Imagine that you are targeting an instruction set architecture that does not have an integer division or modulus operation.

- Write a function in C that takes two signed, 32-bit integers as arguments and returns a 2-element struct containing the quotient and remainder as 32-bit integer fields. For division by zero, return {0,0}.

- Compile your function to a target machine of your choosing and use profiling or debugging tools of your choice to determine the number of instructions executed in a single call to your function.

- If supporting a HW divider increased cycle time by 1%, CPI is exactly 1 for all non-DIV instructions, and DIV operations accounted for 5% of all instructions, how fast would a HW divider have to be to offer a better total program execution time than your SW division?