[clang] [CIR][AMDGPU] Add support for AMDGCN trig_preop builtins (PR #197399)

[Joseph Huber via cfe-commits]

jhuber6 wrote:

> @jhuber6 @arsenm This is one of many PR's that @ayokunle321 has put up to add support for AMDGCN builtins. They're pretty straightforward, so I've been approving them, but I wondered if one of you (or someone else you can recommend with a stake in this target) would want to take a look at what's happening?

These are likely just straight ports of the existing clang builtins, I don't think they require much AMD knowledge beyond checking that the IR looks the same.

https://github.com/llvm/llvm-project/pull/197399