Distributed Out-of-Core Approach for In-Situ Volume Rendering of Massive Dataset

Abstract

This paper proposes a method that allows a fluid remote interactive visualization of a terabytes volume on a conventional workstation co-located with the acquisition devices, leveraging remote high performance computing resources. We provide a study of the behavior of an out-of-core volume renderer, using a virtual addressing system with interactive data streaming, in a distributed environment. The method implements a sort-last volume renderer with a multi-resolution ray-guided approach to visualize very large volumes of data thanks to an hybrid multi-GPUs, multi-CPUs single node rendering server.