<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>分布式 on 硅基 AGI · 智能体学习与测评</title><link>https://guijiagi.com/tags/%E5%88%86%E5%B8%83%E5%BC%8F/</link><description>Recent content in 分布式 on 硅基 AGI · 智能体学习与测评</description><generator>Hugo</generator><language>zh-cn</language><copyright>本站内容采用 CC BY-NC-SA 4.0 国际许可协议授权</copyright><lastBuildDate>Thu, 16 Jul 2026 11:46:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/%E5%88%86%E5%B8%83%E5%BC%8F/index.xml" rel="self" type="application/rss+xml"/><item><title>大模型推理部署架构：从单卡到分布式全方案</title><link>https://guijiagi.com/posts/b1-7a487767/</link><pubDate>Thu, 16 Jul 2026 11:46:00 +0800</pubDate><guid>https://guijiagi.com/posts/b1-7a487767/</guid><description>系统讲解大模型推理部署的架构选型，从单GPU到多卡分布式到弹性云原生方案</description></item><item><title>LLM负载均衡策略</title><link>https://guijiagi.com/posts/llm-load-balancing-strategy/</link><pubDate>Thu, 02 Jul 2026 11:23:00 +0800</pubDate><guid>https://guijiagi.com/posts/llm-load-balancing-strategy/</guid><description>从轮询到感知调度，LLM推理服务负载均衡的完整策略分析</description></item><item><title>Ring Attention：突破百万 Token 上下文的分布式注意力</title><link>https://guijiagi.com/posts/ring-attention-long-context/</link><pubDate>Thu, 25 Jun 2026 12:00:00 +0800</pubDate><guid>https://guijiagi.com/posts/ring-attention-long-context/</guid><description>深度解析 Ring Attention 如何通过环形通信将注意力计算分布到多 GPU，实现百万级 token 上下文窗口。</description></item><item><title>Ring Attention 解析：百万 Token 上下文的秘密</title><link>https://guijiagi.com/posts/ring-attention-explained/</link><pubDate>Thu, 25 Jun 2026 10:00:00 +0800</pubDate><guid>https://guijiagi.com/posts/ring-attention-explained/</guid><description>深入解析 Ring Attention 如何实现百万级 Token 的长上下文训练与推理</description></item><item><title>Agent 系统可扩展性设计：从单机到分布式</title><link>https://guijiagi.com/posts/agent-scalability/</link><pubDate>Wed, 24 Jun 2026 14:00:00 +0800</pubDate><guid>https://guijiagi.com/posts/agent-scalability/</guid><description>系统讲解 AI Agent 系统的可扩展性设计，涵盖水平扩展、状态外部化、无状态 Agent、负载均衡与分布式追踪</description></item></channel></rss>