Entity

Time filter

Source Type


Hoefler T.,National Center for Supercomputing Applications
Computing in Science and Engineering | Year: 2010

Although most large-scale systems are designed with the network as a central component, the interconnection network's energy consumption has received little attention. However, several software and hardware approaches can increase the interconnection network's power efficiency by using the network more efficiently or using throttling bandwidths to reduce the power consumption of unneeded resources. © 2006 IEEE. Source


The study, reported in the journal eLife, represents a major step in understanding the "bacterial brain," said University of Illinois physics professor Klaus Schulten, who led the new research. "On its surface, a bacterium has thousands of receptors that scan the environment and then tell it what to do," he said. This is very much like the sensory input that all animals must process. Of course, bacteria are single-celled organisms and don't have brains, he said. But they nonetheless manage to organize and "remember" sensory signals long enough to respond to them in a way that aids their own survival. The receptors on the surface of a bacterial cell detect light, chemicals, edible things and poisonous things, and transmit that information to a deeper layer of proteins, called kinases, which interpret this data and translate it into a simple choice: "Keep going" or "Change direction!" If the latter decision is made, a kinase hands off a potent chemical signal - a phosphate - to a second kinase, called CheY (KEY why), which then detaches, finds its way to the flagella and activates a process that causes the flagella to reverse their spin. "That makes the bacterium tumble and go in a new, random direction, which may be better than the previous direction," Schulten said. Previous studies have yielded key insights into the structure of the molecular machine that orchestrates this feat, the chemosensory array. Electron microscopy of the inner and outer surfaces of bacterial cells gives some clues, and crystallography - a process that involves stacking purified proteins into crystals so that their three-dimensional characteristics can be measured - provides others. But the fuzzy resolution of the EM snapshots leaves a lot of room for interpretation, and the crystals can resolve only small portions of the array's constituent proteins. Study co-author, experimentalist Peijun Zhang of the University of Pittsburgh, aided this effort by developing a technique to purify the key proteins in the array and combine them in just the right proportions so that they assemble themselves in thin layers - allowing clearer 3-D EM snapshots of their structural conformations and interactions with each other. This vastly improved the resolution of the data. To resolve the picture of the chemosensory array, Schulten and his colleagues used molecular dynamic flexible fitting, a computer modeling approach Schulten's lab developed at Illinois. MDFF simulates the chemical interactions of every atom in a system and makes use of what is known about the structure from EM, crystallography and other experimental data. Such large-scale modeling and simulation requires the heft of a supercomputer, and for this effort the team used Blue Waters at the National Center for Supercomputing Applications at the U. of I. The new study revealed key chemical interactions between the proteins that make up the chemosensory array, and offered new insights into the behavior of these proteins. For example, it revealed for the first time that one region of a kinase called CheA (KEY aye), changes its orientation in relation to the other proteins, in a motion the researchers call "dipping." Further experiments revealed that this part of the kinase is essential to the process that allows a bacterium to respond to its environment and change direction. "A big question in the field is: How does the signal pass from the receptors to the kinases? What is actually happening?" Schulten said. "It has to be a motion. It can't be anything else. But what kind of motion?" More work is needed to determine the relationships and behavior of all of the components of the system, but the new study represents a major gain in comprehension, Schulten said. He compares the process of discovery to that of someone encountering a mechanical clock for the first time. "To know how this mechanical system works, we need to know the structure," he said. "Once we open the clock, see how the gears fit together, then we can start thinking about how the clock actually works. The gears of the bacterial brain are now in place." Explore further: Computational microscope peers into the working ribosome (w/ Video) More information: C Keith Cassidy et al. CryoEM and computer simulations reveal a novel kinase conformational switch in bacterial chemotaxis signaling, eLife (2015). DOI: 10.7554/eLife.08419


