When a small asteroid fell to Earth over the city of Chelyabinsk, Russia, in 2013, the speeding space rock exploded midair and created a shock wave powerful enough to damage thousands of buildings and injure more than 1,200 people. The blast from the Chelyabinsk explosion shattered glass windows up to 58 miles (93 kilometers) away.

Just in time for Asteroid Day (June 30), NASA's Planetary Defense Coordination Officehas released a supercomputer simulation of a Chelyabinsk-like asteroid burning up in Earth's atmosphere. Simulations such as this one "help first responders and other agencies to identify and make better informed decisions for how best to defend against life-threatening asteroid events," NASA officials said in a statement.

Using the Pleiades supercomputer at NASA's Ames Research Facility in California's Silicon Valley, researchers are modeling hypothetical asteroid-impact scenarios like this one to learn more about how dangerous space rocks crumble into pieces after entering Earth's atmosphere. These supercomputer simulations harness NASA's Cart3D aerodynamic design software and the ALE3D modeling software developed by Lawrence Livermore National Lab. [Gallery: Chelyabinsk Explosion of 2013]

"The NASA team was able to run large-scale simulations of the Chelyabinsk asteroid event on Pleiades to produce many impact scenarios quickly, because Cart3D is dozens of times faster than typical 3-D numerical modeling used for aerodynamic analysis," NASA officials said in the statement. "The detailed simulations allowed the team to model the fluid flow that occurs when asteroids melt and vaporize as they break up in the atmosphere."

Experts with the Asteroid Threat Assessment Project at NASA's Ames Research Facility share this research with universities, national labs and government agencies around the world to help people come up with plans for dealing with the threat of asteroid impacts.

Learn more about the Chelyabinsk event from the Science@NASA video series:

Editor's Note:Space.com senior producerSteve Spaletacontributed to this report.

Email Hanneke Weitering at hweitering@space.com or follow her @hannekescience. Follow us @Spacedotcom, Facebookand Google+. Original article on Space.com.

Read the original:

Chelyabinsk-Like Asteroid Gets Incinerated in Supercomputer Simulation (Video) - Space.com

Other notable traits he apparently possesses include modesty, assertiveness and altruism. Presumably, at least some of these characteristics have contributed to his achievements and success.

To arrive at these results, Paysa "gathered speeches, essays, books, the transcripts of interviews and other forms of communication produced by those highlighted above." It put over 2,500 words "through the Watson Personality Insights API."

"Personality Insights extracts personality characteristics based on how a person writes," according to IBM Watson's website. "You can use the service to match individuals to other individuals, opportunities, and products, or tailor their experience with personalized messaging and recommendations."

While it makes sense that former Harvard dropout Zuckerberg's top personality trait is his intellect, others, like his melancholy, may seem more surprising. Either way, the data provide a fascinating look into the mind of a self-made billionaire whose goal has been to "connect the world," one Facebook profile at a time.

Like this story? Like CNBC Make It on Facebook.

See also: Mark Zuckerberg, Satya Nadella and Bill Gates share this crucial personality trait for achieving success

Go here to see the original:

Mark Zuckerberg's top 5 personality traits, according to an IBM supercomputer - CNBC

Seven projects, many of which have applications to Wyoming issues -- including identification of promising catalysts for carbon fuel and chemical production, cloud seeding to enhance precipitation and a study of the Blair-Wallis watershed -- were recently chosen to receive computational time and storage space on the supercomputer in Cheyenne.

University of Wyoming faculty members will head projects that will use the NCAR-Wyoming Supercomputing Center (NWSC). Each project was critically reviewed by an external panel of experts and evaluated on experimental design, computational effectiveness, efficiency of resource use and broader impacts such as how the project involves both UW and National Center for Atmospheric Research (NCAR) researchers; strengthens UWs research capacity; enhances UWs computational programs; or involves research in a new or emerging field.

The Wyoming-NCAR Allocation Panel recently met to evaluate the large allocation requests for the use of computational resources at the NCAR-Wyoming Supercomputing Center, says Bryan Shader, UWs special assistant to the vice president for research and economic development, and a UW professor of mathematics. The projects were granted allocations totaling 65 million core hours. In addition, 6 million core hours were recently awarded to a new faculty member as part of his startup package.

These projects represent new explorations that could not be undertaken without the capabilities of Cheyenne, the recently installed supercomputer at the NWSC, Shader says. Improved computer technology is enabling the study of more complicated or nuanced phenomena.

Over the last year, 23 UW projects used the NWSC. This includes four new projects that were allocated a total of 72 million core hours starting in January 2016. These allocations and use rank UW as first in total allocation and total users; first in total computer charges; and first in active projects among the more than 100 North American universities using the NWSC.

Since the NWSCs opening in October 2012, allocations have been made to 53 UW research projects, including these latest seven.

Cheyenne Projects

A brief description of each of the projects, which begin July 1, is provided below:

-- Maohong Fan, a School of Energy Resources (SER) professor of chemical and petroleum engineering, leads the project, titled Application of Density Functional Theory in Low Carbon Fuel and Chemical Productions. This National Science Foundation (NSF)-funded project will use the Cheyenne supercomputer to identify the most promising catalysts for carbon dioxide or synthetic gas conversion from among a massive set of possibilities. The project should greatly accelerate the application of carbon dioxide conversion techniques that can be used to mitigate the environmental impacts of producing and using carbon-based fuels.

UW faculty members working on the project include: Urszula Norton, a UW associate professor of agroecology in the Department of Plant Sciences; Khaled Gasem, a professor of petroleum engineering; Hertanto Adidharma, an associate professor of petroleum engineering; and Gang Tan, an associate professor from the Department of Civil and Architectural Engineering. The research team also includes collaborators from Jackson State University, the University of Mississippi and the University of Delaware.

-- Jeff French, an assistant professor of atmospheric science, leads the project, titled Numerical Representation of Cloud and Precipitation Growth Processes and of the Effects of Glaciogenic Seeding on Orographic Clouds. Throughout much of the interior western United States, and in many arid regions around the globe, water supplies are derived from precipitation when moist air is lifted as it moves over mountain ranges. This precipitation, known as orographic precipitation, is often in the form of wintertime snowfall that feeds surface runoff and reservoirs, and replenishes subsurface aquifers.

Glaciogenic seeding of clouds -- that is, infusing clouds with chilled air or ice nuclei to encourage formation of ice particles -- is one technology that various Western communities have considered to glean additional water sources. Despite many studies of glaciogenic seeding, there are many important questions that remain unanswered. This project seeks to resolve some of these questions.

