Segmentation fault due to Optimal PME grid and coulomb cut-off

GROMACS version: 2020.2
GROMACS modification: No

Hello everybody,

I am simulating TAPC (organic) molecules in a box, Energy minimization, NVT and NPT equilibiration have been successfully completed, during the md run I received the below messages and then segmentation fault. Can anyone help me what might be the reason or how to solve that ? (I have attached my md.mdp file for complete parameters).

starting mdrun ‘GRM in vacuum’
500000 steps, 250.0 ps.
step 600: timed with pme grid 108 104 96, coulomb cutoff 1.000: 13213.1 M-cycles
step 800: timed with pme grid 100 100 84, coulomb cutoff 1.042: 12038.8 M-cycles
step 1000: timed with pme grid 96 84 72, coulomb cutoff 1.228: 15073.9 M-cycles
step 1200: timed with pme grid 96 96 72, coulomb cutoff 1.177: 10662.1 M-cycles
step 1400: timed with pme grid 96 96 80, coulomb cutoff 1.085: 10134.8 M-cycles
step 1600: timed with pme grid 100 96 80, coulomb cutoff 1.074: 8645.1 M-cycles
step 1800: timed with pme grid 100 100 80, coulomb cutoff 1.059: 8587.2 M-cycles
step 2000: timed with pme grid 104 100 84, coulomb cutoff 1.031: 8858.5 M-cycles
step 2200: timed with pme grid 104 104 84, coulomb cutoff 1.009: 10645.4 M-cycles
optimal pme grid 100 100 80, coulomb cutoff 1.059
step 3600, will finish Fri Jul 10 05:20:51 2020Segmentation fault (core dumped)

Many Thanks,
Mohamedmd.mdp (2.7 KB)

Hi - segfaults should probably never happen. Can you file a bug report here:

and include all the files needed to run grompp? Or post them here and I’ll file the report for you.


Many thanks for your reply,

all the .mdp files I used are attached :)

The problem happened only in the production run using the md.mdp file

nvt.mdp (2.3 KB) npt.mdp (2.5 KB) minim.mdp (915 Bytes) md.mdp (2.7 KB)

Hi - I meant the files that go into the grompp command - you’ve given us the mdp, but not the .top file, any associated .itp files, the input .pdb or .gro, and an index file if you had one.

1 Like

I have added all the requsted files in the below link, please let me know if anything is still needed.


Thanks - I reproduced the crash, filed a bug here