I’m running NetLogo models on a slurm HPC and am experiencing a performance issue that occurs consistently after approximately 4,500,000 ticks. The model starts to exhibit a noticeable and predictable slowdown. I have increased RAM from 30gb to 60gb per model and despite efforts to manage resources by clearing the shortest path cache every 100,000 ticks and handling data exports in batches, clearing lists at the same frequency.