[llvm] [NFC][AMDGPU] Refactor wave reduce test files. (PR #170440)

[via llvm-commits]

easyonaadit wrote:

> LGTM but could you add a description in the commit message? Something along the lines of _"Split wave-reduce floating-point from integer tests. Moved wave-reduce float add/sub/min/max from `llvm.amdgcn.reduce.add/sub/min/max` to `llvm.amdgcn.reduce.fadd/fsub/fmin/fmax`."_

Yupp sure, done.

https://github.com/llvm/llvm-project/pull/170440