<?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>Npu on Damien Dye&#39;s Blog</title>
    <link>https://blogs.damiendye.uk/en/tags/npu/</link>
    <description>Recent content in Npu on Damien Dye&#39;s Blog</description>
    <generator>Hugo</generator>
    <language>en-GB</language>
    <lastBuildDate>Fri, 13 Nov 2026 12:30:00 +0000</lastBuildDate>
    <atom:link href="https://blogs.damiendye.uk/en/tags/npu/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Running An LLM On The Ryzen AI NPU, In A Container</title>
      <link>https://blogs.damiendye.uk/en/hardware/ryzen-ai-npu-llm-in-a-container/</link>
      <pubDate>Fri, 13 Nov 2026 12:30:00 +0000</pubDate>
      <guid>https://blogs.damiendye.uk/en/hardware/ryzen-ai-npu-llm-in-a-container/</guid>
      <description>Putting the XDNA2 NPU in a Ryzen AI 7 PRO 350 to work on Fedora 45: the amdxdna driver on the host and nothing else, FastFlowLM and XRT in a rootless podman container, Open WebUI and Whisper on IPv6 only, started at boot. Plus the measured cost against the Radeon 860M, and why AMD took three years to get here.</description>
    </item>
  </channel>
</rss>
