supercomputing
Latest
Microsoft invests $1 billion in Elon Musk-founded OpenAI
Today, Microsoft announced that it's investing $1 billion in the Elon Musk-founded company OpenAI. The two companies will work together to bring supercomputing technologies and AI to Microsoft Azure. And OpenAI will run its services exclusively in Microsoft's cloud. Ultimately, the partners hope to build artificial general intelligence (AGI), a technology that some say will match or exceed human intellect.
China's supercomputers are the latest target in US trade war
The US and China have been locked in a race for the world's most powerful supercomputer. China was in the lead with its Sunway TaihuLight, which has a 93 petaflop capacity. But the US surpassed that last year, when it released the Summit, which can run at 200 petaflops -- or 200 quadrillion calculations per second. Now, the US is using export restrictions in an attempt to thwart China's supercomputing efforts.
The US again has the world's most powerful supercomputer
The Department of Energy pulled back the curtain on the world's most powerful supercomputer Friday. When Summit is operating at max capacity, it can run at 200 petaflops -- that's 200 quadrillion calculations per second. That smokes the previous record holder, China's Sunway TaihuLight (which has a 93 petaflop capacity). Summit is also about seven times faster than Titan, the previous US record holder which is housed at the same Oak Ridge National Lab in Tennessee. For perspective, in one hour, Summit can solve a problem that it would take a desktop computer 30 years to crack.
IBM's new computers may change how we process big data
IBM HQ. INT. DAY. An oppressive curtain of rain beats down against the window of this small meeting room in IBM's New York HQ. Two IBM scientists are engaging in excited conversation with a representative from the DEPARTMENT OF ENERGY while a third IBM employee sits at one end, removed from the action, chewing their pen and staring out of the window. DOE REPRESENTATIVE: What do you mean $300 million won't get me the two supercomputers I need? SCIENTIST #1: You're generating too much data. There isn't enough bandwidth for it all to go through. DOE REPRESENTATIVE: I don't understand. You're the supercomputer company, make it happen!
IBM pushing System z, Power7+ chips as high as 5.5GHz, mainframes get mightier
Ten-core, 2.4GHz Xeons? Pshaw. IBM is used to the kind of clock speeds and brute force power that lead to Europe-dominating supercomputers. Big Blue has no intentions of letting its guard down when it unveils its next generation processors at the upcoming Hot Chips conference: the company is teasing that the "zNext" chip at the heart of a future System z mainframe will ramp up to 5.5GHz -- that's faster than the still-speedy 5.2GHz z196 that has led IBM's pack since 2010. For those who don't need quite that big a sledgehammer, the technology veteran is hinting that its upcoming Power7+ processors will be up to 20 percent faster than the long-serving Power7, whose current 4.14GHz peak clock rate may seem quaint. We'll know just how much those extra cycles mean when IBM takes to the conference podium on August 29th, but it's safe to say that our databases and large-scale simulations won't know what hit them.
Intel christens its 'Many Integrated Core' products Xeon Phi, eyes exascale milestone
Been wondering when the next big leap in high performance computing would hit? Well, Intel would like you to believe the time is now and the name of that revolution is the Xeon Phi. Formerly codenamed Knights Corner, the Many Integrated Core product is pushing the field of supercomputers into the era of the exaflop by squeezing a teraflop of performance into a package small enough to plug into a PCIe slot. The Phi brand will, at first at least, be applied to specialized coprocessors designed for highly parallel tasks. The chips are built using Intel's 22nm manufacturing process and 3-D TriGate transistors, piling in more that 50 cores in an effort to combat the inroads made by GPU companies like NVIDIA in the supercomputing space. For more info check out the presentation (PDF) and blog post at the source links.
NVIDIA outs a pair of Tesla GPUs to electrify your supercomputer
NVIDIA's announced a pair of Tesla GPUs that'll give some extra pep to your supercomputing tasks. The K10 and K20 units harness the power of Kepler to add more muscle to the company's scientific and technical computing arm that supplies gear to the Barcelona Supercomputing Center and Tokyo's Tsubame 2.0. Internal tests reveal that the hardware is around three times faster than the company's Fermi GPUs -- with the latter card expected to arrive at the end of the year. The company didn't announce pricing, since its aiming them squarely at the big academic institutions, defense contractors and oil explorers -- but if your surname is Buffet or Abramovitch, then they might sell you one at trade.
