Check if a Number is Even or Odd - Common Bit Manipulation Tricks | OneNoughtOne

0/361

Data Structures & AlgorithmsCommon Bit Manipulation Tricks

Common Bit Manipulation Tricks

LevelIntermediate

Duration75 mins

TopicCommon Bit Manipulation Tricks

1 / 5

Check if a Number is Even or Odd

The Simplest Yet Most Fundamental Bit Trick

Before we dive into sophisticated bit manipulation techniques, we must master the most elemental operation in the bit manipulation toolkit: determining whether a number is even or odd. While this might seem trivially simple—after all, we can use the modulo operator n % 2—the bitwise approach reveals the fundamental insight that makes all bit manipulation possible: integers are just sequences of bits, and those bits have meaning.

This page isn't just about replacing modulo with a faster operation. It's about developing the mental model that transforms how you see numbers. When you look at an integer and instantly visualize its binary representation, you've unlocked the key to dozens of optimization techniques that separate competent programmers from exceptional ones.

What You Will Learn

By the end of this page, you will understand why the last bit of any integer determines its parity, how to use the AND operator for this check, why this approach is faster than modulo, and how this simple technique forms the foundation for all bit manipulation patterns.

The Mathematical Foundation

To understand why the least significant bit (LSB) determines parity, we need to revisit how positional notation works in binary.

Every integer in base 2 is expressed as a sum of powers of 2:

$$n = b_k \cdot 2^k + b_{k-1} \cdot 2^{k-1} + \ldots + b_1 \cdot 2^1 + b_0 \cdot 2^0$$

Where each $b_i$ is either 0 or 1.

The critical insight: Every power of 2 greater than $2^0$ is even:

$2^1 = 2$ (even)
$2^2 = 4$ (even)
$2^3 = 8$ (even)
$2^k$ for $k \geq 1$ is always even

Only $2^0 = 1$ is odd.

Therefore: The sum of any combination of $2^1, 2^2, 2^3, \ldots$ will always be even (sum of even numbers is even). The only term that can make the total odd is $b_0 \cdot 2^0 = b_0$.

This means:

If $b_0 = 0$, the number is even
If $b_0 = 1$, the number is odd

The least significant bit alone determines whether a number is even or odd.

Binary Representation and Parity
Decimal	Binary	LSB (b₀)	Parity
0	0000	0	Even
1	0001	1	Odd
2	0010	0	Even
3	0011	1	Odd
4	0100	0	Even
5	0101	1	Odd
6	0110	0	Even
7	0111	1	Odd
8	1000	0	Even
15	1111	1	Odd
42	101010	0	Even
127	1111111	1	Odd

Visual Pattern

Look at the binary representations above. Notice the alternating pattern in the rightmost bit: 0, 1, 0, 1, 0, 1... This is no coincidence—it's the mathematical structure of counting in binary. Every increment toggles the LSB, and every other number has it set.

The Bitwise AND Solution

Now that we understand why the LSB determines parity, we need a way to extract just that bit. This is where the bitwise AND operator becomes our precision tool.

Recall how AND works:

1 AND 1 = 1
1 AND 0 = 0
0 AND 1 = 0
0 AND 0 = 0

AND only produces 1 when both inputs are 1.

The Masking Technique:

To extract the LSB, we AND the number with 1 (binary: 0001):

  n:       b_k b_{k-1} ... b_2 b_1 b_0
& 1:       0   0       ... 0   0   1
────────────────────────────────────────
Result:    0   0       ... 0   0   b_0

Every bit position except the last is ANDed with 0, producing 0. Only the last bit b_0 is ANDed with 1, preserving its value.

The result:

If b_0 = 1, the result is 1 (truthy in most languages)
If b_0 = 0, the result is 0 (falsy in most languages)

even_odd_check.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
def is_even(n: int) -> bool:
    """
    Check if a number is even using bitwise AND.
    
    The expression (n & 1) extracts the least significant bit.
    - If LSB is 0, the number is even → return True
    - If LSB is 1, the number is odd → return False
    
    Time Complexity: O(1) - single bitwise operation
    Space Complexity: O(1) - no additional memory
    """
    return (n & 1) == 0
 
 
def is_odd(n: int) -> bool:
    """
    Check if a number is odd using bitwise AND.
    
