JIT: audit places where minopts and debug codegen diverge

[AndyAyersMS]

Historically we haven't gotten the same amount of usage (especially on 3rd party code) for minopts that we have on debug codegen. With the advent of tiering this is going to change.

We should proactively check the code base to see if there are any other cases like GT_INDEX_ADDR (see #11119) where the IR shapes or IR transformations differ between minopts and debug.

There are a few known areas where behavior will deliberately diverge -- eg debug will preserve NOPs and minopts will not. Other areas of divergence might be good focus areas for testing.

category:implementation
theme:minopts
skill-level:beginner
cost:small

[AndyAyersMS]

Suspect we won't want to perturb things for 3.0, so moving to future.