density fitting in TDDFT didn't work with symmetry

From NWChem

Viewed 289 times, With a total of 3 Posts
Jump to: navigation, search
Forum >> NWChem's corner >> NWChem functionality

Gets Around
Threads 15
Posts 51
1:32:45 AM PDT - Fri, Apr 3rd 2015
Hi, all!

I try to use density fitting in TDDFT calculation of naphtalene (D2h sym), with following input
title "naphtalene TDDFT UV/Vis"
memory stack 100 mb heap 100 mb global 2000 mb
geometry
 symmetry D2h
 H                    -3.36598680     1.24095777     0.00000000
 H                    -3.36598680    -1.24095777     0.00000000
 C                    -2.42448768    -0.70595750     0.00000000
 C                    -2.42448768     0.70595750     0.00000000
 H                    -1.23981036    -2.48124592     0.00000000
 C                    -1.24092183    -1.39739117     0.00000000
 H                    -1.23981036     2.48124592     0.00000000
 C                    -1.24092183     1.39739117     0.00000000
 C                     0.00000000    -0.71416604     0.00000000
 C                     0.00000000     0.71416604     0.00000000
 H                     1.23981036    -2.48124592     0.00000000
 C                     1.24092183    -1.39739117     0.00000000
 H                     1.23981036     2.48124592     0.00000000
 C                     1.24092183     1.39739117     0.00000000
 C                     2.42448768    -0.70595750     0.00000000
 C                     2.42448768     0.70595750     0.00000000
 H                     3.36598680     1.24095777     0.00000000
 H                     3.36598680    -1.24095777     0.00000000
end
basis spherical
 H library def2-SVP
 C library def2-SVP
end
basis "cd basis" spherical
 H library "Weigend Coulomb Fitting"
 C library "Weigend Coulomb Fitting"
end
dft
 xc xbnl07 0.90 lyp 1.00 hfexch 1.00
 cam 0.33 cam_alpha 0.00 cam_beta 1.00
end
tddft
 rpa
 nroots 2
 notriplet
 algorithm 1
end
task tddft energy


But I get "Singularity in Pulay matrix"
   convergence    iter        energy       DeltaE   RMS-Dens  Diis-err    time
 ---------------- ----- ----------------- --------- --------- ---------  ------
 d= 0,ls=0.0,diis     1******************  1.67D+08  1.66D+01  1.14D+14    83.3
  Singularity in Pulay matrix. Error and Fock matrices removed..
 !! scf_movecs_sym_adapt:  180 vectors were symmetry contaminated
  Symmetry fudging
 !! scf_movecs_sym_adapt:  180 vectors were symmetry contaminated
 d= 0,ls=0.0,diis     2****************** -8.43D+07  2.37D+01  1.65D+14    90.9
  Singularity in Pulay matrix. Error and Fock matrices removed..
 !! scf_movecs_sym_adapt:  180 vectors were symmetry contaminated
  Symmetry fudging
 !! scf_movecs_sym_adapt:  180 vectors were symmetry contaminated


When I set symmetry to C1 I get a normal result (same as without CD fitting).
What's wrong with symmetry?

P.S. Nwchem-dev.revision26871-src.2015-02-24
Edited On 2:10:53 AM PDT - Fri, Apr 3rd 2015 by Vladimir

Forum Vet
Threads 3
Posts 835
10:26:05 AM PDT - Fri, Apr 3rd 2015
Many thanks for the feedback

Forum Vet
Threads 3
Posts 835
12:16:35 PM PDT - Fri, Apr 3rd 2015
Vladimir
A fix for this bug has been checked into the NWChem svn repository

Gets Around
Threads 15
Posts 51
11:35:04 PM PDT - Tue, Apr 7th 2015
Many thanks Edoardo, I check out fix from repo, but I have one more related question.
When I calculate DFT energy with this geometry and symmetry D2h, I've got next convergence steps:
   convergence    iter        energy       DeltaE   RMS-Dens  Diis-err    time
 ---------------- ----- ----------------- --------- --------- ---------  ------
 d= 0,ls=0.0,diis     1   -376.8531854784 -8.36D+02  7.42D-03  2.05D+00    35.1
 d= 0,ls=0.0,diis     2   -376.9693001770 -1.16D-01  1.06D-03  2.42D-02    66.4
 d= 0,ls=0.0,diis     3   -376.9728220453 -3.52D-03  4.78D-04  7.42D-03    99.1
 d= 0,ls=0.0,diis     4   -376.9737155828 -8.94D-04  1.81D-04  6.75D-04   135.7
 d= 0,ls=0.0,diis     5   -376.9737947831 -7.92D-05  6.68D-05  1.90D-04   170.5
  Resetting Diis
 d= 0,ls=0.0,diis     6   -376.9738212817 -2.65D-05  1.69D-05  1.58D-06   207.0
 d= 0,ls=0.0,diis     7   -376.9738223506 -1.07D-06  5.13D-06  2.35D-07   242.5
 d= 0,ls=0.0,diis     8   -376.9738224885 -1.38D-07  2.39D-06  6.22D-08   279.1

