<?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>Performance on Tuan Hiep TRAN — The Principal Craft</title><link>https://tuanhiep.github.io/principal/tags/performance/</link><description>Recent content in Performance on Tuan Hiep TRAN — The Principal Craft</description><generator>Hugo -- 0.162.1</generator><language>en-us</language><lastBuildDate>Wed, 01 Apr 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://tuanhiep.github.io/principal/tags/performance/index.xml" rel="self" type="application/rss+xml"/><item><title>From Contention to Throughput (1/5): Move the Work, Not the Latency — The Pre-allocation Paradigm</title><link>https://tuanhiep.github.io/principal/posts/series_2/1.the_law_of_decoupled_computation/</link><pubDate>Wed, 01 Apr 2026 00:00:00 +0000</pubDate><guid>https://tuanhiep.github.io/principal/posts/series_2/1.the_law_of_decoupled_computation/</guid><description>Why Compute-on-Demand fails at scale. We explore the Law of Decoupled Computation, the Decision Framework, and how to eliminate latency from the critical path.</description></item></channel></rss>