Issue with internal unwrapping procedure during runtime

Hello, Recently I encountered an issue related to the internal unwrapping procedure during a simulation run. I noticed that a trajectory collected using dump_exyz or dump_xyz and then unwrapped with OVITO (or another tool) differs from the one obtained with dump_xyz -1 -1 100 dump.xyz unwrapped_position. From a physical standpoint, I expected no diffusion of Br atoms during the simulation. This is indeed the case when I check the OVITO-unwrapped trajectory. However, when I check the GPUMD-unwrapped trajectory, I observe significant migration of Br atoms (which is not due to box drift). dump_gpumd.xyz.txt – GPUMD-unwrapped trajectory dump_ovito.xyz.txt – OVITO-unwrapped trajectory dump_wrapped.xyz.txt – wrapped trajectory The full trajectory is quite large (17 GB, if it is necessary, we can manage how to send all the data), so I am attaching only every 1000th step. A difference in the positions of the lithium atoms in the unwrapped ovito/gpumd trajectories is also observed, although this difference is not as obvious due to the expected lithium diffusion. For a first look, a snapshot of the final step, taken from the same perspective, is included. As a result of this behavior in the unwrapped positions, compute_msd also yields unreliable results. <img width="1907" height="975" alt="Image" src="https://github.com/user-attachments/assets/e491bc29-b5da-450b-a59c-34e0162d3cf7" /> <img width="1882" height="978" alt="Image" src="https://github.com/user-attachments/assets/edabfc41-103a-4fba-9c2e-38f5e62d767a" /> [dump_gpumd.xyz.txt](https://github.com/user-attachments/files/27172014/dump_gpumd.xyz.txt) [dump_ovito.xyz.txt](https://github.com/user-attachments/files/27172015/dump_ovito.xyz.txt) [dump_wrapped.xyz.txt](https://github.com/user-attachments/files/27172013/dump_wrapped.xyz.txt) [run.in.txt](https://github.com/user-attachments/files/27172206/run.in.txt)