<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Yestino - The Signal · GPU][AOTI Support</title><link>https://yestino.com/zh-CN/entities/gpu-aoti-support-7f0a27</link><description>与 GPU][AOTI Support 相关的全部事件</description><language>zh-CN</language><atom:link href="https://yestino.com/zh-CN/entities/gpu-aoti-support-7f0a27/feed.xml" rel="self" type="application/rss+xml"/><item><title>trunk/599f186d3ff3a0e124544ecd62ca63c0a285c32d: [GPU][AOTI] Support host TMA in C++ wrappers (#193608) (#193608)</title><link>https://yestino.com/zh-CN/events/trunk-599f186d3ff3a0e124544ecd62ca63c0a285c32d-gpu-aoti-supp-82f959</link><guid isPermaLink="true">https://yestino.com/zh-CN/events/trunk-599f186d3ff3a0e124544ecd62ca63c0a285c32d-gpu-aoti-supp-82f959</guid><pubDate>Thu, 03 Sep 2026 03:24:41 GMT</pubDate><description>Summary: Host-side TensorDescriptor values created inside a torch.library.triton_op are captured through a non-Dynamo tracing path.
来源：PyTorch Releases</description></item></channel></rss>