normal rcoulomb!=rvdw with verlet

  • ldarre
  • ldarre's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
10 years 3 weeks ago #3565 by ldarre
rcoulomb!=rvdw with verlet was created by ldarre
Hi, I am trying to run CGMD with MARTINI in gromacs 4.6.4 using verlet-pairlist with the recomendation in the example mdp of changing rvdw from 1.2 to 1.1, but grompp gives an error of unsupported rcoulomb different than rvdw. So, would you suggest to reduce rcoulomb to 1.1? has that been tested?

Thanks a lot

Please Log in or Create an account to join the conversation.

More
10 years 2 weeks ago #3573 by jaakko
Replied by jaakko on topic rcoulomb!=rvdw with verlet
Hi,

as far as I know there hasn't been any thorough testing of verlet pairlist scheme. I think switch is also not (yet) supported by the verlet pairlist which will affect things too with Martini.

Interested to hear your experience how it goes!

-Jaakko

Please Log in or Create an account to join the conversation.

More
10 years 2 weeks ago #3577 by djurre
Replied by djurre on topic rcoulomb!=rvdw with verlet
I've recently looked at the effects of using a pure cutoff (LJ) + Reaction-field (Q) with Martini. I've found that if you use the cutoff of 1.1 (instead of 1.2nm) for both LJ and Q, combined with the 'potential modifiers', most forcefield properties are unaltered or change only little. In addition relatively good energy conservation remains if you use a timestep no longer than 30 fs.

If the example mdp has 1.1 for LJ and 1.2 for Coulomb than that is probably a typo from me: both should be 1.1 nm. I'll try to change that asap. I'll hope to submit the results for this issue to some journal soon as well.

Please Log in or Create an account to join the conversation.

  • ldarre
  • ldarre's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
10 years 2 weeks ago #3578 by ldarre
Replied by ldarre on topic rcoulomb!=rvdw with verlet
Hi,
Thanks a lot for your replies.
I'll test then using rcoulomb=rvdw=1.1nm.

Regards

Leonardo

Please Log in or Create an account to join the conversation.

Time to create page: 0.115 seconds