Tayfun Kayhan

Triggered by ongoing AI boom and discussions surrounding it with regards to the nature of intelligence, consciousness and the mind, I made a silly little program (source at the end) over the weekend which, given an input image, tries to guess the exact va…

Motivated by Federico Ponzi’s cool project Bare Metal Space Invaders, in which he’s implemented the classic Space Invaders in Rust on Raspberry Pi 3 system with minimal OS/kernel interaction, I implemented the game on my toy SoC Accellant to conclude the…

Ever since I came across Tom Verbeure’s impressive project in which he’s implemented a real-time Whitted ray tracer on an FPGA, I had a burning desire to have a crack at something similar while adding a thing or two new on top of the idea.

Feeling a lack of intuition about how a fixed-point implementation of a naturally floating-point algorithm would fare against native floating-point performance on modern GPUs, I set out to do a few simple experiments.

I’ve been working with Vulkan API for quite a while now, so I wanted to compile a non-exhaustive list of things that I find are bad practices with regards to performance, functionality and validity for Vulkan workloads.

I finally took the opportunity that comes disguised as a lazy Sunday to write down a quick’n dirty post on how one could convert a super simple scalar function to SIMD for those who want to experiment with SIMD but feel intimidated by all those nasty-look…

If you followed along or just came across my Rasterization in One Weekend blog series without having actually compiled and run the demo workloads, you’d be in for a big surprise if I told you now how slow they turned out to be.

Code for this part can be found here.

Code for this part can be found here.

Code for this part can be found here.