The ABINIT Users Mailing List ( CLOSED )

[matthieu verstraete <matthieu.jean.verstraete@gmail.com>]

I'm afraid the only thing you can do is get more memory: optic is using all of yours, and segfaulting. abinip of course can divide the memory over different processors, whereas optic is not parallelized.

The tetrahedron method is included in several parts of the code, but not in optic to my knowledge (would be a very good idea). There should be a bilinear interpolation scheme - all of this is imported from Sangeeta Sharma's LAPW code (now called elk http://elk.sourceforge.net/).

Matthieu

On Mon, Feb 16, 2009 at 6:08 PM, <gvolo@physics.auth.gr> wrote:

Dear all,

I'm a new user of abinit and quite new to the first principles calculations, so please excuse me in advance.

After successfully compiling abinit and run some band structure calculations on various models, also succesfully, I tried to run some optics and followed the tutorial on optic. My goal is to get the imaginary part of the linear dielectric function for GaAs. I used the input file (the last one - for faster computation) provided in the tutorials to run abinip and tried a convergence study on ecut and kptgrd variables.
Then I run the optic executable using as input the files from the abinip runs, WFK and WF files. When I increase kptgrd more than 16x16x16 in the abinip run and then run optic, I get an error from the optic:

toptic_1.sh: line 83: 24304 Segmentation fault      $PWD/optic <$PWD/toptic_2.files >&log_optic

where the script file is to handle all the file, run abinip and then optic.

For those calculations I'm using the grid infrastructure so the whole run is parallel and uses i.e. 12cpu's for abinip and then since optic (by my understanding) is a sequential run, runs on one of the cpu's the optic run. I have been told that this fault is probably due to insufficient memory, so i was thinking how can i run this job.

The abinip part is finishing successfully even with larger values of kptgrd, so is there a way to continue to the optic part differently?

Also is the use of the tetrahedron method for the kpoint sampling provided by abinit package ?

Thank you in advance,
GV

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dr. Matthieu Verstraete

European Theoretical Spectroscopy Facility (ETSF) 
Dpto. Fisica de Materiales, 
U. del Pais Vasco,            
Centro Joxe Mari Korta, Av. de Tolosa, 72,   Phone: +34-943018393
E-20018 Donostia-San Sebastian, Spain        Fax  : +34-943018390

Mail : matthieu.jean.verstraete@gmail.com
http://www-users.york.ac.uk/~mjv500