Why We Need Exascale Computing

Why We Need Exascale Computing
This post was published on the now-closed HuffPost Contributor platform. Contributors control their own work and posted freely to our site. If you need to flag this entry as abusive, send us an email.

By Paul Messina, Exascale Project Director, Argonne National Laboratory

Exascale computing is the label we use for the next 50- to 100-fold increase in speed over the fastest supercomputers in broad use today. That’s at least a billion billion operations per second. Achieving such speeds is not an easy task. To be useful to a wide spectrum of applications, in addition to peak speed, supercomputers need to have large memories and the ability to store and read vast quantities of data at high speed. In addition, the supercomputer must have a software environment that facilitates the efficient and productive use of the systems.

But every time computing power increases by large factors, new benefits open before us. The benefits of exascale computing--which range from creating novel, more efficient combustion engines and new energy solutions to advances in healthcare, biology, and storm prediction-- could potentially impact every person reading this article.

Reducing Pollution. Exascale computing has the potential to reduce pollution caused by burning fossil fuels. Over 85 percent of the world’s energy is generated by burning fossil fuels. But we can minimize pollution and optimize efficiency by understanding and controlling the chemical process of combustion. Through exascale computing, we expect it will be possible to increase the efficiency of combustion systems in engines and gas turbines for transportation and power generation by potentially 25-50%, and to lower emissions.

Advances in Materials Science. In order to create certain new technologies and inventions, we need to discover new materials with specific properties. One crucial way to make these discoveries is by complex calculations that simulate how materials behave in nature, and the use of massive databases of known compounds to identify combinations that have the desired properties. Deep learning techniques and classical simulation are in use today in this field; exascale computing will enable faster and more complex designs. For example, much more efficient, less expensive, and longer-lived batteries will be developed through discovery of new materials. Batteries are very important, not only for portable consumer electronics and electric automobiles, but to store energy from variable energy sources, such as wind and solar, to use when they’re needed.

New Energy Solutions. Solar energy will also be made more cost-effective through discovery of materials that convert the sun’s rays to electricity more efficiently. Wind turbines can be made more efficient and quieter through computer-based design of their blades, and simulations can also optimize the locations of individual wind turbines in wind farms to yield substantially more energy from the same number of turbines.

Another challenge in the use of alternative energy sources is managing the electric power grid –when the wind dies down in the area of a wind farm, for example. The electric power grid is complex, and adjustments such as firing up a fossil fuel generator have to be made quickly to avert problems; it is not possible to predict accurately such events far ahead of time. Computationally intensive optimization methods can provide the needed guidance, but to get the results in time to be useful requires fast computers.

The benefits of exascale computing will flow from classical simulations but also from large-scale data analysis, deep machine learning, and often the integration of the three methods. Examples of the latter are healthcare (precision medicine) and biology.

Advances in Healthcare. Exascale computing will accelerate cancer research by helping scientists understand the molecular basis of key protein interactions and by automating the analysis of information from millions of cancer patient records to determine optimal treatment strategies. It will also enable doctors to predict the right treatments for the patient by modeling drug responses. These tasks are daunting: for example, drug combination response prediction might require the search of one trillion drug combinations.

Exascale computing applications in biology might enable the prediction of feasible parameter values for dynamic models of metabolism that would enable scientists to design organisms that would perform a variety of tasks. These models might also contribute to the development of treatments for emerging types of infections.

Predicting Severe Weather. Weather prediction models will be able to predict more accurately and quickly the timing and path of severe weather events such as hurricanes by using much higher spatial resolution, incorporating more physics, and assimilating more observational data.

Improving Quality of Life. The use of exascale computing in urban science promises to mitigate health hazards, reduce crime, and improve the quality of life in cities by optimizing infrastructure (such as transportation, energy, housing, healthcare) access and usage choices. Such optimization requires gathering and analysis of numerous types of data from databases, sensors, and simulation results and conducting thousands of potential scenarios. The resulting analyses will be useful for planning new cities or neighborhoods as well as restructuring the infrastructure in existing large urban areas.

There are additional important applications that exascale computing will advance but will require even more computing power to accomplish, such as reverse engineering the human brain to understand complex neural systems; providing new methods to integrate large-scale data; information and simulation dynamics; and ultimately providing societal benefits for health.

Finally, it is worth noting that many planned uses of exascale computing address fundamental scientific questions in fields such as high-energy physics and chemistry. Though they are not explicitly aimed at results that immediately benefit society, historically they have a track record of making discoveries that have major impacts in everyday life. The obvious example is the world wide web, which was invented to support certain aspects of particle physics research.

Exascale computing will potentially benefit society in a myriad of ways, only some of which are mentioned in this article. Novel discoveries, new materials, and solutions to problems that impact everyday life will be possible through new capabilities of exascale computing systems. In addition to benefits to society, key technological advances also hold the promise of maintaining economic security.

Popular in the Community


What's Hot