Mastermind: Weighted-Entropy Heuristic Implementations

This repository contains three standalone programs that implement a weighted-entropy heuristic for the classic Mastermind game, achieving an average of 4.3488 guesses for the full game set.

Program	Purpose	Hardware
kernel.cu	CUDA-accelerated optimiser that computes stage-based weights	NVIDIA GPU
treeGenFixedWeights.cpp	CPU-based generator using fixed weights	any x86-64 CPU
treeGenStageWeights.cpp	CPU-based generator using stage-based weights	any x86-64 CPU

See the author’s paper for a full description of the algorithm and benchmarks.

---

🚀 Quick start (most users)

Pre-compiled binaries are attached to every GitHub release:

Archive	Contents
mastermind-windows.zip	Windows 10/11 64-bit executables (CUDA 12.4, built for compute ≥ 5.2)
mastermind-linux.tar.gz	Linux x86-64 executables (CUDA 11.8, glibc ≥ 2.35)

1. Download & extract.
2. Run mastermind_cuda_windows.exe (Win) or mastermind_cuda_linux (Linux).

---

🛠️ Building from source

Prerequisites

OS	Required tool-chain	GPU driver
Windows	CUDA Toolkit 12.4 + Visual Studio 2022 (MSVC 14.4x)	NVIDIA driver ≥ 531
Linux	CUDA Toolkit 11.8 + GCC 11+	NVIDIA driver ≥ 470

Clone

git clone https://github.com/obsessivecompulsiveaudiophile/mastermind.git
cd mastermind

Windows (Visual Studio Developer Prompt)

cmd
:: CUDA program
nvcc -arch=sm_52 -O3 -o mastermind_cuda_windows.exe kernel.cu
Replace -arch=sm_52 with -arch=sm_86 (or your GPU’s compute capability) for maximum performance on newer cards.

:: CPU programs
cl /EHsc /O2 treeGenFixedWeights.cpp   /Fe:treeGenFixedWeights_windows.exe
cl /EHsc /O2 treeGenStageWeights.cpp   /Fe:treeGenStageWeights_windows.exe

Linux

bash
# CUDA program
nvcc -arch=sm_52 -O3 -o mastermind_cuda_linux kernel.cu

# CPU programs
g++ -O3 -std=c++17 treeGenFixedWeights.cpp   -o treeGenFixedWeights_linux
g++ -O3 -std=c++17 treeGenStageWeights.cpp   -o treeGenStageWeights_linux