The ABINIT Users Mailing List ( CLOSED )

["Anglade Pierre-Matthieu" <anglade@gmail.com>]

Hi,

One of the way to understand what is hapenning is to compile Abinit
yourself (preferably a newer version) with debugging options enabled.
For instance using g95 you could use -g -ftrace=full -fbounds-check.
If you do so, your error message will change drastically from the
elusive
"segmentation fault" to something much clearer like
segmentation fault in file www line xxx try to acess element yyy of array zzz.
In the present case I have tried your input file with Abinit 5.6  and
have obtained no error message. The calculation converge in 30 steps.

regards

PMA

On Fri, Jun 6, 2008 at 11:40 AM, Enrico <enrico.avenati@yahoo.com> wrote:
> Hi,
> i'm trying to do some calculation but i can't get my result converged under
> 10e-4 , not even 10e-1sometimes, because of an annoying error: SEGMENTATION
> FAULT.
>
> I know it's a pretty general error but i don't know nothing else about it. 
> Not
> even how to find out what generates it.
>
> I tried the same calulation on two different PC's (AMD Athlon 64 bit 3.5 
> MHz,
> and AMD Athlon 64 4800+ dual core with 2Gb of RAM)
> and with different version of ABINIT 5.3.4 and 5.4.4
>
> I'm using HGH pseudopotentials.
> My calculation was on LaNi_5, the max Ecut handled is 120 Ha (256 mb of
> occupancy) when i try with 140 Ha the calculation stops and i got the error
> message in the shell.
> I got the same outcome with Dy at 160 Ha or (alpha)Cu_5Dy at 140 Ha ( this
> happens also with TM psp).
>
> I've also checked that the segmentation fault error doesn't occur always at 
> the
> same time if i rerun the calculation.
>
> I hope you could sugget me some ways to give you more specific infos, or 
> how to
> fix it.
>
> Thank you
>
> my input:
>
> #UNIT CELL
> LaNi5
>
>
> acell  5.017 5.017 3.987 Angstr
>
> rprim
> 1 0 0
> -0.5 sqrt(0.75) 0
> 0 0 1
>
> spgroup 191
>
> ntypat 2 # 2 types of atoms
>
> znucl  57 28
>
> natom  6 # atom per cell
>
> typat  1  5*2 # LaNi5
>
> xred
> 0.0 0.0 0.0
> 1/3 2/3 0.0
> 2/3 1/3 0.0
> 1/2 0.0 1/2
> 1/2 1/2 1/2
> 0.0 1/2 1/2
>
>
>
> #RECIPROCAL SPACE INTEGRETION
> kptopt 1
> ngkpt 2 2 2
>
> nshiftk 1
> shiftk 0.0 0.0 0.5
>
> nband 34
> tsmear 0.02
> occopt 4
>
> #PLANE WAVES
> ecut 140
>
> #SCF PROCEDURE
> nstep 90
> ixc 1
> toldfe 1.0d-7
>



-- 
Pierre-Matthieu Anglade