Missing atoms in the .cube files

[alexvakimov]

Describe the bug
Atoms are missing in the .cube file generated by the PSPW code

Orbital isosurface seem to be represented correctly though

Also, when periodic images are added in VMD, atoms seem to be translated to the nearly-identical positions (overlaps) - see one of the screenshots.

Describe settings used
NWChem v7.2.2, Linux operating system
Please see the attached "out" file for more details

Attach log files
please see attached
files.zip

To Reproduce

1. sbatch submit.slm
2. visualize with VMD
   The SLURM submit.slm file and the NWChem input file input.nw are attached.

Expected behavior
All atoms of the unit cell of NaCl are expected to appear, but some are missing;

[alexvakimov]

A bit more analysis into this - checking that the lat_a, lat_b, lat_c might be in a.u. by default (as suggested by @edoapra in email), so setting them all to 10.658 a.u. = 5.64 Angstrom, leads to a pretty expanded structure. I use the VDW rendering in VMD, so expected to have a much more packed structure in this case. The box (as defined by the LimitXYZ) is the [0, 5.64] x[0, 5.64] x[0, 5.64] Angstrom (at least that's what I thought) - see the picture 1.

So, trying the opposite - assume the lat_a, lat_b, and lat_c be actually in Angstrom, but LimitXYZ in a.u. - this leads to better results - the VDW spheres touch and the box is now of the appropriate shape:

also, when the unit cell (both atoms and orbitals) are replicated in x,y,z directions in VMD - the picture makes sense now:

I think the number of atoms printed in the .cube file is reduced to prevent atom overlaps when translating them along the unit cell vectors, so in this regard it is not a bug or a problem.

However, it seems the LimitXYZ sub-section of the "DPLOT" input section is in a.u. by default, not in the Angstrom as the website suggests:

[ebylaska]

You probably need to adjust increase the position tolerance since you're using the limitxyz option

position tolerance = 0.100000E-02

I'll check

[ebylaska]

or possibly the ncell option, e.g.

ncell = 0 0 0

position tolerance = 0.100000E-02

       origin=<   0.000   0.000   0.000 >

[ebylaska]

The position_tolerance didn't fixe the problem. I'll need to debug the code. I should be able to work on this tomorrow. Sorry for the glitch(s).