<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Yestino - The Signal · CUDA Explain</title><link>https://yestino.com/zh-CN/entities/cuda-explain-4d0b4f</link><description>与 CUDA Explain 相关的全部事件</description><language>zh-CN</language><atom:link href="https://yestino.com/zh-CN/entities/cuda-explain-4d0b4f/feed.xml" rel="self" type="application/rss+xml"/><item><title>trunk/1c3d446cda0cb4231a810427809d2b3382395e0b: [CUDA] Explain custom MemPool allocator failures in OOM errors (#199363)</title><link>https://yestino.com/zh-CN/events/trunk-1c3d446cda0cb4231a810427809d2b3382395e0b-cuda-explain-bb5101</link><guid isPermaLink="true">https://yestino.com/zh-CN/events/trunk-1c3d446cda0cb4231a810427809d2b3382395e0b-cuda-explain-bb5101</guid><pubDate>Fri, 02 Oct 2026 00:42:48 GMT</pubDate><description>Human note: make a better error message when custom mempool fails allocation Custom allocators signal allocation failure by returning nullptr.
来源：PyTorch Releases</description></item></channel></rss>