Divide-and-Conquer 3D Convex Hulls on the GPU

TL;DR

This paper presents a GPU-accelerated divide-and-conquer algorithm for computing 3D convex hulls, achieving significant speedups over CPU methods.

Contribution

It introduces a novel parallel algorithm for 3D convex hulls optimized for GPU hardware, demonstrating improved performance.

Findings

Significant speedup over CPU implementations

Effective parallelization of divide-and-conquer approach

Successful implementation on GPU hardware

Abstract

We describe a pure divide-and-conquer parallel algorithm for computing 3D convex hulls. We implement that algorithm on GPU hardware, and find a significant speedup over comparable CPU implementations.