pcg-cpp Documentation

This library provides a C++11 implementation of the PCG (Permuted Congruential Generator) family of random number generators. PCG generators are designed to be fast, statistically excellent, and feature-rich.

The library distinguishes between two types of generators:

1. Normal generators: Standard generators suitable for most use cases.
2. Extended generators: Provide k-dimensional equidistribution and specialized features (often referred to as 'party tricks').

For most applications, normal generators are sufficient.

Understanding the repository layout:

• include: Contains the core header pcg_random.hpp and supporting include files. This is what you need to include in your C++ projects.
• test-high: Contains test code for the high-level API (using shorter, simplified function names).
• sample: Contains human-readable sample code and various usage examples.

You can access PCG generators in two ways, similar to the pattern used by std::mt19937 in the C++ standard library:

1. Convenience Typedefs: The easiest way for most users. For 32-bit numbers, use pcg32. For 64-bit numbers, use pcg64.
2. Underlying Templates: For advanced users who need to specify exact parameters, you can use the templates directly.

Performance Tip: If you are on a 32-bit system and need 64-bit numbers, it may be faster to construct 64 bits using two calls to pcg32_k2 rather than using pcg64.

// Example usage pattern (conceptual)
#include "pcg_random.hpp"

// Using a convenience typedef for 32-bit numbers
pcg32 my_generator(seed);

// Using a convenience typedef for 64-bit numbers
pcg64 my_64bit_generator(seed);

The library is include-only and written in C++11, meaning there is no compiled library to link against; you simply include the headers in your project.

If you wish to build the provided demo programs or run the included tests on a Unix-style system (Linux, macOS), use the following commands:

To build demo programs:

make

To run tests:

make test