DEFENSE

Next week at SC|05, the annual supercomputing show, CEI will introduce EnSight DR, the first commercial visualization application that brings the power of parallel graphics to the user's desktop. "Up until now, only expensive supercomputers have been able to efficiently render extremely large data sets," says Dale Southard, a visualization hardware architect at Lawrence Livermore National Laboratory. Lawrence Livermore, Sandia and Los Alamos National Laboratories helped CEI develop its DR methodology in support of the Department of Energy (DOE) National Nuclear Security Administration's (NNSA) Advanced Simulation and Computing Program. "The availability of a commercial visualization application that harnesses off-the-shelf hardware for visualizing large scientific datasets containing tens of millions and even billions of polygons is critical to the success of our national security work," says Southard. CEI's EnSight DR uses distributed rendering (DR) technology that automatically distributes polygon data into pieces for balanced rendering on multiple graphics processing units (GPUs) running on networked PCs. Once images are rendered, EnSight DR composites the pieces into a single image. Large scenes, high frame rates EnSight DR expands on CEI's SoS (server of servers) capability in EnSight Gold software, used by leading automotive, aerospace and research labs worldwide to visualize some of the world's largest simulation problems. The new software provides a parallel, distributed rendering client that enables users to fully exploit the distributed memory capacity of networked PCs containing high-performance graphics cards. This gives users the ability to visualize the largest datasets on their desktops at interactive rates. EnSight DR also provides distributed, cluster-based rendering for planar multi-tile displays. EnSight DR includes new features for dataset decomposition and dynamic scene distribution to cluster nodes. These features, coupled with a scalable, high-performance image compositing engine and parallel-rendering control, enable EnSight DR to handle datasets and display sizes well beyond other commercial and even research applications. EnSight DR also supports Chromium-based parallel rendering, providing additional flexibility in cluster-computing environments. "The beauty of EnSight DR is that the distributed rendering is transparent to the user - that's a big leap forward," says Kent Misegades, CEI's president. "This is definitely a killer app for those who want to leverage supercomputing-like graphics power on low-cost clusters." Scalability opens visualization frontiers EnSight DR's ability to maintain scalability over a high number of clustered GPUs is highly valued by leading hardware system vendors. "With EnSight DR, CEI has delivered a package that can take full advantage of HP's Scalable Visualization Array," says Winston Prather, vice president of high-performance computing at HP. "EnSight DR accelerates rendering of the vast datasets produced by HP's XC cluster technology, giving users an unprecedented opportunity to visualize data in context." "EnSight DR truly leverages the capabilities found in our high-end visualization platform, which features the NVIDIA nForce Professional paired with NVIDIA Quadro graphics," says Ian Williams, senior applied engineer at NVIDIA. "This combination delivers a new level of application performance on scalable clustered hardware systems." "EnSight DR coupled with a Linux Networx Supersystem enables customers to immediately focus on solving their most important visualization challenges," says Ben Passarelli, vice president of product marketing for Linux Networx. "Users will be able to see and understand engineering and scientific data in new ways while enjoying the systems experience of a Linux Networx Supersystem." CEI's full line of products for extreme visualization and meshing will be shown next week in Booth 1300 at SC/05. The company's products and applications will also be on display in the following booths: HP (#1608), Linux Networx (#1719), SGI (#602), NASA (#1810), Los Alamos National Laboratory (#312) and IBM (#1510).