Archive for Jaguar supercomputer
You are browsing the archives of Jaguar supercomputer.
You are browsing the archives of Jaguar supercomputer.
These were, at least, interesting to me:
Following up on earlier theoretical predictions, MIT researchers have now demonstrated experimentally the existence of a fundamentally new kind of magnetic behavior, adding to the two previously known states of magnetism. Ferromagnetism—the simple magnetism of a bar magnet or compass needle—has been known for centuries. In a second type of magnetism, antiferromagnetism, the magnetic fields of the ions within a metal or alloy cancel each other out. In both cases, the materials become magnetic only when cooled below a certain critical temperature. The prediction and discovery of antiferromagnetism, the basis for the read heads in today’s computer hard disks, won Nobel Prizes in physics for Louis Neel in 1970 and for MIT professor emeritus Clifford Shull in 1994. The experimental work showing the existence of this new state, called a quantum spin liquid (QSL), is reported in the journal Nature. The QSL is a solid crystal, but its magnetic state is described as liquid: Unlike the other two kinds of magnetism, the magnetic orientations of the individual particles within it fluctuate constantly, resembling the constant motion of molecules within a true liquid.
DOE’s Oak Ridge National Laboratory regained the lead in high-performance computing, enjoyed record-setting recognition for its research and became a showpiece for renewable energy technology during 2012. ORNL’s 2012 included achievements in both research and support. ORNL solidified its standing in world-class scientific computing with the upgrade of the Jaguar supercomputer to the 27-petaflop/s Titan, regaining the top spot on the TOP500 list of the world’s supercomputers. Titan also proved to be one of the world’s most energy efficient number crunchers, ranking No. 3 on the Green500 list. The Mars Curiosity rover successfully landed on the Red Planet and began transmitting historic data back to Earth, thanks in part to ORNL’s role in making the radioisotope-fueled generators that power the NASA vehicle and its suite of instruments. ORNL set a record for R&D 100 Awards, often called the Oscars of science and technology. Ten technologies involving ORNL research were named among R&D Magazine’s top 100. The awards reflected the laboratory’s strength in advanced materials research, including technologies related to high-temperature superconducting wire, super-tough protective coatings, advanced absorbents, an advanced rolling mill process and a low-cost, lightweight robotic hand based on additive manufacturing and fluid power.
By showing that tiny particles injected into a liquid crystal medium adhere to existing mathematical theorems, physicists at the University of Colorado Boulder have opened the door for the creation of a host of new materials with properties that do not exist in nature. The findings show that researchers can create a “recipe book” to build new materials of sorts using topology, a major mathematical field that describes the properties that do not change when an object is stretched, bent or otherwise “continuously deformed.” Published online in the journal Nature, the study also is the first to experimentally show that some of the most important topological theorems hold up in the real material world, said CU-Boulder physics department Assistant Professor Ivan Smalyukh, a study senior author. The research could lead to upgrades in liquid crystal displays, like those used in laptops and television screens, to allow them to interact with light in new and different ways. One possibility is to create liquid crystal displays that are even more energy efficient, Smalyukh said, extending the battery life for the devices they’re attached to. The research supports the goals laid out by the White House’s Materials Genome Initiative, Smalyukh said, which seeks to deploy “new advanced materials at least twice as fast as possible today, at a fraction of the cost.”
Nature’s Richard Van Noorden gazes into the crystal ball on a number of science topics, and issues his predictions for the new year. He writes, “Samarium hexaboride might be the next star of materials science, following hints last year that it is a topological insulator — conducting electricity on its surface, but behaving as an insulator inside. Graphene will remain a major celebrity, so expect a flood of reports about copycat materials such as boron nitride, tantalum disulphide and other two-dimensional sheets that can be stacked or sandwiched in precise layers.”
Because modern computers have to depict the real world with digital representations of numbers instead of physical analogues, to simulate the continuous passage of time they have to digitize time into small slices. This kind of simulation is essential in disciplines from medical and biological research, to new materials, to fundamental considerations of quantum mechanics, and the fact that it inevitably introduces errors is an ongoing problem for scientists. Scientists at the DOE’s Lawrence Berkeley National Laboratory have now identified and characterized the source of tenacious errors and come up with a way to separate the realistic aspects of a simulation from the artifacts of the computer method. ”A simulation of a physical process on a computer cannot use the exact, continuous equations of motion; the calculations must use approximations over discrete intervals of time,” says one of the Berkeley Lab researchers, David Sivak. “It’s well known that standard algorithms that use discrete time steps don’t conserve energy exactly in these calculations.”
Silicon wafers destined to become photovoltaic cells can take a bruising through assembly lines, as they are oxidized, annealed, purified, diffused, etched, and layered to reach their destinies as efficient converters of the sun’s rays into useful electricity. All those refinements are too much for 5 percent to 10 percet of the costly wafers. They have micro-cracks left over from incomplete wafer preparation, which causes them to break on the conveyers or during cell fabrication. Scientists at the DOE’s National Renewable Energy Laboratory have developed an instrument that puts pressure on the wafers to find which ones are too fragile to make it through the manufacturing process-and then kicks out those weak wafers before they go through their costly enhancement. NREL’s Silicon Photovoltaic Wafer Screening System is a cube-shaped furnace about 15 inches each side, and can be retrofitted into an assembly line. The loss in revenue due to broken wafers- which increases dramatically as the wafers move closer to completion-is an important barrier to solar energy becoming cost competitive with other energy technologies. Manufacturers need better, less expensive ways to make the cells.
