Chapter 6: Random Number Generation

Introduction

Random number generation is a crucial part of many applications, from simulations to games. In this chapter, we will explore different random number generators (RNGs) in C++, their usage, and best practices for seeding and generating random numbers.

Good Random Generators

Some of the recommended random generators are:

PCG
SplitMix
Truncated Xorshift32* or Xorshift1024*

Comparison of Random Generators

Type name	Family	Period	State size	Performance	Quality	Should I use this?
minstd_rand	Linear congruential generator	2^31	4 bytes	Bad	Awful	No
mt19937	Mersenne twister	2^19937	2500 bytes	Decent	Decent	Probably (see next section)
ranlux24, ranlux48	Subtract and carry	10^171	96 bytes	Awful	Good	No
knuth_b	Shuffled linear congruential generator	2^31	1028 bytes	Awful	Bad	No
default_random_engine	Implementation defined	Varies	Varies	?	?	No
rand()	Linear congruential generator	2^31	4 bytes	Bad	Awful	No

Popular Choices for PRNGs

Non-Cryptographic PRNGs:
- Xoshiro family
- Wyrand
Cryptographic PRNGs:
- Chacha family

Using Mersenne Twister

#include <random> // for std::mt19937
std::mt19937 mt{}; // Instantiate a 32-bit Mersenne Twister
std::cout << mt() << '\t'; // generate a random number

Seeding with the System Clock

#include <chrono> // access to clock

// Seed our Mersenne Twister using steady_clock
std::mt19937 mt{ static_cast<std::mt19937::result_type>(
    std::chrono::steady_clock::now().time_since_epoch().count()
    ) };

// Create a reusable random number generator that generates uniform numbers between 1 and 6
std::cout << die6(mt) << '\t';

Using Random Device for Seeding

#include <random> // for std::mt19937 and std::random_device

std::mt19937 mt{ std::random_device{}() };

std::uniform_int_distribution die6{ 1, 6 }; // for C++14, use std::uniform_int_distribution<> die6{ 1, 6 };
std::cout << die6(mt) << '\t'; // generate a roll of the die here

Common Issues with RNGs

Same Sequence Every Run: If your RNG generates the same sequence of random numbers every time your program is run, you probably didn’t seed it properly (or at all). Make sure you’re seeding it with a value that changes each time the program is run.
Same Number Every Call: If your RNG generates the same number every time you ask it for a random number, then you are probably either reseeding the RNG before generating a random number, or you’re creating a new random generator for each random number.

Conclusion

Understanding random number generators and their proper usage is essential for many applications. This chapter covered various RNGs, how to seed them correctly, and best practices to avoid common pitfalls.