The project, called SNOWIE (Seeded and Natural Orographic Wintertime Clouds: the Idaho Experiment), is a collaborative research program aimed at addressing long-standing questions related to the initiation and growth of precipitation in orographic clouds. The project will apply new and advanced instrumentation, improved understanding of cloud dynamical and microphysical processes, and numerical modeling capabilities to evaluate the potential for orographic precipitation enhancement in ways not possible in decades past.

Besides French, UW researchers involved in the project are Bart Geerts, a UW professor of atmospheric science, and Wei Wu, a postdoctoral student. The project also includes collaborators from the University of Illinois and NCAR.

-- Zach Lebo, a UW assistant professor of atmospheric science, heads the project, titled Modification of Marine Boundary Layer Flow by Topography Along the Western United States Coastline, seeking to provide scientific understanding of how coastal terrain (specifically, that along the western U.S.) contributes to various phenomena in the jet stream that have been observed by a recent UW King Air research expedition.

Tom Parish, a UW professor of atmospheric science and head of the department, along with faculty from the University of Kansas, will collaborate on this NSF-funded project.

-- Xiaohong Liu, a UW professor of atmospheric science and the Wyoming Excellence Chair in Climate Modeling, will study the impacts of smoke aerosols on regional and global weather. The NSF project includes research collaborators from the Georgia Institute of Technology, Auburn University and the Department of Energys Pacific Northwest National Laboratory. It is known that large amounts of aerosols that result from forest fires can change the timing of formation of rain in clouds. This project will investigate how these aerosols affect the spatial distribution of precipitation and will require the new capabilities of the Cheyenne supercomputer.

-- Subhashis Mallick, an SER professor of geology and geophysics, will lead a project that will develop and use new algorithms to characterize the structural and petro-physical properties of a regions subsurface.

-- Wei Wang, a UW Ph.D. student in geology and geophysics, heads a project, titled Application of Full 3-D Waveform Tomography (F3DT) to Image Deep Critical Zone with Ambient-Noise Data. He will study the deep structure of the Blair-Wallis watershed, which is located between Laramie and Cheyenne.

This watershed is part of the Earths critical zone, which is the portion of the Earth that extends from the deepest reach of groundwater to the top of the vegetation. This zone sustains the majority of life on the Earth and, yet, fundamental questions -- such as What are the essential physical, chemical, hydrological and biological processes that control the formation and evolution of the critical zone? -- are still unanswered. The project will provide the data to enable researchers to make reliable inferences regarding how the critical zone evolves under climate, tectonic and anthropogenic events.

By the Numbers

The most recent recommended allocations total 65 million core hours, 147 terabytes of storage space, 75 terabytes of archival storage, and 33,000 hours on data analysis and visualization systems, Shader says.

To provide some perspective on what these numbers mean, Cheyenne can be thought of as 145,152 personal computers that are cleverly interconnected to perform as one computer. The computational time allocated is equivalent to the use of the entire supercomputer for 21 days, 24 hours a day. The 147 terabytes of storage would be enough to store the entire printed collection of the U.S. Library of Congress more than 20 times. Cheyenne consists of about 145,152 processors, also known as cores. An allocation of one core hour allows a project to run one of these processors for one hour, or 1,000 of these for 1/1,000th of an hour.

The supercomputer Cheyenne began operation in March. Its capability to perform 5.34 quadrillion calculations per second places it as the 20th most capable supercomputer in the world. It is more than 2.5 times more capable than its predecessor, Yellowstone.

The NWSC is the result of a partnership among the University Corporation for Atmospheric Research, the operating entity for NCAR; UW; the state of Wyoming; Cheyenne LEADS; the Wyoming Business Council; and Black Hills Energy. The NWSC is operated by NCAR under sponsorship of the NSF.

The NWSC contains one of the worlds most powerful supercomputers dedicated to improving scientific understanding of climate change, severe weather, air quality and other vital atmospheric science and geoscience topics. The center also houses a premier data storage and archival facility that holds historical climate records and other information.

Originally posted here:

Seven UW projects awarded use of Cheyenne supercomputer ... - Gillette News Record

Motherboard

Watch a Supercomputer's Fever Dream of the Chelyabinsk Asteroid Flameout Motherboard On Thursday, NASA released one of the most detailed simulations of the asteroid's dramatic flameout to date (at top), created by the agency's Pleiades supercomputer at the Ames Research Center. The short sequence used the aerodynamics software Cart3D ... Watch a Simulated Asteroid Hit the Atmosphere at 45000 Miles Per HourGizmodo The terrifying simulation that reveals what happens when an asteroid hits Earth's atmosphere at 45000 miles per hourDaily Mail all 85 news articles »

See the rest here:

Watch a Supercomputer's Fever Dream of the Chelyabinsk Asteroid Flameout - Motherboard

Seven projects, many of which have applications to Wyoming issues -- including identification of promising catalysts for carbon fuel and chemical production, cloud seeding to enhance precipitation and a study of the Blair-Wallis watershed -- were recently chosen to receive computational time and storage space on the supercomputer in Cheyenne.

University of Wyoming faculty members will head projects that will use the NCAR-Wyoming Supercomputing Center (NWSC). Each project was critically reviewed by an external panel of experts and evaluated on experimental design, computational effectiveness, efficiency of resource use and broader impacts such as how the project involves both UW and National Center for Atmospheric Research (NCAR) researchers; strengthens UWs research capacity; enhances UWs computational programs; or involves research in a new or emerging field.

The Wyoming-NCAR Allocation Panel recently met to evaluate the large allocation requests for the use of computational resources at the NCAR-Wyoming Supercomputing Center, says Bryan Shader, UWs special assistant to the vice president for research and economic development, and a UW professor of mathematics. The projects were granted allocations totaling 65 million core hours. In addition, 6 million core hours were recently awarded to a new faculty member as part of his startup package.

These projects represent new explorations that could not be undertaken without the capabilities of Cheyenne, the recently installed supercomputer at the NWSC, Shader says. Improved computer technology is enabling the study of more complicated or nuanced phenomena.

Over the last year, 23 UW projects used the NWSC. This includes four new projects that were allocated a total of 72 million core hours starting in January 2016. These allocations and use rank UW as first in total allocation and total users; first in total computer charges; and first in active projects among the more than 100 North American universities using the NWSC.

Since the NWSCs opening in October 2012, allocations have been made to 53 UW research projects, including these latest seven.

