Ludax: A GPU-Accelerated Domain Specific Language for Board Games

2025-06-27

AI Summary

This work addresses the challenge of balancing rule-modelling flexibility and simulation efficiency in board-game AI research. We propose GameDSL, a domain-specific language for board games that enables declarative specification of diverse game rules and automatically compiles them—via a JAX-based framework—into GPU-accelerated, differentiable simulation environments. Our key contribution is the first end-to-end compilation from DSL semantics to fine-grained GPU kernels, achieving both expressive generality and hardware-level execution efficiency. Experiments across six games—including Go and chess—demonstrate 10–50× speedup over conventional CPU-based simulators. Moreover, GameDSL integrates seamlessly into reinforcement learning pipelines, enabling efficient cross-game policy generalization and scalable training.

Technology Category

• Humans and AI: Game Design — Procedural Content Generation & Storytelling
• Search and Optimization: Sampling/Simulation-based Search
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Games have long been used as benchmarks and testing environments for research in artificial intelligence. A key step in supporting this research was the development of game description languages: frameworks that compile domain-specific code into playable and simulatable game environments, allowing researchers to generalize their algorithms and approaches across multiple games without having to manually implement each one. More recently, progress in reinforcement learning (RL) has been largely driven by advances in hardware acceleration. Libraries like JAX allow practitioners to take full advantage of cutting-edge computing hardware, often speeding up training and testing by orders of magnitude. Here, we present a synthesis of these strands of research: a domain-specific language for board games which automatically compiles into hardware-accelerated code. Our framework, Ludax, combines the generality of game description languages with the speed of modern parallel processing hardware and is designed to fit neatly into existing deep learning pipelines. We envision Ludax as a tool to help accelerate games research generally, from RL to cognitive science, by enabling rapid simulation and providing a flexible representation scheme. We present a detailed breakdown of Ludax's description language and technical notes on the compilation process, along with speed benchmarking and a demonstration of training RL agents. The Ludax framework, along with implementations of existing board games, is open-source and freely available.

Problem

• Develop a GPU-accelerated DSL for board games
• Combine game description languages with hardware speed
• Enable rapid simulation for AI research

Innovation

• GPU-accelerated DSL for board games
• Compiles to parallel processing hardware code
• Integrates with deep learning pipelines