Live variable analysis in Compiler Design - Mini Project: Build & Apply

Live Variable Analysis

📖 Scenario: You are working on a simple compiler optimization module. One important step is to find which variables are "live" at each point in a program. A variable is "live" if its current value might be used later before being overwritten.Understanding live variable analysis helps optimize memory and improve program speed by removing unnecessary variable storage.

🎯 Goal: Build a step-by-step live variable analysis for a small program represented as basic blocks with instructions. You will create data structures for the program, set up initial live sets, compute live variables using backward data flow analysis, and finalize the live variable sets for each block.

📋 What You'll Learn

Create a dictionary representing basic blocks with their instructions

Create initial empty live-in and live-out sets for each block

Implement the backward data flow equations to compute live-in and live-out sets

Update live-in and live-out sets until no changes occur (fixed point)

Output the final live-in and live-out sets for each block

💡 Why This Matters

🌍 Real World

Live variable analysis is used in compilers to optimize memory usage by identifying variables that are no longer needed and can be safely overwritten or removed.

💼 Career

Understanding live variable analysis is essential for compiler engineers, performance optimization specialists, and anyone working on low-level code analysis or static analysis tools.

1

Create the program basic blocks

Create a dictionary called basic_blocks with these exact entries:
'B1': ['a = 1', 'b = 2', 'c = a + b'],
'B2': ['b = c + 1', 'd = b + 2'],
'B3': ['e = d + a']

Compiler Design

# Create the dictionary basic_blocks with the given blocks and instructions
# Your code here

Need a hint?

Use a dictionary with keys 'B1', 'B2', 'B3' and lists of strings as values.

2

Initialize live-in and live-out sets

Create two dictionaries called live_in and live_out with keys 'B1', 'B2', 'B3'. Initialize each value as an empty set set().

Compiler Design

basic_blocks = {
    'B1': ['a = 1', 'b = 2', 'c = a + b'],
    'B2': ['b = c + 1', 'd = b + 2'],
    'B3': ['e = d + a']
}
# Initialize live_in and live_out dictionaries with empty sets for each block
# Your code here

Need a hint?

Use dictionary literals with keys 'B1', 'B2', 'B3' and values as empty sets.

3

Implement live variable analysis core logic

Write a function called analyze_live_variables that takes basic_blocks, live_in, and live_out as arguments. Inside, implement a loop that updates live-in and live-out sets for each block until no changes occur. Use these rules:
- live_out[block] is the union of live_in of successor blocks
- live_in[block] is use[block] union (live_out[block] minus def[block])
Assume the control flow graph successors are: {'B1': ['B2'], 'B2': ['B3'], 'B3': []}.
For simplicity, define use and def sets for each block as:
use = {'B1': {'a', 'b'}, 'B2': {'c', 'b'}, 'B3': {'d', 'a'}}
def = {'B1': {'a', 'b', 'c'}, 'B2': {'b', 'd'}, 'B3': {'e'}}

Compiler Design

basic_blocks = {
    'B1': ['a = 1', 'b = 2', 'c = a + b'],
    'B2': ['b = c + 1', 'd = b + 2'],
    'B3': ['e = d + a']
}
live_in = {'B1': set(), 'B2': set(), 'B3': set()}
live_out = {'B1': set(), 'B2': set(), 'B3': set()}
# Define analyze_live_variables function and implement the fixed-point iteration
# Your code here

Need a hint?

Use a while loop with a changed flag. Update live_out by union of successors' live_in. Update live_in by use union (live_out minus def).

4

Run analysis and finalize live variable sets

Call the function analyze_live_variables with basic_blocks, live_in, and live_out. Then add a dictionary called final_live that maps each block to a tuple of (live_in[block], live_out[block]).

Compiler Design

basic_blocks = {
    'B1': ['a = 1', 'b = 2', 'c = a + b'],
    'B2': ['b = c + 1', 'd = b + 2'],
    'B3': ['e = d + a']
}
live_in = {'B1': set(), 'B2': set(), 'B3': set()}
live_out = {'B1': set(), 'B2': set(), 'B3': set()}

successors = {'B1': ['B2'], 'B2': ['B3'], 'B3': []}
use = {'B1': {'a', 'b'}, 'B2': {'c', 'b'}, 'B3': {'d', 'a'}}
defs = {'B1': {'a', 'b', 'c'}, 'B2': {'b', 'd'}, 'B3': {'e'}}

def analyze_live_variables(basic_blocks, live_in, live_out):
    changed = True
    while changed:
        changed = False
        for block in basic_blocks:
            old_live_in = live_in[block].copy()
            old_live_out = live_out[block].copy()
            live_out[block] = set()
            for succ in successors[block]:
                live_out[block] |= live_in[succ]
            live_in[block] = use[block] | (live_out[block] - defs[block])
            if live_in[block] != old_live_in or live_out[block] != old_live_out:
                changed = True

# Call analyze_live_variables and create final_live dictionary
# Your code here

Need a hint?

Call the analysis function with the dictionaries, then create final_live using a dictionary comprehension.