[llvm] [SLP]Initial support for non-power-of-2 (but still whole register) number of elements in operands. (PR #107273)

[Alexey Bataev via llvm-commits]

alexey-bataev wrote:

> For the record, LGTM because I don't spot a functional problem with the code, not because I'm convinced this is always going to be profitable w/o cost model updates.

It is a conservative solution. Relies completely on the current cost model, which always rounds the number of elements up to power-of-2

https://github.com/llvm/llvm-project/pull/107273