Lossless Compression and Decompression on Graphics Hardware

Bachelor-Thesis

Data compression techniques are omnipresent in all areas of Compute Science. They reduce storage space and save bandwidth. Applications running on graphics hardware can benefit from both: Firstly, GPUs posses limited memory capacities and the required data can oftentimes not be held entirely in main memory. Secondly, GPUs exhibit limited bandwidths rates between the host system and graphics hardware, as well as on graphics hardware. Hence, lossless compression techniques have the potential to alleviate these limitations. The goal of this thesis is to investigate the suitability of standard lossless decompression for geometry data (vertex data and attributes) on graphics hardware using OpenGL, OpenCL, and/or CUDA.

Advisor(s)