Chapter 1: Introduction to RAM Optimization

When working on applications that handle significant amounts of data, it's crucial to consider memory usage, especially since RAM is often limited. The most straightforward approach to enhance RAM efficiency involves examining the structures utilized within the program.

Alignment refers to how data is organized in a computer's memory. This organization is determined by how hardware accesses data stored in memory. A data type is said to have n-byte alignment when it is placed at an address that is a multiple of n. Compilers typically align data to be compatible with the CPU architecture, which can result in unallocated memory spaces, a phenomenon known as padding. While this may not affect simpler applications, it can have substantial implications for data-intensive programs that utilize millions of structures, leading to wasted memory.

C++11 introduced the alignof and alignas keywords to facilitate the creation of portable code. However, for many scenarios, simpler methods can yield significant memory savings. By reorganizing code, developers can reclaim gigabytes of memory without needing to know the precise alignment sizes of each type. Instead, leveraging an understanding of the sizes of basic data types can suffice.

To illustrate this, consider the following code which compares three structures. Although S1 and S2 share the same members, their sizes differ due to alignment and padding. Interestingly, S3 contains additional member variables yet occupies the same memory footprint as S2 because it effectively utilizes padding space.

#include <cstdlib> // EXIT_SUCCESS macro

#include <iostream> // cout

#include <typeinfo> // typeid

/---------------------------------------------------------------------------/

struct S1 {

char c1;

int i1;

char c2;

long int li1;

};

struct S2 {

char c1;

char c2;

int i1;

long int li1;

};

struct S3 {

char c1;

char c2;

char padding1;

char padding2;

int i1;

long int li1;

};

/---------------------------------------------------------------------------/

int main() {

std::cout << "Size of char: " << sizeof(char) << std::endl;

std::cout << "Size of int: " << sizeof(int) << std::endl;

std::cout << "Size of long int: " << sizeof(long int) << std::endl;

std::cout << "Size of S1: " << sizeof(S1) << std::endl;

std::cout << "Size of S2: " << sizeof(S2) << std::endl;

std::cout << "Size of S3: " << sizeof(S3) << std::endl;

return EXIT_SUCCESS;

}

To gain a clearer understanding of alignment and padding, refer to the following illustration. Here, char can be stored at any memory address, while int must be positioned at an address that is a multiple of 4, and long int needs to be at a multiple of 8. The green cells represent memory addresses containing data, whereas the yellow cells indicate addresses filled with padding zeroes. The variable names highlighted in green indicate the starting address of each member variable.

For a deeper exploration of padding and alignment in C++, check out the video titled "C++ Weekly - Ep 410 - What Are Padding and Alignment? (And Why You Might Care)."

Chapter 2: Understanding Structural Padding and Packing

In this chapter, we delve into the concepts of structural padding and packing in C and C++. By understanding how these principles apply, we can further optimize memory usage in our applications.

The accompanying video, "Structural Padding & Packing In C & C++," provides valuable insights into these concepts and their significance.