Bond and atom diam image question

I am using the 3 Nov 2022 version of LAMMPS, installed on various multicore machines. I see the same behavior described below on all of them. There are two issues I’d like to resolve with regard to image and movie dump commands that are somewhat related.

First, I find that LAMMPS apparently has some sort of minimum adiam value; if I set adiam below this value, the result no longer depends upon my adiam setting. I am trying to render polymers as thin fibers, but the the imaged results are too thick to see anything except the polymers in the foreground.

Second, I wanted to be sure that any rendering of the bonds is turned off. The online documentation for LAMMPS states that " Using none for both the bond color and width value will turn off the drawing of all bonds." If I use the following dump image syntax, LAMMPS runs without error:

dump            i1 all image 20000 *.jpg type type size 1000 1000 shiny 0.5

However, if I run with this line

dump            i1 all image 20000 *.jpg type type bond none none size 1000 1000 shiny 0.5

I get a very long set of messages (including a segmentation fault), beginning as follows:

[localhost:29250] *** Process received signal ***
[localhost:29247] *** Process received signal ***
[localhost:29248] *** Process received signal ***
[localhost:29248] Signal: Segmentation fault (11)
[localhost:29248] Signal code: Address not mapped (1)
[localhost:29248] Failing at address: 0x8
[localhost:29249] *** Process received signal ***
[localhost:29249] Signal: Segmentation fault (11)
[localhost:29249] Signal code: Address not mapped (1)
[localhost:29249] Failing at address: 0x8
[localhost:29250] Signal: Segmentation fault (11)
[localhost:29250] Signal code: Address not mapped (1)
[localhost:29250] Failing at address: 0x8
[localhost:29247] Signal: Segmentation fault (11)
[localhost:29247] Signal code: Address not mapped (1)

There is no LAMMPS error message.

I would like to resolve both of these issues. I cannot find an example that illustrates the correct bond parameter syntax.

Please construct a small, minimal input deck that can reproduce both issues, so it can be investigated.

I cannot reproduce the segfault with just adding your dump image command to an existing input.

If you don’t want to “see” certain atoms, you need to define a group that does not include them and use that for the dump image output. There is no minimum atom diameter, but if the diameter is such that it would be a fraction of a single pixel, then the output will be at least one pixel. To counter that effect, you would have to create a larger image and then apply some smoothed downscaling (i.e. an antialiasing-like procedure).