News Article
Site: http://www.scientificcomputing.com/rss-feeds/all/rss.xml/all

When researchers need to compare complex new genomes, or map new regions of the Arctic in high-resolution detail, or detect signs of dark matter, or make sense of massive amounts of functional MRI data, they turn to the high-performance computing and data analysis systems supported by the National Science Foundation (NSF). High-performance computing (or HPC) enables discoveries in practically every field of science — not just those typically associated with supercomputers like chemistry and physics, but also in the social sciences, life sciences and humanities. By combining superfast and secure networks, cutting-edge parallel computing and analytics software, advanced scientific instruments and critical datasets across the U.S., NSF's cyber-ecosystem lets researchers investigate questions that can't otherwise be explored. NSF has supported advanced computing since its beginning and is constantly expanding access to these resources. This access helps tens of thousands of researchers each year — from high-school students to Nobel Prize winners — expand the frontiers of science and engineering, regardless of whether their institutions are large or small, or where they are located geographically. Below are 10 examples of research enabled by NSF-supported advanced computing resources from across all of science. Pineapples don't just taste good — they have a juicy evolutionary history. Recent analyses using computing resources that are part of the iPlant Collaborative revealed an important relationship between pineapples and crops like sorghum and rice, allowing scientists to home in on the genes and genetic pathways that allow plants to thrive in water-limited environments. Led by the University of Arizona, Texas Advanced Computing Center, Cold Spring Harbor Laboratory and University of North Carolina at Wilmington, iPlant was established in 2008 with NSF funding to develop cyberinfrastructure for life sciences research, provide powerful platforms for data storage and bioinformatics and democratize access to U.S. supercomputing capabilities. This week, iPlant announced it will host a new platform, Digital Imaging of Root Traits (DIRT), that lets scientists in the field measure up to 76 root traits merely by uploading a photograph of a plant's roots. 2. Designing new nanodevices Software that simulates the effect of an electric charge passing through a transistor — only a few atoms wide — is helping researchers to explore alternative materials that may replace silicon in future nanodevices. The software simulations designed by Purdue researcher Gerhard Klimeck and his group, available on the nanoHUB portal, provide new information about the limits of current semiconductor technologies and are helping design future generations of nanoelectronic devices. NanoHUB, supported by NSF, is the first broadly successful, scientific end-to-end cloud computing environment. It provides a library of 3,000 learning resources to 195,000 users worldwide. Its 232 simulation tools are used in the cloud by over 10,800 researchers and students annually. Earthquakes originate through complex interactions deep below the surface of the Earth, making them notoriously difficult to predict. The Southern California Earthquake Center (SCEC) and its lead scientist Thomas Jordan use massive computing power to simulate the dynamics of earthquakes. In doing so, SCEC helps to provide long-term earthquake forecasts and more accurate hazard assessments. In 2014, the SCEC team investigated the earthquake potential of the Los Angeles Basin, where the Pacific and North American Plates run into each other at the San Andreas Fault. Their simulations showed that the basin essentially acts like a big bowl of jelly that shakes during earthquakes, producing more high-shaking ground motions than the team expected. Using the NSF-funded Blue Waters supercomputer at the National Center for Supercomputing Applications and the Department of Energy-funded Titan supercomputer at the Oak Ridge Leadership Computing Facility, the researchers turned their simulations into seismic hazard models. These models describe the probability of an earthquake occurring in a given geographic area, within a given window of time and with ground motion intensity exceeding a given threshold. Nearly 33,000 people die in the U.S. each year due to motor vehicle crashes, according to the National Highway Traffic Safety Administration. Modern restraint systems save lives, but some deaths and injuries remain — and restraints themselves can cause injuries. Researchers from the Center for Injury Biomechanics at Wake Forest University used the Blacklight supercomputer at the Pittsburgh Supercomputing Center to simulate the impacts of car crashes with much greater fidelity than crash-test dummies can. By studying a variety of potential occupant positions, they're uncovering important factors that lead to more severe injuries, as well as ways to potentially mitigate these injuries, using advanced safety systems. Since Albert Einstein, scientists have believed that when major galactic events like black hole mergers occur, they leave a trace in the form of gravitational waves — ripples in the curvature of space-time that travel outward from the source. Advanced LIGO is a project designed to capture signs