AM05 bug? A single H2 cannot get SCF converged.
Posted: Wed Apr 28, 2010 6:52 am
Dear all,
I am puzzled by the fact that a single H2 molecule in a cubic cell (10x10x10 Angstrom^3) cannot get SCF converged using AM05 functional.
The SCF cycle diverged after the first several nscf calculations.
WARNING: random wavefunctions but no delay for mixing, default for NELMDL
entering main loop
N E dE d eps ncg rms rms(c)
DAV: 1 0.144560336782E+02 0.14456E+02 -0.16806E+03 10 0.745E+03
DAV: 2 0.114315863920E+02 -0.30244E+01 -0.30244E+01 20 0.363E+02
DAV: 3 0.978849484047E+01 -0.16431E+01 -0.16431E+01 20 0.292E+02
DAV: 4 0.814800848570E+01 -0.16405E+01 -0.16405E+01 20 0.299E+02
DAV: 5 0.674689464115E+01 -0.14011E+01 -0.14011E+01 20 0.314E+02 0.520E+00
DAV: 6 0.139680524355E+07 0.13968E+07 -0.55105E+06 15 0.153E+07 0.310E+00
DAV: 7 0.722642742181E+06 -0.67416E+06 -0.19147E+06 20 0.528E+06 0.431E-01
DAV: 8 0.610828518592E+06 -0.11181E+06 -0.13593E+05 20 0.107E+06 0.474E-01
DAV: 9 0.237310349590E+06 -0.37352E+06 -0.48760E+04 20 0.429E+05 0.268E-01
DAV: 10 0.136185088660E+06 -0.10113E+06 -0.34180E+04 20 0.332E+05 0.119E-01
DAV: 11 0.792009257931E+05 -0.56984E+05 -0.72911E+03 20 0.126E+05 0.101E-01
DAV: 12 0.501935416160E+05 -0.29007E+05 -0.22963E+03 15 0.722E+04 0.378E-02
DAV: 13 0.401751074021E+05 -0.10018E+05 -0.10727E+03 20 0.438E+04 0.937E-02
DAV: 14 0.467949977633E+05 0.66199E+04 -0.74147E+02 20 0.413E+04 0.114E-01
DAV: 15 0.348808541930E+05 -0.11914E+05 -0.29775E+02 20 0.203E+04 0.808E-01
DAV: 16 0.777158273880E+05 0.42835E+05 -0.13771E+04 10 0.365E+05 0.941E-01
I compiled vasp.5.2 in IBM as well. cp . The program crashed with this error.
entering main loop
N E dE d eps ncg rms rms(c)
am05_labertw: iteration limit reached.
Should never happen: execution aborted.
I will be appreciated if anyone can point out the workaround.
I am puzzled by the fact that a single H2 molecule in a cubic cell (10x10x10 Angstrom^3) cannot get SCF converged using AM05 functional.
The SCF cycle diverged after the first several nscf calculations.
WARNING: random wavefunctions but no delay for mixing, default for NELMDL
entering main loop
N E dE d eps ncg rms rms(c)
DAV: 1 0.144560336782E+02 0.14456E+02 -0.16806E+03 10 0.745E+03
DAV: 2 0.114315863920E+02 -0.30244E+01 -0.30244E+01 20 0.363E+02
DAV: 3 0.978849484047E+01 -0.16431E+01 -0.16431E+01 20 0.292E+02
DAV: 4 0.814800848570E+01 -0.16405E+01 -0.16405E+01 20 0.299E+02
DAV: 5 0.674689464115E+01 -0.14011E+01 -0.14011E+01 20 0.314E+02 0.520E+00
DAV: 6 0.139680524355E+07 0.13968E+07 -0.55105E+06 15 0.153E+07 0.310E+00
DAV: 7 0.722642742181E+06 -0.67416E+06 -0.19147E+06 20 0.528E+06 0.431E-01
DAV: 8 0.610828518592E+06 -0.11181E+06 -0.13593E+05 20 0.107E+06 0.474E-01
DAV: 9 0.237310349590E+06 -0.37352E+06 -0.48760E+04 20 0.429E+05 0.268E-01
DAV: 10 0.136185088660E+06 -0.10113E+06 -0.34180E+04 20 0.332E+05 0.119E-01
DAV: 11 0.792009257931E+05 -0.56984E+05 -0.72911E+03 20 0.126E+05 0.101E-01
DAV: 12 0.501935416160E+05 -0.29007E+05 -0.22963E+03 15 0.722E+04 0.378E-02
DAV: 13 0.401751074021E+05 -0.10018E+05 -0.10727E+03 20 0.438E+04 0.937E-02
DAV: 14 0.467949977633E+05 0.66199E+04 -0.74147E+02 20 0.413E+04 0.114E-01
DAV: 15 0.348808541930E+05 -0.11914E+05 -0.29775E+02 20 0.203E+04 0.808E-01
DAV: 16 0.777158273880E+05 0.42835E+05 -0.13771E+04 10 0.365E+05 0.941E-01
I compiled vasp.5.2 in IBM as well. cp . The program crashed with this error.
entering main loop
N E dE d eps ncg rms rms(c)
am05_labertw: iteration limit reached.
Should never happen: execution aborted.
I will be appreciated if anyone can point out the workaround.