Data Type Size Optimization

Data Type Size Optimization in C++ (Use Only What You Need)

---

Prerequisites

---

1. Why Data Type Size Matters

Choosing unnecessarily large data types can lead to:

Memory efficiency directly impacts CPU performance.

• Increased memory usage
• Poor cache utilization
• Lower performance

Use the smallest data type that satisfies your requirements

1-1. Array Optimization

❌ Inefficient

int arr[1000000];  // 4MB

64 / 4 = 16 elements per cache line

✔ Optimized

uint8_t arr[1000000];  // 1MB

64 / 1 = 64 elements per cache line

• 4x less memory
• Better cache performance

Cache Impact

CPU cache line ≈ 64 bytes More data per cache line → fewer cache misses

1-2. Struct Optimization

❌ Wasteful

struct Data {
    int a;
    int b;
    int c;
};

✔ Optimized

struct Data {
    uint8_t a;
    uint8_t b;
    uint8_t c;
};

👉 Significant memory reduction when repeated many times

❗ Alignment & Padding

struct A 
{
    uint8_t a;
    int b;
};

Padding added → may reduce benefit

❗ CPU Efficiency

• Some CPUs are optimized for 32-bit or 64-bit operations
• Too small types may require extra instructions

When Optimization Matters

✔ Large arrays / datasets
✔ Performance-critical loops
✔ Embedded systems
✔ Game / simulation engines

2. Trade-offs

✔ DO

• Choose smallest sufficient type
• Use fixed-width types (uint8_t, uint32_t)
• Profile memory-heavy code

❌ DON’T

• Use int by default everywhere
• Over-optimize prematurely
• Ignore alignment and padding

Factor	Small Type	Large Type
Memory	Efficient	Wasteful
Cache	Better	Worse
Range	Limited	Large
CPU ops	Sometimes slower	Often faster