Dear Amber Users,
I'm facing instability when continuing a trajectory using the Berendsen barostat on a GPU cluster.
System: POPC bilayer (288 lipids) in ~8000 choline geranate molecules (box size ~100 × 100 × 200 ų)
What I did:
Performed 5 ns NPT equilibration on CPU (local machine).
Continued for 1 ns on:
*
Local GPU (GTX TITAN, Amber 20) – completed successfully.
*
Cluster GPU (A100 MIG, Amber 24) – crashes almost immediately.
To rule out machine differences, I re-equilibrated the system on the cluster CPU and tried again – same issue.
All runs used the same starting structure and a fixed random seed (ig=100000).
Error on GPU:
ERROR: Calculation halted. Periodic box dimensions have changed too much from their initial values.
Your system density has likely changed by a large amount, probably from
starting the simulation from a structure a long way from equilibrium.
[Although this error can also occur if the simulation has blown up for some reason]
The GPU code does not automatically reorganize grid cells and thus you
will need to restart the calculation from the previous restart file.
This will generate new grid cells and allow the calculation to continue.
It may be necessary to repeat this restarting multiple times if your system
is a long way from an equilibrated density.
Alternatively you can run with the CPU code until the density has converged
and then switch back to the GPU code.
The simulation crashes within the first few steps. mdout shows:
*
Extreme pressure (+87,000 to –72,000 atm)
*
Overflow in temperature and energy values (*********)
This happens even when the restart comes from a well-equilibrated cluster CPU run.
Interestingly, switching to the Monte Carlo barostat avoids the issue completely - no crash on the same GPU.
How can I continue a trajectory using the Berendsen barostat on an A100 GPU, starting from a restart file generated at 100 ns on my local GPU?
Thanks in advance,
Senal
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Jun 29 2025 - 18:00:02 PDT
