From: Strahs, Dr. Daniel Bernard (dstrahs_at_pace.edu)
Date: Fri Apr 15 2022 - 13:04:04 CDT
I'm recently experiencing a weird glitch, and I'm looking for some ideas to throw at it.
I'm currently using an older NAMD: 2.7. Sometimes, when the system restarts a run (either at the conclusion of a 1-ns segment, or restart after a fatal error), the simulation will restart at an earlier structure.
I've checked the outputs: correct filenames are being used. I've tried to diagnose by using VMD to export the final snapshot of a trajectory and comparing those exported coordinates to the final coordinate file completed at the end of a 1-ns segment: the files match. So the structures being used in these erroneous simulations are not coming from the exported coordinates.
Is it possible that the restart files are being altered to an earlier state? The computer system has 12 nodes, 4 cpus/node: could a malfunctioning node cause this problem? Any thoughts/insights would be very welcome.
Thanks in advance!
This archive was generated by hypermail 2.1.6 : Tue Dec 13 2022 - 14:32:44 CST