and it is connected with SUSE 9.2 We just noticed,
> > that the "wait" command does not work in this version and you may replace
> > /bin/tcsh by a proper version (see the recent postings on the ML,
> > www.tcsh.org).
> >
> > Otherwise, one needs a systematic analysis. It does not help to run an scf
> > cycle and find that the moment disagrees. You must run the commands of one
> > cycle "by hand" and check the results (scfXX files) if they are reasonable
> > or not. I tried to explain that already to your colleage.
> >
> > > A colleague of mine (Sergio Palacios) has recently experienced trouble
> > with
> > > FSM calculations, as you can see in a previous thread. I have just
> > downloaded
> > > the newer version of the code, WIEN2k_05.4 and installed it to a Pentium
> > IV
> > > 3.4 GHz PC, with 2 GB RAM, with Ubuntu 5.04 distribution running. I have
> > rerun
> > > an old FSM calculation which worked in a machine with an old Red Hat
> > version
> > > installed (don't know exactly which) and WIEN2k_03. The problem I find is
> > the
> > > same as my colleague, that is, the calculation reaches a normal end, but
> > the
> > > final magnetic moment is not the fixed one. For instance, if the FSM value
> > is
> > > 2.2 it reaches 2.5 and if it is 0.1 it reaches 0.49. Seems as if the FSM
> > code
> > > stops when finding some kind of local minimum, but not the requested
> > global
> > > one (I'm just guessing, I don't really know how this script works). Funny
> > > enough, I have reinstalled in the same machine running Ubuntu the old
> > > WIEN2k_03 code. And it also fails, reaching the same wrong value of
> > magnetic
> > > moment!!. So the problem seems to be in the Linux distribution, something
> > > perhaps to do with a library we have not installed or...?? We have also
> > tried
> > > to run the calculation in a laptop computer running Suse 9.2 and we find
> > the
> > > same problem.
> > > Could anybody tell us if there is a patch or anything we can apply in
> > order
> > > for this script to work properly?
> > > Thanks in advance.
> > > Roberto.
> > >
> > >
> > >
> > > -----------------------------------------------------------
> > > Roberto Iglesias Pastrana
> > > Despacho 17
> > > Área de Física Aplicada
> > > Departamento de Física
> > > Facultad de Ciencias de la Educación
> > > Universidad de Oviedo
> > > C/ Aniceto Sela, s/n 33005 Oviedo
> > > Tel.: 34 (9) 85102898
> > > 34 (9) 85458136
> > > Fax: 34 (9) 85103324
> > > e-mail: roberto at uniovi.es
> > > roberto at pinon.ccu.uniovi.es
> > >
> >
> >
> > 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/
> > --------------------------------------------------------------------------
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
> _______________________________________________
> 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/
--------------------------------------------------------------------------
More information about the Wien
mailing list