Cheyenne Projects

A brief description of each of the projects, which begin July 1, is provided below:

This watershed is part of the Earths critical zone, which is the portion of the Earth that extends from the deepest reach of groundwater to the top of the vegetation. This zone sustains the majority of life on the Earth and, yet, fundamental questions -- such as What are the essential physical, chemical, hydrological and biological processes that control the formation and evolution of the critical zone? -- are still unanswered. The project will provide the data to enable researchers to make reliable inferences regarding how the critical zone evolves under climate, tectonic and anthropogenic events.

By the Numbers

The most recent recommended allocations total 65 million core hours,147 terabytes of storage space, 75 terabytes of archival storage, and 33,000 hours on data analysis and visualization systems, Shader says.

To provide some perspective on what these numbers mean, Cheyenne can be thought of as 145,152 personal computers that are cleverly interconnected to perform as one computer. The computational time allocated is equivalent to the use of the entire supercomputer for 21 days, 24 hours a day. The 147 terabytes of storage would be enough to store the entire printed collection of the U.S. Library of Congress more than 20 times. Cheyenne consists of about 145,152 processors, also known as cores. An allocation of one core hour allows a project to run one of these processors for one hour, or 1,000 of these for 1/1,000th of an hour.

The supercomputer Cheyenne began operation in March. Its capability to perform 5.34 quadrillion calculations per second places it as the 20th most capable supercomputer in the world. It is more than 2.5 times more capable than its predecessor, Yellowstone.

The NWSC is the result of a partnership among the University Corporation for Atmospheric Research, the operating entity for NCAR; UW; the state of Wyoming; Cheyenne LEADS; the Wyoming Business Council; and Black Hills Energy. The NWSC is operated by NCAR under sponsorship of the NSF.

The NWSC contains one of the worlds most powerful supercomputers dedicated to improving scientific understanding of climate change, severe weather, air quality and other vital atmospheric science and geoscience topics. The center also houses a premier data storage and archival facility that holds historical climate records and other information.

Continue reading here:

Seven UW projects awarded space on supercomputer | Newsletter ... - Wyoming Business Report

Microsoft Azure has been taking the world by storm, and rightly so, considering that its not just your typical cloud storage platform. Rather, you can think of it as a framework for rethinking the way enterprise architectures are built and applications are designed. In a way, it gives enterprises all the different components needed to leverage cloud computing as a whole.

In fact, a study by Spiceworks shows that Microsoft Azure is the most popular Infrastructure as a Service (IaaS), and has grown at a phenomenal rate over the last few years. Another report by Forbes states that Azure is the only major platform thats been ranked as a leader consistently in IaaS, PaaS, and SaaS platforms.

Thats just to give you a peek into its popularity.

You may wonder whats the big deal about Azure? Why is it so popular? Well, the simple answer is Azure is way more than what meets the eye.

Its not your typical cloud computing platform because Microsoft is working on it all the time to improve its offerings and to extend its arm to cover every possible domain.

For example, in October 2016, Microsoft CEO Satya Nadella announced at an event in Dublin that Azure cloud will become the first artificial intelligence supercomputer in the world.

Microsoft

This announcement created a wave of excitement in the cloud community because AI is really the future.

But what did Nadella really mean in that announcement? How will Azure transform to an AI supercomputer?

To clarify these hundreds of doubts, Microsoft posted an explanation on its company website, according to which, Azure will bring in a set of apps based on AI technologies. The company believes that AI and machine learning are likely to define the next generation of apps powered by the cloud, and it wants to lead the way.

Following this announcement, Microsoft has been taking many steps to transform Azure into an AI supercomputer. Heres a look at what Microsoft has done so far in this ambitious effort.

The first step to implementing an idea is to form a solid team, and no one knows it better than Microsoft.

In October 2016, Microsoft formed an AI team. It is headed by Harry Shum, a 20-year old in the company who is known for his work in Cortana intelligence and Bing search. He will be assisted by more than 5,000 engineers and scientists who work in Microsofts AI and Research Group.

Gone are the days when CPUs were enough to power computers. With the growing computing power and data volumes, we need something more than CPUs, and this is where GPUs and FGPAs come into place.

Graphics Processing Units (GPUs) are powerful and programmable computing units that work in tandem with CPUs to enhance the processing capabilities of machines. Earlier, GPUs were mostly used for 3D game rendering, but thats changing, as companies realize their ability to handle more computational workloads. In this sense, GPU is a computational powerhouse.

To make the most of this resource, Microsoft is planning to build cloud processing power based on GPUs to meet the needs of the next generation of applications. In fact, the company plans to scale GPUs so they can process tasks in parallel. This will be ideal for high workloads.

Along with GPUs, Microsoft also plans to expand the use of whats called the Field Programmable Gate Arrays (FGPAs). At a high level, FGPAs allow developers to write all kinds of neural network code, spread it across multiple FGPAs, and run all of it at the speed of a silicon chip. To top it, FGPAs can be reprogrammed within seconds to respond to changes in artificial intelligence software or, for that matter, to even meet any unexpected event.

Both GPUs and FGPAs are most conducive for AI applications because they have the computing power, speed and workload-handling capability things that are going to be an integral part of the next generation of AI applications.

Already, Microsoft has started implementing some of these technologies in Azure. For example, its using the combined power of FGPA and GPU to create a virtual machine on Azure that can power 25GB per second with 10-times less latency. Imagine the speed of search with this computing power!

Thats just the beginning. As Microsoft does more research in this area, the speeds are going to increase and latency rates are going to reduce, thereby making it more conducive for AI applications.

GPUs and FGPAs are the computing bed, but thats not all. You need higher level services to build AI applications, and that needs APIs.

These APIs help to perform natural language processing, integrate speech recognition, enhance knowledge exploration, improve search, and more.

Heres a brief look into the APIs developed by Microsoft for boosting AI applications.

Computer vision API This API allows you to analyze an image to identify content, tag, label, and create all kinds of domain-specific models. Content moderator API This API will allow machine-assisted moderation of content, images, and videos. In addition, itll augment human review by providing machine learning models. Emotion API This API detects the motions on a persons face to identify the moods and feelings, so responses can be personalized accordingly. Face API This API helps to detect human faces, compare similar ones, and organize them into groups based on facial similarities. You can use this along with emotion API to provide the best possible response to every customer. Video API With this API, you can process videos intelligently to analyze faces and images, smoothen videos, and do so much more.

You can use these APIs by themselves or in combination with others to build and provide advanced features in your application. These can come particularly handy when you want to create AI-based applications on Azure.

With these strategies in place, Microsoft is all set to advance its ambitions of making Azure the first AI supercomputer. However, it faces intense competition from rivals like IBM, AWS, and Google.

Already, IBM has started offering its AI engine, Watson, as a service. Its also planning to combine Watson with other services such as IBM Data Science Experience to create advanced products thatll give its clients more capabilities than ever before.

