This is tracking issue to refactor float16 operations math functions currently implemented in LLVM libc to be header-only.
Target Functions:
- f16div
- f16div
- f16divf
- f16divf128
- f16divl
- f16mul
- f16mul
- f16mulf
- f16mulf128
- f16mull
- f16sub
- f16sub
- f16subf
- f16subf128
- f16subl
part of: #147386
@bassiounix