The ABINIT Users Mailing List ( CLOSED )

[<ilukacevic@fizika.unios.hr>]

Dear X. Gonze,

Yes (I've included an .in file in my first post, but forgot to reattach it in
the following posts - sorry for that), I'm using ecutsm, with the value of 0.1
Ha (the same one in structural relaxation and rf calc.). I haven't tried to
optimize its value...yet. Do You think its too low?

But I'm not using the same value of ecut in structural relaxation calc. (52 
Ha)
and rf calc. (60 Ha). Could this be the cause of the slow convergence?
I know that because of that I receive an difference in stress tensor 
components
between struct. relax. and rf calc.:




>>>>>>>>> Etotal= -7.21370977356432E+01
 Cartesian components of stress tensor (hartree/bohr^3)
  sigma(1 1)= -1.35956626E-03  sigma(3 2)=  0.00000000E+00
  sigma(2 2)= -1.35956626E-03  sigma(3 1)=  0.00000000E+00
  sigma(3 3)= -1.35956885E-03  sigma(2 1)=  0.00000000E+00
                                                                              
 
                                                <-- struct. relax. calc.
-Cartesian components of stress tensor (GPa)[Pressure=  4.0000E+01 GPa]
- sigma(1 1)= -3.99998150E+01  sigma(3 2)=  0.00000000E+00
- sigma(2 2)= -3.99998150E+01  sigma(3 1)=  0.00000000E+00
- sigma(3 3)= -3.99998915E+01  sigma(2 1)=  0.00000000E+00




>>>>>>>>> Etotal= -7.21382669702764E+01
 Cartesian components of stress tensor (hartree/bohr^3)
  sigma(1 1)= -1.35946560E-03  sigma(3 2)=  0.00000000E+00
  sigma(2 2)= -1.35946560E-03  sigma(3 1)=  0.00000000E+00
  sigma(3 3)= -1.35776261E-03  sigma(2 1)=  0.00000000E+00
                                                                              
 
                                                 <-- rf calc.
-Cartesian components of stress tensor (GPa)[Pressure=  3.9980E+01 GPa]
- sigma(1 1)= -3.99968535E+01  sigma(3 2)=  0.00000000E+00
- sigma(2 2)= -3.99968535E+01  sigma(3 1)=  0.00000000E+00
- sigma(3 3)= -3.99467501E+01  sigma(2 1)=  0.00000000E+00


Could this be also the cause?


Thank You in advance!

Igor Lukacevic