Hi Sacha, I had similar problem at first, just to realise after that
in my grompp.mdp, settings.xml and index.ndx I have not specified
correctly non-bonded groups & correct names. Here is what I did to get
mines working, so mb good to double check yours are all fine too.
So tabulated bond and angle are specified in the topol.top file like
this:
[ bonds ]
1 2 8 1 1 ; corresponds to table_b1.xvg
1 6 8 1 2 ; corresponds to table_b2.xvg
1 5 8 1 2 ; corresponds to table_b2.xvg
2 3 8 1 2 ; corresponds to table_b2.xvg
2 4 8 1 2 ; corresponds to table_b2.xvg
[ angles ]
4 2 3 8 1 1 ; corresponds to table_a1.xvg
5 1 6 8 1 1 ; corresponds to table_a1.xvg
6 1 2 8 1 2 ; corresponds to table_a2.xvg
5 1 2 8 1 2 ; corresponds to table_a2.xvg
3 2 1 8 1 2 ; corresponds to table_a2.xvg
4 2 1 8 1 2 ; corresponds to table_a2.xvg
And non-bonded ones in grompp.mdp like this, here is my example, yours
seem to be simpler:
I have 2 types of energy groups B and M and I have 3 tables of the
interactions: B-B B-M and M-M that are tabulated as table_B_B.xvg
table_B_M.xvg table_M_M.xvg respectively.
energygrps = B M ; groups to write to energy
file
energygrp_table = B B B M M M ; corresponds
That should match what you have in settings.xml:
<non-bonded>
...
<inverse>
...
<gromacs>
<table>table_M_M.xvg</table> similar for B-M and B-B
</gromacs>
</inverse>
</non-bonded>
Additionally you will need a correct index.ndx file that you can make
by
$ make_ndx -f conf.gro
and when it asks you what groups (in my case) i do:
a B*
a M*
q
then I have to go into the index.ndx and manually remove * from [ B* ]
and [ M* ] to make them [ B ] and [ M ]. Mb there is other way around
this part, but thats what works for me.
May be this will help:)
V
On Oct 21, 1:00 pm, Sacha Mould <[email protected]> wrote:
> Hi all,
>
> i'm having problems with my iBI simulation results, the distribution
> function is not converging after 150 iterations:
>
> i'm trying to run coase-graining simulations of polyethylene using IBI
> method. The atomistic system consisted on 10 chain molecules with 50
> DP simulated with LAMMPS using the COMPASS forcefield from where i
> dumped the RDF, bond, and angle distributions. Either RDF and bond
> distribution obtained from atomistic simulation where similar in shape
> with the ones obtained for hexane in "Macromol. Theory Simul. 20, pp
> 472-477 (2011)". For the IBI simulation i want to map 5 monomers into
> a coarse-grained particle resulting on a system with 10 chains with 10
> beads per chain. I generated all the files required for the IBI
> simulation, including exclusions file, using Votca tools. Angle and
> bond interactions (table_a1.xvg and table_b1.xvg) where obtained by
> Boltzmann inverting the distribution functions and then using
> "csg_call convert_potential xvg" script. For non-bonded potential i
> let Votca to do the inversion after calling "csg_inverse". I'm
> assuming that all the particles are of the same type, which means that
> terminal and inner chain particles are equal and the terminal
> hydrogens on the methylenes are neglected. In the grompp file i set a
> time-step of 1 ps and a simulation time of 1ns using stochastic
> dynamics integrator at 450K. The cutoff where set to 1.5 nm. At this
> moment i'm not using any kind of pressure correction and after each
> iteration the distribution function is smoothed twice.
>
> With this setup, after running IBI the results don't converge, the
> distribution function that i'm obtaining at the end of each iteration
> show strange shapes and the convergence values are always in the range
> of 0.36 - 7.
> Other thing that concerns me is on how i should extrapolate to the
> right the bonded potential. The atosmistic results gives me a bonded
> distribution that goes to r=0.29 nm, the bond length of the coarse-
> grained system is much higher than this, so i need to extrapolate the
> bonded potential to higher values of r. I'm using quadratic function
> to extrapolate the table, but i'm not quite sure in which extent these
> extrapolated values are feasible.
>
> I wonder if someone can help me. Is anything wrong with my setup??? Is
> there something missing that i should account with??
>
> Thanks a lot in advance,
>
> Sacha
--
You received this message because you are subscribed to the Google Groups
"votca" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/votca?hl=en.
