Thermostat detects a temperature runaway condition, cannot proceed

[Jennyhu666]

Hi！

Thank for your great work!

The protein has about 300 amino acid residues. I prepared the protein-water molecular system with ambertools and the water box was set to 12.0. I get an error when I execute docker run with root.

DeviceStrategy: setting strategy to [large-molecule / combined]
Preprocessing step already done, skip...
run_preprocess takes 0.028343677520751953 seconds
Loading models...
Start pre-equilibration
Pre-equilibration with 10 eV/A² for 1000 steps
Step 0: Epot = -3279503.931eV Ekin = 1899.233eV Etot = -3277604.698eV
Thermostat detects a temperature runaway condition, cannot proceed.

In addition, I also want to ask if there is pmemd.MPI in the docker image? I modified the preprocess.py script to change sander to mpirun -np 24 --allow-run-as-root pmemd.MPI but it seems mpirun doesn't recognize pmemd.MPI

[ElwynWang]

For the parallel version of Amber calculation tools, you need to get your own license and then change the script accordingly. if you didn't do pre-equilibrium simulations and perform AI2BMD production runs directly, the issue may be happened. It indicted that your system was unstable and thus the simulation failed to begin (usually without sufficient pre-eq simulations or wrong atom orders in the PDB files)

[Jennyhu666]

My protein input file was able to successfully generate topology files with tleap. Does that mean there's nothing wrong with the atomic order?

I can run pmemd.MPI on my local machine. How can I run it on docker? Or can I use the NVT and NPT output files that are already running on the local machine as the input files for AI2BMD?