[llvm] Add `llvm.vector.partial.reduce.fadd` intrinsic (PR #159776)

[Paul Walker via llvm-commits]

https://github.com/paulwalker-arm approved this pull request.

A couple of minor suggestions but otherwise this looks good to me.

We might need to extend the intrinsic's assumed floating-point behaviour (although I still prefer my idea of forcing the use of the actual flags). That said, given the expected use case, and the fact the intrinsic is largely unusable beyond this use case it's something that can be tightened later as we see how things pan out.  We should make sure the details are nailed down before the next release though.

https://github.com/llvm/llvm-project/pull/159776