Composable function systems as a general-purpose rendering framework
2026-06-01 • Graphics
Graphics
AI summaryⓘ
The authors propose a new way to create and control complex visuals and simulations using something called function systems, which need very little memory on GPUs or CPUs. They introduce Quibble, a tool that helps programmers build these visuals directly on a GPU. Their method offers fast performance, can create complex shapes, and works well with other graphics techniques. It also helps artists manage tricky animation tasks, like smoothing frames and shaping point clouds.
function systemsGPUCPUmetaprogrammingvisualizationssimulationspoint cloudslow-framerate animationstopologically non-trivial objectsanimation workflows
Authors
James Schloss
Abstract
Function systems exist as a natural language for the meshless creation and manipulation of complex objects while maintaining minimal memory on the Graphics Processing Unit (GPU) or Central Processing Unit (CPU). This paper proposes a new method for general-purpose (non-fractal) visualizations and simulations with function systems and introduces Quibble, a metaprogramming framework for composing such systems on the GPU. We also discuss several core advantages of this method including runtime performance, the creation of topologically non-trivial objects, and interoperability with other graphical algorithms. Beyond general-purpose imagery and animations, this method can also be used to give artists more control over in-between frames in low-framerate animations, controllably deform point clouds, and metaprogram difficult animation workflows.