<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Abdul Basith</title><description>Notes on systems, performance, and the craft of building software.</description><link>https://abdulbasith.com/</link><item><title>When Doubling the Math Costs Nothing: Overhead-Bound Inference in Robot Models</title><link>https://abdulbasith.com/blog/overhead-bound-inference/</link><guid isPermaLink="true">https://abdulbasith.com/blog/overhead-bound-inference/</guid><description>I switched a robot model from 16-bit to 32-bit and it didn’t get slower. That one result explains why small robot models don’t speed up on bigger GPUs, why lower precision can hurt, and why the same GPU ran 1.65× slower on different machines.</description><pubDate>Sat, 21 Nov 2026 00:00:00 GMT</pubDate></item><item><title>Thirty-two threads, one instruction</title><link>https://abdulbasith.com/blog/thirty-two-threads/</link><guid isPermaLink="true">https://abdulbasith.com/blog/thirty-two-threads/</guid><description>A ground-up introduction to CUDA for programmers who have never written a kernel, where every number was measured on a real GPU.</description><pubDate>Fri, 09 Oct 2026 00:00:00 GMT</pubDate></item></channel></rss>