Hi, I’m Tristan 👋

Today, Anthropic open-sourced their original performance engineering take-home. The task: optimize a kernel running on a custom VLIW SIMD processor simulator. The baseline takes 147,734 cycles. Claude Opus 4.5 got it down to 1,487 cycles - a 99x speedup that beat most humans. I’m Tristan (@trirpi), and I work on AI kernels. Let’s break down how this whole system works. The Architecture at a Glance This is a VLIW (Very Long Instruction Word) SIMD (Single Instruction Multiple Data) processor with a single core (older versions of the take-home had multiple cores). Let me break down what that means. ...

Developing audio plugins can be a daunting at first. Except for knowledge of C++, one has to have knowledge of digital signal processing, GUI development. However, the first obstacle is getting to know a good plugin development framework. JUCE is pretty much the industry standard in this regard. Unfortunately, the JUCE quickstart is not very beginner friendly. It will start to explain the Projucer, a custom system that creates XCode and Visual Studio build files. Although good IDEs, many people prefer VSCode or CLion. Although not clear from the website, these are also supported via CMake. CMake markets itself as the de-facto standard for building C++ code and chances are you’ve come across it while browsing open source C++ projects. ...