Conversion operations These built-in functions convert between vectors of single precision and bfloat16 types.