[llvm] AMDGPU: Implement expansion for f64 exp (PR #182539)

[Jay Foad via llvm-commits]

https://github.com/jayfoad approved this pull request.

I don't really like these large floating point lowerings that we have to take on trust. But I guess this is no worse than lots of others that we already have in the backend.

OK with one more capitalization.

https://github.com/llvm/llvm-project/pull/182539