    The expression (n & 1) directly gives us the parity:
    - 1 means odd
    - 0 means even
    """
    return (n & 1) == 1
 
 
# Demonstration with various numbers
test_numbers = [0, 1, 2, 3, 4, 5, 42, -1, -2, -7, -8, 2**31, 2**31 - 1]
 
print("Number | Binary (last 8 bits) | n & 1 | Even? | Odd?")
print("-" * 60)
for num in test_numbers:
    binary = format(num & 0xFF, '08b')  # Show last 8 bits
    result = num & 1
    print(f"{num:6} | {binary}             | {result}     | {is_even(num)!s:5} | {is_odd(num)}")

Why Not Just Use Modulo?

You might wonder: why bother with n & 1 when n % 2 does the same thing and is more readable? This is a valid question, and the answer involves understanding what happens at the hardware level.

The Modulo Operation:

The modulo operator % computes the remainder after division. For n % 2:

The CPU must perform a division operation (or its optimized equivalent)
Division is one of the most expensive arithmetic operations
Even optimized division circuits take multiple clock cycles

The Bitwise AND Operation:

The & operator performs a parallel bit-by-bit comparison:

All bits are compared simultaneously in a single clock cycle
No division, no multi-step computation
It's literally reading one bit of input

Modulo Approach (n % 2)

•Division required — Even optimized, division is expensive
•Multiple clock cycles — Typically 10-40 cycles depending on architecture
•Negative number complexity — Different languages handle -7 % 2 differently
•Pipeline stalls — Division can stall the CPU pipeline
•Harder to optimize — Compiler must recognize the pattern first

Bitwise Approach (n & 1)

•Single operation — Direct bit extraction
•One clock cycle — Guaranteed O(1) in the truest sense
•Consistent behavior — Works identically for negative numbers (in two's complement)
•No pipeline impact — Trivial for modern CPUs
•Already optimal — No optimization needed

Modern Compiler Optimization

Modern optimizing compilers often convert n % 2 to n & 1 automatically. However, this optimization isn't guaranteed across all compilers, languages, or optimization levels. Using n & 1 explicitly ensures optimal code regardless of compiler behavior, and communicates your intent to other developers: 'I'm doing bit manipulation here.'

The Negative Number Consideration:

One subtle but important difference: modulo behavior varies for negative numbers.

Expression	Python	JavaScript	Java	C++
`-7 % 2`	`1`	`-1`	`-1`	`-1`
`-7 & 1`	`1`	`1`	`1`	`1`

In two's complement representation (used by virtually all modern systems), negative odd numbers still have their LSB set to 1. The bitwise approach gives consistent results across languages, while modulo behavior is language-dependent.

For parity checking, n & 1 is both faster and more predictable.

Deep Dive: Two's Complement Insight

To fully appreciate why n & 1 works for negative numbers, we need to understand two's complement representation—the near-universal method for representing signed integers in binary.

How Two's Complement Works:

For an n-bit signed integer:

Positive numbers are represented normally
To get the negative of a number: invert all bits and add 1

Example with 8-bit integers:

+7 = 00000111

Step 1: Invert all bits
       11111000

Step 2: Add 1
       11111001 = -7

Why the LSB Still Indicates Parity:

Let's trace through the logic:

If n is odd, it ends in 1
Inverting the bits, the LSB becomes 0
Adding 1 sets the LSB back to 1
Result: -n (for odd n) ends in 1
If n is even, it ends in 0
Inverting the bits, the LSB becomes 1
Adding 1 causes a carry, LSB becomes 0
Result: -n (for even n) ends in 0

The parity is preserved through negation!

Two's Complement and Parity (8-bit examples)
Decimal	Binary (8-bit)	LSB	Parity
+7	00000111	1	Odd
-7	11111001	1	Odd
+6	00000110	0	Even
-6	11111010	0	Even
+1	00000001	1	Odd
-1	11111111	1	Odd
+4	00000100	0	Even
-4	11111100	0	Even

Mathematical Proof

This isn't coincidence—it's mathematically guaranteed. The operation of negating in two's complement is equivalent to computing -n = ~n + 1. Since ~n flips the LSB and adding 1 involves a carry chain, the LSB of -n equals the LSB of n. The parity-preserving property is inherent to two's complement arithmetic.

Practical Applications

The even/odd check might seem basic, but it appears in numerous practical scenarios. Understanding when to apply it efficiently is part of developing strong algorithmic intuition.

