In our previous post, we described what a supercomputer is and how it works. Here, we would like to talk about its specific use cases. From weather forecasting to neuroscience, from astrophysics to playing chess, supercomputers can help a wide range of researchers to deal with various computationally intensive tasks. How do scientists apply massive calculation power?
Typically, supercomputers are used for sophisticated, mathematically intensive scientific problems of national or universal significance. But invariably, some supercomputers are designed to do some specific jobs. For example, IBM built the Deep Blue to play chess. The machine searched through a massive database of possible chess moves and evaluated them in relation to the current situation. In 1997, this supercomputer beat chess champion, Garry Kasparov. IBM Watson machine was also designed to play a game, Jeopardy. However, Watson is now used by health insurers to predict patients’ diagnoses and treatments.
Nowadays, supercomputers are more designed for general-purpose. The most important criteria for the problem they can solve is to be amenable to massive parallelization. This goes to say that the task can be split into smaller non sequential chunks. Supercomputers are so powerful that they can provide scientists with insight into phenomena that are too big, too little, too fast, or too slow to observe in laboratories.
In the US, National Oceanic and Atmospheric Administration (NOAA) uses the Weather and Climate Operational Supercomputing System, a supercomputer that can make weather forecast, track oceanic and space weather activities and predict rains, thunderstorms, hurricanes and other weather events. The data from observations is loaded into the computer, which then uses complex mathematical models to predict how the weather conditions might change over time. The output is the basis of almost every forecast on weather applications or broadcast across America. European scientists rely on supercomputers too.
Supercomputers can simulate the natural world digitally. The Earth, space or a human body are studied at a very high resolution, atom by atom. For example, scientists used two supercomputers to run a simulation of human immunodeficiency virus interacting with a cell in human body. One of this is the Blue Waters, in Illinois, and the other — Titan, at Oak Ridge National Laboratory. The simulation produced almost 100 terabytes of date, so researchers needed Blue Waters again just to crunch it.
Scientists at the University of Basel used the Piz Daint, currently the third world’s third most powerful supercomputer, to discover interrelationships in the human genome in search for “memory molecules”. Eventually, the study has to lead to more efficient medical treatment for people who suffer memory disturbance diseases. Neuroscientists also use supercomputers to examine the relationship between the structure and function of the brain as well as its dynamic and physiological structures.
Another illustration of a supercomputer capability is molecular dynamics, the way molecules interact with each other. The simulations allow scientists to dock two molecules together and study their interaction. With powerful machines like the supercomputer, researchers can determine the shape of a molecule’s surface and generate an atom-by-atom picture of its geometry. This cannot just be done in a laboratory.
Asides from modeling natural phenomena, supercomputers can perform a wider range of functions, like figuring out how a bomb or chemical weapon would disperse its harmful elements around. The IBM’s Sequoia machine, for example, is used for nuclear weapon security.
Additionally, Astrophysicists use supercomputers as “time machines” to investigate the past and the future of our universe. In 2000, the Blue Horizon, a supercomputer deployed in San Diego, powered a simulation of the collision of two galaxies: Andromeda and our own Milky Way. Although this is not expected to happen in the next three billion years, the results can be seen now.
In spite of these functions, supercomputers are not always used in the best way they are supposed to. Russian engineers were busted for mining crypto-currencies with the supercomputer of the Federal Nuclear Center, a high-security organization.
Distributed General Supercomputing
The good thing about any computer is that it is a general-purpose machine, which means you can use it in different ways. You can send emails, play games or edit photos by running a different program. Theoretically, a general-purpose supercomputer can be used for absolutely anything. The demand for computing is growing at a fast pace. At the same time, our desktop computers and laptops can be used for a significant portion of the day.
With the block-chain technology, some companies now propose the solution or project — , where participants can lend and borrow computing resources and make money in the process. An example is the Altumea, a decentralized platform where researchers, scientists or artists, who needs to do high-performance calculation, can buy computing power directly from individual GPU owners, who do not use them, round-the-clock. This allows executing large parallel applications significantly cheaper in comparison to a traditional supercomputer. The similar model is used at distributed computing platforms like SETI@Home or Golem.
Of course, this not an ultimate list of supercomputer’s application. Quantum mechanics, oil and gas exploration, molecular modeling, cryptanalysis and physical simulations such as airplane and spacecraft aerodynamics are just a few other areas mention. We’ll probably see more and more of them in the nearest future.
