NVIDIA 发布 CUDA Python 1.0，提供稳定 API 与全平台访问
CUDA Python 1.0: Stable APIs, One Foundation, Full Platform Access
NVIDIA 发布 CUDA Python 1.0，主打稳定 API、统一基础和全平台访问。文章回顾了 Python 开发者使用 GPU 的两条旧路径，一是自己学 CUDA C++ 写扩展并维护 Python 绑定，二是依赖 PyTorch、CuPy、RAPIDS 等上层库，后者虽然撑起了 Python GPU 生态，但存在局限。
CUDA Python 1.0 把稳定 API 与全平台访问打包，读者可据此判断 Python 直连 GPU 的路径是否成熟。
For years, a Python developer who needed a GPU had two realistic choices: Learn NVIDIA CUDA C++ well enough to write an extension, set up a build toolchain, and...
For years, a Python developer who needed a GPU had two realistic choices: Learn NVIDIA CUDA C++ well enough to write an extension, set up a build toolchain, and maintain bindings back to Python, which most people never did; or move up the stack and let someone else’s library do it, namely PyTorch, CuPy, or RAPIDS. The second option is why the Python GPU ecosystem thrives. But it has limits.
来源：NVIDIA Developer Blog · developer.nvidia.com