Thank you @akohlmey
I understand. At the end of the day it’s just a matter of convenience: for analysis purposes, I am more interested in the content than in the format. Having the output in a Paraview-readable format would just allow me to bypass writing a Python script to convert from one format to another, when I want to visualise it.
I was wondering if dump grid/vtk would work with 3D density grid. I tried running with
fix myDensFix all ave/grid 100 10 1000 100 100 1 density/mass
instead, but I don’t know how to define the grid to pass to dump grid/vtk maybe I should use a compute instead? The only discussion I found on the topic is: https://matsci.org/t/problem-using-fix-ave-grid-in-lammps/48011, and when I run:
compute myGrid all property/grid 100 100 1 id ix iy iz
fix myDensFix all ave/grid 100 10 1000 100 100 1 density/mass c_myGrid:grid:data[*]
I also get Fix ave/grid cannot operate on per-atom and per-grid values which I don’t understand since it should indeed not operate on per-atom values, only grid (right?).