<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Milliseconds.dev — .NET Benchmarks</title>
    <description>Real .NET vs Python (and more) performance benchmarks. Real hardware. Real numbers.</description>
    <link>https://milliseconds.dev</link>
    <atom:link href="https://milliseconds.dev/rss" rel="self" type="application/rss+xml"/>
    <language>en-us</language>
    <lastBuildDate>Sun, 31 May 2026 19:37:33 GMT</lastBuildDate>
    
    <item>
      <title><![CDATA[NumPy vs TensorPrimitives: Softmax at LLM Vocab Scale]]></title>
      <description><![CDATA[⚡ .NET is 1.53× faster. At GPT-2 scale .NET and NumPy are tied. Scale to Llama-3's 128k vocab and .NET pulls 1.53× ahead — because TensorPrimitives reuses a single SIMD buffer while NumPy keeps allocating.]]></description>
      <link>https://milliseconds.dev/blog/01-aurora</link>
      <guid isPermaLink="true">https://milliseconds.dev/blog/01-aurora</guid>
      <pubDate>Sun, 31 May 2026 00:00:00 GMT</pubDate>
    </item>
  </channel>
</rss>