[llvm] [AMDGPU] Support 'F' and 'A' conversion specifiers with ocl printf (PR #148784)

[Vikram Hegde via llvm-commits]

vikramRH wrote:

> I thought we were moving away from this implementation

That's my plan too, however this is needed urgently for correctness

https://github.com/llvm/llvm-project/pull/148784