Gromacs 2023: Segmentation fault with gmx anaeig (but no problem with Gromacs 2022.4)

User discussions
1

GROMACS version: 2023
GROMACS modification: No
Here post your question
I appreciate your help. Thanks in advance.
I get error with Gromacs 2023. However, I get the command done successfully in previous Gromacs versions (including 2022.4) for the same system. Files are added (link provided).

echo C-alpha C-alpha | gmx anaeig -s md_sim.tpr -f md_sim_1.xtc -proj -first 1 -last 6
:-) GROMACS - gmx anaeig, 2023 (-:

Executable: /mnt/l/Software/Gromacs-2023/bin/gmx
Data prefix: /mnt/l/Software/Gromacs-2023
Working dir: /mnt/c/Users/USER/Downloads/anaeig-test
Command line:
gmx anaeig -s md_sim.tpr -f md_sim_1.xtc -proj -first 1 -last 6

Segmentation fault

Link for files:
Gromacs-2023-gmx-anaeig-segmentation-fault

2

Hi! Thank you for reporting.

That’s indeed a bug in GROMACS 2023. I opened an issue to track it, Segfaults due to calling fn2ftp(nullptr) (#4756) · Issues · GROMACS / GROMACS · GitLab.

We’ll fix it in 2023.1 or 2023.2. I expect you don’t need an urgent workaround and can just use 2022.x for your analysis?

3

Thanks a lot.
I was feeling very happy with Gromacs 2023!
I am getting 2X simulation speed (~300 ns/day for the system I mentioned) compared to Gromacs 2022.4.
I will wait for a fix.

4

As a workaround, you can add -3d temp3d.gro to your command. It will cause extra work, especially for long trajectories, but it should not crash.

5

Thank you very much!
I am asking for another help. What if I want 2d projection of vector1 and vector2?
I was using the following command:
echo C-alpha C-alpha | gmx anaeig -s md_sim.tpr -f md_sim_fit.xtc -2d -first 1 -last 2