SCF Convergence

2,864 views
Skip to first unread message

SeokYun

unread,
Nov 21, 2012, 5:35:06 PM11/21/12
to cp...@googlegroups.com
Dear CP2K users,

I am running my job and I have this SCF convergence problem.

There is no SCF convergence issue initially, but usually after 50 steps, there is a convergence problem.

Rather than approaching EPS_SCF, the values are either staying constant or increasing.

  65 OT DIIS     0.15E+00    4.5     0.00002065      -131.7084955753 -4.55E-06

  Trace(PS):                                  100.0000000000
  Electronic density on regular grids:       -100.0000006853       -0.0000006853
  Core density on regular grids:              100.0000001227        0.0000001227
  Total charge density on r-space grids:       -0.0000005626
  Total charge density g-space grids:          -0.0000005626

    66 OT DIIS     0.15E+00    4.2     0.00002066      -131.7085001385 -4.56E-06

  Trace(PS):                                  100.0000000000
  Electronic density on regular grids:       -100.0000006853       -0.0000006853
  Core density on regular grids:              100.0000001227        0.0000001227
  Total charge density on r-space grids:       -0.0000005626
  Total charge density g-space grids:          -0.0000005626

    67 OT SD       0.15E+00    4.0     0.00002064      -131.7085030685 -2.93E-06

  Trace(PS):                                  100.0000000000
  Electronic density on regular grids:       -100.0000006853       -0.0000006853
  Core density on regular grids:              100.0000001227        0.0000001227
  Total charge density on r-space grids:       -0.0000005626
  Total charge density g-space grids:          -0.0000005626

    68 OT SD       0.15E+00    4.2     0.00002064      -131.7085041649 -1.10E-06

  Trace(PS):                                  100.0000000000
  Electronic density on regular grids:       -100.0000006853       -0.0000006853
  Core density on regular grids:              100.0000001227        0.0000001227
  Total charge density on r-space grids:       -0.0000005626
  Total charge density g-space grids:          -0.0000005626

    69 OT SD       0.15E+00    4.1     0.00002064      -131.7085052613 -1.10E-06

  Trace(PS):                                  100.0000000000
  Electronic density on regular grids:       -100.0000006853       -0.0000006853
  Core density on regular grids:              100.0000001227        0.0000001227
  Total charge density on r-space grids:       -0.0000005626
  Total charge density g-space grids:          -0.0000005626

    70 OT SD       0.15E+00    4.4     0.00002065      -131.7085063577 -1.10E-06

  Trace(PS):                                  100.0000000000
  Electronic density on regular grids:       -100.0000006853       -0.0000006853
  Core density on regular grids:              100.0000001227        0.0000001227
  Total charge density on r-space grids:       -0.0000005626
  Total charge density g-space grids:          -0.0000005626

    71 OT SD       0.15E+00    4.3     0.00002065      -131.7085074543 -1.10E-06

  Trace(PS):                                  100.0000000000
  Electronic density on regular grids:       -100.0000006853       -0.0000006853
  Core density on regular grids:              100.0000001227        0.0000001227
  Total charge density on r-space grids:       -0.0000005626

Also here is my input file:

