SegFault in MARTINI minimization
Summary
Reported by a user working through a MARTINI tutorial: https://gromacs.bioexcel.eu/t/segmentation-fault-core-dumped/6695
Exact steps to reproduce
Use martini3_worked
directory from http://cgmartini.nl/images/stories/workshop2021/tutorials/tutorial_2/tutorial_protein_part1.zip.
The only file needed: minimization-vac.tpr
$ gmx mdrun -s minimization-vac.tpr
:-) GROMACS - gmx mdrun, 2023.2-dev-20230622-644f53588d (-:
Executable: /home/aland/gromacs/build/clion-debug/bin/gmx
Data prefix: /home/aland/gromacs (source tree)
Working dir: /home/aland/martini-fail/tutorial_2/martini3_worked
Command line:
gmx mdrun -s minimization-vac.tpr
Back Off! I just backed up md.log to ./#md.log.1#
Reading file minimization-vac.tpr, VERSION 2018.6 (single precision)
Note: file tpx version 112, software tpx version 129
Update groups can not be used for this system because an incompatible virtual site type is used
Using 1 MPI thread
Using 8 OpenMP threads
Segmentation fault (core dumped)
Looks like gmx::constr_vsiten<(gmx::VSiteCalculatePosition)0, (gmx::VSiteCalculateVelocity)1>
is getting called with empty velocity array. So, it looks similar to #4774 (closed), but it reproduces even with the fix applied.
For developers: Why is this important?
Segfaulting is never good
If this is a bug, (1) what happens, and (2) what did you expect to happen?
- GROMACS segfaults
- GROMACS minimizes the system (supposedly, behavior with 2018) or displays an error if something relevant has changed since.
Relevant input files, logs and/or screenshots
The TPR is attached above.
Possible fixes
???
Edited by Andrey Alekseenko