Acceleration

Hi everyone,
I hope you are doing well.
I am applying an external force using acceleration. The way I calculated force was F=ma and m was Nc * Mw. Subsequently, P is ma/(LyLz). Note there is a unit conversion as well which I did not mention but overall it is a simple calculation.
For better clarification, here is a brief on system geometry:
Ly = 3.1613 nm
Lz = 5.05434 nm
Nc = 624
Mw = 12.0107 g/mol
I get the pressure for an acceleration equaling to 1 nm/ps^2, and then I use it to get to my desired pressure.

In my simulation setup, I used these parameters as below:
freezegrps = P2 P1
freezedim = N Y Y N Y Y
acc-grps = P1 P2
accelerate = 0.06419455 0 0 -0.00013009 0 0

0.06419455 and 0.00013009 are equivalent to 50 MPa and 0.1 MPa, respectively.
I read through posts here, that acceleration was going to be abandoned in 2021-2022. Then, I noticed, it is not abandoned at all in latest versions of gromacs, such as 2023.5.
I have done two types of simulations using the same setup; one with the gromacs 2016.4, and the other with gromacs 2023.5. There is a noticeable difference between them. basically, the latter resulted in much more filtered water (~350), whereas the former led to ~80 water molecules.
So, I checked my simulation log file, and noticed both of them have one thing in common:
acc: 0.0617287 0 0 -0.00259593 0 0 -0.00246584 0 0
nfreeze:N Y Y N Y Y N N N

First thing that is questionable here is that the first component of the first two acc sets is not the same as what I put in the mdp file. Could anyone please help me with this to understand better?
The next mysterious thing is that what is this -0.00246584 0 0? Is it for the other groups which are not under acceleration and freezing? If yes, why it is negative? I mean it should not be negative because the right hand of the system is under 50 MPa, while the other side is under 0.1 MPa.

My guess is that you have center of mass motion removal on (which is default). Thus to keep the COM stationary, the COM needs to be accelerated by a force equal to the sum of the forces on your two groups and with opposite sign. This is what you see. If you don’t want this you need to turn off center of mass motion removal.

Hi,
thanks for the hint.
Indeed when comm-mode is none, then I see the exact acceleration I applied on my graphene sheets.