The ABINIT Users Mailing List ( CLOSED )

[gmatteo@pcpm.ucl.ac.be]

Dear Juan Maria, 

I have noticed that in your input file nbandkss is set to -1 
This means that the code is trying to diagonalize a 
huge matrix (32327*32327) and there might an out of memory  
during the allocation of particular arrays inside the outkss routine. 

I don't think your GW calculation really needs such a large number of states 
also because diagonalizing this matrix would be more CPU demanding 
than the GW calculation itself.  

As a rule of thumb, I suggest you generate a KSS file containing 
a number of states equal to (10*number of valence states).
There should be, indeed, enough empty states to perform all the 
converge tests as explained in the GW-tutorial, while both the CPU-time 
and the memory required should be reasonable.
  
Hope this helps
Best Regards, 
Matteo Giantomassi

> Dear Abinit forum members,
> 
> I have problem to generate the KSS file with the latest version of
> ABINIT (I had also problems with previous versions). I have run the
> calculation in several plataforms, but it always stops without any error
> message. This is the input file:
> 
> INPUT FILE
> 
> acell 17.598 9.059 6.643 angstrom  
>                   
> enunit 1
> 
> nbandkss -1
> 
> istwfk 1 1
> 
> symmorphi 0
> 
> ntypat 5          
> znucl 78 17 7 6 1 
>                   
> nband 200
> 
> natom  92
> natrd  12         
> typat 1 2 3 4 4 4 4 4 5 5 5 5  
> xred              
>               0.0000000000E+00  9.6676409262E-01  2.5000000000E-01
>               9.0333387142E-02  7.8432314128E-01  2.5000000000E-01
>               7.3025422589E-02  1.3533939083E-01  2.5000000000E-01
>               1.4813222070E-01  1.2225211910E-01  2.5000000000E-01
>               1.9485628558E-01  2.4350167592E-01  2.5000000000E-01
>               1.6261311096E-01  3.8229164940E-01  2.5000000000E-01
>               8.4976076507E-02  3.9592559651E-01  2.5000000000E-01
>               4.0852452778E-02  2.7033082504E-01  2.5000000000E-01
>               5.9040593898E-02  5.0560926362E-01  2.5000000000E-01
>               1.9795018781E-01  4.8105832937E-01  2.5000000000E-01
>               2.5652968255E-01  2.3009169426E-01  2.5000000000E-01
>               1.6890816530E-01  1.0086529108E+00  2.5000000000E-01
> 
> spgroup 63       
> chkprim 0
> 
> ecut 20.0   
>       
> kptopt 1
> ngkpt 1 1 2
> shiftk 0.0 0.0 0.0
> 
> nstep 90          
> toldfe 1.0d-6     
> diemac 1.0     
> diemix 0.5 
> 
> END OF THE INPUT FILE
> 
> And this is the output point in which it stops:
> 
> OUTPUT FILE
> 
>  At SCF step   13, etot is converged :
>   for the second time, diff in etot=  9.266E-08 < toldfe=  1.000E-06
> 
>  Cartesian components of stress tensor (hartree/bohr^3)
>   sigma(1 1)=  3.65720097E-04  sigma(3 2)= -1.38904380E-10
>   sigma(2 2)=  3.80621182E-04  sigma(3 1)=  0.00000000E+00
>   sigma(3 3)=  3.37704677E-04  sigma(2 1)=  0.00000000E+00
> 
> 
> ======================================================================
>  Calculating and writing out Kohn-Sham electronic Structure file  Using
> diagonalized wavefunctions and energies (kssform=1)  number of Gamma
> centered plane waves  32327
>  number of Gamma centered shells   8279
>  number of bands  30510
>  maximum angular momentum components      4
>  Diagonalized eigenvalues
>    k    eigenvalues [eV]
> 
> END OF THE OUTPUT FILE
> 
> Do you know anything about this error? Thank you in advavnce,
> 
> Dr. Juan Maria Garcia Lastra
> ETSF. San Sebastian. Spain.
> 



----------------
Matteo Giantomassi
PCPM/FSA/UCL

-------------------------------------------------
This mail sent through IMP: atlas.pcpm.ucl.ac.be