[CPMD-list] ADJMU| BISECTION COULD NOT
Eung-Gun Kim
eunggun.kim at chemistry.gatech.edu
Tue Mar 1 17:57:35 CET 2005
Dear Axel,
Thanks very much for your kind reply. I tried KOHN-SHAM ENERGIES as you
suggested, and unfortunately this workaround didn't work for me. Now I am
trying to adjust the box size a little bit to see if I can get around with the
error.
Regards,
EG
Quoting Axel Kohlmeyer <axel.kohlmeyer at theochem.ruhr-uni-bochum.de>:
>
> On Thu, 24 Feb 2005, Eung-Gun Kim wrote:
>
> EK> Dear List Subscribers,
>
> dear eung-gun,
>
> [...]
> EK> lately I found myself in the following situation when certain values of
> KPOINTS
> EK> are used:
> EK>
> EK> 1) All calculations go smoothly with ODIIS, independent of KPOINTS.
> EK> 2) Restarting with the LANCZOS diago goes fine all the way through the
> very
> EK> last k point, and then crashes with
> EK> "...SUBROUTINE ADJMU| BISECTION COULD NOT CONVERGE...".
> EK> 3) This error sometimes goes away when k point grid ratios are
> adjusted, say,
> EK> from 2 8 14 to either 2 8 16 or 1 4 8.
> EK> 4) Similar errors were also encountered in a system with different
> contents
> EK> and supercell size.
> EK> 5) The error message complains about a small number of STATES. I feed
> here
> EK> with the total number of filled states, and a larger number is not
> allowed
> EK> by the code when switching from ODIIS to LANCZOS. And I don't see
> a
> EK> reason why I need a larger STATES for my organic crystals with
> sizable
> EK> band gaps.
>
>
> i cannot tell you why you see this problem. i've seen it myself and
> worked around it by simply running a KOHN-SHAM ENERGIES run after
> the initial wavefunction optimization. IIRC, for lanczos you should add
> about 10-20% virtual states to just provide the corresponding number on
> the line after KOHN-SHAM ENERGIES.
>
> for all subsequent calculations you can then use the resulting restart
> with the desired larger number of STATES given in the &SYSTEM section.
> but since you seem only interested in the eigenvalues, the KS-energies
> run should provide you with everthing you want.
>
>
> regards,
> axel.
> EK>
> EK> I am attaching portions of the input and output files of a representative
> diago
> EK> run below for your reference. Thanks very much for reading.
> EK>
> EK> Regards,
> EK>
> EK> EG Kim
> EK>
> EK>
>
###############################################################################
> EK> input file
> --------------------------------------------------------------------
> EK>
> EK> &CPMD
> EK> RESTART COORDINATES WAVEFUNCTION
> EK> OPTIMIZE WAVEFUNCTION
> EK> LANCZOS DIAGONALISATION
> EK> LANCZOS PARAMETER
> EK> 200 8 0 1.D-8
> EK> STORE WAVEFUNCTIONS OFF
> EK> MAXSTEP
> EK> 1
> EK> &END
> EK>
> EK> &DFT
> EK> FUNCTIONAL BLYP
> EK> &END
> EK>
> EK> &SYSTEM
> EK> ANGSTROM
> EK> SYMMETRY
> EK> ORTHORHOMBIC
> EK> CELL ABSOLUTE DEGREE
> EK> 7.92534743 10.6994 6.8 90.0 90.0 90.0
> EK> CUTOFF
> EK> 70
> EK> CHARGE
> EK> 0
> EK> STATES
> EK> 92
> EK> KPOINTS MONKHORST-PACK
> EK> 6 4 6
> EK> ......
> EK> ........
> EK>
> EK> output file
> -------------------------------------------------------------------
> EK>
> EK> ........
> EK> ......
> EK>
> EK> <<36:36<<<<<<<<<<<< LANCZOS DIAGONALIZATION <<<<<<<<<<<<<<<<<<<<
> EK> >> TIME FOR INITIAL SUBSPACE DIAGONALIZATION: .70
> EK> >> CYCLE NCONV B2MAX B2MIN #HPSI TIME
> EK> 1 92 2.248E-13 1.633E-15 1.00 .11
> EK> ADJMU| EIGENVALUES:
> EK> 1 1 -0.744303257893736525
> EK> 1 2 -0.744207855529486761
> EK> 1 3 -0.744131382450865297
> EK> 1 4 -0.744046961929959005
> EK> 1 5 -0.713127925040847876
> EK> 1 6 -0.713072937556081698
> EK> 1 7 -0.712971653074167544
> EK> 1 8 -0.712869568517556229
> EK>
> EK> ....
> EK> ....
> EK> ....
> EK>
> EK> 36 92 0.123545374312123368
> EK> AMU1= -0.745159946143003404 RHINT1= -183.944444444444457
> EK> AMU2= 0.133314578748302381 RHINT2=
> -0.960653778747655451E-11
> EK> DAMU= 0.878474524891305730 RHINT = -183.944444444444457
> EK> RHINT-NEL= -183.944444444444457
> EK> ADJMU! THE NUMBER OF STATES [ 92 ] IS TOO SMALL
> EK>
> EK>
> EK> PROGRAM STOPS IN SUBROUTINE ADJMU| BISECTION COULD NOT CONVERGE [PROC=
> 0]
> EK>
> EK>
> EK> _______________________________________________
> EK> CPMD-list mailing list
> EK> CPMD-list at cpmd.org
> EK> http://cpmd.org/mailman/listinfo/cpmd-list
> EK>
> EK>
>
> --
>
> =======================================================================
> Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de
> Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673
> Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045
> D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
> =======================================================================
> If you make something idiot-proof, the universe creates a better idiot.
>
>
--
+ -
| Eung-Gun Kim, Ph.D.
| Postdoctoral Fellow
| School of Chemistry and Biochemistry
| Georgia Institute of Technology
| Atlanta, GA 30332-0400
| U.S.A.
+ -
| Phone: +1.404.894.6456
| Email: eunggun.kim at chemistry.gatech.edu
+ -
More information about the CPMD-list
mailing list