Interface for swapping inputs and outputs based on cost model + working permute example (#18929)

Summary:

I have seen the following pattern quite a bit so far where we have

wrapper(x0), wrapper(x1), x2 -> op -> wrapper(o0), o1

for example, and it is semantically equivalent to instead run
x0, x1, wrapper(x2) -> op -> o0, wrapper(o1)

In case 1, we had 3 wrapper ops, and in case 2, we have 2, so it's better to run 2. This interface formalizes this idea. I also implemented the pass which uses this interface for the permute case.

What is really nice about this interface is that in a single pass, we can do a LOT of cleanup. See the docs in the diff for a case for adds.

Differential Revision: D100917820

Author: Andrew Grebenisan