Big data makes supercomputers usher in a second spring

"Big Data" requires more powerful computers to handle, which is good news for Clay, October 21 news.

Clay, a supercomputer pioneer in the 1970 's, was on the verge of being eliminated 20 years ago, but now it's back to its former glory. In the past 12 months, Clay's share price has risen by about 1 time times.

The explosive growth of data and the urgent desire to understand the meaning of data require more than the processing power of standard computers. "It is widely believed that supercomputers have been eliminated 5 years ago and that ' I can finish the simulation work on my Notebook," said Barry Burding Barry Bolding, vice president of Clay. In the case of a small amount of data, this perception may be correct. However, many of the data currently applied are measured in kilobytes (equivalent to 10 of 18 bytes). Larger data volumes require more powerful computers, and more powerful computers generate more data. ”

Experts estimate that the world generates 2.5 bytes of data a day--2,500,001,000 megabytes--and a doubling of data storage capacity every 40 months, which is Clay's strength.

The basic configuration of the clay computer price of 500,000 dollars, the volume and a refrigerator basically equivalent. Big customers can assemble 200 or more of these basic units into more powerful, millions of-dollar supercomputers, such as the Titan of the US Department of Energy Oak Ridge Laboratory. Titan, which was built last year, covers an area equivalent to a basketball court and is the third-highest supercomputer in the world, capable of more than 20000 trillion operations per second.

Of course, most enterprises do not need such a strong processing power, or connect multiple computers into a computer network to complete the work of data processing, the latter is cheaper and more energy-efficient. The advantage of supercomputers is that it can complete a large number of interconnect computations and is more suitable for complex simulation and data mining.

Weather applications on smartphones, for example, are based on a large number of models that research institutions run on supercomputers, and through the use of supercomputer models, financial institutions can detect cyber fraud or security incidents in a few seconds. "Big data is emerging, but supercomputers are designed to deal with big data, which has always been a strength for clay," said Steve Conway, an analyst at IDC, a marketing research firm, Steve Camvil. ”

Cray (Seymour Cray), a "father of Supercomputing", founded Clay in 1972 and changed hands several times. At present, more than 900 employees, the market value of about 940 million U.S. dollars.

With the recent revival of supercomputers, Clay has gained Wall Street attention. This June, Clay sold a XC30 supercomputer to the European medium-term weather center. For a long time, the European medium-term weather Forecast Center has been an IBM customer.

Wall Street expects clay revenues to reach $519 million trillion this year-a 23% increase from 2012 and a gross margin of about 34%. Clay shares rose 91% in the past 12 months, with rival SGI shares up 90%. Clay P/E is 36 times times higher than SGI 19 times times.

The computer market, which is priced at more than $500,000 trillion, has been rising, from 2.7 billion U.S. dollars in 2008 to more than 1 time to $5.6 billion in 2012, IDC said. By the end of 2017, the high-performance computing market will grow by 7% a year, much higher than the enterprise server market.

Last year, Clay's two-thirds revenue was directly or indirectly related to the U.S. government. Clay is fighting for new customers who are interested in big data. Last year, Clay set up a new department called Yarcdata, the main business direction is to analyze the massive data, discover the new model. "Unstructured databases are becoming increasingly popular," Bodeen said. Solving any complex problem requires establishing a connection between datasets. ”

Yarcdata helped the US government find fraud in the medical plan. Its private clients cover medical, financial, life sciences, telecommunications and other fields.

Ragen, analyst Sid Paraca of the fund company McAdams Wright, said that Clay's new efforts had yielded results and that, over time, revenues would grow further, "supercomputers are not ordinary goods and manufacturers need time to accumulate experience." It's not hard to make supercomputers, but it's not easy to produce a supercomputer that works.

(Responsible editor: The good of the Legacy)