Likewise, Google has been working on advanced machine learning plans to bring in more customers to the Google Cloud Platform. It is also looking to tap into an open-source library called Tensorflow to further its AI plans.

Amazon is not to be left behind, too. A few months back, it unveiled a plan for its GPU-powered cloud computing service to offer AI-based services, genomics, molecular modeling, and more.

In the light of this competition, its important for Microsoft to move ahead faster with its plan to get a big slice of the AI market before others start closing in.

Nadella announced that the company is planning to make the Azure platform as the first AI supercomputer in the world. Over the last few months, a lot of progress has happened in this area that includes the use of FGPAs and GPUs for computing and the development of APIs for communicating with AI applications.

Though theres more work needed in this area, Microsoft is nevertheless on the right path to fulfilling its ambitious plans.

Exciting days are ahead for both Microsoft and for the tech industry as a whole.

Post Views: 69

Link:

Microsoft's lofty cloud goal: Make Azure the first AI supercomputer - TechGenix (blog)

June 27, 2017 Example of hypothetical seismic wave propagation with mountain topography using the new EDGE software. Shown is the surface of the computational domain covering the San Jacinto fault zone between Anza and Borrego Springs in California. Colors denote the amplitude of the particle velocity, where warmer colors correspond to higher amplitudes. Credit: Alex Breuer, SDSC

Record-setting seismic simulations run earlier this year on the Cori supercomputer at Lawrence Berkeley National Laboratory's National Energy Research Scientific Computing Center (NERSC) were the subject of two presentations at the ISC High Performance conference in Frankfurt, Germany this week.

One of the presentations details a new seismic software package, developed by researchers at the San Diego Supercomputer Center (SDSC) in conjunction with Intel, that enabled the fastest seismic simulation ever run to date: 10.4 petaflops/s. The largest simulation used 612,000 Intel Xeon Phi Knights Landing (KNL) processor cores on the Cori KNL system. The simulations, which mimic possible large-scale seismic activity in Southern California, were done using a new software system called EDGE (Extreme-Scale Discontinuous Galerkin Environment), a solver package for fused seismic simulations.

"In addition to using the entire Cori KNL supercomputer, our research showed a substantial gain in efficiency in using the new software," said Alex Breuer, a postdoctoral researcher from SDSC's High Performance Geocomputing Laboratory (HPGeoC) who presented the findings at ISC17. "Researchers will be able to run about two to almost five times the number of simulations using EDGE, saving time and reducing cost."

This research was enabled by the NERSC Director's Reserve. The research team also helped NERSC test, debug and validate the Cori KNL system at scale soon after Cori was installed, noted Richard Gerber, head of NERSC's high performance computing department.

A second HPGeoC paper presented at ISC17 covers a new study of the AWP-ODC software that has been used by the Southern California Earthquake Center (SCEC) for years. The software was optimized to run in large-scale for the first time on the latest generation of Intel data center processors, called Intel Xeon Phi x200.

These simulations, which also used the Cori KNL system, attained competitive performance to an equivalent simulation on the entire GPU-accelerated Titan supercomputer, located at Oak Ridge National Laboratory, and has been the resource used for the largest AWP-ODC simulations in recent years. Additionally, the software obtained high performance on Stampede-KNL at the Texas Advanced Computing Center at The University of Texas at Austin.

Both projects are part of a collaboration announced in early 2016 under which Intel opened a parallel computing center (PCC) at SDSC to focus on seismic research, including the ongoing development of computer-based simulations that can be used to better inform and assist disaster recovery and relief efforts.

Such detailed computer simulations allow researchers to study earthquake mechanisms in a virtual laboratory. "These two studies open the door for the next-generation of seismic simulations using the latest and most sophisticated software," said Yifeng Cui, founder of the HPGeoC at SDSC, director of the Intel PCC at SDSC and PI on this research. "Going forward, we will use the new codes widely for some of the most challenging tasks at SCEC."

Explore further: SDSC joins the Intel Parallel Computing Centers program

The San Diego Supercomputer Center (SDSC) at the University of California, San Diego, is working with semiconductor chipmaker Intel Corporation to further optimize research software to improve the parallelism, efficiency, ...

A software toolkit developed at the Department of Energy's Lawrence Berkeley National Laboratory (Berkeley Lab) to better understand supercomputer performance is now being used to boost application performance for researchers ...

For the past year, staff at the Department of Energy's National Energy Research Scientific Computing Center (NERSC) have been preparing users of 20 leading science applications for the arrival of the second phase of its newest ...

U.S. Department of Energy (DOE) high-performance computer sites have selected a dynamic fusion code, led by physicist C.S. Chang of the DOE's Princeton Plasma Physics Laboratory (PPPL), for optimization on three powerful ...

The U.S. Department of Energy's (DOE) National Energy Research Scientific Computing (NERSC) Center and Cray Inc. announced today that they have finalized a new contract for a Cray XC40 supercomputer that will be the first ...

Researchers at the San Diego Supercomputer Center (SDSC) at the University of California, San Diego, are the recipient of NVIDIA's 2015 Global Impact Award for their collaborative work in developing an accelerated GPU (graphics ...

A new and highly virulent outbreak of malicious data-scrambling software appears to be causing mass disruption across the world, hitting companies and governments in Europe especially hard.

After a seven-year legal battle, European authorities came down hard on Google on Tuesday for taking advantage of its dominance in online searches to direct customers to its own businesses, fining the tech giant a record ...

While doing research at the Woods Hole Marine Biological Laboratory in Massachusetts, Sindy Tang learned of a remarkable organism: Stentor coeruleus. It's a single-celled, free-living freshwater organism, shaped like a trumpet ...

Mobile phone carriers scooped up airwaves no longer needed by television broadcasters last March in a $19-billion auction designed by UBC and Stanford University researchers.

Inside a cavernous northern Utah warehouse, hydraulic engineers send water rushing down a replica of a section of a dam built out of wood, concrete and steeltrying to pinpoint what repairs will work best at the tallest ...

Paris' Cathedral of Notre Dame has a ghost orchestra that is always performing, thanks to a sophisticated, multidisciplinary acoustics research project that will be presented during Acoustics '17 Boston, the third joint meeting ...

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Read more here:

Record-setting seismic simulations run on the Cori supercomputer - Phys.Org

IBM and the US Air Force have announced that theyre teaming up to build a unique supercomputer based on IBMs TrueNorth neuromorphic architecture. The new supercomputer will consist of 64 million neurons and 16 billion synapses, while using just 10W of wall power.

