lammps program stops without error

Hello,

I have installed lammps of latest version with REAX optional package on my newly built beowulf computer cluster with 2 nodes. When I run a NVT simulation for 20000 steps with timestep 1fs, it works well. But When I run the simulation for 150000 step, the simulation always stops at some a step like 84201 without giving any error message in the output file. It is strange that if I firstly ssh to my node2, when the simulation crashed, I was kicked out from the ssh shell with an error "...connection reset by peer... ", this didn't happen to my node1. And other program such as VASP run well on this cluster. I really do not know what is going on. Hope for help!

Best,

Han

Hello,

I have installed lammps of latest version with REAX optional package on my newly built beowulf computer cluster with 2 nodes. When I run a NVT simulation for 20000 steps with timestep 1fs, it works well. But When I run the simulation for 150000 step, the simulation always stops at some a step like 84201 without giving any error message in the output file. It is strange that if I firstly ssh to my node2, when the simulation crashed, I was kicked out from the ssh shell with an error "...connection reset by peer... ", this didn't happen to my node1. And other program such as VASP run well on this cluster. I really do not know what is going on. Hope for help!

you need to talk to the people running the machine that you are using.
this very likely looks like an issue with the setup there, not at all
like a LAMMPS problem.

axel.

I have no idea. I don’t think anyone can debug the problem
with this limited info.

Steve

hi,

i think if you have no error in your log file, it’s probably not really LAMMPS itself crashing, but rather the environment it runs in. you should see error messages in the terminal, though.

did you make sure your cluster is running properly?

best,

nikita