When is the next monumental breakthrough in science and where is it coming from?  Engineers, researchers, analysts and scientists have been using supercomputers and high performance computing (HPC) for decades with ever-evolving degrees of progress.  Recent advancements in HPC have positioned many of these domain experts on the very edge of making breakthroughs in physics, chemistry and biology a weekly or even daily occurrence.

When creative brains are empowered with innovative tools, amazing discoveries occur.  Although advanced computing has been around for quite some time, it is all relative, and science has been restricted in the ability to run rich simulations that accurately model the real world. Previous generations of computing technology have always been relatively handicapped by the amount of data that they can process and the time that it takes to complete the job, so although we evolve, our results always benefit from advances in data resolution, accuracy and performance. 

Even if a compute engine was powerful enough in the past to gobble up all the data injected, it might not be economical to tie up an expensive resource to wait for the results. HPC system directors and administrators constantly have to make tough decisions on which jobs to allocate to their machines to balance the value of the desired results with the opportunity lost to addressing a different problem or set of problems in the same amount of compute time. Now that the world’s faster supercomputers are operating at sustained petascale performance and can scale to better handle highly parallel computations on massive datasets, the resolution is getting finer, and the results are getting better. Better system performance means more calculations per time period or more accurate simulations faster. The result – an acceleration in the rate of scientific discovery.

One great example of the benefits of these HPC enhancements is the worldwide interest in timely and accurate weather/climate prediction. Humans on every continent use weather forecasts to plan their work and play, to the point that we demand them updated every couple of hours. The enormous datasets processed by today’s faster supercomputers enable better prediction on a finer, more detailed regional space, as well as reproducing the calculations quickly enough to respond to quick changing climate events.

Beyond the hourly weather data collection and re-calculations, supercomputers are used to hunt for extreme weather events, with scientists helping automate the search for disturbances like hurricanes in huge datasets. “We’re using state-of-the-art methods in data mining and high performance computing to locate and quantify extreme weather phenomena in the very large datasets generated by today’s climate models,” says Prabhat, a scientific visualization expert in Berkeley Lab’s Computational Research Division. “We want to help answer the question: How will climate change impact the frequency of extreme weather?”

Prabhat, and others at Berkeley Labs as well as Oak Ridge National Laboratory, are using the AMD Operton^TM processor powered ‘Hopper’ Cray XE6 supercomputer stationed at the National Energy Research Scientific Computing Center (NERSC) to automate the analysis of enormous climate simulation datasets to track the weather events that build and contribute to extreme events like heavy precipitation and cyclones. You can learn more by reading the case study: ‘Hopper’ Helps Speed ID-ing of Extreme Weather Events

Additionally, adaptive supercomputing advancements, where we build a very scalable, energy efficient system interconnect and infrastructure, have also led to the ability to integrate different processor technologies so that we can have the right processor for the right application. This hybrid approach optimizes the computing capabilities to deliver the ultimate balance of traditional high performance CPU architectures and enormously parallel accelerator engines or coprocessors. On large scientific codes there are invariably sections that execute well on sequential processors while other functions can benefit from the performance acceleration of offloading to a highly parallelized coprocessor.

Fluid dynamic simulations enable a unique way to study, predict and ultimately improve the behavior of ‘soft matter’.  Large scale experimentation can be costly, time consuming or even impossible in some scientific domains, so modeling behavior on supercomputers can provide breakthroughs in chemistry or physics that we might not have addressed otherwise for decades to come. Understanding and controlling the phase separation of liquid mixtures can assist in numerous earthly applications – imagine the practical benefits of improving mixtures and suspensions like engine lubricants, paints, chemicals and even ice cream.

“A crucial factor in obtaining accurate results is the size of the physical system that we can simulate,” says lead ‘Ludwig’ code author Kevin Stratford of the Edinburgh Parallel Computer Centre (EPCC). “To model complex problems in large systems, we need efficient, scalable application performance over large numbers of nodes.”