IBM demonstrates Deep Thunder weather app on iPad
Almost everyone knows about Deep Blue, the chess playing supercomputer. But have you heard of its counterpart Deep Thunder? If you haven't, you will, because it may be coming to iOS as an iPad app, says IBM in an IdeaLab article on the Talking Points Memo website. Deep Thunder, like Deep Blue, is a parallel processing supercomputing system. Instead of chess, Deep Thunder is designed to assist with weather forecasting. It takes weather data from a variety of sources like the NOAA and the USGS and crunches that data to provide a detailed weather forecast. IBM says Deep Thunder "allows for an unprecedented granular look at incoming weather over an 84-hour period in a specific location, down to within a square mile, much more detailed than any other current weather forecast can provide." Not only would this be fantastic for iOS users, it would also be useful for landscapers and other people whose jobs are influenced by the weather. There's no timetable for release, but Talking Points Memo did see a working, but rudimentary version of the app.
VT nears completion of HokieSpeed, world's 96th most powerful supercomputer
If basking in the presence of a powerful supercomputer is on your list of "must-haves" when selecting a proper university, then you may wish to fire off an admissions application to the Hokies at Virginia Tech. The school's HokieSpeed system is now in its final stages of testing, which combines 209 separate computers, each powered by dual six-core Xeon E5645 CPUs and two NVIDIA M2050 / C2050 448-core GPUs, with a single-precision peak processing capability of 455 teraflops. To put things in perspective, HokieSpeed is now the 96th most powerful computer in the world, and yet it was built for merely $1.4 million in loose change -- the majority of which came from a National Science Foundation grant. As a further claim to fame, HokieSpeed is the 11th most energy-efficient supercomputer in the world. Coming soon, the system will drive a 14-foot wide by four-foot tall visualization wall, which is to consist of eight 46-inch Samsung 3D televisions humming in unison. After all, with virtually limitless potential, these scientists will need a fitting backdrop for all those Skyrim sessions. The full PR follows the break, complete with commentary from the system's mastermind, Professor Wu Feng.
Intel's gettin' kinda heavy, it's got the power, gonna break your heart
Intel spent most of yesterday standing in the bar at the SC11 conference showing off the size of its computing prowess. Admittedly, it's on a roll; the Xeon E5 processor powers ten of the Top 500 supercomputers in the world, and Intel chips as a whole are found inside 85 percent of all the machines on the list -- not to mention the E5's newly minted PCI Express 3.0 compatibility. The Santa Clara chip maker is also building a new Exascale lab at the Barcelona Computing Center, a ten petaflop "Stampede" machine at the Texas Advanced Computing center and several other machines for Government agencies like NASA. The Seattle shindig was a great excuse for Intel's Rajeeb Hazra to show off the new Knights Corner co-processor, built with a new Tri-Gate 3D 22nm process that packs 50 cores into one strip of silicon. In fact, the only thing Intel wasn't bragging about was being bested by the record breaking Fujitsu K, the mere mention of which caused everyone to go a little bit too quiet and begin to glower (we assume).
Barcelona readies hybrid ARM-based supercomputer, uses NVIDIA GPUs for heavy lifting
NVIDIA has announced that it'll be providing CUDA GPUs for Barcelona's Supercomputing Center, with the facility looking to substantially boost its energy efficiency with these later this week at the SC11 Conference in Seattle. While the words "low power" and "energy efficiency" are a bit of a buzz kill in the high-octane high-MFLOP world of supercomputing, the BSC thinks it'll use between 15 to 30 times less power than current systems. Titled the Mont Blanc Project, it's aiming to multiply those energy savings by four to ten times by 2014. While other supercomputers eat their way though megawatts of the electric stuff, hopefully a drop in power demands won't affect this machine's supercomputing scores.