Common Use Cases

•Alternating Patterns — Rendering stripe patterns in graphics, alternating row colors in tables, checkered patterns in game boards
•Index Parity in Algorithms — Many algorithms branch based on whether an index is even or odd (e.g., merge sort merge phase, certain matrix operations)
•Median Finding — When finding the median, behavior differs for even vs. odd length arrays
•Bit Counting Optimizations — Some bit counting algorithms process pairs of bits differently
•Cryptographic Operations — Certain cryptographic algorithms have parity-dependent operations
•Game Development — Turn-based games often use parity to determine which player's turn it is
•Memory Alignment — Checking if addresses are word-aligned often involves parity checks

practical_examples.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
# Example 1: Alternating pattern generation
def generate_checkerboard(size: int) -> list[list[str]]:
    """
    Generate a checkerboard pattern using bitwise parity check.
    Cell color depends on whether (row + col) is even or odd.
    """
    board = []
    for row in range(size):
        board_row = []
        for col in range(size):
            # (row + col) & 1 gives alternating pattern
            if (row + col) & 1:
                board_row.append("⬛")  # Odd sum → black
            else:
                board_row.append("⬜")  # Even sum → white
        board.append(board_row)
    return board
 
# Display 5x5 checkerboard
for row in generate_checkerboard(5):
    print(" ".join(row))
print()
 
 
# Example 2: Median calculation with parity check
def find_median(sorted_arr: list[int]) -> float:
    """
    Find the median of a sorted array.
    Behavior differs based on array length parity.
    """
    n = len(sorted_arr)
    mid = n // 2
    
    if n & 1:  # Odd length: single middle element
        return float(sorted_arr[mid])
    else:  # Even length: average of two middle elements
        return (sorted_arr[mid - 1] + sorted_arr[mid]) / 2.0
 
# Test
odd_arr = [1, 3, 5, 7, 9]
even_arr = [1, 3, 5, 7]
print(f"Median of {odd_arr}: {find_median(odd_arr)}")
print(f"Median of {even_arr}: {find_median(even_arr)}")
print()
 
 
# Example 3: Efficient modular exponentiation (preview)
def power_mod(base: int, exp: int, mod: int) -> int:
    """
    Compute (base^exp) % mod using binary exponentiation.
    Uses parity check to decide multiply vs square.
    """
    result = 1
    base = base % mod
    
    while exp > 0:
        if exp & 1:  # If exp is odd, multiply result
            result = (result * base) % mod
        exp >>= 1      # Divide exp by 2
        base = (base * base) % mod
    
    return result
 
# Test: 2^10 mod 1000 = 1024 mod 1000 = 24
print(f"2^10 mod 1000 = {power_mod(2, 10, 1000)}")
# Test: 3^13 mod 7 = 1594323 mod 7 = 3
print(f"3^13 mod 7 = {power_mod(3, 13, 7)}")

Common Pitfalls and Edge Cases

While the even/odd check is straightforward, there are subtle issues that can trip up developers. Understanding these edge cases deepens your mastery.

Watch Out For

•Zero is Even — Don't forget that 0 & 1 = 0, so zero is correctly identified as even. This is mathematically correct but sometimes surprises developers.
•Floating Point Numbers — Bitwise operations don't work on floats. In JavaScript, 3.7 & 1 gives 1 because the number is truncated to an integer first.
•Very Large Numbers — In languages with arbitrary precision integers (Python), bitwise operations work correctly. In languages with fixed-width integers (Java, C++), ensure you're using appropriate types.
•Boolean Trap — Don't confuse n & 1 (gives 0 or 1) with n && 1 (logical AND, different behavior).
•Operator Precedence — In expressions like n & 1 == 0, remember that == has higher precedence than & in many languages. Use parentheses: (n & 1) == 0.

edge_cases.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
# Edge Case 1: Zero is even
print(f"0 & 1 = {0 & 1}")  # Output: 0
print(f"Is 0 even? {(0 & 1) == 0}")  # Output: True
 
# Edge Case 2: Very large numbers (Python handles gracefully)
huge = 2**1000 + 1
print(f"(2^1000 + 1) is odd? {(huge & 1) == 1}")  # Output: True
 
huge_even = 2**1000
print(f"(2^1000) is even? {(huge_even & 1) == 0}")  # Output: True
 
# Edge Case 3: Operator precedence trap
n = 5
# WRONG: This evaluates as n & (1 == 0), which is n & False = 0
# Python doesn't have this issue due to its operator precedence,
# but C/C++/Java do!
# 
# In C/C++/Java, always write: (n & 1) == 0
# Not: n & 1 == 0
 
# Demonstration of correct usage
for num in [0, 1, 2, -1, -2, 2**100, 2**100 + 1]:
    is_even = (num & 1) == 0
    print(f"{num} is {'even' if is_even else 'odd'}")

JavaScript 32-bit Limitation

JavaScript converts numbers to 32-bit signed integers for bitwise operations. For numbers larger than 2³¹ - 1, results may be unexpected. Use BigInt for reliable bit operations on large numbers: (BigInt(n) & BigInt(1)) === BigInt(0).

