RVV implementation

[camel-cdr]

Hi, I've written a shishua like implementation using the risc-v scalable vector isa: https://gist.github.com/camel-cdr/495937d851b39a85ac3be515e4dd132e

Would this be a thing worth including in this repo, and if so, do there need to be any changes?

It's written in assembly, because it uses a technique not exposed by the C intrinsics: Setting elements in a vector by accessing it through multiple lower LMUL registers. It seems to be possible but quite verbose.

Feedback in the implementation would also be appreciated.