IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

UC San Diego to House 'Comet' Supercomputer

Comet is a landmark project made possible by the National Science Foundation, which gave the center $24 million to build and run the machine.

SDSC Center
The San Diego Supercomputer Center
(TNS) -- What would you get if everyone in the world simultaneously began doing 250,000 math calculations per second?

You’d come up with the computing power of Comet, a new machine that UC San Diego will use to open the world of supercomputing to a much larger mix of the nation’s scientists.

As many as 10,000 researchers a year will use Comet, the latest supercomputer to be added to the university’s renowned San Diego Supercomputer Center, which will celebrate its 30th birthday Wednesday.

The machine, which is roughly the size of two school buses, will tackle questions such as: What impact does high-frequency trading have on financial markets? What happens to brain cells during an epileptic seizure? How do earthquake faults behave over long periods of time?

“Historically, supercomputers have mostly been used by physicists, engineers and others from the hard sciences”, said Robert Sinkovits, director of scientific computing applications at the center.

“Those researchers will use Comet. But we’re also expanding into the social sciences, finance and ecology. They have good data sets and the algorithms to analyze them,” he said. “We want to serve the long tail of science.”

Comet is a landmark project made possible by the National Science Foundation, which gave the center $24 million to build and run the machine. It is seven times more powerful than Gordon, a supercomputer that the center introduced only three years ago.

As technology improves, the center periodically replaces or supplements older machines with newer ones.

Comet is coming online amid a huge and growing demand for high-performance computers, especially in the life sciences. Advances in technology are enabling scientists to generate and analyze extraordinary amounts of information.

One of the biggest advances came from San Diego’s Illumina, which slashed the cost of sequencing a person’s genome to $1,000 from $10 million in less than a decade. The company’s machines are so powerful that just one of them could do all of the sequencing that was done by the national laboratories during the Human Genome Project.

Genetic sequencers are revolutionizing personalized medicine.

J. Craig Venter of La Jolla, the famed biologist who played a major role in the Human Genome Project, created Human Longevity Inc. to sequence hundreds of thousands of people’s genomes to help scientists explore patterns in disease and aging.

On Tuesday, Human Longevity is scheduled to introduce Human Nucleus, a platform for doing personalized medical work-ups that include everything from examining a person’s microbiome to sequencing his or her genome.

The microbiome is the collection of microbes — both good and bad — that live on people’s skin and in their mouths and guts.

Many scientists are studying the DNA of these microbes in hopes of finding ways to predict a person’s risk for developing disease.

The field is led by Rob Knight, who was recruited to UC San Diego this year to expand the university’s work on the microbiome. Knight said he took the position in part because of the school’s strength in supercomputing.

His research produces almost unfathomable amounts of data.

“You’d need 1,000 DVDs to store all of the genetic information contained in a single teaspoon of fecal matter,” said Knight, who came from the University of Colorado.

Such work fuels an intense demand for computing power.

“We’d have plenty of subscribers if we could build a second Comet,” Sinkovits said.

Larry Smarr is even more blunt.

“The demand for these kind of computers just seems to be inexhaustible,” said Smarr, a UC San Diego physicist and computer scientist. (Smarr profile).

Smarr would know. He played a key role in creating a worldwide market for supercomputers, which he refers to as “time shrinkers” because of their ability to rapidly process data.

In 1983, when Smarr was a young faculty member at the University of Illinois, he co-wrote a proposal that asked the National Science Foundation to underwrite a network of supercomputers that would be placed at universities across the country.

Many scientists considered Smarr’s idea to be audacious and unlikely to be approved. But the foundation agreed to establish supercomputing centers at a handful of sites, including UC San Diego, which opened its center in 1985.

At the time, consumers knew little about such machines. But they were awakening to the promise of the personal computers, partly due to the boldness of Apple. The company introduced the country to its MacIntosh personal computer in January 1984 by airing the now-classic “1984” TV ad during the Super Bowl. The message: Save humanity from conformity.

Apple helped to ignite the personal computer revolution, which was later stoked by the rise of the Internet and, more recently, by the runaway growth in social media and smartphones. Consumers also are snapping up wireless, wearable sensors such as Fit, a personal activity tracker.

Seemingly everything that can be measured is being measured, from people’s expressions in computer dating videos to political rhetoric on social media to the flight patterns of California condors and the burn patterns of wildfires.

The boom in data processing represents a blessing and curse for scientists.

Warren Im, a University of Kansas researcher who uses Comet to study cells and membranes, said: “This computer is so powerful, it lets us explore questions that we might not have been able to ask.”

The downside? Storage.

“We’ve seen basic researchers learning Linux and cloud computing to keep up with it all,” said Aaron Chang, director of the Center for Computational Biology and Bioinformatics at UC San Diego. “I’ve had people show up at my door with USB drives filled with data. It has to be pushed on to the cloud so that it is accessible anytime, anywhere.”

Smarr will soon have some fresh data to add to the cloud. He’s collaborating with Knight on a project that examines the microbes of hundreds of people. The number crunching will be handled by Comet, which hums away, day and night, in a corner of the supercomputing center.

“It’ll take Comet a few months to do the work,” Smarr said. “On a single modern desktop computer, operating around the clock, it would take a century.”

©2015 The San Diego Union-Tribune Distributed by Tribune Content Agency, LLC.