[llvm] [TTI][Vectorize] Migrate masked/gather-scatter/strided/expand-compress costing (NFCI) (PR #165532)

[David Green via llvm-commits]

davemgreen wrote:

What is the benefit of this approach, over having separate functions? It would seem that adding intrinsic-specific information would be simpler to pass to individual cost functions if needed.

There is also getInterleavedMemoryOpCost that is conceptually similar but doesn't have an intrinsic at the moment. (Although I think it might be worth adding one).

https://github.com/llvm/llvm-project/pull/165532