[llvm] [LoopVectorize] Vectorize select-cmp reduction pattern for increasing integer induction variable (PR #67812)

[Mel Chen via llvm-commits]

Mel-Chen wrote:

> Although IFindLastIV can be fixed to FFindLastIV when parsing instructions, could we reject IFindLastIV with FCmpInst and AddReductionVar for FFindLastIV explicitily?

This is possible, but it may cause going through the pattern identification process one more time.

https://github.com/llvm/llvm-project/pull/67812