[Octopus-users] user_defined species

[Emily]

Hi Xavier,

I'm still having this problem with a segmentation fault when using both
an atomic species and a user-definied species.  If I have either alone,
I have no problem.  I just upgraded to the newest version of octopus,
and the problem persists.  I am using ifort and linux on a 32 bit
machine.  Do you have any ideas for me?

Thanks, Emily

My inp file is:

CalculationMode = gs
units = 'eVA'

Nitrogen_mass = 14.0
Nitrogen_z = 7

%Species
  'Vext' |0.0 | user_defined | 0 | "a*(x+y+z)"
  'N' | Nitrogen_mass | spec_ps_hgh | Nitrogen_z | 1 | 0
%

% Coordinates
 'Vext' | 0.5 | 0.5 | 0.5 | no
 'N' | 0 | 0 | 0 | no
%

ExtraStates = 1
%Occupations
  2 | 1 | 1 | 1
%

BoxShape = sphere
radius = 5.0
spacing = 0.18

TypeOfMixing = broyden
debuglevel=2

---------------------------------
and the output file contains:
---------------------------------

  <><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
                    Running octopus, version 2.1.0
              build time - Tue Dec  4 14:31:54 EST 2007
                            svn revision:

                           Compiler: ifort
Compiler flags: -u -zero -fpp1 -nbs -pc80 -pad -align -unroll -O3 -ip
-tpp7 -xW

                The octopus is swimming in n49 (Linux)
            Calculation started on 2007/12/10 at 11:07:19

************************** Calculation Mode **************************
Input: [CalculationMode = gs]
**********************************************************************


****************************** Species *******************************
Specie "Vext" is an user-defined potential.
   Potential = a*(x+y+z)^@
(END)   

----------------------------
and the debug output contains:
----------------------------

* O      0.014781    0.000000   14384 | ..|..|..|specie.specie_read
* I      0.014991    0.010000   14384 | ..|..|..|specie.specie_init
* I      0.015189    0.010000   14384 |
..|..|..|..|specie.specie_iwf_fix_qn
forrtl: severe (174): SIGSEGV, segmentation fault occurred

Stack trace terminated abnormally.  
----- Original message -----
From: "Xavier Andrade" <xavier at tddft.org>
To: "Emily" <listemily at eml.cc>
Cc: "octopus users group" <octopus-users at tddft.org>
Date: Thu, 15 Nov 2007 13:58:00 +0100 (CET)
Subject: Re: [Octopus-users] user_defined species

Hi Emily,

I didn't manage to reproduce the segfault, neither with trunk nor 2.1.
What
version, compilers and os are you using? Can you run it with
debuglevel=2
and post the last lines before the segmentation fault?

Cheers,

Xavier


On Wed, 14 Nov 2007, Emily wrote:

> I'm trying to add my own potential to the existing potential of a
> species already defined in Octopus.  So I'm using a user_defined
> Species.  It seems everything works fine when I use only this species to
> define some potential.  But when I try to add some other species also,
> it does not work.
>
> Here's an example of an inp file that works fine:
> ___________________________________________________
> a = 2
> CalculationMode = gs
> Dimensions = 1
> NonInteractingElectrons = yes
>
> radius = 10
> spacing = 0.1
>
> % Species
>  "HO" | 0.0 | user_defined | 2 | "0.5*x^2"
>  "Vext" |0.0 |user_defined| 0
>  |"(x+5)*(1+tanh(a*(x+5)))-(x-5)*(1+tanh(a*(x-5)))"
> %
>
> % Coordinates
>  "HO" | 0 | 0 | 0 | no
>  "Vext" | 0.5| 0 | 0 | no
> %
>
> Output = wfs + density + potential
> OutputWfsNumber = 1-4
> OutputHow = gnuplot + axis_x
> ___________________________________________________
> However another simple file that doesn't work is here:
>
> ___________________________________________________
> a=0 (a=1 also does not work)
> CalculationMode = gs
> units = 'eVA'
>
> Nitrogen_mass = 14.0
> Nitrogen_z = 7
>
> %Species
>  'Vext' |0.0 | user_defined | 0 | "a*(x+y+z)"
>  'N' | Nitrogen_mass | spec_ps_hgh | Nitrogen_z | 1 | 0
> %
>
> % Coordinates
> 'Vext' | 0.5 | 0.5 | 0.5 | no
> 'N' | 0 | 0 | 0 | no
> %
>
> ExtraStates = 1
> %Occupations
>  2 | 1 | 1 | 1
> %
>
> BoxShape = sphere
> radius = 5.0
> spacing = 0.18
>
> TypeOfMixing = broyden
> ___________________________________________________
>
> There seems to be a problem in the parser, because the file
> status/out.oct ends before it should, a couple of lines after the
> species block is defined.  In the output I get the errors
> forrtl: severe (174): SIGSEGV, segmentation fault occurred
> Stack trace terminated abnormally.
>
> Can anybody tell me if they have used the user_defined species with
> another kind of species?
>
> Thanks,
> Emily
>
> ----- Original message -----
> From: "Xavier Andrade" <xavier at tddft.org>
> To: "Emily" <listemily at eml.cc>
> Cc: "octopus users group" <octopus-users at tddft.org>
> Date: Fri, 19 Oct 2007 23:51:09 +0200 (CEST)
> Subject: Re: [Octopus-users] UserDefinedTDPotential
>
> Hello Emily,
>
> On Fri, 19 Oct 2007, Emily wrote:
>
>> Hello,
>>
>> I tried entering a formula, as in UserDefinedTDPotential = x, but it
>> complained I used the variable x without defining it.
>>
> You have to pass a string as the potential, so it should be
>
> UserDefinedTDPotential = "x"
>
>> Also, may I use this variable to add my own potential even in a
>> groundstate run, or only for time-dependent calculations?
>
> I think that for the gs you have to use a user defined specie. There is
> a
> lot redundancy there, probably we should fix this for 3.0.
>
> Cheers,
>
> Xavier
> _______________________________________________
> Octopus-users mailing list
> Octopus-users at tddft.org
> http://www.tddft.org/mailman/listinfo/octopus-users
>