Newer versions of `mesa 22.3.x` generate "overlapping triangles" visualization problems
Many thanks for this software. I've seen newer versions of mesa
(versions greater than 22.3.x
) having some visualization issues in programs that have forever worked ok in mesa
. For instance, one of these programs is Visual Molecular Dynamics.
Visual Molecular Dynamics (VMD) is a free program that allows the visualization of molecules and proteins.
As with many other visualization programs, it relies on mesa
.
Running vmd
just implies running in terminal:
$ vmd file.pdb
where file.pdb
is just an ASCII file (typically with extension .pdb
(Protein Data Bank)),
which just contains the cartesian XYZ coordinates of all the atoms.
The current mesa
release 22.3.3
prevents vmd
from visualizing the structures correctly.
I've contacted VMD
developers and it seems there's a bug in the current mesa
release (below I show their views).
- If I downgrade to
Mesa 22.2.1
, the problem still exists. - If I downgrade to
Mesa 22.1.7
, the problem is solved.
Please let me show you a minimal working example (I'm running Linux).
Installation of VMD:
Is very straightforward:
Download it from here:
https://www.ks.uiuc.edu/Development/Download/download.cgi?PackageName=VMD
and install it:
https://www.ks.uiuc.edu/Research/vmd/current/ig/node6.html
Minimal working example:
Download the pdb
file from here: https://github.com/DavidCdeB/vmd-newcartoon-problem/blob/main/6eid.pdb
(Otherwise you can also find it here: https://www.rcsb.org/structure/6EID )
Run vmd
over the pdb structure:
$ vmd 6eid.pdb
In the VMD Main
window prompt, select: Graphics > Representations
In the Draw style
tab, select:
Coloring Method = Secondary Structure
and
Drawing Method = NewCartoon
or NewRibbons
It shows something like this - the visualization shows overlapping triangles, it is completely wrong...
I've tested it with different versions of mesa
:
- If I downgrade to Mesa 22.2.1, the problem still exists.
- If I downgrade to Mesa 22.1.7, the problem is solved.
I've contacted VMD and here is their official reply:
https://www.ks.uiuc.edu/Research/vmd/mailing_list/vmd-l/34066.html
which I have copied here:
The disconnected triangles being shown indicate some kind of problem
with interpretation of mesh connectivity of the triangle vertex arrays
being drawn. Even in the case that there was some kind of bad secondary
structure calculation result from STRIDE, the low-level VMD code that
generates triangle meshes for the NewRibbon and NewCartoon
representations has been reliable for more than 10 years now.
Since this issue has only arisen now, I have to conclude that
these recent Mesa drivers for your hardware are buggy.