The ABINIT Users Mailing List ( CLOSED )

[Vincent Chevrier <vincent.chevrier@dahn.phys.dal.ca>]

Hi Nicola,

The ecut is in hartree (40 Rydberg). I'll try a higher ecut. I have tried
varying inpt, ntheta, nphi with no luck.

I also ran gdb on the dumped core files. All their backtraces yield the
following stack:

#0  0x0000002a95abefb4 in _int_free () from
/usr/local/openmpi/lib/libopen-pal.so.0
#1  0x0000002a95abddb5 in free () from
/usr/local/openmpi/lib/libopen-pal.so.0
#2  0x000000000045d716 in __hpf_dealloc ()
#3  0x0000000000416449 in drvaim_ ()
#4  0x000000000040fb88 in MAIN_ ()
#5  0x000000000040e3c0 in main ()

Not sure what this means, seems the be machine specific but I get the same
error on two different machines (the other one doesn't use openmpi). Again
if anyone is interested in repeating the error all the files needed are
here (takes < 20 min):
http://dahn.phys.dal.ca/AIM_bader_files.tar.gz

thanks

Vincent
Physics Dept
Dalhousie University

>
>  Hello!
>        Is encut 20 in Rydberg? That is not much, even in PAW, for Bader
> analysis. I see you don't set inpt, ntheta, nphi. Have you tried
changing that?
>  All the best,
>               Nicola Seriani
>               University of Vienna
>
>
>
> ------------------------ Ursprüngliche Nachricht
------------------------
> Betreff: [abinit-forum] AIM - Bader - Segmentation Fault
> Von:     "Vincent Chevrier" <vincent.chevrier@dahn.phys.dal.ca>
> Datum:   Mi, 29.10.2008, 16:39
> An:      forum@abinit.org
> --------------------------------------------------------------------------
>
> Hi,
>
> I'm trying to use AIM for bader charge integration. I have a well
converged DEN file generated using PAW datasets (ecut 20, kptrlen 84).
Below is the content of my .in file. I've tried to vary the values of
lgrad2, lstep2, dpclim, maxatd, maxcpd as suggested by the tutorial with
no success, as well as the values of dpclim, nsa, nsb and nsc. The
calculation dies close to the end of surface determination with the
cluster indicating a segmentation fault (and a core dump). I've
> successfully run the AIM code on other similar structures.
>
> I don't think it's a lack of memory as I've dedicated over 2G to this
process and it indicates using very little.
> I've tried this on two different machines having different versions of
Abinit (5.4.4 and 5.5) with identical results.
>
> Any help or suggestions would be greatly appreciated.
>
> ****** .in file ******
> atom 5
>
> crit 2
> surf 1
> irho 1
> ivol 1
> dpclim 5.0d-2
> dltyp 1
> gpsurf 1
> nsa 2
> nsb 2
> nsc 2
> **********************
>
> For the truly interested my .in .files .fc (from Abinit website) and DEN
files are available at
> http://dahn.phys.dal.ca/AIM_bader_files.tar.gz
>
> Thanks
>
> Vincent
>
> Physics Dept
> Dalhousie University
>
>
>
>
>
>
>
>