GPU benchmarks based on strange attractors

Abstract

The main purpose of presented GPU benchmark is to generate complex polygonal mesh structures based on strange attractors with fractal structure. Attractors have to be created as 4D objects using quaternion algebra. Polygonal mesh can have different numbers of polygons because of iterative application of this system. The major complexity of every mesh would provide efficient results using multiple methods such as ray-tracing, anti aliasing and anisotropic filtering to evaluate GPU performance. Our main goal is to develop new faster algorithm to generate 3D structures and apply its complexity for GPU benchmarking.