Its common, when discussing CPUs, to compare them with the human brain. Superficially, the two seem similar brains, like CPUs, receive inputs, perform calculations based on those inputs, and then return a result. But while brains and conventional CPUs may seem similar at an extremely high level, that similarity disintegrates as soon as you start examining either system in any detail.

Transistors are binary (theyre either on or off), and they can only change the behavior of other transistors that theyre directly connected to. Neurons, in contrast, have both an analog and a binary aspect. The dendrites the receiving arms of a neuron have analog function in that they give a little electrical ripple called a graded potential whenever they get a ping from an upstream neuron. If they send enough graded potentials to the cell body, the latter then sends a binary off/on pulse train down the axon. The axon of a nerve sort of speaks in binary, even though it has to accommodate both binary and analog input.

Scientists have spent decades creating software models that more closely resemble the way brains process information. But theres an enormous efficiency gap when attempting to simulate something as different as a brain on modern silicon. While modern CPUs may be millions of times faster at certain calculations than any human, the human brains power efficiency is orders of magnitude better than the most efficient conventional CPU we can build.

IBMs TrueNorth project is an attempt to build a neuromorphic, or brain-like CPU directly in hardware. The goal is to design superior neural nets and create artificial intelligence in power envelopes that could conceivably operate outside of data centers or fixed installations.

IBM isclaiming that the TrueNorth Neurosynaptic System (thats the official moniker) can convert and process data from multiple sources in parallel, while simultaneously pairing with more conventional processors to analyze the data.

The US Air Force Research Laboratory (AFRL) was the earliest adopter of TrueNorth for converting data into decisions, said Daniel S. Goddard, director, information directorate, U.S. Air Force Research Lab. The new neurosynaptic system will be used to enable new computing capabilities important to AFRLs mission to explore, prototype, and demonstrate high-impact, game-changing technologies that enable the Air Force and the nation to maintain its superior technical advantage, he said in a statement.

The new system will fit in a 4U standard server rack with 512 million neurons in total per rack. IBM claims this represents an 800-percent annual increase over the last six years, as the first systems contained just 256 neurons.

Now read:How neuromorphic brain chips will begin the next era in computing

Excerpt from:

IBM, US Air Force Are Building a Neuromorphic Supercomputer ... - ExtremeTech

Courtney Maums Touch and the inevitability of a technological singularity.

Ray Kurzweil says that a technological singularity will occur around 2045.He said earlier this year that the process has already begun: By 2029, computers will have human-level intelligence putting them inside our brains, connecting them to the cloud, expanding who we are. Today, thats not just a future scenario Its here, in part, and its going to accelerate.

So how do we bide our time for the next 12 years? Everyone hates being on Facebook. I hate this website has been a common refrain on Twitter lately, overused and then appropriated and cycled in and out of abyssal irony. So even though Kurzweil and Softbank CEO Masayoshi Son are convinced that the singularity will occur in the mid-2040s,its easy to imagine people unplugging or going off the grid (pick your clich) in the next couple of years. I could see flip phones coming back and a mass exodus from Facebookhowever temporary, people are eager to reclaim something vague they feel was lost in the last decade.

Facebook feels like a prison: you have to be on your best behavior, or your grandmother might not be happy, or your old boss or your ex or that friend you havent seen in ages and would like to keep it that way. Writing a Facebook status is so formal and airlessits where you go to talk about your engagement, your crises, and cancer diagnosis. Mothers and Fathers Day are the worstcompletely unusable, youre just bombarded by earnest posts and old photos and hopeless laments of loss and people thatll never be the same. Its beyond information overload: its a sneak peek at the impending merger of all of humanity into one super-consciousness, totally zen and spaced out and floating in some Kubrickian cosmic womb.

But were not quite there yet. The valley is not yet canny. Its all still pretty messy. Touch, Courtney Maums second novel, follows trend-forecaster Sloane Jacobsen, fortysomething tech wizard with a pseudo-supernatural ability to see where the cultures going and what people will want in five or 10 years (most famously, she invented the ubiquitous swipe on tablets and smartphones). Shes bogged down by her bumbling husband Roman, an oblivious and unintentionally hilarious Frenchman whos taken to wearing skintight Zentai suits and caused a minor sensation: Traipsing about town in his metallic gold one, riding the metro, contemplating the Seine. The alleged elegance and nonchalance with which Roman appropriated fetish custom thrilled the bougie masses. Overnight, his Instagram account became supercharged. Two hundred thousand, four hundred thousand: Sloane had stopped checking before she saw it reach five.

Meanwhile, Sloane struggles with her latest project of predictions: she sees people reproducing less, with products marketed toward singles and couples that de-stigmatize remaining childless. Shes brought on to work at a typically sexist tech company called Mammoth run by an idiot caricature named Daxter (hed fit right in at Uber). Sloanes gut and gift drift away from the childless and impersonal to the exact opposite: forgoing smartphones and social media accounts in favor of hug hotspots and greater interpersonal interaction as revolution against phone zombies. Everyone feels too connected today, but for a myriad of reasons (family, work, addiction), maintaining a Facebook, Instagram, or Twitter account remains a necessity. Totally trapped, unable to do anything about Mark Zuckerbergs pathetic attempt at being a politician and the very real possibility that he could be a presidential nominee in the next 15 years.

Eventually this is all realized: Roman writes an op-ed on spec called After God Goes Sex about the death of penetrative sex: I have a theory that were entering a period that is post-sex. People want beyond that, yes? They want moreit is too simple, the I-put-this-into-you, you-enter-me, the pounding? So many people are finding their virtual sex lives so much richer than their real ones. Having recently been fucked properly by a younger co-worker Jin (her first vaginal orgasm in ten years), Sloane is horrified by her doofus husband and his tacit acknowledgment of their frigid relationship. Then the op-ed gets picked up by The New York Times on spec, goes viral, and Sloane about loses her mind, finding comfort in one of the novels more endearing and sweet characters: her loyal, driverless company car Anastasia, a piece of AI thats actually functional and emotionally intelligent unlike Siri, who seems stillborn, introduced the same month that Steve Jobs died.

Sloane has her own moment via an admittedly overused device: she goes off in a Mammoth board meeting in front of Daxter and all of her co-workers, defiantly quitting and promising that Romans neo-sensualist bullshit is a flash in the pan, and that people will soon cut as many cords as they can and go out into the world and talk and touch and be touched.

