Skip to Content.
Sympa Menu

forum - Re: [abinit-forum] Etotal versus etotal

forum@abinit.org

Subject: The ABINIT Users Mailing List ( CLOSED )

List archive

Re: [abinit-forum] Etotal versus etotal


Chronological Thread 
  • From: TORRENT Marc <marc.torrent@cea.fr>
  • To: forum@abinit.org
  • Subject: Re: [abinit-forum] Etotal versus etotal
  • Date: Thu, 03 Sep 2009 14:12:03 +0200
  • Organization: CEA-DAM

Dear Vincent,

Some explanations (but not all).

etotal is used to converge the scf cycle and contains an entropy energy computed with tsmear.
It is obtained by direct scheme if iscf<10, by double-counting scheme if iscf>=10:

iscf=7 :  etotal=e_intern      - tsmear*entropy
iscf=17:  etotal=eDC_intern - tsmear*entropy

Etotal (EtotalDC) is just printed at the end of the SCF cycle and recomputed at that level.
It uses tphysel if not zero, else tsmear:

If tphysel<>0:
Etotal=e_intern      - tphysel*entropy
EtotalDC=eDC_intern - tphysel*entropy

If tphysel=0:
Etotal=e_intern      - tsmear*entropy
EtotalDC=eDC_intern - tsmear*entropy

So, it explains the differences between etotal and Etotal.
Why is it implemented like this: I don't know.
The developer who introduced tphysel should be able to explain that.


Concerning the difference between Etotal and EtotalDC when tphysel<>0... it sounds strange.
But I cannot explore further if I'm not sure of the convergence of your SCF cycle.
toldfe is not the right criterium.
you should use tolvrs and put it to a very low value (i.e 1.d-15 or lower).

Regards,

Marc



Etotal and Etotal

Vincent Chevrier a écrit :
60945.142.68.35.223.1251977332.squirrel@dahn.phys.dal.ca" type="cite">
Dear Bernard,
thank you for your reply. I've rerun with toldfe=1e-9. But etotal does not
always agree with Etotal(DC). See below, in all cases tsmear=0.01 and the
runs only differ in the parameters stated.

toldfe=1e-5, tphysel=0
    >>>>>>>>> Etotal= -2.24785447837035E+02
    >>>> Etotal (DC)= -2.24785888234037E+02
    etotal            -2.2478510911E+02

toldfe=1e-5, tphysel=0.002131
    >>>>>>>>> Etotal= -2.24785745210972E+02
    >>>> Etotal (DC)= -2.24784864246079E+02
    etotal            -2.2478486425E+02

toldfe=1e-9, tphysel=0
    >>>>>>>>> Etotal= -2.24785888288957E+02
    >>>> Etotal (DC)= -2.24785888278458E+02
    etotal            -2.2478510947E+02

toldfe=1e-9, tphysel=0.002131
    >>>>>>>>> Etotal= -2.24784859172166E+02
    >>>> Etotal (DC)= -2.24784864956616E+02
    etotal            -2.2478486496E+02

when tphysel=0.002131 then etotal = Etotal(DC), but when tphysel=0 then
etotal != Etotal(DC). Why is this, or more specifically what is:

1 - the cause of the difference between Etotal(DC) and etotal
2 - the effect of tphysel on the difference between Etotal(DC) and etotal

Thanks
Vincent

  
Dear Vincent,

With iscf=17 (the default value of iscf),  ETOT
is equal to the so-called "double counting" _expression_ of the energy
    
written "Etotal (DC)" in the .out file (you could check it in your .out
file).
  
Moreover, Etotal and Etotal (DC) (both written in the .out file) are
    
expected to agree if
  
your calculation is well converged.
In your case, your convergence criterion
is weak (toldfe 1.d-05) so Etotal and Etotal (DC)
are not equal. You could
increase you convergence criterion to check this.

Best regards
Bernard


Vincent Chevrier a écrit :
    
Hi,
I am wondering what the difference is between Etotal and etotal for PAW
      
calcs. I can't remember ever seeing a difference stated on the ABINIT
website. I'm especially interested in knowing how tphysel affects this
difference.
  
I completed two runs with the same structure the only difference being
      
tphysel=0 and tphysel=0.00213.
  