IBM developing largest data drive ever, with 120 petabytes of bliss
So, this is pretty... big. At this very moment, researchers at IBM are building the largest data drive ever -- a 120 petabyte beast comprised of some 200,000 normal HDDs working in concert. To put that into perspective, 120 petabytes is the equivalent of 120 million gigabytes, (or enough space to hold about 24 billion, average-sized MP3's), and significantly more spacious than the 15 petabyte capacity found in the biggest arrays currently in use. To achieve this, IBM aligned individual drives in horizontal drawers, as in most data centers, but made these spaces even wider, in order to accommodate more disks within smaller confines. Engineers also implemented a new data backup mechanism, whereby information from dying disks is slowly reproduced on a replacement drive, allowing the system to continue running without any slowdown. A system called GPFS, meanwhile, spreads stored files over multiple disks, allowing the machine to read or write different parts of a given file at once, while indexing its entire collection at breakneck speeds. The company developed this particular system for an unnamed client looking to conduct complex simulations, but Bruce Hillsberg, IBM's director of storage research, says it may be only a matter of time before all cloud computing systems sport similar architectures. For the moment, however, he admits that his creation is still "on the lunatic fringe."
IBM and NCSA end their Blue Waters affair, go back to just being supercomputing friends
It seems that IBM and the National Center for Supercomputing Applications at the University of Illinois have hit a snag in their once fruitful relationship. After nearly four years of partnering for the Blue Waters petascale supercomputer, the NCSA's recently released a joint statement explaining that IBM's "terminated" its involvement with the project. If you'll recall, IBM was supplying its latest Power7 rigs to get all that data flowing, but the company's now decided that Ol' Blue will require more resources than initially anticipated. Apparently, there were talks to try and keep the spark alive, but since those fell through both have decided to return each others CDs and hoodies assets involved with the project (per contract terms): IBM gives back the money, while the NCSA returns any hardware supplied. The two plan on remaining in touch for future endeavours, and the NCSA doesn't appear to be too down either as it'll be consulting with the National Science Foundation to keep Blue Waters afloat. You'll find the official statement in the source link below, but we've included a video rendition of how we'd like to imagine it past the break. [Photo credit: kosheahan]
US federal government to close 800 data centers, walk into the cloud
Sure, it's been just a few months since the National Security Agency asked for a $900 million supercomputing complex – you know, to help out with all that internet wiretapping. But concern about deficit spending will mean shuttering 800 other federal data centers in the US, or 40 percent of total government capacity. The closures are part of a larger push toward greater efficiency and consolidation, with an estimated savings of $3 billion a year; moving services to the cloud will mean more savings in licensing fees and infrastructure. Single-digit savings might sound like chump change when you realize the federal information technology budget runs around $80 billion a year, but hey, it's a start, right? [Photo thanks to Adrian Levesque and Adam Koniak.]
Intel plans exascale computing by 2018, wants to make petaflops passé
Sure, Fujitsu has a right to be proud of its K supercomputer -- performing over 8 petaflops with just under 70,000 Venus CPUs is nothing to sneeze at. Intel isn't giving up its status as the supercomputing CPU king, however, as it plans to bring exascale computing to the world by the end of this decade. Such a machine could do one million trillion calculations per second, and Intel plans to make it happen with its Many Integrated Core Architecture (MIC). The first CPUs designed with MIC, codenamed Knights Corner, are built on a 22nm process that utilizes the company's 3D Tri-Gate transistors and packs over 50 cores per chip. These CPUs are designed for parallel processing applications, similar to the NVIDIA GPUs that will be used in a DARPA-funded supercomputer we learned about last year. Here we thought the war between these two was over -- looks like a new one's just getting started. PR's after the break.
NSA wants $896.5 million to build new supercomputing complex
The federal government may be cutting corners left and right, but that hasn't stopped the NSA from requesting nearly $900 million to help beef up its supercomputing capabilities. According to budget documents released by the Department of Defense yesterday, the NSA is looking to construct a massive new High Performance Computing Center in Maryland, designed to harness plenty of supercomputing muscle within an energy efficient framework. As with many other data centers, the NSA's $896.5 million complex would feature raised floors, chilled water systems and advanced alarm mechanisms, but it would also need about 60 megawatts of power -- the same amount that powers Microsoft's gargantuan, 700,000 square-foot data center in Chicago. According to the DoD, however, the NSA would use that juice judiciously, in the hopes of conserving enough water, energy and building materials to obtain LEED Silver certification. Another chunk of the funding, not surprisingly, would go toward fortifying the facility. The NSA is hoping to pour more than $35 million into building security and perimeter control, which would include a cargo inspection facility, advanced surveillance, and systems designed to detect any radiological, nuclear, or chemical threats. If all goes to plan, construction would wrap up by December 2015.
