<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Sequence-Parallelism on 111qqz的小窝</title><link>https://111qqz.com/tags/sequence-parallelism/</link><description>Recent content in Sequence-Parallelism on 111qqz的小窝</description><generator>Hugo -- gohugo.io</generator><language>zh-cn</language><managingEditor>hust.111qqz@gmail.com (111qqz)</managingEditor><webMaster>hust.111qqz@gmail.com (111qqz)</webMaster><copyright>© 2011-2026 111qqz</copyright><lastBuildDate>Thu, 29 Oct 2026 10:00:00 +0800</lastBuildDate><atom:link href="https://111qqz.com/tags/sequence-parallelism/index.xml" rel="self" type="application/rss+xml"/><item><title>Sequence Parallelism 与 Context Parallelism：为什么沿序列切分还不够？</title><link>https://111qqz.com/2026/10/sequence-context-parallelism/</link><pubDate>Thu, 29 Oct 2026 10:00:00 +0800</pubDate><author>hust.111qqz@gmail.com (111qqz)</author><guid>https://111qqz.com/2026/10/sequence-context-parallelism/</guid><description>&lt;p&gt;&lt;a href="https://111qqz.com/2026/10/tensor-parallelism-matrix-split/" &gt;上一篇 TP&lt;/a&gt; 写完之后，我对 Megatron 的 Column Parallel + Row Parallel 已经比较清楚了：前一层按输出特征切分，后一层按输入特征切分，两层之间不需要 AllGather，只在最后做一次 AllReduce 把部分和加起来。&lt;/p&gt;</description></item></channel></rss>