High performance computing (HPC) is the use of aggregated computing power to do a lot of computation, fast. Like a petaflop per second fast. It has become a competitive advantage for organizations in fields where processing more data, quickly really matters – fields such as financial services, scientific and medical research, and energy exploration.
To out-compute is to out-compete
In a worldwide IDC study, 97 percent of companies that had adopted high performance computing said they could no longer compete or survive without it. Suzy Tichenor and Albert Reuther explain: “More-capable HPC resources often translate into faster time-to-market (in some cases more than 50 percent faster), reduced costs, and superior product quality.”
For certain types of activities, high performance computing can be a powerful competitive advantage. Here are five:
High performance computing for high frequency trading
One of the best-known high performance computing use cases is high frequency trading. HPC gives traders the capability to execute trades within microseconds or milliseconds (that is extremely low latency). High frequency trading (HFT) has made billions for the firms that do it – exploiting latency advantages has been estimated to account for $21 billion in profit per year. Now regulators are looking to high performance computing to monitor HFT. In 2011, the Securities and Exchange Commission proposed the development of systems to monitor high frequency trading in real time and provide early warnings for market anomalies such as those that preceded the Flash Crash in 2010.
HPC for fraud analysis
Another use case for high performance computing in the financial sector is fraud analysis. Consider, for example, PayPal. The company was managing 13 million financial transactions per day and needed to detect fraud in real time as millions of transactions were processed between disparate systems; find suspicious patterns in related data sets; and create and deploy new fraud models into event flows quickly and with minimal effort. The company built an HPC environment and implemented an HPC data analysis program andreportsfirst-year savings of $710 million in detected fraud that it wouldn’t have been able to detect before.
HPC for genomic sequencing
High performance computing is also used in medical research applications such as genomic sequencing. Researchers at the Translational Genomics Research Institute(TGen) in Phoenix, for example, rely on cutting-edge HPC technology for genomic sequencing in their search for cures for rare diseases. High performance computing allows the researchers to get insights and results more quickly. James Lowey, TGen’s vice president of technology, explains: “Without that high-performance computing infrastructure, the time to discovery is just too long. If you tried to do some of these types of things on your desktop computer, it can take weeks.”
HPC for computational fluid dynamics
Organizations in many fields rely on computational fluid dynamics (CFD);Boeing is one example. The aircraft manufacturer uses CFD to optimize aerodynamic design, which is an important driver of fuel efficiency for airplanes. High performance computing has enabled faster set-up and run times, increased capability, and improved accuracy of Boeing’s computational fluid dynamics modeling. The results: more efficient aircraft with less need for wind tunnel and flight testing.
HPC for seismic modeling
In the energy sector, high performance computing is used for seismic modeling. In a new oilfield, for example, seismic modeling tells the company where to drill to maximize output and minimize cost and risk. In existing oilfields, seismic modeling can help the company know which wells can yield more oil. As one IT analyst told us, energy companies use a lot of computing power to get the last drops out of their assets in a given field.
For these five use cases – in which high performance computing capabilities are a matter of competitive advantage – the capacity of the data center to support such capabilities is no longer “just” an IT issue. It’s a competitiveness issue.
IO colocation data centers support these high performance computing use cases, and many more. Highly efficient and capable of supporting densities up to 30 kW per rack, a modular data center can be just the environment an organization needs to out-compute, and out-compete.Learn more.