[Pw_forum] Wired and large (4Ry) total energy difference on different machines

[Fenggong Wang]

Dear everyone:
       Recently we got quite different total energy using exactly the same
input files for pwscf cacluation when running on different machines.
On one machine, the total energy of the first ionic step is:
!    total energy              =   -3188.41593157 Ry
     Harris-Foulkes estimate   =   -3188.41560923 Ry
     estimated scf accuracy    <       0.00005695 Ry

     The total energy is the sum of the following terms:

     one-electron contribution =  -55000.39131345 Ry
     hartree contribution      =   27705.53744978 Ry
     xc contribution           =    -530.25923399 Ry
     ewald contribution        =   24636.69787808 Ry
     electric field correction =       0.00107603 Ry
     smearing contrib. (-TS)   =      -0.00178802 Ry
while on the other one, it is:
!    total energy              =   -3184.04215443 Ry
     Harris-Foulkes estimate   =   -3184.04248811 Ry
     estimated scf accuracy    <       0.00007177 Ry

     The total energy is the sum of the following terms:

     one-electron contribution =  -55000.21479459 Ry
     hartree contribution      =   27705.35347136 Ry
     xc contribution           =    -525.87732678 Ry
     ewald contribution        =   24636.69787808 Ry
     electric field correction =       0.00042900 Ry
     smearing contrib. (-TS)   =      -0.00181150 Ry

As we can see, the energy difference is mainly from the xc part. During the
ionic relaxation, the large energy difference maintains. This is wired and
we have no idea of what the problem is. Since the partial core correction
is used for the Al pseudopotential, does the large difference have
something related to the Al psps? Could any one have any idea or hint about
this? Thank you.
Our input files are enclosed as follows for your reference:
&control
      calculation = 'relax'
      prefix='lao_sto_1+10ucs_surf',
      nstep=200,
      pseudo_dir ='./'
      outdir='./'
      verbosity = 'high'
      nstep = 200
      dipfield = .true.
      tefield = .true.
      restart_mode = 'from_scratch'
 /
 &system
      ibrav= 6, celldm(1)=7.279889198, celldm(2)=1, celldm(3)=15.5
      nat=60, ntyp= 5,
      ecutwfc = 50, occupations='smearing', degauss = 0.01
      nspin = 1
      edir = 3, eopreg=0.03, emaxpos=0.860
 /
 &electrons
    mixing_mode = 'plain'
    mixing_beta = 0.10
    conv_thr = 1.0e-4
    startingwfc='random'
    startingpot='atomic'
 /
 &ions
    ion_dynamics = 'bfgs',
    pot_extrapolation = 'atomic',
    upscale = 100,
    trust_radius_ini = 0.5,
 /

ATOMIC_SPECIES
La  138.9055  La_LDA.upf
Al   26.9815  Al_LDA.upf
Sr   87.6200  Sr_LDA.upf
Ti   47.8670  Ti_LDA.upf
O    15.9994  O_LDA.upf

K_POINTS {automatic}
6 6 1 0 0 0ATOMIC_POSITIONS (angstrom)
O        1.926180751   0.000000000   1.932851169
O        1.926180751   0.000000000   5.806416120
O        1.926180751   0.000000000   9.690288605
O        1.926180751   0.000000000  13.570435914
O        1.926180751   0.000000000  17.444898624
O        1.926180751   0.000000000  21.316455327
O        1.926180751   0.000000000  25.188371707
O        1.926180751   0.000000000  29.064620363
O        1.926180751   0.000000000  32.947258569
O        1.926180751   0.000000000  36.826561486
O        1.926180751   0.000000000  40.718241080
O        1.926180751   0.000000000  44.413343364
O        1.926180751   1.926180751  -0.022898252
O        1.926180751   1.926180751   3.855269167
O        1.926180751   1.926180751   7.733476374
O        1.926180751   1.926180751  11.614174551
O        1.926180751   1.926180751  15.491455293
O        1.926180751   1.926180751  19.364662145
O        1.926180751   1.926180751  23.236913267
O        1.926180751   1.926180751  27.111740990
O        1.926180751   1.926180751  30.990883340
O        1.926180751   1.926180751  34.870585517
O        1.926180751   1.926180751  38.750367991
O        1.926180751   1.926180751  42.575880071
O        0.000000000   1.926180751   1.932851169
O        0.000000000   1.926180751   5.806416120
O        0.000000000   1.926180751   9.690288605
O        0.000000000   1.926180751  13.570435914
O        0.000000000   1.926180751  17.444898624
O        0.000000000   1.926180751  21.316455327
O        0.000000000   1.926180751  25.188371707
O        0.000000000   1.926180751  29.064620363
O        0.000000000   1.926180751  32.947258569
O        0.000000000   1.926180751  36.826561486
O        0.000000000   1.926180751  40.718241080
O        0.000000000   1.926180751  44.413343364
Ti       1.926180751   1.926180751   1.835084398
Ti       1.926180751   1.926180751   5.714812422
Ti       1.926180751   1.926180751   9.595819437
Ti       1.926180751   1.926180751  13.476028107
Ti       1.926180751   1.926180751  17.351541485
Ti       1.926180751   1.926180751  21.223872514
Ti       1.926180751   1.926180751  25.096507617
Ti       1.926180751   1.926180751  28.972779773
Ti       1.926180751   1.926180751  32.853697401
Ti       1.926180751   1.926180751  36.735490165
Ti       1.926180751   1.926180751  40.630380050
Al       1.926180751   1.926180751  44.503275725
Sr       0.000000000   0.000000000  -0.079993855
Sr       0.000000000   0.000000000   3.786562857
Sr       0.000000000   0.000000000   7.660339507
Sr       0.000000000   0.000000000  11.541763403
Sr       0.000000000   0.000000000  15.419292965
Sr       0.000000000   0.000000000  19.292333803
Sr       0.000000000   0.000000000  23.163762621
Sr       0.000000000   0.000000000  27.037320220
Sr       0.000000000   0.000000000  30.916125648
Sr       0.000000000   0.000000000  34.794891468
Sr       0.000000000   0.000000000  38.659475210
La       0.000000000   0.000000000  42.684568538




-- 
Dr. Fenggong Wang
Postdoctoral Researcher

Department of Chemistry
University of Pennsylvania
231 South 34th Street, BOX 314
Philadelphia, PA 19104-6323
e-mail: wangfenggong at gmail.com; fenggong at sas.upenn.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20120917/573edb66/attachment.html>