forum@abinit.org
Subject: The ABINIT Users Mailing List ( CLOSED )
List archive
- 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 VincentDear Vincent, With iscf=17 (the default value of iscf), ETOT is equal to the so-called "double counting" _expression_ of the energywritten "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) areexpected to agree ifyour 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 PAWcalcs. 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 beingtphysel=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 etotalThanksVincent 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 Bohrrprim 1.0 0.0 0.00.0 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+002.7080350417E-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 |
- [abinit-forum] Etotal versus etotal, Vincent Chevrier, 09/02/2009
- Re: [abinit-forum] Etotal versus etotal, Bernard Amadon, 09/03/2009
- Re: [abinit-forum] Etotal versus etotal, Vincent Chevrier, 09/03/2009
- Re: [abinit-forum] Etotal versus etotal, TORRENT Marc, 09/03/2009
- Re: [abinit-forum] Etotal versus etotal, matthieu verstraete, 09/03/2009
- Re: [abinit-forum] Etotal versus etotal, Vincent Chevrier, 09/03/2009
- Re: [abinit-forum] Etotal versus etotal, TORRENT Marc, 09/04/2009
- Re: [abinit-forum] Etotal versus etotal, Vincent Chevrier, 09/04/2009
- [abinit-forum] problem with optic, pablo picasso, 09/04/2009
- Re: [abinit-forum] Etotal versus etotal, matthieu verstraete, 09/04/2009
- Re: [abinit-forum] Etotal versus etotal, matthieu verstraete, 09/04/2009
- Re: [abinit-forum] Etotal versus etotal, TORRENT Marc, 09/04/2009
- Re: [abinit-forum] Etotal versus etotal, Vincent Chevrier, 09/04/2009
- REĀ : [abinit-forum] Etotal versus etotal, Marc.TORRENT, 09/06/2009
- Re: [abinit-forum] Etotal versus etotal, TORRENT Marc, 09/04/2009
- Re: [abinit-forum] Etotal versus etotal, Vincent Chevrier, 09/03/2009
- Re: [abinit-forum] Etotal versus etotal, matthieu verstraete, 09/03/2009
- Re: [abinit-forum] Etotal versus etotal, TORRENT Marc, 09/03/2009
- Re: [abinit-forum] Etotal versus etotal, Vincent Chevrier, 09/03/2009
- Re: [abinit-forum] Etotal versus etotal, Bernard Amadon, 09/03/2009
Archive powered by MHonArc 2.6.16.