<?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>Ilp on System &amp; Software Engineering</title>
    <link>https://riyaneel.github.io/tags/ilp/</link>
    <description>Recent content in Ilp on System &amp; Software Engineering</description>
    <generator>Hugo -- 0.160.0</generator>
    <language>en-us</language>
    <lastBuildDate>Sun, 05 Apr 2026 23:17:00 +0100</lastBuildDate>
    <atom:link href="https://riyaneel.github.io/tags/ilp/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>The Instruction Count Lie: Latency, Throughput, and the Dependency Graph</title>
      <link>https://riyaneel.github.io/posts/latency-vs-throughput/</link>
      <pubDate>Sun, 05 Apr 2026 23:17:00 +0100</pubDate>
      <guid>https://riyaneel.github.io/posts/latency-vs-throughput/</guid>
      <description>Counting instructions is not measuring execution time. A deep dive into how Read-After-Write hazards serialize the Out-of-Order engine — and how restructuring dependencies unlocks the superscalar machine.</description>
    </item>
  </channel>
</rss>
