Velocity distribution and

User discussions
1

GROMACS version: 2023.2
GROMACS modification: No

Hi,

I wanted to get more familiar with the GROMACS code base and decided to check the velocity generation code and noticed a couple of things. I hope it’s ok if I ask a few questions.

  1. It considers all atoms as adding up 3 DoF, but if holonomic constraints are placed, some may not count toward the total DoF. Hence, the subsequent scaling of the velocities may not give the correct temperature.
  2. The center of mass rot+transl velocities are then subtracted from the velocities of the particles; I’m guessing the impact on the temperature and the kinetic energy should be small.

In both scenarios, does the code expect the thermostat to fix those deviations in the first steps?

And last question: after the first step and the application of holonomic constraints, will the resulting velocities still be Maxwellian distributed?

Thanks for any info