[llvm] [ConstantFolding] Fold vector reduce `fmax` and `fmin` intrinsics (PR #143904)

[Yingwei Zheng via llvm-commits]

dtcxzyw wrote:

> Can you please provide Alive2 proof to show minnum/maxnum are safe to reassociate?

It looks ok to reassociate the operands (ignoring SNaN handling): https://alive2.llvm.org/ce/z/8jan7w


https://github.com/llvm/llvm-project/pull/143904