of these events. Since gravitational waves are expected to travel at the speed of light, detecting them requires two gravitational wave observatories, located 1,865 miles apart and working in unison, that can triangulate the gravitational wave signals and determine the source of the wave in the sky. In addition to being an astronomical challenge, Advanced LIGO is also a "big data" problem. The observatories take in huge volumes of data that must be analyzed to determine their meaning. Researchers estimate that Advanced LIGO will generate more than one petabyte of data a year, the equivalent of 13.3 years' worth of high-definition video. To achieve accurate and rapid gravity wave detection, researchers use Extreme Science and Engineering Discovery Environment (XSEDE) — a powerful collection of advanced digital resources and services — to develop and test new methods for transmitting and analyzing these massive quantities of astronomical data. Advanced LIGO came online in September, and advanced computing will play an integral part in its future discoveries. What happens when a supercomputer reaches retirement age? In many cases, it continues to make an impact in the world. The NSF-funded Ranger supercomputer is one such example. In 2013, after five years as one of NSF's flagship computer systems, the Texas Advanced Computing Center (TACC) disassembled Ranger and shipped it from Austin, TX, to South Africa, Tanzania and Botswana to give root to a young and growing supercomputing community. With funding from NSF, TACC experts led training sessions in South Africa in December 2014. In November 2015, 19 delegates from Africa came to the U.S. to attend a two-day workshop at TACC as well as the Supercomputing 2015 International Conference for High Performance Computing. The effort is intended, in part, to help provide the technical expertise needed to successfully staff and operate the Square Kilometer Array, a new radio telescope being built in Australia and Africa which will offer the highest resolution images in all of astronomy. In September 2015, President Obama announced plans to improve maps and elevation models of the Arctic, including Alaska. To that end, NSF and the National Geospatial-Intelligence Agency (NGA) are supporting the development of high-resolution Digital Elevation Models in order to provide consistent coverage of the globally significant region. The models will allow researchers to see in detail how warming in the region affects the landscape in remote areas, and allow them to compare changes over time. The project relies, in part, on the computing and data analysis powers of Blue Waters, which will let researchers store, access and analyze large numbers of images and models. To solve some of society's most pressing long-term problems, the U.S. needs to educate and train the next generation of scientists and engineers to use advanced computing effectively. This pipeline of training begins as early as high school and continues throughout the careers of scientists. Last summer, TACC hosted 50 rising high school juniors and seniors to participate in an innovative new STEM program, CODE@TACC. The program introduced students to high-performance computing, life sciences and robotics. On the continuing education front, XSEDE offers hundreds of training classes each year to help researchers update their skills and learn new ones. High-performance computing has another use in education: to assess how students learn and ultimately to provide personalized educational paths. A recent report from the Computing Research Association, "Data-Intensive Research in Education: Current Work and Next Steps," highlights insights from two workshops on data-intensive education initiatives. The LearnSphere project at Carnegie Mellon University, an NSF Data Infrastructure Building Blocks project, is putting these ideas into practice. 9. Experimenting with cloud computing on new platforms In 2014, NSF invested $20 million to create two cloud computing testbeds that let the academic research community develop and experiment with cloud architectures and pursue new, architecturally-enabled applications of cloud computing. CloudLab (with sites in Utah, Wisconsin and South Carolina) came online in May 2015 and provides researchers with the ability to create custom clouds and test adjustments at all levels of the infrastructure, from the bare metal on up. Chameleon, a large-scale, reconfigurable experimental environment for cloud research, co-located at the University of Chicago and The University of Texas at Austin, went into production in July 2015. Both serve hundreds of researchers at universities across the U.S. and let computer scientists experiment with unique cloud architectures in ways that weren't available before. The NSF-supported "Comet" system at the San Diego Supercomputer Center (SDSC) was dedicated in October and is already aiding scientists in a number of fields, including domains relatively new for supercomputer integration, such as neuroscience. SDSC recently received a major grant to expand the Neuroscience Gateway, which provides easy access to advanced cyberinfrastructure tools and resources through a web-based portal, and can significantly improve the productivity of researchers. The gateway will contribute to the national BRAIN Initiative and deepen our understanding of the human brain.


