[llvm] [LoopVectorize] Don't scalarize predicated instruction with optsize (PR #129265)

[David Sherwood via llvm-commits]

https://github.com/david-arm commented:

I like this patch - it seems very sensible! However, as I mentioned in one of the comments I think it would be good to split the NFC parts out into their own patch because they are a useful clean-up by themselves.

Do you have any performance results for benchmarks (aside from the issue described in https://github.com/llvm/llvm-project/issues/66652) when building with -Os? It would be interesting to see if there is any impact. I imagine there might be some improvements too?

https://github.com/llvm/llvm-project/pull/129265