Its a pipedream, and Maum gives Sloane one paragraph of doubt: Maybe she was wrong, and humanity would never right itself, leaving machine and Homo sapiens to meld into one, tourists now preferring to visit the Grand Canyon virtually because they were too consumed by social anxiety to go that far away. (Again, the shadow of Infinite Jest looms above.) Touch is a Pollyanna fantasy sent into the world right before the deluge hits and were all cyborgs with nanobots running through our blood. Sloanes predictions mirror similar delusions about the severity of climate change: But maybejust maybepeople were more ready to than she thought to shove off the shackles of psychosomatic loneliness. Maybe the revolution could come earlier, if they bucked the trends.

I think we all know that the levees are about to break on both fronts. The question is which comes first? Apocalyptic floods or the transcendent super-consciousness? Wheres that thing going to be stored? Even the arctic seed bank has been compromised. What if we reach a glorious singularity only to be fried by the tide? No one left to put us in a bag of rice overnight. Lets hurry up and jump into the matrix and avoid the ashen misery of The Road. Dont forget to bring a friend.

Follow Nicky Smith on Twitter: @MUGGER1992

Go here to see the original:

Sympathy for the Supercomputer - Splice Today

In the recent ISC17 conference in Frankfurt, Germany, the HPC communitys attention was inevitably focused on big supercomputers, new processors, and other technologies. Of course, there were a number of voices also calling for attention to the non-hardware aspects of HPC, especially software and people. However, most of those discussions centered on programmers, sysadmins, and researchers using HPC, as is customary in the HPC community.

Source: TACC

In a similar manner, training courses at HPC centres are almost exclusively aimed at programmers and new users. Many people reading the title of this article might think of the future leading computational scientists, that is, users, or computer science/technology researchers.

But, there are many people whose primary role is running HPC centres, fighting for funding, architecting and delivering HPC services to users, to enable that computational science and engineering output. Most of these HPC leaders have evolved into the role from a background as a user or as service staff. Unfortunately, the development and training opportunities to help future HPC service managers learn essential skills are scarce.

A number of HPC leaders in industry have recognised this problem, and have called for an opportunity for their HPC systems and computational sciences staff to learn about the business processes required to deliver a successful HPC environment. They believe this training will be essential in developing future HPC leaders and improving the collaboration with the researchers they support.

Some readers will know that the author of this article has been actively trying to address this issue for several years, with tutorials (e.g., SC13, SC16, OGHPC17) webinars and conference talks. Due to this demand from the HPC community, TACC and NAG have combined to create the first HPC for Managers training program in Austin Texas, September 12-14. Topics covered will include strategic planning, technology options, procurement and deployment, people issues, total cost of ownership, performance measurement, and more. A broad scope of HPC is covered from department scale clusters to the largest supercomputers, and consideration of cloud.

This training is a major step forward in professional training for HPC managers and developing the future leaders in the community. Registration is open now. Details and a full agenda can be found at: https://www.tacc.utexas.edu/education/institutes/hpc-concepts-for-managers

Andrew Jones can be contacted via twitter (@hpcnotes), via LinkedIn (https://www.linkedin.com/in/andrewjones/), or via the TOP500 editor.

View original post here:

Where Will Future HPC Leaders Come From? - TOP500 News

The 8th Irish Supercomputer List was released last week, featuring two new Top500-class machines. The new entrants, from an undisclosed software company, feature at spots 196 and 197 on the 49th Top500 Supercomputer List, each with a Linpack Rmax of 819.16 TFlop/s. This more than doubles the Irish HPC capacity, up from 1.46 PFlop/s to 3.01 Pflop/s.

Ireland has ranked on the Top500 list 29 times over a history of 23 years with a total of 18 machines. Over half of these machines (11) and rankings (18) have been in the last 6 years, representing Irelands increasing pace of HPC investment.

The software company that owns the machines, known as Software Company M on the Top500 list, occupies 6 of the 7 top ranks (1-5 & 7) on the current Irish list. All six are HP Cluster Platform family installations. These are the 2nd and 3rd highest Irish Top500 rankings of all time, the top Irish Top500 performer being Stokes, the ICHEC (Irish Centre for High-End Computing) machine which was ranked 117 on the Top500 in November 2008. Stokes also holds the Irish Top500 longevity record, ranking on four consecutive Top500 lists, with a performance of 25.11 Tflop/s. This is 3.1% of the Rmax performance of the new machines. The current number 6 machine is Fionn, maintained by ICHEC, with a performance of 140.4 Tflop/s.

Sign up for our insideHPC Newsletter

Read more from the original source:

HPC Capacity Doubles on Irish Supercomputer List - insideHPC

BBC News

BBC - Future - Meet the world's most powerful computer BBC News Two decades ago China did not have a single supercomputer and now it has the world's most powerful. and more »

See the rest here:

BBC - Future - Meet the world's most powerful computer - BBC News

WHICH Championship side will be leading the pack after the first five games? A super computer may have the answer.

WHICH Championship side will be leading the pack after the first five games? A super computer may have the answer.

1 / 24

The Championship fixture list was released earlier this week and excitement is already beginning to build for the season ahead.

Aston Villa, Norwich and Middlesbrough are all thought to be the teams most likely to compete for the title next year, according to the bookies.

But there are plenty of sides below them who should be attempting to make it into the play-off spots.

Leeds begin the campaign with new manager Thomas Christiansen.

The Elland Road side will be hoping to do even better than last season after missing out on the play-offs due to a poor run of results late on.

And fans of the Yorkshire club will be delighted to see that talkSPORTs super computer has predicted their club to start the season brightly.

But which team will be leading the table after five matches?

Click through the gallery above to see the super computers Championship predicted table after five matches.

Go here to see the original:

Super computer predicts how Championship will look after five matches Leeds doing well! - Daily Star

To put it frankly, China has been absolutely killing the competition in the TOP500 supercomputer rankings for the last couple of years. Its Sunway TaihuLight has been sitting at the top of the pile well ahead of the Tianhe-2, another Chinese supercomputer. Now, the country is saying that it has a new computer which will be ready in 12 months.

The new Sunway exascale computer being developed by the National Supercomputer Center (NSC) and the National Research Center of Parallel Computer Engineering and Technology (NRCPC) will be able to execute a quintillion calculations per second, making it eight times faster than the Sunway TiahuLight, which scored 93 petaflops to make first place on the TOP500 list.

The new supercomputer has already gone into production in Jinan. Once completed, the new computer will support further research and scientific applications in fields including marine environments, biological information, aviation, and aerospace. As we get better supercomputers, computationally intensive tasks such as predicting the weather and climate change will become easier to perform and the results will be more accurate.

