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

[Mel Chen via llvm-commits]

Mel-Chen wrote:

@ayalz @fhahn Ping.
Could you clearly specify the shortcomings of this PR that prevent it from being approved?
For example, do I need to complete the FindLast reduction and not only handle the induction variable, or any other reasons? This could help speed up the progress of this PR. Thank you.

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