<?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>Gpu on CctoctoFX</title>
    <link>https://pillumina.github.io/tags/gpu/</link>
    <description>Recent content in Gpu on CctoctoFX</description>
    <image>
      <title>CctoctoFX</title>
      <url>https://pillumina.github.io/imgs/icon_head.png</url>
      <link>https://pillumina.github.io/imgs/icon_head.png</link>
    </image>
    <generator>Hugo -- 0.148.2</generator>
    <language>en</language>
    <lastBuildDate>Mon, 22 Jun 2026 09:04:00 +0800</lastBuildDate>
    <atom:link href="https://pillumina.github.io/tags/gpu/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>LLM 系统分析方法论（五）：训练显存估算</title>
      <link>https://pillumina.github.io/posts/aiinfra/llm-computation-methodology/part-5/</link>
      <pubDate>Mon, 22 Jun 2026 09:04:00 +0800</pubDate>
      <guid>https://pillumina.github.io/posts/aiinfra/llm-computation-methodology/part-5/</guid>
      <description>训练显存完整估算：从单卡四笔账（权重/优化器/梯度/激活）出发，叠加 TP/PP/DP/CP/EP 并行折扣，结合 ZeRO/FSDP、Gradient Checkpointing、Offload 建立训练显存体系。含 M3 完整案例和多模态/LoRA 微调场景。</description>
    </item>
  </channel>
</rss>
