"real-time" when it comes to kernels doesn't necessarily imply "fast" and Nanos only targets virtualized environments as well. RT is more for situations like "If I hit the brakes I want the car to stop". Nanos' sweet spot is more for cloud workloads.

Having said that with enough interest I think we could look at project such as https://projectacrn.org/ but it's not really a focus at the moment and would probably become a 'flavor'.

As for scheduling itself though, we recently added support for UMCG: https://nanovms.com/dev/tutorials/user-mode-concurrency-grou... .

RT linux kernels would probably have lower throughput.

With a chart showing 1.579019x the throughput, a more appropriate claim would be "60% faster than Linux".