Re: [ESPResSo-users] Rigid Bond

[Rudolf Weeber]

Hi,
On Mon, Jan 22, 2018 at 01:13:00AM +0330, Amir Khosravani wrote:
> Dear ESPResSo users,
> I want to simulate a polymer with nearly constant length. The polymer
> experiences some large forces so I need almost rigid bonds to simulate it.
> The first idea is using a harmonic bond with large stiffness, and the
> second is using a combination of Weeks-Chandler-Andersen potential and FENE
> with large stiffness. But when I was reading the user guide documentation I
> founded there is a rigid bond potential in Espresso. My question is about 3
> parameters which this potential need. Can you tell me a little bit about
> this potential, parameters, and efficiency of that? Is this potential
> faster than those two initial ideas? Because the speed of simulation is
> important in my work. It would be my pleasure if you have any other
> suggestion?
With regards to the rigid bond, see
http://espressomd.org/html/doc/inter.html#rigid-bonds
Positional tolerance applies to the length of the distance vector between bound 
particles, velocity tolerance to the projection of the velocity difference onto 
the distance vector between the bound particles.
For an application to a chain see the testcase:
https://github.com/espressomd/espresso/blob/python/testsuite/rigid_bond.py

Whether harmonic/fene bonds or rigid bonds via rattle are faster, you probably 
need to measure. The first needs a lower time step for the bonds not to break, 
the latter requires extra calculations to maintain the constraints on the bond 
vectors.

When performance is of importance, the most important things to do is using a 
minimal set of features in myconfig.hpp and tuning the skin via
system.cell_system.tune_skin()
http://espressomd.org/html/doc/espressomd.html#espressomd.cellsystem.CellSystem.tune_skin

Regards, Rudolf


--
Dr. Rudolf Weeber
Institute for Computational Physics
Universität Stuttgart
Allmandring 3
70569 Stuttgart
Germany
Phone: +49(0)711/685-67717
Email: address@hidden
http://www.icp.uni-stuttgart.de/~icp/Rudolf_Weeber