Looking for vectorize support for C++ backend

[xxr3376]

I found that C++ backend doesn't support vectorize annotation as WASM backend does. In this case C++ backend emits the same output no matter vectorized or not.

There's even no vectorize support in python TUI, which is very different from the report in the paper.

Is there any difficulty in emitting vectorized code as WASM backend do? Or do I miss something?

P.S. loop tool is brilliant! It's really fun to optimize code like a video game!