Mount Moran Supercomputer Helps UW Scientists Reach Peak Performance

(NOVEMBER 04, 2015) — Grab a pencil and write the number 150. Then add 12 zeros. As in trillions.

That’s how many calculations per second the University of Wyoming’s Mount Moran supercomputer can use to crunch problems like how life forms evolve, modeling the Snowy Range’s water circulatory system and how electrical transmission line capacity is affected by to what region electricity is sent.

That research, being conducted for the U.S. Department of Energy, has more than 530 million pieces of information stemming from hourly readings for five years from about 180 electrical production sites in the Rocky Mountain Power area.

Feel intimidated? Don’t. In 2014, the fastest supercomputers in the world still took more than 40 minutes to simulate one second’s activity in a human brain.

Still, supercomputers are downright handy when munching mountains of data, like the U.S. Department of Energy project. That research is a collaboration among Roger Coupal in the College of Agriculture and Natural Resources, faculty members in the College of Engineering and Robert Godby in the College of Business.

They are studying all electricity production outlets and sources – wind, solar, coal, hydro – and then predicting what happens to prices, costs and production when transmission capacity is increased in a certain direction, to Denver and the Front Range, for example.

The team’s faithful, fast desktop computer took up to five days to chug through the program and, if there were an error, the team had to run the entire program again, which could take another five days.

“It was taking way too long on a fast, single computer because we program the simulation model, then we see what happens,” said Coupal, head of the Department of Agricultural and Applied Economics. “If we missed something, we had to figure it out and then wait from 18 hours to up to five days. Tweaking, fixing the model if you see problems, takes an incredibly long time to do.”

Mount Moran takes about 20 minutes.

Coupal’s team plans to next put the data into a large regional economic model that encompasses the economies of Colorado and Wyoming.

“If it’s going to take a week for every run to debug the simulation model, it could take months,” said Coupal. “Mount Moran turns this around quickly.”

About 30 UW research teams use the supercomputer, said Dane Skow, director of the Advanced Research Computing Center in Information Technology, and probably Mount Moran’s chief advocate and shepherd.

The ARCC’s potent data munching power sitting on the floor below Skow’s office is a reality because the university made supercomputing a priority and the Wyoming Legislature invested in the infrastructure, he said.

Skow noted Wyoming’s situation – scientists having access to the high-performance National Center on Atmospheric Research’s Yellowstone supercomputer and Mount Moran – is remarkable for such a sparely populated state.

Yellowstone is rated 50th fastest in the world.

“Staying at the top of the game takes continuous improvement and steady investment,” Skow said. “This is a marathon, not a sprint. I think that this is one of the gems of the university we really ought to make sure everybody understands and knows is available.”

Clocking supercomputer speeds are like new car tests: professional drivers on simulated courses – machines perform a standard test without any other program running to affect performance.

Mount Moran hums along normally at about 80 percent.

“We have run up to 99 percent,” said Skow. “But when that happens, it’s kind of like congestion on a freeway. Everything tends to bog up. The least disturbance causes problems.”

Run less than 80 percent and the calculating power Mount Moran brandishes is wasted.

Research is evolving and oftentimes influenced by what scientists think they can do today with the resources they have, he said.

“The home run for us is when we can help people do things they didn’t think about before because they didn’t think it was possible, by creating a capability they didn’t know they had,” said Skow. “That’s the big payoff. That’s our Nobel Prize.”