
  <rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
      <title>compusential</title>
      <link>https://compusential.com/blog</link>
      <description>Journal of a passionate technologist building production AI Tools (RAG platforms, intelligent query systems, and agentic workflows). Hacker of local models across an NVIDIA DGX Spark (128GB Unified VRAM) &amp; an RTX 4090, rangefinder shooter (Leica M10 with Summicron 35 &amp; Summilux 50 ASPH), and trilingual builder based in England.</description>
      <language>en-GB</language>
      <managingEditor>Yang Pei</managingEditor>
      <webMaster>Yang Pei</webMaster>
      <lastBuildDate>Thu, 20 Aug 2026 00:00:00 GMT</lastBuildDate>
      <atom:link href="https://compusential.com/tags/gpu-computing/feed.xml" rel="self" type="application/rss+xml"/>
      
  <item>
    <guid>https://compusential.com/blog/the-cuda-moment</guid>
    <title>The CUDA Moment: From Bricking a RIVA 128 at 13 to the 8800 GTX</title>
    <link>https://compusential.com/blog/the-cuda-moment</link>
    <description>From bricking an early NVIDIA RIVA 128 at age 13 in Beijing to walking to work at NVIDIA when CUDA 1.0 launched on the GeForce 8800 GTX — the origin of Compusential.</description>
    <pubDate>Thu, 20 Aug 2026 00:00:00 GMT</pubDate>
    <author>Yang Pei</author>
    <category>cuda</category><category>nvidia</category><category>gpu-computing</category><category>retro-hardware</category><category>origins</category>
  </item>

  <item>
    <guid>https://compusential.com/blog/the-local-silicon-equation</guid>
    <title>The Local Silicon Equation: 128GB VRAM, Bandwidth Limits, and Cloud Illusions</title>
    <link>https://compusential.com/blog/the-local-silicon-equation</link>
    <description>Why spending £4,000 on a 128GB NVIDIA DGX Spark alongside an RTX 4090 makes sense: the friction of the cloud hourly meter, memory bandwidth vs capacity trade-offs, and running 131k context on local silicon.</description>
    <pubDate>Thu, 20 Aug 2026 00:00:00 GMT</pubDate>
    <author>Yang Pei</author>
    <category>hardware</category><category>gpu-computing</category><category>local-llm</category><category>production-ai</category><category>vllm</category>
  </item>

    </channel>
  </rss>