News Article
Site: http://phys.org/technology-news/

High-performance computing (or HPC) enables discoveries in practically every field of science—not just those typically associated with supercomputers like chemistry and physics, but also in the social sciences, life sciences and humanities. By combining superfast and secure networks, cutting-edge parallel computing and analytics software, advanced scientific instruments and critical datasets across the U.S., NSF's cyber-ecosystem lets researchers investigate questions that can't otherwise be explored. NSF has supported advanced computing since its beginning and is constantly expanding access to these resources. This access helps tens of thousands of researchers each year—from high-school students to Nobel Prize winners—expand the frontiers of science and engineering, regardless of whether their institutions are large or small, or where they are located geographically. Below are 10 examples of research enabled by NSF-supported advanced computing resources from across all of science. Pineapples don't just taste good—they have a juicy evolutionary history. Recent analyses using computing resources that are part of the iPlant Collaborative revealed an important relationship between pineapples and crops like sorghum and rice, allowing scientists to home in on the genes and genetic pathways that allow plants to thrive in water-limited environments. Led by the University of Arizona, Texas Advanced Computing Center, Cold Spring Harbor Laboratory and University of North Carolina at Wilmington, iPlant was established in 2008 with NSF funding to develop cyberinfrastructure for life sciences research, provide powerful platforms for data storage and bioinformatics and democratize access to U.S. supercomputing capabilities. This week, iPlant announced it will host a new platform, Digital Imaging of Root Traits (DIRT), that lets scientists in the field measure up to 76 root traits merely by uploading a photograph of a plant's roots. Software that simulates the effect of an electric charge passing through a transistor—only a few atoms wide—is helping researchers to explore alternative materials that may replace silicon in future nanodevices. The software simulations designed by Purdue researcher Gerhard Klimeck and his group, available on the nanoHUB portal, provide new information about the limits of current semiconductor technologies and are helping design future generations of nanoelectronic devices. NanoHUB, supported by NSF, is the first broadly successful, scientific end-to-end cloud computing environment. It provides a library of 3,000 learning resources to 195,000 users worldwide. Its 232 simulation tools are used in the cloud by over 10,800 researchers and students annually. Earthquakes originate through complex interactions deep below the surface of the Earth, making them notoriously difficult to predict. The Southern California Earthquake Center (SCEC) and its lead scientist Thomas Jordan use massive computing power to simulate the dynamics of earthquakes. In doing so, SCEC helps to provide long-term earthquake forecasts and more accurate hazard assessments. In 2014, the SCEC team investigated the earthquake potential of the Los Angeles Basin, where the Pacific and North American Plates run into each other at the San Andreas Fault. Their simulations showed that the basin essentially acts like a big bowl of jelly that shakes during earthquakes, producing more high-shaking ground motions than the team expected. Using the NSF-funded Blue Waters supercomputer at the National Center for Supercomputing Applications and the Department of Energy-funded Titan supercomputer at the Oak Ridge Leadership Computing Facility, the researchers turned their simulations into seismic hazard models. These models describe the probability of an earthquake occurring in a given geographic area, within a given window of time and with ground motion intensity exceeding a given threshold. Nearly 33,000 people die in the U.S. each year due to motor vehicle crashes, according to the National Highway Traffic Safety Administration. Modern restraint systems save lives, but some deaths and injuries remain—and restraints themselves can cause injuries. Researchers from the Center for Injury Biomechanics at Wake Forest University used the Blacklight supercomputer at the Pittsburgh Supercomputing Center to simulate the impacts of car crashes with much greater fidelity than crash-test dummies can. By studying a variety of potential occupant positions, they're uncovering important factors that lead to more severe injuries, as well as ways to potentially mitigate these injuries, using advanced safety systems. Since Albert Einstein, scientists have believed that when major galactic events like black hole mergers occur, they leave a trace in the form of gravitational waves—ripples in the curvature of space-time that travel outward from the source. Advanced LIGO is a project designed to capture signs of these events. Since gravitational waves are expected to travel at the speed of light, detecting them requires two gravitational wave observatories, located 1,865 miles apart and working in unison, that can triangulate the gravitational wave signals and determine the source of the wave in the sky. In addition to being an astronomical challenge, Advanced LIGO is also a "big data" problem. The observatories take in huge volumes of data that must be analyzed to determine their meaning. Researchers estimate that Advanced LIGO will generate more than 1 petabyte of data a year, the equivalent of 13.3 years' worth of high-definition video. To achieve accurate and rapid gravity wave detection, researchers use Extreme Science and Engineering Discovery Environment (XSEDE)—a powerful collection of advanced digital resources and services—to develop and test new methods for transmitting and analyzing these massive quantities of astronomical data. Advanced LIGO came online in September and advanced computing will play an integral part in its future discoveries. What happens when a supercomputer reaches retirement age? In many cases, it continues to make an impact in the world. The NSF-funded Ranger supercomputer is one such example. In 2013, after five years as one of NSF's flagship computer systems, the Texas Advanced Computing Center (TACC) disassembled Ranger and shipped it from Austin, Texas to South Africa, Tanzania and Botswana to give root to a young and growing supercomputing community. With funding from NSF, TACC experts led training sessions in South Africa in December 2014. In November 2015, 19 delegates from Africa came to the U.S. to attend a two-day workshop at TACC as well as the Supercomputing 2015 International Conference for High Performance Computing. The effort is intended, in part, to help provide the technical expertise needed to successfully staff and operate the Square Kilometer Array, a new radio telescope being built in Australia and Africa which will offer the highest resolution images in all of astronomy. In September 2015, President Obama announced plans to improve maps and elevation models of the Arctic, including Alaska. To that end, NSF and the National Geospatial-Intelligence Agency (NGA) are supporting the development of high-resolution Digital Elevation Models in order to provide consistent coverage of the globally significant region. The models will allow researchers to see in detail how warming in the region affects the landscape in remote areas, and allow them to compare changes over time. The project relies, in part, on the computing and data analysis powers of Blue Waters, which will let researchers store, access and analyze large numbers of images and models. To solve some of society's most pressing long-term problems, the U.S. needs to educate and train the next generation of scientists and engineers to use advanced computing effectively. This pipeline of training begins as early as high school and continues throughout the careers of scientists. Last summer, TACC hosted 50 rising high school juniors and seniors to participate in an innovative new STEM program, CODE@TACC. The program introduced students to high-performance computing, life sciences and robotics. On the continuing education front, XSEDE offers hundreds of training classes each year to help researchers update their skills and learn new ones. High-performance computing has another use in education: to assess how students learn and ultimately to provide personalized educational paths. A recent report from the Computing Research Association, "Data-Intensive Research in Education: Current Work and Next Steps," highlights insights from two workshops on data-intensive education initiatives. The LearnSphere project at Carnegie Mellon University, an NSF Data Infrastructure Building Blocks project, is putting these ideas into practice. Experimenting with cloud computing on new platforms In 2014, NSF invested $20 million to create two cloud computing testbeds that let the academic research community develop and experiment with cloud architectures and pursue new, architecturally-enabled applications of cloud computing. CloudLab (with sites in Utah, Wisconsin and South Carolina) came online in May 2015 and provides researchers with the ability to create custom clouds and test adjustments at all levels of the infrastructure, from the bare metal on up. Chameleon, a large-scale, reconfigurable experimental environment for cloud research, co-located at the University of Chicago and The University of Texas at Austin, went into production in July 2015. Both serve hundreds of researchers at universities across the U.S. and let computer scientists experiment with unique cloud architectures in ways that weren't available before. The NSF-supported "Comet" system at the San Diego Supercomputer Center (SDSC) was dedicated in October and is already aiding scientists in a number of fields, including domains relatively new for supercomputer integration, such as neuroscience. SDSC recently received a major grant to expand the Neuroscience Gateway, which provides easy access to advanced cyberinfrastructure tools and resources through a web-based portal, and can significantly improve the productivity of researchers. The gateway will contribute to the national BRAIN Initiative and deepen our understanding of the human brain. Explore further: Innovative new supercomputers increase nation's computational capacity and capability


