CPHF Module - ikain not converged


Clicked A Few Times
Dear NWChem developers and users,

We are currently having troubles with the CPHF Module during the frequency analysis.
The Iterative solution of linear equations are not converging, which results in the error message "ikain not converged".
We managed to set the number of iterations higher with "set cphf:maxiter 2500", but they are simply continuing without converging. This happens for a range of structures/isomers. How can we modify the input to help the convergence?


This is our input (without the "set cphf:maxiter 2500"):
start Sn10P2_1SiPBE0

memory total 32000 mb

charge 0

geometry units angstrom noautoz noautosym
Sn                    0.26925119     1.31076216     1.34259267
Sn                    1.57791812    -0.72646357    -0.78694794
Sn                   -0.31621535    -1.76669855     1.51215804
Sn                   -2.60153299    -1.06867122    -0.61711425
Sn                   -2.09472757     1.92404765    -0.88224526
Sn                   -0.92864827     0.03956535    -2.89486175
Sn                   -2.39715028     0.52585329     1.90018541
Sn                   -0.32598565    -2.74446405    -1.43359287
Sn                    2.63361744    -0.31579865     2.21931902
Sn                    2.87506804     2.35376327     0.01632841
P                     3.89433998    -0.01255075     0.01213677
P                     0.58277208     1.63770465    -1.28199310
end

basis
* library def2-tzvpp
end

ecp
Sn library def2-ecp
end

dft
vectors input test.movecs output dft.movecs
xc pbe0
direct grid xfine
mult 1
convergence ncydp 0 dampon 1d99 dampoff 1d-7 damp 70
iterations 750
noio
end



task dft freq

Which results in the following (partial) output:
                                NWChem CPHF Module
                                ------------------
  scftype          =     RHF
  nclosed          =      125
  nopen            =        0
  variables        =    70125
  # of vectors     =       36
  tolerance        = 0.10D-03
  level shift      = 0.00D+00
  max iterations   =       50
  max subspace     =      360

Grid integrated density:     249.999995405115
Requested integration accuracy:   0.10E-13
SCF residual:   1.406252467695138E-004


Iterative solution of linear equations
  No. of variables    70125
  No. of equations       36
  Maximum subspace      360
        Iterations       50
       Convergence  1.0D-04
        Start time  11581.9


   iter   nsub   residual    time
   ----  ------  --------  ---------
     1     36    7.89D-02   13507.1
     2     72    4.62D-02   15430.3
     3    108    1.35D-02   17354.8
     4    144    1.78D-02   19278.3
     5    180    9.51D-03   21202.9
     6    216    2.64D-03   23127.5
     7    252    2.01D-03   25052.1
     8    288    1.81D-03   26976.3
     9    324    2.33D-03   28900.8
    10    360    1.20D-03   30825.6
    11    360    1.16D-03   32750.1
    12    360    1.32D-03   34674.6
    13    360    1.95D-03   36598.9
    14    360    2.83D-03   38523.6
    15    360    9.89D-03   40449.2
    16    360    1.21D-02   42374.0
    17    360    7.45D-03   44298.7
    18    360    5.96D-03   46223.9
    19    360    6.84D-03   48148.9
    20    360    4.36D-03   50074.1
    21    360    4.84D-03   51999.1
    22    360    5.85D-03   53924.5
    23    360    2.57D-03   55849.6
    24    360    3.52D-03   57775.6
    25    360    3.40D-03   59700.1
    26    360    7.79D-03   61624.8
    27    360    1.20D-02   63549.0
    28    360    6.15D-03   65473.2
    29    360    6.80D-03   67398.7
    30    360    4.00D-03   69323.5
    31    360    4.38D-03   71248.2
    32    360    5.43D-03   73173.4
    33    360    8.65D-03   75099.5
    34    360    3.89D-03   77024.0
    35    360    2.65D-03   78949.3
    36    360    3.01D-03   80873.5
    37    360    6.64D-03   82797.5
    38    360    5.16D-03   84722.5
    39    360    4.49D-03   86647.1
    40    360    4.64D-03   88572.1
    41    360    3.54D-03   90497.1
    42    360    2.21D-03   92421.8
    43    360    4.54D-03   94347.5
    44    360    3.59D-03   96272.7
    45    360    6.50D-03   98198.2
    46    360    4.00D-03  100122.7
    47    360    4.88D-03  102047.4
    48    360    5.20D-03  103972.0
    49    360    3.35D-03  105898.4
    50    360    3.00D-03  107824.8
------------------------------------------------------------------------
lkain: not converged                   0
lkain: not converged                   0
lkain: not converged                   0
lkain: not converged                   0
------------------------------------------------------------------------
lkain: not converged                   0
------------------------------------------------------------------------
lkain: not converged                   0
------------------------------------------------------------------------
lkain: not converged                   0
------------------------------------------------------------------------
------------------------------------------------------------------------
------------------------------------------------------------------------
------------------------------------------------------------------------
------------------------------------------------------------------------
------------------------------------------------------------------------
  current input line :
  current input line :
  current input line :
  current input line :
     0:
     0:
     0:
------------------------------------------------------------------------
     0:
------------------------------------------------------------------------
------------------------------------------------------------------------
------------------------------------------------------------------------
------------------------------------------------------------------------
This type of error is most commonly associated with calculations not reaching
convergence criteria
lkain: not converged                   0
This type of error is most commonly associated with calculations not reaching
convergence criteria


Thank you in advance!

Forum Vet
The following input gives a complete frequency run.
start Sn10P2_1SiPBE0

memory stack 1200 mb heap 100 mb global 3650 mb noverify

charge 0

geometry units angstrom #noautoz noautosym
Sn                    0.26925119     1.31076216     1.34259267
Sn                    1.57791812    -0.72646357    -0.78694794
Sn                   -0.31621535    -1.76669855     1.51215804
Sn                   -2.60153299    -1.06867122    -0.61711425
Sn                   -2.09472757     1.92404765    -0.88224526
Sn                   -0.92864827     0.03956535    -2.89486175
Sn                   -2.39715028     0.52585329     1.90018541
Sn                   -0.32598565    -2.74446405    -1.43359287
Sn                    2.63361744    -0.31579865     2.21931902
Sn                    2.87506804     2.35376327     0.01632841
P                     3.89433998    -0.01255075     0.01213677
P                     0.58277208     1.63770465    -1.28199310
end

basis spherical
* library def2-tzvpp
end

ecp
Sn library def2-ecp
end

dft
vectors input atomic output dft.movecs
xc pbe0
direct grid fine
mult 1
smear
convergence ncydp 0 dampon 1d99 dampoff 1d-7 damp 45 energy 1d-9
iterations 750
end
driver
tight
end
set quickguess t
task dft optimize


task dft freq

Clicked A Few Times
Thank you, that did indeed do the trick. The calculations are properly converging now.

What exactly does "set quickguess t" do?


Forum >> NWChem's corner >> Running NWChem