<?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>weekly-papers-2026-06-20 on Sparse Notes</title>
    <link>https://sparsenotes.com/series/weekly-papers-2026-06-20/</link>
    <description>Recent content in weekly-papers-2026-06-20 on Sparse Notes</description>
    <image>
      <url>https://sparsenotes.com/images/og-default.png</url>
      <link>https://sparsenotes.com/images/og-default.png</link>
    </image>
    <generator>Hugo -- gohugo.io</generator>
    <lastBuildDate>Sat, 20 Jun 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://sparsenotes.com/series/weekly-papers-2026-06-20/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>ARGUS: Production-Scale Tracing and Performance Diagnosis for 10,000&#43; GPU Clusters</title>
      <link>https://sparsenotes.com/posts/2026/06/papers/2026-06-20-argus-gpu-tracing/</link>
      <pubDate>Sat, 20 Jun 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/06/papers/2026-06-20-argus-gpu-tracing/</guid>
      <description>An always-on, fine-grained tracing system for &amp;gt;10,000-GPU LLM training clusters that keeps overhead under 2% by decomposing observation into three independent hierarchies and compressing kernel traces ~3,700x.</description>
    </item>
    
    <item>
      <title>The Bi-Channel Networking Paradigm for Database Systems in the Cloud</title>
      <link>https://sparsenotes.com/posts/2026/06/papers/2026-06-20-bi-channel-networking/</link>
      <pubDate>Sat, 20 Jun 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/06/papers/2026-06-20-bi-channel-networking/</guid>
      <description>Cloud database systems are now CPU-bound on the kernel TCP stack as NIC bandwidth has outrun core speed by 20×; this paper proposes splitting communication into a fast user-space data path and a reliable kernel control path.</description>
    </item>
    
    <item>
      <title>The Google File System (2003)</title>
      <link>https://sparsenotes.com/posts/2026/06/papers/2026-06-20-google-file-system/</link>
      <pubDate>Sat, 20 Jun 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/06/papers/2026-06-20-google-file-system/</guid>
      <description>Ghemawat, Gobioff and Leung&amp;#39;s twenty-year-old paper on the file system that powered MapReduce, Bigtable and almost everything else inside early-2000s Google — and the architectural ideas (single master, append-mostly mutations, control/data-plane split) that quietly underwrite modern object storage.</description>
    </item>
    
    <item>
      <title>Weekly CS Paper Digest — Jun 14 – Jun 20, 2026</title>
      <link>https://sparsenotes.com/posts/2026/06/papers/</link>
      <pubDate>Sat, 20 Jun 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/06/papers/</guid>
      <description>Three blog-length entries from this week&amp;#39;s CS paper digest: The Google File System (2003) as Seminal pick, plus top picks on ARGUS — an always-on tracing system for 10,000-GPU LLM training clusters — and the Bi-Channel Networking Paradigm for cloud databases.</description>
    </item>
    
  </channel>
</rss>
