[CPMD-list] Segmentation faults

[Juerg Hutter]

Hi

I wanted to add to Axel's answer to your questions.

The segfault is most likely due to a bug in fnonloc.F

at line 526 it should be

      IF (TIVAN.OR.TISGP) THEN
        CALL FNONLOC(C2,F,NSTATE,AUXC,DDIA,IKIND,ISPIN)
        CALL TIHALT('   FNONLOC',ISUB)
        RETURN
      ENDIF

The third line has to be added. When this line is missing
you get memory corruption and segfaults that occur during
MD. At which step it occurs is compiler and hardware specific.

----------------------------------------------------------
Juerg Hutter                   Phone : ++41 1 635 4491
Physical Chemistry Institute   FAX   : ++41 1 635 6838
University of Zurich           E-mail: hutter at pci.unizh.ch
Winterthurerstrasse 190
CH-8057 Zurich, Switzerland
----------------------------------------------------------


On Fri, 30 Apr 2004, Holger Hesske wrote:

> Dear CPMD-users,
>
> As I wrote, i got discontinous behaviour in restarting a MD-calculation. But
> after I got a lot of advices (thanks for the quick replies), I try to change
> several things.
> I run MD-simulation for small isolated systems containing methylamine and
> H3O+ or comparable molecules. For test purposes I run very short times (100
> steps of 5 a.u.) and then try to RESTART it. But when I use RESTART CELL the
>  program stops with Segmentation fault error before even one step of
> dynamics is done.
> The last written output is
>
> ***  ATOMWF | THE NEW SIZE OF THE PROGRAM IS  63225 Bytes ***
>
>
> Irritated with this results I started a new long MD run (10000 steps) but
> also SEGMENTATION FAULT arises after 611 steps of the Simulation (every kind
> of system!)
> Maybe it is my system (CPMD 3.7.2 at an SGI ORIGIN / IRIX 6.5.22) but I did
> a lot of geometric optimsations before. And it had worked fine.....
> For Clearance I Put the Input in this mail. Hope anybody could help......
>
> Thanks in advance
>
> Holger
>
> INPUT
>
> &CPMD
> FILEPATH
> /usr/local/CPMD/OUTPUTS/METHYLAMINE/
>
> MOLCEULAR DYNAMICS
> TIMESTEPS ELECTRONS
> 4
> TIMESTEP IONS
> 5
> SPLINE POINTS
> 1500
> TEMPERATURE
> 298
> TEMPCONTROL IONS
> 298 5
> MAXSTEP
> 10000
> TRAJECTORY XYZ SAMPLE
> 5
> &END
>
> &SYSTEM
> ANGSTROM
> CELL ABSOLUTE DEGREE
> 10.0 10.0 10.0 90.0 90.0 90.0
> SYMMETRY
> 0
> POINT GROUP
> AUTO
> CUTOFF
> 22
> 400eV
>
> CHARGE
> 1
> SCALE
> TESR
> 1 1 1
> &END
> &DFT
> FUNCTIONAL BLYP
> &END
>
> &ATOMS
> *H_SG_BLYP Kleinman-Bylander
> LMAX=S
> 8
> ..
> ..
> *O_SG_BLYP Kleinman-Bylander
> LMAX=P
> 1
> ..
> *N_SG_BLYP Kleinman-Bylander
> LMAX=P
> 1
> ..
> *C_SG_BLYP Kleinman-Bylander
> LMAX=P
> 1
> ...
>
> ISOTOPE
> 2.000
> 15.999
> 14.008
> 12.000
>
> CONSTRAINTS
> FIX COORDINATES
> 3
> 1 0 0 1
> 9 0 0 0
> 10 0 0 0
> END CONSTRAINTS
> &END
>
> --
> "Sie haben neue Mails!" - Die GMX Toolbar informiert Sie beim Surfen!
> Jetzt aktivieren unter http://www.gmx.net/info
>
> _______________________________________________
> CPMD-list mailing list
> CPMD-list at cpmd.org
> http://www.cpmd.org/mailman/listinfo/cpmd-list
>