Section: New Results
Out-of core visualization of structured grids
We have presented in  and  a volume roaming system dedicated to oil and gas exploration. Our system combines probe-based volume rendering with data processing and computing. The daily oil production and the estimation of the world proven-reserves directly affect the barrel price and have a strong impact on the economy. Among others, production and correct estimation are linked to the accuracy of the subsurface model used for predicting oil reservoirs shape and size. Geoscientists build this model from the interpretation of seismic data, i.e. 3D images of the subsurface obtained from geophysical surveys. Our system couples visualization and data processing for the interpretation of seismic data. It is based on volume roaming along with efficient volume paging to manipulate the multi-gigabyte data sets commonly acquired during seismic surveys. Our volume rendering lenses implement high quality pre-integrated volume rendering with accurate lighting. They use a generic multimodal volume rendering system that blends several volumes in the spirit of the ``stencil'' paradigm used in 2D painting programs. In addition, our system can interactively display non-polygonal isosurfaces painted with an attribute. Beside the visualization algorithms, automatic extraction of local features of the subsurface model also take full advantage of the volume paging.
Sort-last parallel rendering is an efficient technique to visualize huge datasets on COTS clusters. The dataset is subdivided and distributed across the cluster nodes. For every frame, each node renders a full resolution image of its data using its local GPU, and the images are composited together using a parallel image compositing algorithm. In this paper, we present a performance evaluation of standard sort-last parallel rendering methods and of the different improvements proposed in the literature. This evaluation is based on a detailed analysis of the different hardware and software components. We have presented in  a new implementation of sort-last rendering that fully overlaps CPU(s), GPU and network usage all along the algorithm. We present experiments on a 3 years old 32-nodes PC cluster (see Figure 6 -A) and on a 1.5 years old 5-nodes PC cluster, both with Gigabit interconnect, showing volume rendering at respectively 13 and 31 frames per second and polygon rendering at respectively 8 and 17 frames per second on a 1024×768 render area, and we have shown that our implementation outperforms or equals many other implementations and specialized visualization clusters. We have also experimented applications to CAD/CAM visualization (Figure 6 -B), integrating results from R. Toledo's Ph.D thesis on efficient visualization of higher-order primitives on the GPU.
Protein docking is a fundamental biological process that links two proteins. This link is typically defined by an interaction between two large zones of the protein boundaries. Visualizing such an interface is useful to understand the process thanks to 3D protein structures, to estimate the quality of docking simulation results, and to classify interactions in order to predict docking affinity between classes of interacting zones. Since the interface may be defined by a surface that separates the two proteins, it is possible to create a map of interaction that allows comparisons to be performed in 2D. We have presented in  a highly efficient algorithm that extracts an interface surface and creates a valid and low-distorted interaction map. Another benefit of our approach is that a pre-computed part of the algorithm enables the surface to be updated in real-time while residues are moved. Applications of this method to biological problems have been published in  ,  . Figure 7 shows our tools used in the reality center.