NVIDIA 推出 CUTLASS Python 新功能
NVIDIA 给 CUTLASS Python 加了新工具，能让开发者更快写 GPU 内核，还加了错误检查功能。
NVIDIA 在 PyTorch Conference North America 2026 上展示了 CUTLASS Python 的最新改进。CUTLASS Python 是用于编写最新 GPU 代速度级内核的技术栈，支持当今最关键的 AI 工作负载。新功能包括 CuTe 扩展、CUTLASS 原语、任务调度器、编译器诊断和静态检查。
High-Velocity GPU Kernel Authoring with CUTLASS Python
Join Guray Ozen from @nvidia at PyTorch Conference North America 2026, where he will present recent improvements to CUTLASS for Python.
CUTLASS for Python is the technology stack used to write speed-of-light kernels for the latest GPU generations, powering today's most critical AI workloads. In his presentation, Guray will introduce several new productivity features for developers and agents, including CuTe extensions, CUTLASS primitives, a task scheduler, compiler diagnostics, and static checks to help catch errors before launch.
Join us on October 20-21 in San Jose: https://t.co/jBApW8nESi
#PyTorchCon
