Parallel Computing Reveals Cosmic Riddles

"This shows what happens to gas in the solar atmosphere, or corona, where it heats up," explained Dmitruk, a research scientist with the Bartol Research Institute at UD. "Magnetic fields contribute to intense turbulence, which enhances energy dissipation, making the sun's corona much hotter than its surface."

To learn how coronal heating ultimately affects satellites in space and life on Earth, Bartol faculty member Bill Matthaeus said, computer simulations are essential. Studying activity on the sun's surface and in the corona also provides insights into the fundamental nature of space physics, plasma and the universe, according to Dmitruk.

"The sun is like a big laboratory for understanding plasma properties, which influence so many other events in space," Dmitruk added.

Unfortunately, questions about coronal heating and other complex cosmic problems can be tough or impossible to tackle with single-processor computers.

On Oct. 1, the National Science Foundation (NSF) awarded $500,000 to support a new parallel computing facility for Bartol. The "major research infrastructure" (MRI) grant will support a parallel system based on 100 linked processors, each of which will run at speeds up to 600 megahertz, connected by fast Ethernet hardware, reported Matthaeus, who will serve as principal investigator for the project.

New equipment will support simulations in the fields of astrophysics, space sciences, condensed matter physics, turbulence theory, solar physics and more, Matthaeus said. With NSF funding available through Sept. 30, 2002, Bartol/UD researchers are investigating a system similar to the Avalon-Beowulf Cluster, developed by the Los Alamos Center for Nonlinear Studies and Goddard Space Flight Center. (See http://cnls.lanl.gov/Internal/Computing/Avalon/.)

Bartol's latest award follows two other NSF grants in 1996 and 1997, which paid for a complete modernization and expansion of the institute's computing facilities. The new NSF award "will give us a truly world-class, parallel computing facility," Matthaeus said. "It should be a significant drawing card for students and faculty."

To run parallel computing jobs, researchers from many campus units--including Bartol, physics and astronomy, electrical and computer engineering (ECE), computer and information sciences (CIS) and mechanical engineering (ME)--must first develop suitable "code" or instructions for the machines, Matthaeus said. As part of his doctoral thesis work in Argentina, Dmitruk already has developed parallel code for the coronal heating simulation.

Simulations allow researchers to visualize and understand events they otherwise couldn't see because X-ray and white-light images "tell us only part of the story," Matthaeus said. "For example, there are no diagnostics available for detecting the small-scale, intense electric currents that may be responsible for heating the plasma to one- or two million degrees in the lower corona."

And, coronal heating doesn't follow conventional rules: While heat normally dissipates as it moves outward, the sun's lower atmosphere is much hotter than its visible surface, the photosphere.

On the sun, Dmitruk explained, magnetic fields confine and direct motions of plasma, which consists of protons, electrons and ionized particles. Disturbances in the corona involve interactions of the plasma and the magnetic field. These motions can't always be observed, but they can be accurately calculated using computer models.

By comparing his computer simulation of those disturbances--using a model called reduced magnetohydrodynamic (RMHD)--with data from observations and mathematical calculations, Dmitruk can better assess heating and energy loss within different parts of the corona.

The bottomline, Dmitruk said, is that "all the mixing, or disturbance in the sun's surface and corona, triggers a very rapid energy loss." This loss is characterized by a cascade of energy at smaller and smaller scales, where it is efficiently dissipated, resulting in a turbulent state. "Now," he said, "we must determine how that activity accelerates and controls the solar wind, which affects the protective bubble or magnetosphere around the Earth."

The NSF grant will promote new knowledge of a host of cosmic questions. Those slated to use the parallel system include Gary Zank, David Seckel, Stan Owocki and Dermott Mullan of Bartol; Krzysztof Szalewicz, physics and astronomy; and Lori Pollock,CIS. Shiyi Chen, a former Bartol student now with Johns Hopkins University; and Sean Oughton, another Bartol graduate with University College, London, also will collaborate with the UD/Bartol team.

Additional UD collaborators include Guang Gao, ECE; Gagan Agrawal and David Saunders both of CIS; and Lian-Ping Wang, ME.

###

To view the coronal heating simulation, see http://www.udel.edu/PR/NewsReleases/corona.html.