Researchers ran the “Ludwig” parallel computing code to simulate soft matter systems on the Jaguar Cray XT5 supercomputer at Oak Ridge National Laboratory, and later scaled the investigation higher, running on a 10-cabinet hybrid Cray XK6 system, a supercomputer which integrates AMD Opteron processors with many-core accelerator technology. Together, over the last 12 years, EPCC and the University of Edinburgh have collaborated to advance soft matter system simulation. The recent advances in scalability and, increased parallel computing power without compromising intercommunication performance, have enabled the collective team to accelerate their rate of scientific investigation.

 “Large-scale computer simulation has become a central tool for research in materials physics,” says University of Edinburgh professor and head for the Soft Matter group Michael Cates. “And thanks to the Cray XK6 system and other similar innovations, the rate of increase in computational capability is breathtaking.”

The case study here can provide additional insights: The Science of Smooth Ice Cream

It’s not just researchers and scientists, labs and universities making the scientific breakthroughs either. Commercial industry is also using high performance computing to push the envelope…literally. Boeing engineers have used the Jaguar Cray supercomputer at ORNL to validate aerodynamic codes for airplane design, specifically investigating lift during various takeoff and landing scenarios.  This results in a faster, cheaper and safer outcome than building numerous full sized prototypes in the commercial and military airframe world.

“Jaguar provided us with a unique opportunity to learn that it’s possible to achieve completely converged solutions to steady state computational fluid dynamics equations and about high lift aircraft configurations in as little as 2 to 4 hours of wall clock time, saving an order of magnitude over our experiences from use of smaller-scale systems,” said Boeing’s John Bussoletti. “Now we’re confidently exploring what impact that can have on our airplane development cycle, both in terms of performance improvements and cycle time reduction.”

Boeing uses this research to model airplanes more accurately with the end goals to make them safer and stronger as well as increase their fuel efficiency. The company validated and improved several aerodynamics codes, saving Boeing both time and money. Experimental techniques such as wind tunnels require the use of mostly empirical methods to ‘extrapolate’ airplane characteristics, while aerodynamics via HPC simulation was far more efficient. Takeoff with the Boeing Uses ‘Jaguar’ to Improve Aerodynamics Codes case study.

Cray and AMD – Helping Accelerate Scientific Discovery

Cray and AMD have a long history of collaboration. Over the years, the relationship has produced some of the world’s most productive supercomputers for scientific and commercial research. Throughout the relationship, AMD has made several major technological leaps in processor architecture and design. Processors have gone from dual-core to quad-core to six-core over the last several years. The launch of the AMD Opteron™ 6200 Series processor, which had gone by the code name “Interlagos”, introduced the world’s first 16-core x86 processor. The processor’s architecture is very flexible and can be applied effectively to a variety of workloads and problems.

Cray has implemented the AMD processor technology with HPC infrastructure which scales to perform at ‘sustained’ petascale levels, far more important to advancing science than spikes of unsustainable ‘peak’ performance. In fact, Cray has leveraged that petaflop proven technology to deliver HPC configurations and price points across the performance spectrum, including departmental/divisional platforms starting at just $200K USD.

Now scientists can afford to utilize the same compatible technology that drives the biggest and fastest supercomputers in the world. For existing Cray customers, the entry-level supercomputers powered by AMD Opteron processors provide investment protection. Over the years, Cray users have been able to upgrade as new technologies become available. This includes upgrades to processors, blades, and storage, which allows an organization to leverage an initial investment in a Cray system and scale the system over time.

The entry-level HPC systems are just the latest result in this partnership between Cray and AMD, enabling, as well as encouraging, new, innovative and advanced scientific discovery.

To see more case study examples of HPC breakthroughs in science, browse: www.cray.com/Products/XE/Resources

To learn more about the Cray supercomputers powered by AMD Opteron processors, visit: www.cray.com/ownacray