[llvm] [LLVM] Add `llvm.vector.reduce.fdot` intrinsic (PR #192174)

[Matt Arsenault via llvm-commits]

https://github.com/arsenm commented:

Why does this need a new intrinsic? You can just represent this as `fadd %acc, vector.reduce.fmul(%a, %b)`

https://github.com/llvm/llvm-project/pull/192174