Jamaica may be able to benefit from newly found deposits of rare-earth elements that are key ingredients for smartphones, computers and numerous other high-tech goods, the Caribbean island’s top mining official reports. Science, Technology, Energy & Mining Minister Philip Paulwell says Japanese researchers believe they have found “high concentrations of rare-earth elements” in the country’s red mud, or bauxite residue. In a statement to Jamaica’s Parliament, Paulwell said researchers from Japan’s Nippon Light Metal Co. believe rare-earth elements can be efficiently extracted in Jamaica, where a once-flourishing bauxite industry has fallen on hard times. Paulwell touted the discovery as a potentially significant boon for the Caribbean island’s chronically sputtering economy. A pilot program will establish the scope of any potential commercial project on Jamaica, which is about the size of the state of Connecticut. The environmental and planning agency has already authorized the pilot program but other government agencies still need to examine it. Nippon Light Metal has agreed to invest $3 million in buildings and equipment for the pilot project while also being responsible for operating costs. Any rare-earth elements produced during this phase will be jointly owned by Jamaica and the Japanese company. Negotiations for commercialization are expected to occur at a later date.
A nanoscale coating that’s at least 95 percent air repels liquid and causes it to recoil from treated surfaces. In addition to super stain-resistant clothes, the coating could lead to breathable garments that protect soldiers and scientists from chemicals, and advanced waterproof paints that dramatically reduce drag on ships. Droplets of solutions that would normally damage either your shirt or your skin recoil when they touch the new “superomniphobic surface.” Of more than 100 liquids, only two chlorofluorocarbons were able to penetrate the coating. Chlorofluorocarbons are chemicals used in refrigerators and air conditioners. The “superomniphobic surface” repelled coffee, soy sauce, and vegetable oil, as well as toxic hydrochloric and sulfuric acids that could burn skin. The coating is also resistant to gasoline and various alcohols.
Okay, I have to confess that I don’t really know if the Materials Genome Initiative is going to be discussed in the event I am about to describe. As a matter of fact, I am not sure exactly what’s going to be discussed in what may be one of the worst-publicized-but-possibly-most-interesting-online-science-events this year, but I will be shocked if the MGI isn’t frequently mentioned.
Here is what I know for sure, based on an obscure notice titled “Challenges and Opportunities in Big Data” on Science360 Live, an online science project of the NSF:
On Thursday, March 29, 2012, from 2-3 pm ET, federal government science heads from OSTP, NSF, NIH, DOE, DOD, DARPA and USGS will outline how their agencies are engaged in Big Data research. Their remarks will be followed by a panel of thought leaders from academia and industry, moderated by Steve Lohr of the New York Times. The event will take place at the American Association for the Advancement of Science in Washington, D.C.
If you are interested in watching the webcast live, apparently all you have to do is click on the Science360 Live link above at the stated time. I suspect a recorded version will be made available in a few days.
I first learned about this event this morning after reading a Gigaom blog post. The post claims that the speakers will include John Holdren, assistant to the president and director of OSTP; Subra Suresh, director of NSF; and Marcia McNutt, director of USGS. Gigaom says the event is sponsored by OSTP; a CTOsite post (the only other original post I could find on this event) says it is sponsored by both OSTP and NSF.
I tried to fact-check this by searching through the media advisories from the White House, NSF and the OSTP pressroom and blog, but couldn’t find any announcements, whatsoever. But this thing sounds legit — and potentially interesting. I think the CTOsite has it right when it says, ”[T}his will be a GREAT event for those who would like to learn how to explain Big Data concepts. This type of evolution can be very helpful to enhancing cross-government sharing of lessons learned and concepts. I believe this will be very positive. This initiative will discuss R&D activities. Although at a high level, federal R&D focus is usually expressed in terms of frameworks and broad goals, these goals normally translate into budgets so the research community outside of the government should really pay attention.”
The MGI, I think, is one of the few topics that the group will be able to explain with relative ease, and that the probable audience can get their arms around. Besides, I seriously doubt the group wants the discussion to go in the direction of talking about the Hugely Big Data story (NSA’s new yottaflop-scale “Matrix” project), the spin-offs of which are actually going to help enable MGI and other enormous modeling and data management projects.
The goal is to integrate existing nuclear energy and nuclear national security modeling and simulation capabilities with high-performance computing to simulate radiation in order to support the design and safety of nuclear facilities, improve reactor core designs and nuclear fuel performance and ensure the safety of nuclear materials, such as spent nuclear fuel.
John Wagner, technical integration manager for nuclear modeling at ORNL says, “We’re now simulating entire nuclear facilities, such as a nuclear power reactor facility with its auxiliary buildings and the ITER fusion reactor, with much greater accuracy than any other organization that we’re aware of.”
“Software for modeling radiation transport has been around for a long time,” he adds, “but it hadn’t been adapted to build on developments that have revolutionized computational science. There’s no special transformational technology in this software; but it’s designed specifically to take advantage of the massive computational and memory capabilities of the world’s fastest computers.”
The project has been awarded eight million processor hours on Jaguar for the purpose of developing a “uniquely detailed simulation of the power distribution inside a nuclear reactor core.” This is expected to cut years off the process of designing new and better reactors.