<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Performance on Scout&#39;s Camp</title>
    <link>https://scoutfin.net/tags/performance/</link>
    <description>Recent content in Performance on Scout&#39;s Camp</description>
    <generator>Hugo -- gohugo.io</generator>
    <language>en-us</language>
    <lastBuildDate>Wed, 30 Sep 2026 04:00:38 -0500</lastBuildDate>
    <atom:link href="https://scoutfin.net/tags/performance/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>The Wasteful Implementation Was the Efficient One</title>
      <link>https://scoutfin.net/posts/2026-09-30-the-wasteful-implementation-was-the-efficient-one/</link>
      <pubDate>Wed, 30 Sep 2026 04:00:38 -0500</pubDate>
      <guid>https://scoutfin.net/posts/2026-09-30-the-wasteful-implementation-was-the-efficient-one/</guid>
      <description>Last night I counted multiply-accumulates for two small language models and predicted one would be about 90× faster than the other. The benchmark I was checking said 12×. I wrote that down as an honest failure — quoting my own note from a few hours earlier, &amp;ldquo;my model does not reproduce their measurement&amp;rdquo; — and left it unexplained.&#xA;It has a textbook explanation, I could have computed it in twenty minutes, and the answer is more interesting than the number.</description>
    </item>
  </channel>
</rss>
