Sign up for our newsletter and get the latest HPC news and analysis.

Cornell Medical College's high resolution CAVE

Last week visualization display company Christie announced a new CAVE installation they’ve completed with the Weill Cornell Medical College in New York City.

The David A. Cofrin Center for Biomedical Information at the HRH Prince Alwaleed Bin Talal Bin Abdulaziz Alsaud Institute for Computational Biomedicine is powered by eight Christie Mirage 3-chip DLP projectors with active stereo capability.  The Christie projectors deliver a resolution of 1920×1920 (3.68 megapixels) per wall – 334% higher resolution than most previous CAVEs.  The results are superior 3D images that set new standards in molecular modeling and other avenues of biomedical research.

CAVE imageA CAVE (which is short for Cave Automatic Virtual Environment according to the Wikipedia) is an immersive 3D environment with images projected on up to 5 walls of a cube (the “cave”) that a user steps in to explore a virtual world. In the image at right, the CAVE user has stepped into the environment (facing away from you) and is interacting with a dataset. The user typically wears at least stereo glasses, and may also wear a dataglove, or carry a wand or other interaction device.

Vanessa Borcherding, Systems Administrator for the Institute for Computational Biomedicine, Weill Cornell Medical College, remarked, “ The ICB is focused on enabling computational methodologies in the biological sciences, including genome studies, molecular modeling and modeling of cell and organ systems.  The Cofrin Center’s Facility allows us to take many pieces of data from a computer and reconstruct it in an immersive 3D environment to help researchers make decisions and gain insights quickly and intuitively.”

Borcherding added, “Pixel density is key to visualizing the vast amount of data we need to analyze. We chose Christie because they were the only company to propose a genuine high definition CAVE solution, which no one else could offer.  We were especially impressed with their forward-thinking, innovative approach, their expertise in building visualization facilities, and their ability to perfectly integrate this technology to meet not only our technical needs, but our rigid engineering requirements as well.”

Anyway, as you can see from the image the walls of the CAVE are pretty big, and the projectors used in CAVEs have been historically fairly low resolution, making them a good solution for datasets with a strongly spatial component or complex spatial relationships if you were willing to accept some smoothing. Of course, as datasets continue to grow along with the HPC machines that produce them, this tradeoff is increasingly undesirable. Thus the relevance of the Christie announcement. 3.68 MP per wall is a nice improvement, but of course we are going to need much higher resolution very soon as trans-petaFLOPS machines become more common.

Resource Links: