Dear All, I am writing to seek your expertise in a computational chemistry problem that I have been encountering. I am using the cp2k.popt version of the cp2k program and the Multiwfn software to generate an input file (cp2k.inp) for a molecular dynamics (MD) simulation. The simulation is performed using the NPT_F ensemble and is run on a cluster with 64 cores using mpirun. Unfortunately, the MD simulation stops updating after approximately 5 steps, while the CPU remains occupied. The cluster has dual AMD 7763 CPUs and 512 GB of memory, which should be sufficient for the simulation. However, I have observed that the simulation does not report any errors or warnings in the output file, and the memory usage does not exceed 40 GB. I have also attempted to troubleshoot the issue by testing different settings such as thermostats (Nose, CSVR), orbital transformation (OT) or diagonalization, including DFTD3 dispersion correction, and trying different pseudopotentials and basis sets, but these have not improved the simulation. I am unsure what the issue may be and would greatly appreciate any insight or advice you could provide on how to resolve this problem. Thank you for your time and consideration. Best regards，Peng Xiao -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to <a href="mailto:cp2k+unsubscribe@googlegroups.com">cp2k+unsubscribe@googlegroups.com</a>. To view this discussion on the web visit <a href="https://groups.google.com/d/msgid/cp2k/f743d44b-9b82-458b-9fce-62b4728001b9n%40googlegroups.com?utm_medium=email&utm_source=footer">https://groups.google.com/d/msgid/cp2k/f743d44b-9b82-458b-9fce-62b4728001b9n%40googlegroups.com</a>.