Supercomputing ages ‘like fine wine’

Major League Baseball, potential cures for cancer and internet search engines all have something in common. They are influenced by big data and its application in analytics.

WSU alumnus Peter Ungaro, who graduated in 1990, spoke about this influence and more during a lecture titled “Fusion of Supercomputing, Scientific Applications and Big Data” yesterday afternoon in the CUB Junior Ballroom.

The presentation kick-started the Creighton Distinguished Lecture series, which is sponsored by the WSU Institute of Shock Physics.

Yogendra Gupta, the director of the institute, invited Ungaro to share what he has learned since 2003 as the president and CEO of Cray Inc.: “The Supercomputing Company.”

“Cray is the best-known supercomputing company,” Gupta said. “It invented supercomputing in the ‘70s.”

 “Like fine wine,” Ungaro said, “some things improve with age.”

When Cray first started, the company developed every component of its supercomputers. Today, it develops some components that can be used with other hardware.

In the 1990s, supercomputing was a huge component of intellectual, military and nuclear application. Currently it is used primarily for scientific analysis and research purposes, Ungaro said.

“From our perspective, we want to model things,” he said. “The complexity of the algorithms is changing the architecture of computers in response to big data.”

Big data is a big player when it comes to analytics, which is applied in every field that relies on science for improvement.

Jack Creighton, former CEO of the UAL Corporation which runs United Airlines, acknowledged how useful this technology would have been in his past positions.

“We could have used computers with the lost bags,” he said.

Creighton has served as the chairman of the WSU Foundation and is one of the people this lecture series was named.

Ungaro described his field of study as a needle in a haystack. Data is the haystack. A question is the needle.

Corporations like Yahoo and Google contain their servers in large warehouses where systems are often stored on eight-wheelers or on the floor, Ungaro said. They store a large amount of data but obtain it at a lower speed.

Supercomputers are intended to simplify the extraction process by providing more confined servers. They supply fast data from the big data at the expense of being more costly and aesthetically pleasing, Ungaro explained.

“Supercomputers are very pretty,” Ungaro said, “unlike Google and Yahoo’s warehouse systems.”

Cray has assisted in providing national security and threat analytics, information about Medicaid and Medicare fraud and consumer information for Virgin Media.

The same methods he used in those fields have been used to link developing data with future data during research of new cancer treatments.

Researchers need their technology to adapt with their methods of study. As they work, they take data and plug it into their machines. Once the information has been submitted, the researchers can form a hypothesis, test it and resubmit it for further study.

This process is used in professional athletics as well, influencing the box score and play-by-play procedures of baseball. These analytics have led to a new area of study called PITCHf/x and HITf/x.

For each PITCHf/x researchers analyze 20 pieces of data, including the rotation and angle of the pitch. For each HITf/x they analyze five pieces of data that include the style and quality of both hit and pitch.

These analytics are generally used for video games developed by Xbox and Playstation. Ungaro said Cray is working with the first MLB team to apply this information on the field. He declined to name the team.

The company is pushing to take the technology responsible for these developments and create total integration between the methods of big and fast data retrieval, Ungaro said.

Update: Three paragraphs in this story were reworded to accurately identify Jack Creighton, and a quote was properly attributed to Ungaro.