When I calculate TDDFT energy the convergence steps is:
   convergence    iter        energy       DeltaE   RMS-Dens  Diis-err    time
 ---------------- ----- ----------------- --------- --------- ---------  ------
 d= 0,ls=0.0,diis     1   -376.8531372345 -8.36D+02  7.50D-03  2.05D+00    96.9
 !! scf_movecs_sym_adapt:  108 vectors were symmetry contaminated
  Symmetry fudging
 !! scf_movecs_sym_adapt:    5 vectors were symmetry contaminated
 d= 0,ls=0.0,diis     2   -376.9692666795 -1.16D-01  1.08D-03  2.42D-02   185.5
 !! scf_movecs_sym_adapt:  106 vectors were symmetry contaminated
  Symmetry fudging
 !! scf_movecs_sym_adapt:    6 vectors were symmetry contaminated
 d= 0,ls=0.0,diis     3   -376.9727907722 -3.52D-03  4.97D-04  7.42D-03   275.1
 !! scf_movecs_sym_adapt:   48 vectors were symmetry contaminated
  Symmetry fudging
 !! scf_movecs_sym_adapt:    2 vectors were symmetry contaminated
 d= 0,ls=0.0,diis     4   -376.9737147876 -9.24D-04  2.30D-04  6.76D-04   376.3
 !! scf_movecs_sym_adapt:   13 vectors were symmetry contaminated
  Symmetry fudging
 d= 0,ls=0.0,diis     5   -376.9737940057 -7.92D-05  7.10D-05  1.88D-04   476.4
 !! scf_movecs_sym_adapt:    2 vectors were symmetry contaminated
  Symmetry fudging
  Resetting Diis
 d= 0,ls=0.0,diis     6   -376.9738202880 -2.63D-05  1.67D-05  1.63D-06   575.9
 !! scf_movecs_sym_adapt:    2 vectors were symmetry contaminated
  Symmetry fudging
 d= 0,ls=0.0,diis     7   -376.9738213538 -1.07D-06  5.21D-06  2.43D-07   675.0
 !! scf_movecs_sym_adapt:    2 vectors were symmetry contaminated
  Symmetry fudging
 d= 0,ls=0.0,diis     8   -376.9738215016 -1.48D-07  2.58D-06  6.36D-08   775.6

And TDDFT with DF-Weigend Coulomb Fitting (after applying the fix only) is:
   convergence    iter        energy       DeltaE   RMS-Dens  Diis-err    time
 ---------------- ----- ----------------- --------- --------- ---------  ------
 d= 0,ls=0.0,diis     1   -376.8534828617 -8.36D+02  7.46D-03  2.05D+00    89.9
 !! scf_movecs_sym_adapt:  108 vectors were symmetry contaminated
  Symmetry fudging
 !! scf_movecs_sym_adapt:    7 vectors were symmetry contaminated
 d= 0,ls=0.0,diis     2   -376.9696746812 -1.16D-01  1.07D-03  2.42D-02    98.3
 !! scf_movecs_sym_adapt:  106 vectors were symmetry contaminated
  Symmetry fudging
 !! scf_movecs_sym_adapt:    4 vectors were symmetry contaminated
 d= 0,ls=0.0,diis     3   -376.9732008776 -3.53D-03  4.83D-04  7.41D-03   106.1
 !! scf_movecs_sym_adapt:   48 vectors were symmetry contaminated
  Symmetry fudging
 !! scf_movecs_sym_adapt:    4 vectors were symmetry contaminated
 d= 0,ls=0.0,diis     4   -376.9740934347 -8.93D-04  1.88D-04  6.74D-04   115.0
 !! scf_movecs_sym_adapt:   13 vectors were symmetry contaminated
  Symmetry fudging
 d= 0,ls=0.0,diis     5   -376.9741727345 -7.93D-05  6.55D-05  1.89D-04   124.0
  Resetting Diis
 d= 0,ls=0.0,diis     6   -376.9741992716 -2.65D-05  1.71D-05  1.59D-06   132.9
 !! scf_movecs_sym_adapt:    2 vectors were symmetry contaminated
  Symmetry fudging
 d= 0,ls=0.0,diis     7   -376.9742003410 -1.07D-06  5.09D-06  2.36D-07   142.2
 !! scf_movecs_sym_adapt:    2 vectors were symmetry contaminated
  Symmetry fudging
 d= 0,ls=0.0,diis     8   -376.9742004805 -1.39D-07  2.42D-06  6.22D-08   151.2


Why symmetry is "fudging" in two last cases?

P.S. It is observed on Nwchem-dev.revision26871-src.2015-02-24 and Nwchem-dev.revision26971-src.2015-04-01 with fix
Edited On 12:19:59 AM PDT - Thu, Apr 9th 2015 by Vladimir


Forum >> NWChem's corner >> NWChem functionality



Who's here now Members 0 Guests 1 Bots/Crawler 0

AWC's: 2.5.10 MediaWiki - Stand Alone Forum Extension
Forum theme style by: AWC
Retrieved from "http://www.nwchem-sw.org/index.php/Special:AWCforum/st/id1636/density_fitting_in_TDDFT_didn%27t_....html"