NVIDIA 发布 CUDA Rust，支持用 Rust 编写 GPU 内核
原标题：Introducing CUDA Rust: Two Tracks for Writing GPU Kernels
AI 摘要
NVIDIA 于 2026 年 9 月宣布推进 CUDA Rust，支持用 Rust 原生编写 GPU 内核并编译为 PTX，提供 SIMT 和 Tile 两种编程模型。首个工具 cuda-oxide 作为 rustc 代码生成后端，可将 Rust 内核编译到 PTX，并提供了完整的向量加法示例。此举旨在满足 AI 系统层对 Rust 的需求，并计划在 2027 年及以后持续完善。
正文节选
In September 2026, NVIDIA announced it is leaning into native GPU programming in Rust. CUDA C++ and CUDA Python are mature, enterprise-grade toolchains, and NVIDIA will be growing and maturing CUDA Rust into 2027 and beyond The systems layer of AI spans inference engines, serving infrastructure, drivers, and agent runtimes, and it churns constantly as models and techniques change. More and more of it is written in Rust, which catches whole classes of bugs at compile time without giving up perfor