China unveils 2.5-petaflop supercomputer

IDG News Service - China is unveiling a new supercomputer on Thursday that incorporates thousands of graphics chips and can reach a sustained performance of 2.5 petaflops, making it one of the fastest systems in the world.

Located at the National Supercomputing Center in Tianjin, the Tianhe-1A supercomputer has 7,168 Nvidia Tesla M2050 GPUs (graphics processing units), each with 448 processor cores. It also has 14,336 six-core Intel Xeon CPUs.

The supercomputer was built by China's National University of Defense Technology and is "the fastest system in China and in the world today," Nvidia claimed in a press release.

Besides a sustained performance of 2.5 petaflops measured by the Linpack benchmark, it has a theoretical performance of 4.669 petaflops when all the GPUs are operational, according to an Nvidia spokesman. The benchmarks were provided by the National Supercomputing Center in China, he said.

China has been moving up the supercomputing ranks in recent years. The last Top500 list of the fastest supercomputers, issued in June, lists the Nebulae supercomputer in Shenzhen as the world's second fastest. That system also combines Intel CPUs and Nvidia GPUs.

The fastest supercomputer in the world according to the June list is the Jaguar system at the U.S. Department of Energy's Oak Ridge National Laboratory, which can deliver 1.76 petaflops of sustained performance.

The Tianhe-1A was announced two weeks before the release of the next Top500 list, so it is too early to say if it will be the fastest system on the list.

China is looking to boost its computing resources and is doing a lot with hybrid supercomputer designs, said Nathan Brookwood, principal analyst at Insight 64.

A number of supercomputers have combined GPUs with CPUs to boost performance. GPUs are specialized co-processors that are faster than traditional CPUs at executing certain tasks, like those used in scientific and mathematical applications.

IBM's Watson could usher in new era of medicine

Computerworld - The game-show-playing supercomputer Watson is expected to do much more than make a name for itself on Jeopardy.

IBM's computer could very well herald a whole new era in medicine.

That's the vision of IBM engineers and Dr. Eliot Siegel, professor and vice chairman of the University of Maryland School of Medicine's department of diagnostic radiology.

Siegel and his colleagues at the University of Maryland and at Columbia University Medical Center are working with IBM engineers to figure out the best ways for Watson to work hand in hand with physicians and medical specialists.

Siegel, who refers to the computer not as the champ of Jeopardy but as "Dr. Watson," says he expects the computer, which can respond to questions with answers rather than with data and spread sheets, to radically improve doctors' care of their patients.

"There is a major challenge in medicine today," Siegel told Computerworld. "There's an incredible amount of information in a patient's medical record. It's in the form of abbreviations and short text. There's a tremendous amount of redundancy, and a lot of it is written in a free-form fashion like a blog or text.

"As a physician or radiologist, it might take me 10 or 20 or 60 minutes or more just to understand what's in a patient's medical record," he said.

Within a year, Siegel hopes that "Dr. Watson" will change all of that. Watson is expected to be able to take a patient's electronic medical records, digest them, summarize them for the doctor and point out any causes for concern, highlighting anything abnormal and warning about potential drug interactions.

"It offers the potential to usher in a whole new generation of medicine," Siegel said. "If all Dr. Watson did was allow me to organize electronic medical records and bring to my attention what's most important and summarize it, that would be incredibly valuable to me.

"Even small things that Watson can do will change the way I, and my colleagues, practice medicine," he said.

The ability to deliver a single, precise answer from these documents could go a long way in transforming the healthcare industry. Watson, the IBM computing system designed to play Jeopardy!, could deliver such a solution.

Richard F. Doherty, research director at analysis firm Envisioneering Group, said he's excited to have a computer organize his medical history for his physician.

"That sounds excellent," Doherty said. "I think we've all been through the situation of filling out forms for new doctors, and then they don't have the time to read through it all, and they just say, 'What? You have a sore throat?' Having Watson help attend to our needs sounds like a great application of [the computer]."

But organizing and summarizing patient histories isn't all Watson is expected to do.

Siegel, who also works with the National Cancer Institute, said he's hoping that Watson will also be able to take patient and treatment information from hundreds, if not thousands, of hospitals and pull it all together.

Then, when a doctor is considering treating a patient with a particular drug or treatment, he first can ask Watson how that treatment has worked on patients with similar diagnoses and backgrounds.

"Watson can ingest information efficiently and rapidly," Siegel said. "It'll have an encyclopedic knowledge and suggest diagnostic and therapeutic possibilities based on databases much larger than one physician can possibly hold in his head.

"This technology brings a potential to have a renaissance of medical diagnosis," he added. "It offers the potential for us in the next five or 10 years to routinely deploy computers when working with our patients."

Jennifer Chu-Carroll, an IBM researcher on the Watson project, said the computer system is a perfect fit for the healthcare field.

"There's so much electronic information out there, and it's projected to continue to grow," Chu-Carroll said. "Nobody can possibly ingest all that information. Without a tool, there's no way to leverage it."

She also said she believes that at some point Watson will have a speech-recognition capability so it can actually go into an exam room and listen to a patient talk about his symptoms while it runs through his medical records.

