<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Sean William Carroll</title><description>Sean Carroll&apos;s personal website.</description><link>https://www.seanwilliamcarroll.com/</link><item><title>Four Hash Maps, Three Key Distributions, and One Catastrophe</title><link>https://www.seanwilliamcarroll.com/posts/hash-map-baseline/</link><guid isPermaLink="true">https://www.seanwilliamcarroll.com/posts/hash-map-baseline/</guid><description>We wrote four hash map implementations, predicted how they&apos;d perform, and declared open addressing the winner. Then we changed the key distribution and watched the winner become 11,000x slower.</description><pubDate>Tue, 24 Mar 2026 05:01:00 GMT</pubDate></item><item><title>When BFS Isn&apos;t Enough: From SPFA to Dijkstra on Network Delay Time</title><link>https://www.seanwilliamcarroll.com/posts/network-delay-spfa-to-dijkstra/</link><guid isPermaLink="true">https://www.seanwilliamcarroll.com/posts/network-delay-spfa-to-dijkstra/</guid><description>I accidentally implemented SPFA while thinking I was doing BFS with relaxation, then had to build a min-heap from scratch before I could do Dijkstra properly.</description><pubDate>Wed, 18 Mar 2026 05:01:00 GMT</pubDate></item><item><title>Benchmarking Tries, Part 2: What &quot;Arena&quot; Actually Means</title><link>https://www.seanwilliamcarroll.com/posts/benchmarking-tries-part-2/</link><guid isPermaLink="true">https://www.seanwilliamcarroll.com/posts/benchmarking-tries-part-2/</guid><description>My first &quot;arena&quot; trie wasn&apos;t an arena at all — fixing that revealed real cache locality effects, disassembly surprises, and data-oriented design wins.</description><pubDate>Mon, 16 Mar 2026 05:01:00 GMT</pubDate></item><item><title>Benchmarking Tries: Arena vs unique_ptr</title><link>https://www.seanwilliamcarroll.com/posts/benchmarking-tries/</link><guid isPermaLink="true">https://www.seanwilliamcarroll.com/posts/benchmarking-tries/</guid><description>Implementing a trie led me into cache effects, const_cast, and what the compiler can and can&apos;t optimize away.</description><pubDate>Sun, 15 Mar 2026 05:01:00 GMT</pubDate></item><item><title>Benchmarking Union-Find&apos;s Two Optimizations</title><link>https://www.seanwilliamcarroll.com/posts/benchmarking-union-find/</link><guid isPermaLink="true">https://www.seanwilliamcarroll.com/posts/benchmarking-union-find/</guid><description>A completely opaque data structure turned out to be a flat array of ints with two elegant optimizations — I benchmarked each one to understand what it actually contributes.</description><pubDate>Sat, 14 Mar 2026 05:01:00 GMT</pubDate></item><item><title>Building a Sampling Profiler with Claude</title><link>https://www.seanwilliamcarroll.com/posts/building-bench/</link><guid isPermaLink="true">https://www.seanwilliamcarroll.com/posts/building-bench/</guid><description>How I built a ptrace-based sampling profiler for Linux/AArch64 — what I wrote, what Claude wrote, and where it helped and didn&apos;t.</description><pubDate>Wed, 11 Mar 2026 13:00:00 GMT</pubDate></item><item><title>Advent of Code 2024</title><link>https://www.seanwilliamcarroll.com/posts/aoc_2024/</link><guid isPermaLink="true">https://www.seanwilliamcarroll.com/posts/aoc_2024/</guid><description>How I completed all 50 stars of Advent of Code 2024 using C++20.</description><pubDate>Sat, 10 May 2025 18:10:11 GMT</pubDate></item><item><title>Wer bin ich? (Who am I?)</title><link>https://www.seanwilliamcarroll.com/posts/wer_bin_ich/</link><guid isPermaLink="true">https://www.seanwilliamcarroll.com/posts/wer_bin_ich/</guid><description>A short post about learning German — written in both German and English.</description><pubDate>Fri, 09 May 2025 14:47:27 GMT</pubDate></item></channel></rss>