[lammps-users] Segmentation fault in terms of dihedral

LAMMPS LAMMPS Mailing List Mirror
1

Dear all,

Thank you very much for answering to my previous question, Axel.
I could solve it!
Now I got "segmentation fault" when I add Dihedral information.
I tried with just one dihedral types and dihedral and checked the
dihedral_coeff style again, but it seems it still doesn't work...
Would you tell me which part of input script I should check?
Following is my input script and data file.
Thank you very much.

Best regards,
Nanako Takahashi

2

This was a one-line bug in the charmm dihedral that didn't
detect you were doing something unusual: running bonds,
etc without a pair potential. The CHARMM dihedral
uses pair settings for 1-4 interactions in the dihedral.
You also had the weighting for these set to 0, so you
don't actually need a pair potential, but the code assumed
you had one.

I'll post a patch.

Steve