A visualization of highly complex supercomputer-generated physics simulations

So How Sharp Is Sharp?

June / July 2005 By: Margaret Lovell Volume 3 Number 3

Print this Article

E-mail this Article

The extraordinary computing horsepower, commonly known as supercomputing, developed by the national laboratories in collaboration with the biggest computer manufacturers, like Cray and IBM, can have some unexpected applications. A group of artists and scientists of imaging and visualization gathered at New York University last December to begin to create a photographic system capable of capturing and displaying a gigapixel —€”one billion pixels—€”of visual information in a single image.

The first Big Picture Summit was organized by artist-photographer Clifford Ross and co-hosted by Sandia and the Interactive Telecommunications Program at NYU's Tisch School of the Arts.

Ross says his goal in bringing together top imaging experts from leading scientific institutions is to bring closer to reality his desire to create a "you are there" photographic experience for those who have not personally witnessed the sublime beauty of natural scenes.

In the early 15th century, the impulse to render flesh more realistically drove the artist Jan van Eyck to perfect the recently developed technique of oil painting. "The same sort of impulse is driving me, except that I'm trying to capture a mountain. Pixels are simply 21st century oil paint," Ross says.

The scientists have different but complementary goals. Computational scientists at Sandia National Laboratories believe a display system of the magnitude proposed by Ross will enhance the ability of its scientists to visualize and gain insight from massively complex data sets that can be understood only through human intuition. "We have a lot in common with an artist like Clifford Ross and his quest to make extremely detailed images that evoke a powerful emotional response," says Sandian Carl Diegert. "We want to understand from an intuitive standpoint what it is that enables viewers to gain insight —€” for example, a visual metaphor that makes a human viewer comfortable and thus better able to interact with an image. Computer science alone is not likely to invent a means for scientists to intuitively comprehend highly complex problems."

Ross' newly patented R1 camera system (www.cliffordross.com), which broke through the gigapixel barrier, has achieved some of the highest resolution single-shot images ever created. The quality of the first landscape images (the "Mountain" series) created with the R1 convinced many of the scientists involved in the summit to join in the effort, says Diegert.

The project could have major implications for all industries that rely on precise imaging, including environmental science, space exploration, telecommunications and homeland security, says Diegert. The project has two parts. The first is to design and build a new camera, expanding on concepts embodied in the R1, that can capture a gigapixel of digital information at a speed of 1/15th of a second or faster.

The second part is to create the display system, which Ross likens to building an "electronic Sistine ceiling." It will have 16 times greater data display capabilities than one currently in use at Sandia, which is among the world's most advanced. The display would provide an overall view of images at a very large scale while allowing viewers to perceive extremely fine detail.

As Sandia continues to bring supercomputing on line and develops a greater understanding of its capabilities, industrial, academic, and other government customers will find new and possibly imaginative ways to use this exceptional computing strength, which can process information at a lightning rate of tens of teraflops, where each teraflop is one trillion operations per second.

A photographic camera and display system being sought by artists and scientists involved in the Big Picture Summit would have 16 times greater data display capability than this display at Sandia National Laboratories, which scientists use to visualize highly complex supercomputer-generated physics simulations.

Margaret Lovell is a senior technical writer at Technically Write.