Dear gromacs users,
I am running a simulation of acetonitrile-water binary mixture and using the amber forcefield. The acetonitrile was parametrized using the generalized amber forcefield or gaff and I have used tip3p water model. After running the energy minimization, I am running 200 ns of equilibration MD with velocity rescale thermostat (tau-t = 0.1 ps) and Berendsen barostate (tau-p = 1 ps). Initial box length was 5.10 nm (cubic box), during NPT equilibration box length increased to 5.60 nm.
When I am trying to run the production MD with Parrinello-Rahman barostat (tau-p = 5 ps), then I am many encountering LINCS warning after a few seconds of starting the production MD simulation. I checked that the box length extends to a very high value of hundreds of nm.
However, when I am running the production MD with Berendsen barostat (tau-p = 5 ps), I am not encountering any LINCS warnings or errors and the simulation is running smoothly and box length is fluctuation around an average of 5.61 nm.
What is the reason for the system to blow up when I am using the Parrinello-Rahman barostat? Any help will be much appreciated