Temperature of system not attained to desired set value

GROMACS version: 18.5
GROMACS modification: No
Here post your question
I have set temperature of my system to 305 K, during the NVT and NPT equilibration run, system attained 305 K, but during production run, system attained temperature of 314 K even though value is set at 305 K. Couldn’t figure out why this is happening. Please help.
md.mdp (2.2 KB)
npt.mdp (2.2 KB)
nvt.mdp (2.2 KB)


You mdp parameters seem reasonable.

What kind of system is this?

Thank you Berk Hess for the reply.

This is stack of three lipid bilayers one above other. Without any water. There are three types of molecules in bilayer. I have highlighted head atoms with red, to highlight bilayer boundary, it is not different molecule.

I assume you quickly get the incorrect temperature, right?

Could you check if you get the same issue with the 2021 version?

Yes, somewhat quickly, temperature increased within first 100 ps.

Here in figure, with blue color, I have shown last 40 ps of equilibration run (using npt.mdp) and then with continuation = yes in md.mdp, I am further extending the simulation.

Sure, I will try 2021 version, and get back to you.


You have a nasty resonance with a period of about 2.4 ps in your system. My guess is that this is due to the second order Parrinello-Rahman pressure coupling coupling to some vibrational mode in your system. You have tau_t set to 5 ps, but your compressibility will be off from the one of water, as you have a pure membrane system. So I guess your actual tau_t is around 2.4 ps. You can try tau_t is 10 or 20 ps and see if that removes the resonance. Even better would be using the c-rescale barostat that we added in the 2021 version of GROMACS.

Thanks Berk Hess for the solution, it worked!

I tested both suggestions, tried tau_p = 10 with PR pressure coupling and tau_p = 1 with c-rescale pressure coupling. In both cases, system temperature equilibrated around set temperature.

Thanks again