用 NVIDIA Warp 与 MJWarp 加速机器人仿真学习
原标题：How to Use NVIDIA Warp and MjWarp to Accelerate Robotics Simulation and Learning Workflows
AI 摘要
Hugging Face 博客发布教程，介绍如何用 NVIDIA Warp 和 MJWarp 加速机器人仿真与学习工作流。文章以 SO-101 机械臂为例，展示如何将兼容的 MuJoCo 模型迁移到最多 2048 个并行 MJWarp 环境，并说明 Warp 的 GPU 内核编译、可微分与确定性执行等能力。其价值在于通过批量并行提升强化学习等场景的聚合吞吐量，而非单环境步进速度。
正文节选
MuJoCo Warp (MJWarp), built on NVIDIA Warp, takes compatible MuJoCo models into that GPU-scale regime. In this article, we will move an SO-101 follower arm from a familiar MuJoCo workflow to as many as 2,048 parallel MJWarp environments and examine the technology and validation steps that make the transition possible. Figure 1. How MJWarp connects Python to GPU simulation. MuJoCo loads and compiles the MJCF model; MJWarp implements the physics in NVIDIA Warp, which compiles CUDA kernels to advan