Pseudo-Random Number Generator (PRNG)

This repository contains a C++ implementation of the ran2 pseudo-random number generator, based on Chapter 7 of Numerical Recipes in C. It includes uniform distribution tests, automated visualization using gnuplot, and a Python testing suite to verify correctness.

Overview

The ran2 algorithm is a long-period ($> 2 \times 10^{18}$) random number generator designed by L'Ecuyer. It incorporates a Bays-Durham shuffle and added safeguards to ensure high-quality pseudo-random numbers. It returns a uniform random deviate between 0.0 and 1.0 (exclusive of the endpoint values).

Features

• High-Quality PRNG: Implementation of the classic ran2 algorithm.
• Helper Interface:
  • Randomize(): Dynamically seeds the generator using time(NULL).
  • Random(): Returns a uniform double in the range [0.0, 1.0).
  • Random(int N): Returns a uniform integer in the range [0, N).
• Distribution Analysis: Automatically bins the generated sequence into 20 sub-intervals to test for uniform distribution.
• Auto-plotting: Generates a GNUplot configuration (temp.txt) and exports a distribution plot (PRNG1.png).

---

Requirements

To compile and run the application, you need:

1. A C++ compiler (e.g., g++ or clang++).
2. gnuplot installed on your system (optional, required to generate the plot image).

Installing Gnuplot (Optional)

• macOS (via Homebrew):

  brew install gnuplot

• Ubuntu/Debian:

  sudo apt-get install gnuplot

---

Getting Started

1. Compile the Code

Compile the source code using the following shell command:

g++ -o PRNG.out PRNG.cpp -Ofast

2. Run the Program

Run the compiled binary:

./PRNG.out

3. Execution flow

1. The program asks you to specify the number of random numbers to generate:

   chand adad random mikhayd? (yek adade integer vared konid)=
   

   (Enter an integer, e.g., 10000)
2. It prints the generated random sequence.
3. It prints the frequency distribution counts (p[i]) across 20 bins.
4. It prints the normalized frequencies.
5. It compiles a gnuplot script (temp.txt) and runs it to save the plot as PRNG1.png.

---

Automated Testing

An automated Python test script is provided in the repository to compile, run, and validate the C++ code's functionality and statistical behavior.

To run the tests:

python3 test_prng.py

What the test checks:

• Successful compilation of PRNG.cpp.
• Boundary check of all generated random numbers (must be in [0.0, 1.0)).
• Integrity of the frequency binning (sum of bins must equal the total requested samples).
• Verification of normalized frequencies.
• Verification of output files (temp.txt and PRNG1.png).
• Statistical sanity check (the mean of the generated uniform distribution must be close to 0.5).

---

License & Reference

This implementation is based on the ran2 routine from Numerical Recipes in C. Please refer to Numerical Recipes for their licensing terms.