EECS665 – Compiler Construction

Written work 9

Due on November 5th @ 3:00 PM (in class, to Drew, or at Engineering front office)

Not accepted late

Question 1

In class, we have used the two registers (the base pointer and the stack pointer) to track the boundaries of an activation record. A mysterious cloaked figure claims to have implemented a Levi compiler that only needs to use the stack pointer. Is this possible for Levi? How about for other languages?

Question 2

Assume that your nemesis has invented a language with the following aspects:

Users may only define 10 functions
Dynamic memory is allowed, but is always placed on the heap
Functions may not call themselves

Your nemesis claims that you will never run out of memory on a function call, and has therefore designed a runtime environment in which programs only have 100 memory slots. Describe at least 2 scenarios in which a function call will cause you to run out of memory. You may write pseudocode if you wish.

Question 3

Consider the following 3AC snippet. What code is generated from each quad to create X64 code?
Q1: [b] := 1
Q2: ifz [b] goto Q4
Q3: [b] := 0
Q4: nop

Question 4

We have discussed two memory regions that "grow" and "shrink" dynamically: the heap and the stack. What is the difference between how these regions are used?