Pressure scaling more than 1%. This may mean your system is not yet equilibrated

GROMACS version: 2025.2
I am running NPT simulations for the TIP4P/2005 water model, where Tcoupl=Nose-Hoover (tau_t=1.0ps), Pcoupl=Parrinello-Rahman (tau_p=2.0ps), and a time step of dt=1fs. However, when using the new GROMACS version (2025.2), the system quickly becomes unstable and crashes, accompanied by the warning: “Step 147001 Pressure scaling more than 1%. This may mean your system is not yet equilibrated. Use of Parrinello-Rahman pressure coupling during equilibration can lead to simulation instability, and is discouraged”. In contrast, the simulation runs normally with the older version (2022.1).

The initial configurations I have chosen have already been equilibriated. I have tested multiple versions and found that this issue occurs starting from GROMACS 2023 onward, but I don’t konw what the reason is. Below are the simulation results obtained using versions 2025.2 and 2022.1:

Hi @YGoing ,

a possible issue that comes to my mind is the change of the default values for nstpcouple. In GMX 2022.1 the default value is 10, in GMX 2025.2 the default is 100. Take a look at this thread for more information.

Best,

Marius

Indeed. I just found out that you need about 500 integration steps per period for the Parrinello-Rahman barostat. So that would mean nstpcouple=2 for with tau_p=2 ps. Note that a period of 2 ps is very short. I am currently investigating this issue.

1 Like

Thank you! I set nstpcouple=10 and nsttcouple=10, the results are normal now!

Thank you! By setting the nstcouple value to be consistent with the old version, I’m now getting normal results

I opened an issue, with high priority: Too large nstpcouple value for Parrinello-Rahman barostat increases volume fluctuations (#5424) · Issues · GROMACS / GROMACS · GitLab

I did tests for the worst case: a small water box. I uploading a fix for the release 2025 branch. This gives nstpcouple=10 with tau_p=5.

1 Like