<?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 Stephen Krings</title>
    <link>https://stephenkrings.com/tags/gpu/</link>
    <description>Recent content in Gpu on Stephen Krings</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Tue, 14 Jul 2026 19:58:00 -0600</lastBuildDate>
    <atom:link href="https://stephenkrings.com/tags/gpu/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Making Hugging Face Image Generation Fit on a 12GB RTX 5070</title>
      <link>https://stephenkrings.com/posts/hugging-face-image-generation-rtx-5070-cpu-offload/</link>
      <pubDate>Tue, 14 Jul 2026 19:58:00 -0600</pubDate>
      <guid>https://stephenkrings.com/posts/hugging-face-image-generation-rtx-5070-cpu-offload/</guid>
      <description>&lt;p&gt;Goal: wiring Hugging Face models into my local llm (Ollama) app, testing prompt-following, and seeing how far I can push a 12GB RTX 5070 before the system starts leaning hard on the CPU.&lt;/p&gt;
&lt;p&gt;The short version: it works, but only because the stack is allowed to spill parts of the model back to system memory.&lt;/p&gt;
&lt;h2 id=&#34;the-current-setup&#34;&gt;The current setup&lt;/h2&gt;
&lt;p&gt;The image UI lives on my web server, but the GPU does not. The request path looks like this:&lt;/p&gt;</description>
    </item>
    <item>
      <title>Best of Both Worlds: Using a Gaming PC&#39;s RTX 5070 from a separate Linux workstation</title>
      <link>https://stephenkrings.com/posts/rtx-5070-linux-gpu-server/</link>
      <pubDate>Tue, 14 Jul 2026 01:05:00 -0600</pubDate>
      <guid>https://stephenkrings.com/posts/rtx-5070-linux-gpu-server/</guid>
      <description>&lt;p&gt;I do all my development on a Linux box with no GPU. My RTX 5070 lives across the room in a Windows gaming machine. Rather than dual-booting or doing native Windows dev (no), I turned the gaming box into a headless Linux GPU server I can SSH into, and it still plays games.&lt;/p&gt;
&lt;h2 id=&#34;the-setup&#34;&gt;The setup&lt;/h2&gt;
&lt;p&gt;&lt;img alt=&#34;Diagram: Linux dev box connecting over ssh to the Windows gaming PC on port 2345 via netsh portproxy, which relays into WSL2 Ubuntu 24.04 with full GPU passthrough&#34; loading=&#34;lazy&#34; src=&#34;https://stephenkrings.com/images/gpu-setup-diagram.png&#34;&gt;&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
