[CPMD-list] Help optimization of slab!

[=?gb2312?B?c2h5bWE=?=]

Dear CPMD user:
     I met some problems when I optimized geometry of (001) slab of TiN, 5 layers + 5 layers of vacuum. I imitate the input-file in the femd folder of cpmd-test and modifed it. But it displayes like following part in my out-file:
      <<32:32<<<<<<<<<<<< LANCZOS DIAGONALIZATION <<<<<<<<<<<<<<<<<<<<
 >> TIME FOR INITIAL SUBSPACE DIAGONALIZATION:               3.99
 >> CYCLE     NCONV        B2MAX        B2MIN     #HPSI      TIME
        1         0    4.318E-09    1.996E-15      6.00     17.49
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !! UPDRHO| THE NUMBER OF STATES IS TOO SMALL                  !!
 !!         F(  32, 12) =    2.62419E-06                       !!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !! UPDRHO| THE NUMBER OF STATES IS TOO SMALL                  !!
 !!         F(  32, 13) =    2.62419E-06                       !!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !! UPDRHO| THE NUMBER OF STATES IS TOO SMALL                  !!
 !!         F(  32, 26) =    2.62401E-06                       !!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !! UPDRHO| THE NUMBER OF STATES IS TOO SMALL                  !!
 !!         F(  32, 28) =    2.03191E-06                       !!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !! UPDRHO| THE NUMBER OF STATES IS TOO SMALL                  !!
 !!         F(  32, 29) =    2.03191E-06                       !!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 !! UPDRHO| THE NUMBER OF STATES IS TOO SMALL                  !!
 !!         F(  32, 31) =    2.62401E-06                       !!
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 ==------------------------------------------------------------==
 ==     NFI=     44      ETOT=    -68.988408   TCPU=    830.91 ==
 == DRHOMAX= 5.862E-03  DETOT=    -7.459E-04    THL= 0.000E+00 ==
 ==------------------------------------------------------------==

What's the meaning of it? Does it mean I should increase the number of states in my input-file? Or need I correct some part of my input-file? If it continue on runing, will I get the right values? My input-file follows and please check and correct it if you find something wrong with it:

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
&INFO
 (001) slab of TiN, 5 layers + 8 layers of vacuum
&END

&CPMD
 OPTIMIZE GEOMETRY
 UNITE HESSIAN
 BFGS
 FREE ENERGY FUNCTIONAL
 LANCZOS DIAGONALISATION
 LANCZOS PARAMETERS
 1  6  20  1.0D-18
 TROTTER FACTOR
 0.001
 BOGOLIUBOV CORRECTION OFF
 GRAM-SCHMIDT ORTHOGONALISATION
 CONVERGENCE ORBITAL
 1.0D-5
 CONVERGENCE GEOMETRY
 1.0D-4
 MAXSTEP
 600
 BROYDEN MIXING
 0.15  200  0.01  0  8
 ALEXANDER MIXING
 1.1
 ELECTRON TEMPERATURE
 1000
 MIRROR
 COMPRESS WRITE32
 STRUCTURE BONDS
 PRINT FORCES COORDINATES ON
&END

&SYSTEM
 POINT GROUP
 AUTO
 SYMMETRY
 8
 ANGSTROM
 SCALE
 CELL
 3.182  1.0  9.2  0.0  0.0  0.0
 CUTOFF
 80
 STATES
 32
 TESR
 3
 KPOINTS MONKHORST-PACK FULL
 8  8  1
&END

&DFT
 FUNCTIONAL HCTH
&END

&ATOMS
*Ti_MT_HCTH_NLCC.psp KLEINMAN-BYLANDER RAGGIO=1.6
 LMAX=D LOC=S SKIP=P
 5
 0    0    0
 1/2  1/2  1/13
 0    0    2/13
 1/2  1/2  3/13
 0    0    4/13

*N_MT_HCTH.psp  KLEINMAN-BYLANDER
 LMAX=P
 5
 1/2  1/2  0
 0    0    1/13
 1/2  1/2  2/13
 0    0    3/13
 1/2  1/2  4/13
&END 
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Another question is that I found some cpmdtest files always specify the method (GAUSS-HERMIT or KLEINMAN-BYLANDER) to be used for the calculation of the nonlocal parts of the pseudopotential, I want to know are there any difference between the two method in nature and how to decide to specify which method to be used in the calculation? Deponding on pesudopential?

Thanks in advance and expecting your reply!
Best whishes!
                                                                           shyma
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://cpmd.org/pipermail/cpmd-list/attachments/20050513/9a9b9ea0/attachment.html