This thread brings to mind an issue I’ve wondered about: For hard real-time systems, do people use specialized CPU’s and hardware? Caches can introduce hard-to-predict timing issues. I’m sure there are other aspects of modern hardware that present similar problems; large pipelines? Speculative instruction execution? Does Intel Time-Coordinated Computing address some of these hardware issues? Are there embedded CPU’s from TI, Intel, AMD, etc. that are specifically designed to support predictable and understandable “lockstep” hard real-time time behavior?