In November, we reported that Japan would be gunning for the top spot on the supercomputer rankings. The Japanese Ministry of Economy, Trade and Industry had planned to spend $173 million on developing a new supercomputer called ABCI or AI Bridging Cloud Infrastructure. The country hopes that it can edge in front of China by scoring 130 petaflops on the Linpack benchmark. With the latest news from China, Japan might not have the opportunity to move ahead after all.

Source: CGTN

Follow this link:

China plans new supercomputer for June 2018 - Neowin

EMERGING TECH

IBM and the Air Force Research Laboratory have partnered to develop an artificial intelligence-based supercomputer with what they call abrain-inspired, neural network design.

Based on a 64-chip array, the company and AFRL are designing the newIBM TrueNorth Neurosynaptic System to recognize patterns and carry outintegrated sensory processing functions. IBM first developed a TrueNorth platform for a Defense Advanced Research Projects Agency program in partnership with Cornell University.

Both IBM and AFRL envision TrueNorth as able to convert data such as images, video, audio and text from multiple, distributed sensors into symbols in real time. AFRL seeks to combine that so-called "right-brain" function with "left-brain" symbol processing capabilities in conventional computer systems.

The goal is to enable multiple data sources to run in parallel against the same neural network and help independent neural networks form an ensemble to also run in parallel on the same data.

Once complete, the new TrueNorth platform's processing power would aim to equal that of 64 million neurons and 16 billion synapses as the processor component consumes energy equal to that of a 10-watt light bulb.

AFRL is investigating potential uses of the system in embedded, mobile and autonomous settings where limitations exist on the size, weight and power of platforms.

About the Author

Ross Wilkers is a senior staff writer for Washington Technology. He can be reached at rwilkers@washingtontechnology.com. Follow him on Twitter: @rosswilkers. Also find and connect with him on LinkedIn.

Go here to see the original:

IBM, Air Force to collaborate on brainy supercomputer -- Washington ... - Washington Technology

Today IBM announced they are collaborating with the U.S. Air Force Research Laboratory (AFRL) on a first-of-a-kind brain-inspired supercomputing system powered by a 64-chip array of the IBM TrueNorth Neurosynaptic System. The scalable platform IBM is building for AFRL will feature an end-to-end software ecosystem designed to enable deep neural-network learning and information discovery. The systems advanced pattern recognition and sensory processing power will be the equivalent of 64 million neurons and 16 billion synapses, while the processor component will consume the energy equivalent of a dim light bulb a mere 10 watts to power.

AFRL was the earliest adopter of TrueNorth for converting data into decisions, said Daniel S. Goddard, director, information directorate, U.S. Air Force Research Lab. The new neurosynaptic system will be used to enable new computing capabilities important to AFRLs mission to explore, prototype and demonstrate high-impact, game-changing technologies that enable the Air Force and the nation to maintain its superior technical advantage.

IBM researchers believe the brain-inspired, neural network design of TrueNorth will be far more efficient for pattern recognition and integrated sensory processing than systems powered by conventional chips. AFRL is investigating applications of the system in embedded, mobile, autonomous settings where, today, size, weight and power (SWaP) are key limiting factors.

The IBM TrueNorth Neurosynaptic System can efficiently convert data (such as images, video, audio and text) from multiple, distributed sensors into symbols in real time. AFRL will combine this right-brain perception capability of the system with the left-brain symbol processing capabilities of conventional computer systems. The large scale of the system will enable both data parallelism where multiple data sources can be run in parallel against the same neural network and model parallelism where independent neural networks form an ensemble that can be run in parallel on the same data.

The evolution of the IBM TrueNorth Neurosynaptic System is a solid proof point in our quest to lead the industry in AI hardware innovation, said Dharmendra S. Modha, IBM Fellow, chief scientist, brain-inspired computing, IBM Research Almaden. Over the last six years, IBM has expanded the number of neurons per system from 256 to more than 64 million an 800 percent annual increase over six years.

The system fits in a 4U-high (7) space in a standard server rack and eight such systems will enable the unprecedented scale of 512 million neurons per rack. A single processor in the system consists of 5.4 billion transistors organized into 4,096 neural cores creating an array of 1 million digital neurons that communicate with one another via 256 million electrical synapses. For CIFAR-100 dataset, TrueNorth achieves near state-of-the-art accuracy, while running at >1,500 frames/s and using 200 mW (effectively >7,000 frames/s per Watt) orders of magnitude lower speed and energy than a conventional computer running inference on the same neural network.

The IBM TrueNorth Neurosynaptic System was originally developed under the auspices of Defense Advanced Research Projects Agencys (DARPA) Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE) program in collaboration with Cornell University. In 2016, the TrueNorth Team received the inaugural Misha Mahowald Prize for Neuromorphic Engineering and TrueNorth was accepted into the Computer History Museum. Research with TrueNorth is currently being performed by more than 40 universities, government labs, and industrial partners on five continents.

Sign up for our insideHPC Newsletter

Continued here:

AFRL Taps IBM to Build Brain-Inspired AI Supercomputer - insideHPC - insideHPC

If the OddsShark Super Computer becomes a sentient being, were all doomed. Its gone 9-1 against the spread and 8-2 straight up during the 2017 NFL postseason and will more than likely become our robot overlord sooner rather than later if it keeps improving at the rate it has. All Matrix theories aside, the computer is cleaning up and is back one more time this season for its Super Bowl 51 pick and its siding with the underdog Atlanta Falcons.

The computer has been riding the Falcons all through the playoffs and correctly predicted that theyd blow out both the Seahawks and Packers. This pick is a little different, however, as the Dirty Birds will now have to face the consensus best team in football for the right to lift the Lombardi Trophy.

With a predicted score line of 29.3-21.1 for the Falcons, the computer is very confident in Atlanta and I have to agree. I wrote about the three reasons why the Falcons are going to win the Super Bowl, so it appears the computer and I are wired quite similarly. Weve been on the same page on just about everything this postseason so maybe when the machines rise up, theyll keep me around as a pet or something.

Another significant note regarding that projected score is that it would not come anywhere close to breaching the record-setting total that opened at 58.5. The public is heavily on the side of the OVER and given how these two offenses have been playing, its hard to disagree.

A loss and failure to cover for the Patriots here would be just their fourth ATS loss of the season. Win or lose, that's an incredible record and if you've beenbacking them, congratulations you probably don't even need a win here.

Although the Falcons specific trends dont exactly shine a great light on Atlantas SB odds, the underdog has won and covered the last five years at the Super Bowl. I know thats not incredibly specific to either of these teams but itdefinitely paints a telling picture. Vegas sets lines very carefully to get the most money on the side they think will lose and opening the Patriots as a small favorite has definitely done that.

Over 60% of the public is on the Patriots and if the computers right, the majority of people betting on a side will be very disappointed when the dust settles on Super Bowl Sunday.

