<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Gpu-Performance on 111qqz的小窝</title><link>https://111qqz.com/tags/gpu-performance/</link><description>Recent content in Gpu-Performance on 111qqz的小窝</description><generator>Hugo -- gohugo.io</generator><language>zh-cn</language><copyright>© 2015-2026 111qqz</copyright><lastBuildDate>Sun, 23 Aug 2026 12:45:00 +0800</lastBuildDate><atom:link href="https://111qqz.com/tags/gpu-performance/index.xml" rel="self" type="application/rss+xml"/><item><title>从 FLOPs 到 Latency：GPU 推理性能到底由什么决定？</title><link>https://111qqz.com/2026/08/flops-to-latency/</link><pubDate>Sun, 23 Aug 2026 12:45:00 +0800</pubDate><guid>https://111qqz.com/2026/08/flops-to-latency/</guid><description>&lt;h2 class="relative group"&gt;起因
 &lt;div id="起因" class="anchor"&gt;&lt;/div&gt;
 
 &lt;span
 class="absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100 select-none"&gt;
 &lt;a class="text-primary-300 dark:text-neutral-700 !no-underline" href="#%e8%b5%b7%e5%9b%a0" aria-label="锚点"&gt;#&lt;/a&gt;
 &lt;/span&gt;
 
&lt;/h2&gt;
&lt;p&gt;最近在做 CS336（Language Modeling from Scratch）的 resource accounting 作业，里面要求逐个列出 Transformer forward pass 中所有的 matmul，然后按 \(2MKN\) 计算 FLOPs。&lt;/p&gt;</description></item></channel></rss>