Performance Benchmarking

While we've discussed the theoretical performance advantages of bitwise AND over modulo, let's examine actual benchmarks. In practice, modern compilers often optimize n % 2 to n & 1, but understanding the raw difference helps appreciate the optimization.

What to expect:

In optimized builds, performance should be nearly identical
In unoptimized builds or interpreted languages, bitwise can be 2-5x faster
The difference matters most in tight inner loops processing millions of iterations

benchmark.py
Python
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
import timeit
import random
 
# Generate test data
test_data = [random.randint(-10**9, 10**9) for _ in range(10_000_000)]
 
def modulo_check():
    """Count even numbers using modulo."""
    count = 0
    for n in test_data:
        if n % 2 == 0:
            count += 1
    return count
 
def bitwise_check():
    """Count even numbers using bitwise AND."""
    count = 0
    for n in test_data:
        if (n & 1) == 0:
            count += 1
    return count
 
# Verify both give same result
assert modulo_check() == bitwise_check()
 
# Benchmark
print("Benchmarking 10 million even/odd checks...")
print()
 
modulo_time = timeit.timeit(modulo_check, number=3) / 3
print(f"Modulo (n % 2 == 0):  {modulo_time:.4f} seconds")
 
bitwise_time = timeit.timeit(bitwise_check, number=3) / 3
print(f"Bitwise (n & 1 == 0): {bitwise_time:.4f} seconds")
 
print()
speedup = modulo_time / bitwise_time
print(f"Speedup: {speedup:.2f}x")
print()
print("Note: In Python, the difference may be small due to interpreter")
print("overhead. In compiled languages (C/C++), the difference is more")
print("pronounced in unoptimized builds.")

When Does Performance Matter?

For a single even/odd check, the performance difference is negligible. But in scenarios like image processing (checking millions of pixel coordinates), game physics (per-frame calculations across thousands of objects), or data analysis (processing billions of records), these micro-optimizations compound significantly.

The Bigger Picture: Foundation for Advanced Techniques

The even/odd check using n & 1 is the gateway technique to bit manipulation. Once you internalize this pattern—using AND with a mask to extract specific bits—you've learned the conceptual foundation for:

Extracting bit ranges:

n & 0b1111  // Extract the last 4 bits
n & 0xFF   // Extract the last 8 bits (one byte)

Checking specific bit positions:

n & (1 << k)  // Check if bit at position k is set

Creating bit masks dynamically:

(1 << k) - 1  // Mask with k rightmost bits set

Every technique in this module builds on this foundation. By mastering the simple case of extracting the LSB, you've taken the first step toward fluency in bit manipulation.

Key insight: The expression n & 1 is not just a trick—it's the application of the fundamental principle that AND + mask = bit extraction. This principle, once understood, unlocks an entire category of optimization techniques.

Mental Model Check

You should now be able to look at any integer and instantly visualize: 'Is the rightmost bit a 0 or a 1?' If you can do this, you've developed the mental model needed for the techniques in this module. If not, spend more time with the binary representations until they become second nature.

Summary and Key Takeaways

We've explored the simplest yet most fundamental bit manipulation technique in depth. Let's consolidate what we've learned:

Key Takeaways

•The LSB determines parity — In binary, only the 2⁰ term can be odd. All other powers of 2 are even.
•AND extracts specific bits — n & 1 isolates the least significant bit, giving 0 for even and 1 for odd numbers.
•Bitwise is faster than modulo — While modern compilers often optimize this, bitwise AND is guaranteed to be optimal.
•Two's complement preserves parity — Negative numbers maintain the same parity relationship through negation.
•Use parentheses — Write (n & 1) == 0 to avoid operator precedence issues.
•This is the foundation — The AND + mask pattern you've learned here applies to all bit extraction operations.

What's Next:

In the next page, we'll learn to extract the rightmost set bit from a number—a slightly more complex operation that introduces the interaction between a number and its two's complement negation. This technique is fundamental to many efficient algorithms including Brian Kernighan's bit counting algorithm.

Page Complete

You now understand the fundamental bit manipulation technique for parity checking. This foundation—visualizing numbers in binary and using AND with masks to extract bits—will serve you throughout the rest of this module and beyond. Proceed to the next page to learn about extracting the rightmost set bit.