a key performance optimization for Linux workloads that could be done in silicon is branch prediction for epoll_wait and io_uring_wait_cqe. prior art can be seen in IBM mainframes, and Apple's M1 objc_msgSend branch prediction.
Conversation
in the same way, branch prediction for WaitForMultipleObjects could be a significant performance improvement on Windows
1
8
i suspect we will see these kinds of optimizations in hardware *designed* to run Fuschia.
2
5
It shipped in production as a replacement for CastOS on the Nest Hub.
I think it's quite feasible that ChromeOS moves to using it from Linux in the next few years. It would still use Linux for the Android VM for a long time. I'm sure they eventually want it for mobile, etc. too.
They make it clear that it's not a replacement for ChromeOS and Android but it's quite clear that it does have the ambition of being deployed in those places. It likely wouldn't be shipped as a competitor to them but rather a transparent replacement without user-facing impact.