[llvm] [LLVM][CodeGen][SVE] Lower bfloat ISD::VECREDUCE_SEQ_FADD operations. (PR #199732)

[Paul Walker via llvm-commits]

paulwalker-arm wrote:

Thanks @davemgreen and sorry for wasting your time @sdesmalen-arm. I've updated the cost model so scalable vector ordered bfloat reductions now return an invalid cost.

https://github.com/llvm/llvm-project/pull/199732