IBM touts new Power7 systems, still no mass market Watson
IBM's latest announcement probably won't get us any closer to securing our very own in-house version of Watson, but the firm is boasting a new line of Power7 products that includes an upgraded version of the supercomputer's server. First up are the BladeCenter PS703 and PS704, sporting 16 cores and 32 cores, respectively -- the PS704 touts a 60 percent increase in speed over its predecessors. The Power 750, the same system that gave Watson the stuff to slaughter those humans on Jeopardy!, is getting an upgrade that supports as many as 32 cores and can run up to 128 simultaneous threads, while the Power 755 offers up high-performance computing with 32 cores of its own. The cheapest version of the Power 750 Express rings in at about $30,000. So, no, we won't be battling Watson in a Jeopardy! Home Edition showdown anytime soon, but we're happy to see that our favorite supercomputer could be even smarter -- or at the very least, faster -- the next time it shows up on the boob tube. Full PR after the break.
IBM breakthrough brings us one step closer to exascale computing, even more intense chess opponents
The path to exascale computing is a long and windy one, and it's dangerously close to slipping into our shunned bucket of "awesome things that'll never happen." But we'll hand it to IBM -- those guys and gals are working to create a smarter planet, and against our better judgment, we actually think they're onto something here. Scientists at the outfit recently revealed "a new chip technology that integrates electrical and optical devices on the same piece of silicon, enabling computer chips to communicate using pulses of light (instead of electrical signals), resulting in smaller, faster and more power-efficient chips than is possible with conventional technologies." The new tech is labeled CMOS Integrated Silicon Nanophotonics, and if executed properly, it could lead to exaflop-level computing, or computers that could handle one million trillion calculations per second. In other words, your average exascale computer would operate around one thousand times faster than the fastest machine today, and would almost certainly give Garry Kasparov all he could stand. When asked to comment on the advancement, Dr. Yurii A. Vlasov, Manager of the Silicon Nanophotonics Department at IBM Research, nodded and uttered the following quip: "I'm am IBMer, and exascale tomfoolery is what I'm working on."* *Not really, but you believed it, didn't you?
MIT app turns your Android phone into a supercomputer... of sorts
Oh, sure -- a few people have called Google's Nexus One a "superphone," but suddenly, that nickname has taken on a whole new level of meaning. A team of talent from MIT has put its head down in order to concoct a new Android application that can come darn close to solving complex computational problems in just a fraction of the time that it'd take a bona fide supercomputer. The goal here is to let researchers and scientists convert to Google's mobile OS, but if you aren't falling for that one, it's also designed to "let engineers perform complicated calculations in the field, and to better control systems for vehicles or robotic systems." Of course, the models that are hosted on the phone do require a supercomputer to create, but once certain formulas are embedded, the app can then compute approximations in mere seconds rather than hours. Best of all, rbAPPmit is available for download as well speak in the source link below, but we'd probably wait for the (presumably thick) user guide to surface before diving in headfirst. [Thanks, Alasdair]
Virginia Tech's HokieSpeed supercomputer to rely on CPU and GPU synergies
Virginia Tech's no stranger to housing supercomputers -- those folks strung together 324 Mac Pros back in 2008 just for kicks, giggles and "research" -- but their latest computing monolith is quite the shift from the ordinary. A cool $2 million is floating over to Blacksburg in order to create HokieSpeed, a "versatile new supercomputing instrument" that'll soon be primed and ready to handle not just one or two tasks, but a variety of disciplines. Wu Feng, associate professor of computer science at the university, calls this magnificent monster a "new heterogeneous supercomputing instrument based on a combination of central processing units (CPUs) and graphical processing units (GPUs)," with expected performance to be orders of magnitude higher than their previous claim to fame, System X. One of its first assignments? To give end users the ability "to perform in-situ visualization for rapid visual information synthesis and analysis," and during the late hours, hosts a campus-wide Quake deathmatch. Just kidding on that last bit... maybe.
