In advance of his much anticipated keynote at the upcoming EnterpriseHPC event happening in Carlsbad, California from September 7-10, Cray CEO, Peter Ungaro spoke with HPCwire’s Tom Tabor. In addition to offering a look ahead at the company’s technology and business focus for the coming year, Ungaro shed some light on what attendees can expect when he takes the stage.
Please put into context how commercial HPC has impacted Cray’s top line business. Where was it in the past and what is the foreseeable impact on revenue?
Commercial supercomputing is increasingly impacting our business. Several years ago, our company was doing almost no business in the commercial segments and many analysts we predicting that commercial needs for supercomputing were limited. The growth of big data and the expansion of the “internet of things” have brought about an increased need for supercomputing resources to analyze, more efficiently model, and interpret these growing datasets. Much of this growth has been in the commercial segments such as energy, manufacturing, life sciences and finance, just to name a few. Our growth in the commercial business over the past few years has seen more than 100+ percent annual growth and we expect that over time, we see it as increasing to be around a third of our overall business.
From a zero base point to nearly 30% of your overall business, that’s a big shift. Are their specific segments where you’re seeing greater acceptance and traction?
We are seeing traction in both traditional commercial R&D and in non-traditional Cray segments. We have has also seen commercial growth from all our product areas, including supercomputing, storage and analytics. The energy segment has been very strong, especially in selling supercomputing and storage for next generation oil and gas exploration. Some of the non-traditional segments include finance, telco and even professional sports for our analytics solutions.
You mention Big Data … is the Big Data boom playing a major role in this growth?
Absolutely. The growth in the amount and variety of data has directly impacted how commercial companies can effectively use that data to understand their products and customers and grow their businesses. The dramatic increase in the amount of sensors, customer and social data has led to an increasing need for more advanced simulation, improved machine learning, more accurate image processing and interpretation, and data discovery analytics. While many people just think about analytics when someone mentions the words “big data,” without a doubt commercial demand for supercomputing is directly linked to the big data boom. At Cray, we refer to this as data-intensive computing.
Can you give us some examples of the fastest commercial application you’re running? Or maybe better stated, give us examples of significant job efficiency improvements?
Applications performance is not just about speed, but also about size and accuracy. Commercial applications can require supercomputing-class scaling to achieve the performance demands of enterprise customers. A great example of this is the work we have done with NCSA and Ansys to improve the scaling of their Fluent computational fluid dynamics application to tens of thousands of cores, with a goal of decreasing simulation times from 25 days to 11 hours. In one simulation they achieved 80 percent efficiency of scaling to 20,000 cores on a Cray supercomputer. In the area of life sciences, a research team from the University of Chicago is using a Cray supercomputer to analyze 240 full genomes in just two days! In fact, one of the largest XC30 systems we have ever sold is at a Fortune 50 customer today – something I couldn’t have imagined 10 years ago.
With this move into the commercial sector, I’m sure you’re working on solutions for things you never thought Cray would be doing. What is the most unusual application area you’re serving?
I must say, that I am a huge sports fan. Even though my Canadian parents had me in hockey skates about 10 minutes after I could walk, baseball was my strongest sport. I am very excited that a Major League Baseball team is using a Cray system for deep analytics, basically a “Moneyball 2.0” type of application. Their application, combining the use of immense amounts of baseball statistical data to provide insights into optimal hitter/pitcher matchups, is an excellent example of how our data-intensive world is impacting our company in the area of analytics. Never in my lifetime did I think a professional sports team would have a Cray supercomputer! Of course, if I could get the Vancouver Canucks hockey team to successfully use one, I’d be a hero in my family!
Moneyball 2.0, that’s a great story. As it goes, the president of the Society of American Baseball Researchers will be speaking at the event. Will you be discussing more of these topics during your keynote at Enterprise HPC ’14?
Yes, and I am looking forward to being a part of the conference and sharing our perspective on the convergence of supercomputing and big data in the enterprise space.
We’re looking forward to seeing you there as well … thanks for your time Pete.
There is still some limited space for attendance to the Enterprise HPC summit. If you’re an end user of commercial HPC or advanced technical computing, consider requesting an invite–qualified attendees will be shipped out to Carlsbad with 5-star hotel covered for the duration of the exclusive summit. http://www.enterprisehpc.com