[CPMD-list] about pseudopotential and an error

[Axel Kohlmeyer]

On Wed, 22 Sep 2004, whzhang wrote:

WZ> cpmd-lDear CPMD users and authors
WZ>    I just begin to use CPMD. I want to simulate surface NiAl(110). I
WZ> only found one kind of pseopotential for atom Ni(Ni_MT_BLYP_NLCC.psp)
WZ> and then I use Al_MT_GIA_BLYP_NLCC.psp for atom Al. When I perform
WZ> wavefunciton optimization the program was stopped .

[...]
WZ> 
WZ>   ATRHO| WARNING! THE NUMBER OF GENERATED ATOMIC STATE (39)
WZ>   ATRHO| WARNING! IS LESS THAN THE NUMBER OF STATE (40)
WZ>   ATRHO| WARNING!     1 STATES ARE EQUAL TO ZERO
WZ>   ATRHO| USE SECTION BASIS TO SPECIFY ATOMIC BASIS
WZ>   ATRHO| CHARGE(R-SPACE):   39.000000      (G-SPACE):   39.000000
WZ>  ATOMWF| WARNING! RANDOMIZATION OF     1 STATES
WZ> 

ok, you have a warning here, that should be taken seriously,
if you don't do that intentionally.

WZ>  ****************************************************************
WZ>  *                      ATOMIC COORDINATES                      *
WZ>  ****************************************************************
WZ>        1      Ni           7.880671      15.761003       0.000000
WZ>        2      Ni           7.880671      15.761003     189.132041
WZ>        3      Ni           7.880671      15.761003      94.566020
WZ>        4      Al           0.000000       0.000000       0.000000
WZ>        5      Al           0.000000       0.000000     189.132041
WZ>        6      Al           0.000000      15.761003      94.566020
WZ>  ****************************************************************
WZ> 
WZ> 
WZ>  DEGREES OF FREEDOM FOR SYSTEM:                                15
WZ> 
WZ>  CPU TIME FOR WAVEFUNCTION INITIALIZATION:         188.11 SECONDS
WZ>  ***    RWFOPT| PROGSIZE DATA= 527.04 MByte  /  STACK=   0.04 MByte
WZ> 
WZ> 
WZ>  PROGRAM STOPS IN SUBROUTINE K_UPDWF| KPOINT+VIRTUAL STATES only with DIAGONALIZATION methods [PROC =
WZ>    0]

that is the final piece of useful information.
your input cannot work with cpmd. in fact, there are 
two points to consider:

- you use k-points and virtual states, that means that you have
to use LANCZOS DIAGONALIZATION with cpmd.

- you have a metallic system, that means, that you have to 
use the FREE ENERGY FUNCTIONAL feature.

please have a look at the 'femd' subdirectory in the CPMD-tests
package, and in the hints section of the CPMD manual for examples
and tips of how to do such a calculation. you should also be
aware of the fact, that CPMD is not so well tuned for this
kind of calculation  (in contrast to gamma-point only car-parrinello
md with norm-conserving pseudopotentials).

regards,
	axel kohlmeyer.


WZ> MPI Application rank 0 exited before MPI_Finalize() with status 231
WZ> 
WZ> 
WZ> 
WZ> 
WZ> 
WZ> best regards 
WZ> 
WZ> 
WZ> ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡whzhang
WZ> ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡whzhang at ustc.edu
WZ> ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡2004-09-22
WZ> 

-- 


=======================================================================
Dr. Axel Kohlmeyer                        e-mail: axel.kohlmeyer at rub.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/
=======================================================================