A machine learning cluster is experiencing intermittent kernel panics under heavy load. How do you debug this at Advanced Micro Devices?
