Simulation with a 3 fs timestep

GROMACS version: 2024.2-plumed_2.9.1
GROMACS modification: Yes/No
Here post your question

Hi all,

I have started using Gromacs in the past few weeks (starting MD as a general rule) to perform enhanced sampling on a flexible apo protein in water. Implementing the tips from this very informative page, I managed to improve my simulation speed from 17 ns / day to 90 ns/day which is awesome. However I wonder how to validate it, especially as I get a a stable run with a timesep of 3fs (without lincs warnings),which is not so much a typical timestep value I believe. To do so I had to fix “constraints= h-bonds”. "mass-repartition-factor = 3’ in the mdp file (as stated in the page)

I want to check if my physics is still good. I am planning to run a NVE and check the energy drift. I also wanted to check the stability of the temperature and pressure as I am doing a NPT simulation. (with the md integrator, V-rescale, Parrinello-Rahman). Would you consider it a good validation protocol or would you recommend something else ?

Thank you very much for your insights.
Have a great day !