<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Ethan Puyaubreau · Writing</title><description>Write-ups on high-performance computing, GPU computing, infrastructure, and the projects behind them, by Ethan Puyaubreau.</description><link>https://ethan-puyaubreau.github.io/</link><language>en-US</language><item><title>Tearing the GPU node down from Proxmox to bare Debian</title><link>https://ethan-puyaubreau.github.io/blog/gpu-debian-nvidia</link><guid isPermaLink="true">https://ethan-puyaubreau.github.io/blog/gpu-debian-nvidia</guid><description>The GPU node ran Proxmox. I wiped it for bare Debian 13 so the GPU would sit directly under the host kernel, then spent the evening in the NVIDIA driver gauntlet Trixie hands you. The part that caught me was Secure Boot. (It has since gone back to Proxmox.)</description><pubDate>Thu, 18 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Charging GPU energy to the code that spent it</title><link>https://ethan-puyaubreau.github.io/blog/kokkos-gpu-energy</link><guid isPermaLink="true">https://ethan-puyaubreau.github.io/blog/kokkos-gpu-energy</guid><description>A profiler tells you where GPU code spends time. I wanted to know where it spends joules, so I built a Kokkos Tools connector that samples power on a side thread and integrates it over each profiled region. On ArborX DBSCAN, the faster implementation saves more energy than time: 19% less time, 25% less energy.</description><pubDate>Fri, 12 Jun 2026 00:00:00 GMT</pubDate></item><item><title>A galaxy in a browser tab, computed on your GPU with WebGPU</title><link>https://ethan-puyaubreau.github.io/blog/nbody-webgpu</link><guid isPermaLink="true">https://ethan-puyaubreau.github.io/blog/nbody-webgpu</guid><description>A real-time gravitational N-body galaxy, a brute-force O(N²) sum recomputed every frame in a WebGPU compute shader. 65k bodies, no server, no precompute.</description><pubDate>Tue, 09 Jun 2026 00:00:00 GMT</pubDate></item></channel></rss>