CEI’s parallel rendering compositor achieved a speed of 3.17 billion polygons per second using 76 standard dual-processor PCs running on a high-performance networking system. The scene used for testing contained 684 million triangles and ran at a rate of 4.6 frames per second.

CEI’s record results have major implications for engineers and scientists who increasingly desire to visualize computational problems that require very high-fidelity data, such as hydrodynamics, large-scale structural analysis, and airflow testing using finely detailed automotive and aerospace models.

“We’ve demonstrated technology that enables commercial software to run on clusters of inexpensive, commodity visualization hardware, interactively handling visualization problems so large that they could not be efficiently processed with traditional supercomputers,” says Randall Frank, senior developer for CEI. “This is important due to the rapid increase in the size of simulation models, coupled with the need for interactivity.”

CEI’s high-efficiency parallel compositor is an enhancement to the company’s SoS (server of server) capability that is already in use worldwide to visualize some of the world’s largest simulation problems. The compositor will be a part of CEI’s new DR (distributed rendering) technology that will be incorporated this fall into the company’s EnSight, EnLiten and EnVideo visualization products.

CEI’s DR line of systems will be demonstrated at Supercomputing 2005 this November in Seattle, and at CEIViz ’06 in February in Orlando.

About CEI