Pgx: Hardware-accelerated parallel game simulation for reinforcement learning

03/29/2023

∙

We propose Pgx, a collection of board game simulators written in JAX. Thanks to auto-vectorization and Just-In-Time compilation of JAX, Pgx scales easily to thousands of parallel execution on GPU/TPU accelerators. We found that the simulation of Pgx on a single A100 GPU is 10x faster than that of existing reinforcement learning libraries. Pgx implements games considered vital benchmarks in artificial intelligence research, such as Backgammon, Shogi, and Go. Pgx is available at https://github.com/sotetsuk/pgx.

READ FULL TEXT

Pgx: Hardware-accelerated parallel game simulation for reinforcement learning

Parallel Q-Learning: Scaling Off-policy Reinforcement Learning under Massively Parallel Simulation

GPU-Accelerated Atari Emulation for Reinforcement Learning

Deep RTS: A Game Environment for Deep Reinforcement Learning in Real-Time Strategy Games

Assessing the Potential of Classical Q-learning in General Game Playing

Neural Game Engine: Accurate learning of generalizable forward models from pixels

Neural Game Engine: Accurate learning ofgeneralizable forward models from pixels

TAPA: A Scalable Task-Parallel Dataflow Programming Framework for Modern FPGAs with Co-Optimization of HLS and Physical Design

Pgx: Hardware-accelerated parallel game simulation for reinforcement learning

Related Research

Parallel Q-Learning: Scaling Off-policy Reinforcement Learning under Massively Parallel Simulation

GPU-Accelerated Atari Emulation for Reinforcement Learning

Deep RTS: A Game Environment for Deep Reinforcement Learning in Real-Time Strategy Games

Assessing the Potential of Classical Q-learning in General Game Playing

Neural Game Engine: Accurate learning of generalizable forward models from pixels

Neural Game Engine: Accurate learning ofgeneralizable forward models from pixels

TAPA: A Scalable Task-Parallel Dataflow Programming Framework for Modern FPGAs with Co-Optimization of HLS and Physical Design