INCITE collaboration boosts combustion simulation

(page 2 of 2)

“The code pretty much scales up linearly with an increasing number of processors,” says Jacqueline Chen, a combustion researcher at Sandia, a DOE lab.  “These guys have been really instrumental in assisting us with scalar and vector optimization of S3D.”

Now the Sandia researchers are turning to computer scientists to help sort through 30 terabytes of data the simulation generated.  “Conventional tools used to visualize and analyze the data simply don’t work on such large datasets,” Chen says, so they’re collaborating with Kwan-Liu Ma, computer science professor at the University of California, Davis, to develop new tools.  With Ma, they hope to create programs to automatically extract and track interesting features and to visualize results.

It’s the kind of teamwork that’s necessary if researchers hope to attack fundamental problems with major computer power.  “No one group can wrap their brains around all these different areas,” Chen says.  “Performing simulations at this scale requires expertise in combustion and turbulence as well as expertise in computer science for extracting and tracking salient regions of interest in the flow and in visualization or optimization on a particular architecture.”

« Previous       1   |   2   |   Print