[llvm] [NVPTX] Support llvm.{exp2, log2} for f32 and vector of f32 (PR #120519)

[Princeton Ferro via llvm-commits]

Prince781 wrote:

Updates:

- Added f16 and bf16 variants which promote to f32
- Support is **off** by default. User turns on with either `-nvptx-approx-{log2,exp2}f32` or `-enable-unsafe-fast-math`
- Added expected failure tests when support is not requested.

https://github.com/llvm/llvm-project/pull/120519