Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning
This provides a high-performance solution for robotics researchers and practitioners, enabling faster training of complex tasks, though it is incremental as it optimizes existing simulation and training workflows.
The paper tackles the problem of slow training times in robot learning by introducing Isaac Gym, a GPU-based physics simulation platform that integrates simulation and policy training on the GPU, achieving 2-3 orders of magnitude faster training compared to conventional CPU-based methods.
Isaac Gym offers a high performance learning platform to train policies for wide variety of robotics tasks directly on GPU. Both physics simulation and the neural network policy training reside on GPU and communicate by directly passing data from physics buffers to PyTorch tensors without ever going through any CPU bottlenecks. This leads to blazing fast training times for complex robotics tasks on a single GPU with 2-3 orders of magnitude improvements compared to conventional RL training that uses a CPU based simulator and GPU for neural networks. We host the results and videos at \url{https://sites.google.com/view/isaacgym-nvidia} and isaac gym can be downloaded at \url{https://developer.nvidia.com/isaac-gym}.