Dietze M.C.,Boston University | Lebauer D.S.,Urbana University | Kooper R.,National Center for Supercomputing Applications
Plant, Cell and Environment | Year: 2013

The potential for model-data synthesis is growing in importance as we enter an era of 'big data', greater connectivity, and faster computation. Realizing this potential requires that the research community broaden its perspective about how and why they interact with models. We provide a review and perspective on the statistics and informatics of model-data fusion in plant biology. Overall we promote a community-based paradigm to model-data synthesis and highlight some of the tools and techniques that facilitate this approach. The potential for model-data synthesis is growing in importance as we enter an era of 'big data', greater connectivity and faster computation. Realizing this potential requires that the research community broaden its perspective about how and why they interact with models. Models can be viewed as scaffolds that allow data at different scales to inform each other through our understanding of underlying processes. Perceptions of relevance, accessibility and informatics are presented as the primary barriers to broader adoption of models by the community, while an inability to fully utilize the breadth of expertise and data from the community is a primary barrier to model improvement. Overall, we promote a community-based paradigm to model-data synthesis and highlight some of the tools and techniques that facilitate this approach. Scientific workflows address critical informatics issues in transparency, repeatability and automation, while intuitive, flexible web-based interfaces make running and visualizing models more accessible. Bayesian statistics provides powerful tools for assimilating a diversity of data types and for the analysis of uncertainty. Uncertainty analyses enable new measurements to target those processes most limiting our predictive ability. Moving forward, tools for information management and data assimilation need to be improved and made more accessible. © 2013 John Wiley & Sons Ltd. Source

Discover hidden collaborations