| Hey Florian, Hello again Ulf,
Thanks, again, for a great answer :) I’ll be starting to modify the code soon enough by modifying the existing update functions of the VV.
Have a great week, Evyatar.
I think an explicit copy of the previous forces is much more clean than a "hackish" solution with artificial particles. If you need some inspiration, there are algorithms in the code that need the old positions or velocities of particles and make copies of those. You could do it the same way with the forces. Also I wanted to add that Espresso also has a one step integrator which can be called by integrate_sd, which does Borwnian or Stokesian Dynamics, albeit I'm not sure if this code is production quality. Also the integration code in integrate.cpp is already quit messy and will undergo major refactoring by me within the next two months or so. You have been warned. As to the benefits of a different thermostat / integrator I assume you have good reasons to believe that your simulations will benefit from that and that it is worthwhile to do that work. Technically speaking and without having looked at your equations I'd say that adding an other two-step integrator isn't that hard. This should work as you proposed by updating the existing update functions of the VV.
Cheers,
|