Project

General

Profile

Feature #362692

wish list of integrators

Added by Felix Höfling over 2 years ago. Updated over 2 years ago.

Status:
New
Priority:
Low
Assignee:
-
Target version:
-
Start date:
Due date:
% Done:

0%


Description

Some integrators that would be nice to have in HALMD (with low priority). Generally, the integrators fall in two categories: 1st order (Langevin) equations and 2nd order (Newtonian) equations

  • Brownian dynamics (=overdamped Langevin equation), can be mimicked by abusing the Boltzmann thermostat with timestep-dependent masses ... Better: Euler-type integrator based on the velocities, forces enter via some mobility: Stokes (easy) or Rotne-Prager (difficult)
  • Brownian dynamics plus imposed flow field (for poor-man's shear simulations, e.g., for confined systems)
  • DPD/Lowe thermostat (conserves momentum locally)

Checklist

  • Brownian dynamics
  • DPD thermostat

History

#1 Updated by Nicolas Höft over 2 years ago

There is one Brownian Dymamics implementation:
https://github.com/jb--/halmd

But there would be a need for review.. I wonder if this can simplified by reusing the Boltzmann module

#2 Updated by Felix Höfling over 2 years ago

A benefit of a separate Brownian integrator over (ab)using the Boltzman
module is that the random velocities don't need to be stored to and read
back from global memory. Second, the masses keep their actual meaning
(timestep-dependent masses are weired).

Also available in: Atom PDF