&GLOBAL

  PROJECT C20H14O

  PRINT_LEVEL MEDIUM

  RUN_TYPE GEO_OPT

&END GLOBAL

&MOTION

 &GEO_OPT

  MINIMIZER CG

  MAX_ITER 200

  MAX_FORCE 5.0E-3

 &END GEO_OPT

&END MOTION

&FORCE_EVAL

  METHOD Quickstep

  &DFT

    BASIS_SET_FILE_NAME ./BASIS_SET

    POTENTIAL_FILE_NAME ./POTENTIAL

    &MGRID

      CUTOFF 300

    &END MGRID

    &QS

      EPS_DEFAULT 1.0E-10

    &END QS

    &SCF

      EPS_SCF 1.0E-6

      MAX_SCF 100

      SCF_GUESS RESTART

      &OT

       MINIMIZER DIIS

      &END OT

    &END SCF

    &XC

      &XC_FUNCTIONAL Pade

      &END XC_FUNCTIONAL

    &END XC

&END DFT

  &SUBSYS

    &CELL

      ABC 40.000 40.000 40.000

    &END CELL

&COORD

C    15.02103029        21.43572257     20.18000314

C    12.13797029        19.86029257     20.10217314

C    9.387700286        21.39858257     20.17763314

H    12.31361029        18.71215257     20.04348314

C    9.392820286        24.70136257     20.34288314

H    8.494900286        20.88161257     20.15168314

C    12.16548029        26.11910257     20.41407314

H    8.512790286        25.23974257     20.36974314

C    14.93316029        24.32580257     20.32459314

C    17.27508029        19.49524257     20.06628314

H    15.45990029        24.76075257     20.34681314

C    18.34824029        16.68641257     19.89707314

C    22.15934029        16.83150257     19.89931314

C    22.84514029        19.70026257     20.05987314

O    20.12023029        21.82217257     20.06602314

C    17.05926029        14.17188257     19.77462314

C    22.56556029        11.35336257     19.61229314

C    18.37166029        11.22565257     19.61264314

C    23.67063029        14.35575257     19.78200314

H    12.11085029        26.74907257     20.92021314

H    16.27474029        14.03204257     19.82687314

H    24.60208029        14.24160257     19.67745314

H    22.92407029        10.57705257     19.51190314

C    24.88894029        21.35826257     19.98780314

C    24.78771029        24.12058257     19.80812314

C    27.56932029        19.53722257     20.15068314

C    30.06265029        21.06305257     20.09477314

C    30.04334029        23.99254257     19.90571314

C    27.61230029        25.90008257     19.74150314

H    27.60669029        26.98852257     19.67128314

H    30.96030029        24.58552257     19.88310314

H    30.90325029        20.48235257     20.14647314

H    27.66110029        18.53225257     20.21699314

H    23.86707029        24.53232257     19.76595314

H    17.89108029        10.23014257     19.46794314

    &END COORD

    &KIND H

      BASIS_SET DZVP-GTH-PADE

      POTENTIAL GTH-PADE-q1

    &END KIND

    &KIND C

      BASIS_SET DZVP-GTH-PADE

      POTENTIAL GTH-PADE-q4

    &END KIND

    &KIND O

      BASIS_SET DZVP-GTH-PADE

      POTENTIAL GTH-PADE-q6

    &END KIND

&END SUBSYS

&END FORCE_EVAL



I think my input file is correct so I'm not so sure what the problem might be.

Can anyone give me an advice as to what the problem might be?

Thank you. I really appreciate it.

Sandeep Kumar Reddy

unread,
Nov 21, 2012, 10:29:24 PM11/21/12
to cp...@googlegroups.com
Hi,

include  OUTER SCF loop in SCF section.. 


   &SCF
       SCF_GUESS RESTART
       EPS_SCF 1.0E-7
       MAX_SCF 50
#       MIXING 0.3
       &OUTER_SCF
          EPS_SCF 1.0E-7
          MAX_SCF 500
       &END
       &OT
          MINIMIZER CG
          PRECONDITIONER FULL_ALL
          ENERGY_GAP 0.001
       &END

    &END SCF

This will change pre-conditioner frequently so that scf convergence is faster. If above option didn't work, try include these lines in DFT--XC section.


 &XC_GRID
            XC_DERIV  SPLINE2
            XC_SMOOTH_RHO NN50
 &END XC_GRID


And re-run the job.


-Sandeep







--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To view this discussion on the web visit https://groups.google.com/d/msg/cp2k/-/IUGj7T_6Vo4J.
To post to this group, send email to cp...@googlegroups.com.
To unsubscribe from this group, send email to cp2k+uns...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/cp2k?hl=en.

shoutian sun

unread,
Nov 21, 2012, 10:36:08 PM11/21/12
to cp...@googlegroups.com
Hi SeokYun,
I have some suggestions for your input file.
 
1. The conformer you used for optimization is very poor, even using atomic unit,
    the C-C distance of the ring is about 3.33 anstrom (kinds be wrong) or 3.33 bohr
    (about 1.76 angstrom, also a bit larger, isn't it?).  So, my suggestion is rebuilding
    your conformer.
 
2. use the POISSON_SOLVER (I think your systerm is a cluster, right?)
    &POISSON
         PERIODIC none
         POISSON_SOLVER wavelet
    &END POISSON
3. use  MINIMIZER CG for OT sec. (If SCF doesn't converge, sometimes, it works)
 
4. XC_FUNCTIONAL Pade is not a good choice. (BLYP, I always used.)
 
5. sometime, SCF doesn't converge, you have to use smooth method in the &XC section
    , as follows:
      &XC_GRID
         XC_SMOOTH_RHO NN10
         XC_DERIV SPLINE2_SMOOTH
      &END XC_GRID

6. cell parameter:      ABC 40.000 40.000 40.000
    I think the default unit is angstrom, right?
    40.0 angstrom for 35 atoms conformer is quite big.
    And very time consuming, Don't you think so?
    (PERIODIC none for clusters.)
 
7. The coordinates of your conformer. I think your comformer
    is build in atomic unit. And you do not modify it and just put
    it in the &COORD.....&END COORD section. Am I right?
    However, the problem is the default unit of atom coordinates
    for cp2k is angstrom, not bohr. So, please modify it, or add
    key word: UNIT Bohr in the following sec.
 
    &COORD
        C    15.02103029        21.43572257     20.18000314
        C    12.13797029        19.86029257     20.10217314
            ..................
       &END COORD
 
8. using POISSON_SOLVER wavelet, you have to use the following key words
    to put the cluster in the center of the box.
 
    &TOPOLOGY
        &CENTER_COORDINATES 
        &END
    &END
 
I rebuild your comformer and do the optimization with my input file, it converge well.
For detail, please refer to cp2k input reference.
 
Hope, it works for you.
 
Regards,
sun
Chemistry, CUHK, Hong Kong
 
 
 
 

SeokYun Kim

unread,
Nov 22, 2012, 1:05:02 AM11/22/12
to cp...@googlegroups.com
Thank you very much.

I really appreciate it.

I'll try to modify and rerun my job.


--
You received this message because you are subscribed to the Google Groups "cp2k" group.
Reply all
Reply to author
Forward
0 new messages