high performance computing on graphics processing units: hgpu.org

The recent improvements in data-acquisition methods have resulted in the emergence of increasingly larger volumetric datasets. The design of GPU volume rendering solutions must have into account this trend while dealing with the limited available memory in a graphics card. In this work, we present a pipeline for volume rendering that stores a compressed version of the dataset in the GPU memory. During visualization, the volume is divided in bricks, which are decompressed and rendered with a specific level of resolution depending on their distance to the camera. As the main tasks of our pipeline are executed entirely on the GPU, we minimize the communication between the CPU and the graphics hardware. We obtain competitive results with other GPU implementations of compressed volume rendering, with a refresh rate between 30 and 60 FPS for volumes of size in the range between 256^3 and 512^3.