"Think of some version of Watson being a physician's assistant," Chu-Carroll said. "In its spare time, Watson can read all the latest medical journals and get updated. Then it can go with the doctor into exam rooms and listen in as patients tell doctors about their symptoms. It can start coming up with hypotheses about what ails the patient."

She added, "The physician will make the decisions, but Watson can help."

Doherty said that having a supercomputer that can ingest and analyze loads of data and then answer questions much as a human would could radically change not only medical diagnostics, but also medical research and pandemic recognition and management.

"Spotting trends could save lives and save money," he said. "What humans can't always see, Watson may be able to.

"I think we're on the cusp of a revolution," Doherty added.

Sharon Gaudin covers the Internet and Web 2.0, emerging technologies, and desktop and laptop chips for Computerworld. Follow Sharon on Twitter at  @sgaudin or subscribe to Sharon's RSS feed . Her e-mail address issgaudin@computerworld.com.

Read more about Mainframes and Supercomputers in Computerworld's Mainframes and Supercomputers Topic Center.

IBM's Watson Wins Jeopardy! Next Up: Fixing Health Care

The final round in the epic man vs. machine battle that's been playing out on Jeopardy! all week was fought tonight. The winner: Watson, IBM's supercomputer, who soundly defeated flesh-and-blood opponents Ken Jennings and Brad Rutter, ending a three-night tournament and beginning a long period of social and technological commentators trying to figure out what it all meant.

To mark the event, IBM held a viewing party this evening at popular New York ping-pong spot Spin NYC with some of the engineers who helped create the supercomputer. IBM also took the opportunity to announce a collaboration with speech-technology company Nuance to bring Watson like computing and analysis to the healthcare world.

"We're moving beyond Jeopardy!" Dave Ferrucci, the principal engineer behind Watson, said at the event. "With the Watson technology, we're going to look at creating a commercial offering in the next 24 months that will help empower doctors to do higher quality decision making and diagnoses."

Trebek opened the show by talking about what he learned over the past two days, remarking on Watson's propensity to wager seemingly random amounts on Daily Doubles and Final Jeopardy, and joking that Toronto is now a U.S. city, a reference to Watson's bizarre answer in the previoius night's final round. Watson pulled out to an early lead, though Jennings and Rutter soon responded. Jennings came from behind on the first round's Daily Double, pulling ahead to $7,200.

Round Two saw Watson cement its lead, but it also revealed again the computer's tendency to bet strange totals on Daily Doubles, wagering just $367 at one point. By the end of the round, though, Watson led the trio with $23,440.

With final Jeopardy, it was Jennings' last chance to win, since Rutter had fallen behind. His correct question of "Who is Bram Stoker" to an answer about 19th Century novelists was accompanied by a message: "I for one welcome our computer overlords." The message was prescient. Watson also had the correct answer, though, winning it the game with $77,973 total. At Spin NYC, the bar erupted into applause at the win.

Ferrucci explained Watson's strange wagers: "They seem random to us mere mortals. What's actually going on there is that the team has trained on human betting patterns. It's considering its confidence. It's also considering where it is in the game, and how much more there is to go. It's a very complex calculation, with very precise results. We could have rounded it, but we figured just give the number."

What IBM does think Watson is good for is data analysis and aiding decision making, which is why the company's approaching the health care field first. The technology has the ability to scan and analyze data from far more sources than a human ever could in a short period of time, potentially aiding doctors in diagnosing complex but urgent conditions.

Going a step further, Ferrucci speculated that at least part of the technology might someday make its way into mobile devices, bringing Watson-like analysis directly to consumers.

"It's so much better on mobile devices to answer succinctly. That could be very helpful. But ultimately it's more of a business intelligence kind of interaction."

Despite the Jeopardy win and the promise of Watson, Ferrucci is careful to point out that his creation is still no substitute for human decision making. After all, when Watson gets a question wrong—as in the Toronto example—it gets it extremely wrong.

"With Jeopardy! these are human questions written for humans, whereas all the computer has are words. It can't rely on human context to determine things.

"I hope people stop and scratch their heads, and think about the limits of computation, and what does it mean to be intelligent," Ferrucci said. "When you deconstruct this, and look at the machine, is any part of this really understanding the question? No. We don't want computers in my opinion making value judgments about what it means to be human. Only humans can do that."

The Future of Watson

Final Jeopardy! and the Future of Watson

The IBM team who designed Watson has achieved another milestone in the history of computer science. After the Jeopardy! challenge concludes, the team faces the task of developing real world solutions based on this technology.

The impact of a machine like Watson will be felt throughout business, government and society. Join the conversation to find out how the IBM team achieved this historic feat and chat live with IBM Watson Principal Investigator Dr. David Ferrucci, IBM Fellow and CTO of IBM’s SOA Center for Excellence Kerrie Holley and Columbia University Professor of Clinical Medicine Dr. Herbert Chase, hosted by "Man v. Machine" author Stephen Baker.

To submit questions to the panel, sign on or join Twitter and use the hashtags #ibmwatsonand #askwatson.

Tune in for the webcast on Feb. 17 at 11:30 AM EST.