Dear Lammps users,
I am trying to perform the same simulations of a simple alkane in the force field GROMOS using GROMACS and LAMMPS. In Gromacs, I use the LINCS algorithm to fix constant lengths of chemical bonds between C-C atoms along the polymer chain. In LAMMPS, I would like to use the fix rigid procedure for fixing constant chemical bonds lengths. In your opinion, can I make the bonds between CC atoms “tougher” in the model I use in LAMMPS by applying procedure fix rigid in order to results could be compared with GROMACS? I wanted to clarify whether I understand correctly that in order to do this I need to create a group containing two atoms for each bond and separately carry out integration for each of these groups? If I have several thousand of such bonds, will I need to fix each bond as a separate group and integrate them separately or I can create a needed number of groups and use only one integrator?
Best regards,
Victor
You cannot. Rigid objects may not have overlapping atoms (and LAMMPS only supports adding 31 groups anyway), so that cannot work. Please check out fix poems from the POEMS package instead. There so-called joints, i.e. rigid groups that overlap by one atom are allowed. You would have to read the rigid grouping from a file in order to make this work.
An alternative to that would be using a large force constant for the bonds and run_style respa with appropriate settings, so that only the bonded interactions need to be computed frequently due to the required small (inner) time step.
Axel.
Thank you, Prof. Kohmeyer.
Victor
| Без вирусов. www.avast.ru |
сб, 4 июл. 2020 г. в 12:13, Axel Kohlmeyer <[email protected]>:
Given the algorithmic differences between LINCS in GROMACS and POEMS in LAMMPS, would not it be easier to just use SHAKE to fix the covalent bonds?
Regards
Tatiana Kuznetsova
Fix shake in LAMMPS cannot be used to make chains of bonds with more than 3 atoms rigid.