[clang] [CIR][AMDGPU] Add support for AMDGCN cos builtins (PR #197806)

[via cfe-commits]

================
@@ -0,0 +1,26 @@
+#include "../CodeGenCUDA/Inputs/cuda.h"
----------------
skc7 wrote:

For bf16 AMDGCN math builtins, upstream/OGCG only tests them in OpenCL, which is what the incubator mirrored. `builtins-amdgcn-gfx1250.hip` is a fine name for it. 

https://github.com/llvm/llvm-project/pull/197806