>From the .out file:
For tphysel=0:
    Kinetic energy  =  1.72081044902126E+01
    Hartree energy  =  1.17823500013938E+01
    XC energy       = -3.59810665254476E+01
    Ewald energy    = -1.07436906857906E+02
    PspCore energy  =  8.93535605840166E+00
    Loc. psp. energy= -4.47984809673791E+01
    Spherical terms = -7.44950150296981E+01
    >>>>> Internal E= -2.24785658830422E+02
    -kT*entropy     =  2.10993387856660E-04
    >>>>>>>>> Etotal= -2.24785447837035E+02
and
    etotal  -2.2478510911E+02
For tphysel=0.002131
    Kinetic energy  =  1.72075689848333E+01
    Hartree energy  =  1.17845516746048E+01
    XC energy       = -3.59818457081452E+01
    Ewald energy    = -1.07436906857906E+02
    PspCore energy  =  8.93535605840166E+00
    Loc. psp. energy= -4.48011025224259E+01
    Spherical terms = -7.44958705044103E+01
    >>>>> Internal E= -2.24788248875047E+02
    -kT*entropy     =  2.50366407550064E-03
    >>>>>>>>> Etotal= -2.24785745210972E+02
and
    etotal  -2.2478486425E+02
below is the input file. Can anybody help me understand:
1 - the cause of the difference between Etotal and etotal
2 - the effect of tphysel on the difference between Etotal and etotal
      
Thanks
  
Vincent
Input file:
    ndtset   2
   tphysel1  0.002131
   tphysel2  0.0
    occopt   4
    tsmear   0.01
 pawecutdg   15
      ecut   15
    kptopt   1
   nshiftk   1
    shiftk   0.5 0.5 0.5
  kptrlatt   6  0  0
             0  3  0
             0  0 10
     nstep   50
    toldfe   1.0d-5
    ntypat   2
     znucl   3 14
     natom   34
     typat   1  1  1  1  2  1  1  1
             2  1  1  1  2  1  1  1
             2  1  1  1  1  2  1  1
             1  2  1  1  1  2  1  1
             1  2
     acell   1.5065427965E+01  2.8615352526E+01  8.4083802112E+00 Bohr
      
rprim   1.0 0.0 0.0
  
             0.0 1.0 0.0
             0.0 0.0 1.0
      xred   7.4996953092E-13  5.0000000000E-01  0.0000000000E+00
             9.6480532055E-02  1.9420479791E-01  0.0000000000E+00
      
2.7080350417E-01  3.4610078252E-01  0.0000000000E+00
1.5292044305E-01  2.5787557735E-02  0.0000000000E+00
4.1748605305E-01  1.6031572532E-01  0.0000000000E+00
2.5955975937E-01  9.5181187368E-02  5.0000000000E-01
4.0978302750E-01  2.5635768786E-01  5.0000000000E-01
9.6925526499E-02  3.9393640703E-01  5.0000000000E-01
4.2837868374E-01  4.3177429281E-01  5.0000000000E-01
9.0351946795E-01  8.0579520209E-01  0.0000000000E+00
7.2919649583E-01  6.5389921748E-01  0.0000000000E+00
8.4707955695E-01  9.7421244227E-01  0.0000000000E+00
5.8251394695E-01  8.3968427468E-01  0.0000000000E+00
7.4044024063E-01  9.0481881263E-01  5.0000000000E-01
5.9021697250E-01  7.4364231214E-01  5.0000000000E-01
9.0307447350E-01  6.0606359297E-01  5.0000000000E-01
5.7162131627E-01  5.6822570719E-01  5.0000000000E-01
5.0000000000E-01  0.0000000000E+00  0.0000000000E+00
4.0351946795E-01  6.9420479791E-01  0.0000000000E+00
2.2919649583E-01  8.4610078252E-01  0.0000000000E+00
3.4707955695E-01  5.2578755773E-01  0.0000000000E+00
8.2513946947E-02  6.6031572532E-01  0.0000000000E+00
2.4044024063E-01  5.9518118737E-01  5.0000000000E-01
9.0216972503E-02  7.5635768786E-01  5.0000000000E-01
4.0307447350E-01  8.9393640703E-01  5.0000000000E-01
7.1621316266E-02  9.3177429281E-01  5.0000000000E-01
5.9648053206E-01  3.0579520209E-01  0.0000000000E+00
7.7080350417E-01  1.5389921748E-01  0.0000000000E+00
6.5292044305E-01  4.7421244227E-01  0.0000000000E+00
9.1748605305E-01  3.3968427468E-01  0.0000000000E+00
7.5955975937E-01  4.0481881263E-01  5.0000000000E-01
9.0978302750E-01  2.4364231214E-01  5.0000000000E-01
5.9692552650E-01  1.0606359297E-01  5.0000000000E-01
9.2837868374E-01  6.8225707188E-02  5.0000000000E-01
  
    




  




Archive powered by MHonArc 2.6.16.

Top of Page