Does anyone know if an operating system is coming out with this as a kernel level implementation. I think its hugely important with the next generation of multicore architectures which could have 16 cores easily to take the task of code vectorization off of the Application Programmers. Yeh for me it might not be that difficult to add a Thread here and there to spread out the work, but then again what happened to that big thing called Hardware Abstraction. I shouldn't have to worry about how many cores a computer has. By programming for the oldest platform, all my code should work on new OS models.

I've heard of some compilers that will try to vectorize your code, but I think this should be handled by the OS at image load time.