Hello

The performance of the molecular dynamics package will not change considerably given that the simulation parameters do not change. In fact, the 100 ns simulation is huge, and I doubt if any home computer can carry that out. Depending on your system size it may take days of course.

Hi Ching Feng Weng

Assuming your potential parameters are finalized and you don't have access to additional computational resources, there are a few things you can try to do (though it will only provide some speed-up).

1) If you are sure that you're dynamics are slow enough for your particles to not move beyond the skin radius for one timestep, then you can reduce the frequency of building the neighbor list -- build it once in a few steps instead of every step. You can increase the skin radius (slight increase in computational cost) to decrease the frequency of the neighbor list build-up further (greater decrease in computational expense). Note: Your cutoff+skin should less than half the box size.

2) If you have a shorter simulation to test how significant your long range interactions -- assuming they aren't you can always reduce your cut-off radius (though in most cases this is not a good solution).

Hi

this range of time is not too huge for simulation, because some coarse-grained simulations can be calculated in nano to micro seconds range of time, but it depends on your system that has how many cores and its CPU or other properties. If you want to use such period of time, super computer source with 24 cores or more should be used.

Increase the time step. And if your package allows, you can divide the calculation into several shorter ones, using commands to save the state of the system (coordinates, particle velocities), that is, your calculation can be done in stages.

Hy

Its totally depends upon you that how you wanted to run the simulation the fastest and easiest one is Gromacs totally based on command where you can change the md.mdp file as per your need like how much time u wanted to run the simulation