[llvm] AMDGPU/GlobalISel: Start legalizing minimumnum and maximumnum (PR #140900)

[Tim Gymnich via llvm-commits]

tgymnich wrote:

fyi: I still have the SDAG implementation for `G_FMINIMUM` and `G_FMAXIMUM` ported to GISel laying arround. Just waiting on `GISelValueTracking::computeKnownFPClass`

https://github.com/llvm/llvm-project/pull/140900