DPLOT errors, movecs question

From NWChem

Viewed 508 times, With a total of 3 Posts
Jump to: navigation, search

Clicked A Few Times
Threads 2
Posts 7
I am running a TDC calculation on benzene, trying to get the transition density on a grid of points.

My code executes successfully log file, and says it prints a .cube file, but no .cube file is produced (civecs and movecs files exist with the correct names).
However, while this silently fails on benzene, it succeeds on other small molecules like urea and alanine, with the exact same nw file.

	vectors benzene_TDC_29.movecs
	civecs benzene_TDC_29.civecs_singlet
	root 1
		-7.6	6.4	71
		-5.8	6.2	61
		-9.2	5.2	72
		output benzene_TDC_29_root1.cube

                                NWChem Input Module

  Limits (a.u.) specified for the density plot:

        From        To      # of spacings
X   -14.36192    12.09425        71
Y   -10.96041    11.71630        61
Z   -17.38548     9.82658        72

  Total number of grid points =                325872

  1-st set of MOs      : benzene_TDC_29.movecs
  CIVECS file     : benzene_TDC_29.civecs_singlet
  Output is written to : benzene_TDC_29_root2.cube
  Type of picture      : CHARGE DENSITY
  Format used          : Gaussian9x Cube
  Spin                 : TOTAL   
  The density is computed using all orbitals
  The density is computed on the specified grid
  Root:                     2
  *** tmom(2)***:  -0.14486905029922717     
  *** tmom(3)***:   0.24651580880316629     
  *** tmom(4)***:    5.5080313350135503E-002
  max element    6.0011393754093471E-002

  Sum of elements      =                        0.00230
  Integration volume   =                        0.05235
  Integrated Charge    =                        0.00012

 Task  times  cpu:        0.4s     wall:        1.8s

                                NWChem Input Module

 Summary of allocated global arrays
  No active global arrays

                         GA Statistics for process    0

Also, a separate question:

I wish to calculate the transition density cube manually, to compare with nwchem. To do this, I need to compute the explicit transition orbital function. The debug print command lets me get the coefficients to the atomic orbital functions, so I can use the published expressions for the basis functions to manually compute the AOs. I can then compute the MOs if I have the MOs in terms of the AOs, and then use the transition density matrix and the occupied->virtual transitions for each root to calculate the transition density.

However, I'm having a problem computing the MOs in terms of the AOs. Print debug in tddft prints MO vectors 6-33 of 36 in terms of the atomic orbitals, but doesn't print 1-5 or 34-36. I'm also having problems compiling mov2asc to get them from the .movecs file (undefined reference to ma_init__; libma.a file not produced when compiling tools). What is the best way to get the missing MOs?

                       DFT Final Molecular Orbital Analysis

 Vector    6  Occ=2.000000D+00  E=-9.932632D+00
              MO Center= -2.5D-01, -3.8D-01, -2.6D+00, r^2= 2.8D-02
   Bfn.  Coefficient  Atom+Function         Bfn.  Coefficient  Atom+Function  
  ----- ------------  ---------------      ----- ------------  ---------------
    19      0.991232   7 C  s         

 Vector    7  Occ=2.000000D+00  E=-7.975795D-01
              MO Center= -2.3D-01,  2.1D-01, -9.5D-01, r^2= 2.3D+00
   Bfn.  Coefficient  Atom+Function         Bfn.  Coefficient  Atom+Function  
  ----- ------------  ---------------      ----- ------------  ---------------
     3      0.314196   2 C  s                32      0.299189  11 C  s         
     8      0.294147   3 C  s                14      0.258416   5 C  s         
    26      0.240243   9 C  s                20      0.183374   7 C  s         

 Vector    8  Occ=2.000000D+00  E=-7.043076D-01
              MO Center= -2.4D-01,  1.1D-01, -1.2D+00, r^2= 3.2D+00
   Bfn.  Coefficient  Atom+Function         Bfn.  Coefficient  Atom+Function  
  ----- ------------  ---------------      ----- ------------  ---------------
    32     -0.362863  11 C  s                14      0.357271   5 C  s         
    26     -0.334599   9 C  s                 8      0.321964   3 C  s         

Gets Around
Threads 1
Posts 147
Using your input file (taken from the log file you provided), I was able to produce the desired cube files without issue. There is a slight bug in the released versions of NWChem in that DPLOT writes cube files to the working directory rather than the permanent or scratch directories (unless of course one or both of those coincide with the working directory).

To have NWChem print all of the orbitals to the output file, add

print "final vectors analysis"

to the DFT block.

Clicked A Few Times
Threads 2
Posts 7
print "final vectors analysis" works perfectly- thanks. It seems that print debug doesn't print MO vectors that are >=99% composed of a single atomic orbital.

However, a "find / -name benzene_TDC_29_root1.cube" fails to find any files, so it is not an issue of the file ending up in the wrong directory.
Executing the same nw code on other molecules like urea and alanine seems to intermittently succeed and fail for no apparent reason. One thing I have noticed is that the .db files are not produced in precisely the same runs where the .cube files are not produced.

Are there any other log files or nwchem configuration information I could post that might help?
Edited On 9:18:59 AM PDT - Wed, Jul 27th 2016 by Checksum

Gets Around
Threads 1
Posts 147
That is puzzling. Were there any issues during compilation?

If you use

print "final vectors"

in the DFT block instead, the code will print the matrix of all of the MO coefficients so you can obtain the small contributions that the orbital analysis output skips.

Forum >> NWChem's corner >> Running NWChem

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

AWC's: 2.5.10 MediaWiki - Stand Alone Forum Extension
Forum theme style by: AWC