The computer has completely disregarded the Patriots' 4-0 SU and ATS record against the Falcons in their last four meetings and the Dirty Birds' 0-5 SU and ATS record in their last five games as underdogs in the playoffs.

For a more human breakdown of the biggest pro sporting event in North America, check out the following links and make sure to check out our YouTube channel for all our video content:

See original here:

Super Bowl 51 Super Computer Picks | Odds Shark

IBM and the USAF announced on Friday that the machine will run on an array of 64 TrueNorth Neurosynaptic chips. The TrueNorth chips are wired together like, and operate in a similar fashion to, the synapses within a biological brain. Each core is part of a distributed network and operate in parallel with one another on an event-driven basis. That is, these chips don't require a clock, as conventional CPUs do, to function.

What's more, because of the distributed nature of the system, even if one core fails, the rest of the array will continue to work. This 64-chip array will contain the processing equivalent of 64 million neurons and 16 billion synapses, yet absolutely sips energy -- each processor consumes just 10 watts of electricity.

Like other neural networks, this system will be put to use in pattern recognition and sensory processing roles. The Air Force wants to combine the TrueNorth's ability to convert multiple data feeds -- whether it's audio, video or text -- into machine readable symbols with a conventional supercomputer's ability to crunch data.

This isn't the first time that IBM's neural chip system has been integrated into cutting-edge technology. Last August, Samsung installed the chips in its Dynamic Vision Sensors enabling cameras to capture images at up to 2,000 fps while burning through just 300 milliwatts of power.

Read this article:

The Air Force and IBM are building an AI supercomputer - Engadget

Six of the seven most powerful computers in Ireland are owned by one company, with new entries on the list more than doubling the countrys HPC capacity.

Investment in high-performance computers (HPCs) in Ireland is continuing apace, with two new machines in recent months storming into the worldwide top 200.

Known only as Company M, a software company, rather than research centre, has seen its latest toys enter the global ranking of supercomputers at 196 and 197, respectively.

These supercomputers represent the second- and third-highest positions ever recorded by Irish computers on the global Top500 list they are both a Linpack Rmax of 819.16teraflops.

In 2008, a Xeon quad core machine operated by the Irish Centre for High-End Computing (ICHEC) reached 117th on the list, falling out of the top 100 within two years.

ICHEC still has one of Irelands most powerful machines, though, with Fionn the only computer outside of Company Ms array that makes it into the top seven domestically (sixth).

This more than doubles the Irish HPC capacity, which is up from 1.46 petaflops in November 2016, to 3.01 petaflopstoday.

Ireland has ranked on the Top500 list 29 times over a history of 23 years, with a total of 18 machines. More than half of these machines (11) and rankings (18) have been in the last six years, representing Irelands increasing pace of HPC investment.

The continued growth of the Irish Supercomputer List reflects an exciting period of high-performance computing expansion, said Dr Brett Becker of the School of Computer Science, University College Dublin.

With emerging technologies in data analytics, AI and machine learning driving the proliferation of high-performance computing globally, it is important that Ireland continues to invest in high-performance computing, said Becker, who also maintains the Irish Supercomputer List.

Participating as close to the top of the overall global computing list is important, he said, in order to remain globally competitive in todays emerging technologies that promise to drive the future economy and to improve the quality of peoples lives.

Irelands history in the global Top500 supercomputer ranking. Click to enlarge. Image: Irish Supercomputer List

Two Chinese supercomputers and an upgraded supercomputer in Switzerland rank ahead of the US now in the overall global list, released earlier this week.

Chinalast yearrevealed the most powerful machinein the world, the Sunway TaihuLight, with 93 petaflops of processing power. It is this machine that still reigns supreme.

Now, the supercomputer arms race is heating up once again, with news that the US Department of Energy is pumping $258m into research in this field across six American tech companies: IBM, Intel, HP Enterprise, Nvidia, Cray and AMD.

The purpose of the PathForward programme, the department said, is to maximise the energy efficiency and overall performance of future large-scale supercomputers.

The rest is here:

What is the most powerful supercomputer in Ireland? - Siliconrepublic.com

One of the more unusual pieces of news at this years ISC High Performance conference was the announcement by the National Supercomputing Center in Wuxi that it will be offering a cut-down version of the Sunway TaihuLight supercomputer for more mainstream HPC users.

TaihuLight is the reigning champ on the TOP500 list, delivering a whopping 93 petaflops on the Linpack benchmark. Besides being the number one system, its other big claim to fame is that it is constructed almost entirely from Chinese-made componentry. In particular, the system is powered by the 260-core ShenWei processor, known as the SW26010. Each of TaihuLights 40,960 ShenWei chips delivers three teraflops of peak performance.

The commercial version they announced at ISC is called the Sunway Micro and is based a dual-socket SW26010 server node. The system is aimed at a broad spectrum of industrial and research applications including deep learning, oil & gas exploration, climate modeling, etc.

Source:National Supercomputing Center in Wuxi

The two-processor design means each node delivers a very respectable six peak teraflops. Unlike the TaihuLight supercomputer, whose single-socket nodes were outfitted with a scant 32 GB of memory, the Sunway Micro can be equipped with 64 GB to 256 GB. That gives Micro buyers the option to have lot more local memory to feed these high-flying ShenWei chips. Each node is also equipped with 12 GB of local storage of undefined type and origin.

While talking with some of the folks at the Wuxi booth during the ISC exhibition, they revealed that the Micro nodes can be clustered together via a network based on InfiniBand technology, which apparently is similar, but not identical to the TaihuLight network implementaion. Given that these servers will be used in relatively small clusters, they didnt have to develop a network for supercomputer-level scalability.

One of the most unusual aspects of the Sunway Micro is that it is being sold by the National Supercomputing Center in Wuxi. That might seem like an odd thing for a supercomputing center to do, given its public mission. But since the center supplies the system software and developer toolset for these ShenWei-based machines, they basically act as system integrators for the commercial offering. As for the TaihiLight, the Micro was developed by the National Research Center of Parallel Computer Engineering & Technology (NRCPC).

Software support includes C/C++ and Fortran compilers for the ShenWie, as well as supporting runtime libraries. For parallel software development, Wuxi includes MPI, OpenACC and Athread implementations targeted to the ShenWei platform. An integrated development environment, with a debugger and performance monitor, are also included.

Besides selling the standard version of the Micro, the Wuxi center will also provide customized solutions. Pricing for the system was not made public.

Go here to read the rest:

Makers of TaihuLight Supercomputer Offer Commercial Version - TOP500 News