Petaflops. Something about the word brings a smile to our faces – maybe because it sounds like a vaguely dirty word? But the term refers to processing power rather than a part of the human anatomy. A petaflop computer performs 1,000 trillion operations per second, and the new XK6 supercomputer announced today by Cray promises to scale up to 50 petaflops. Now, math isn't one of our strong points, but we're pretty sure we'd be able to rock a pretty awesome game of Dwarf Fortress on one of those puppies.
The second day of Watson's man-vs-machine Jeopardy! challenge has come to close, and IBM's super-computer dealt a resounding defeat to its pitiful, meat-filled opponents. In light of that fact, we thought now might be a good time to learn a little more about our future overlord, so we found an expert on the subject to answer some of our questions.
Stephen Baker is the author of Final Jeopardy: Man vs. Machine and the Quest to Know Everything, a new book about how Watson came to be. He was gracious enough to answer some of our burning questions. Read on to find out what he said!
With more than a little help from Nvidia, Amazon hopes to bring supercomputing to the masses through a new Elastic Computer Cloud (EC2) offering called "Cluster GPU Instances."
By "masses," we're referring to enterprises and start-ups, not everyday Joes looking to run Crysis or improve their Folding@Home score (go Team 11108). Amazon's new Cluster GPU Instance is a server with two quad-core Intel Xeon X5570 processors, two Nvidia Tesla M2050 GPUs (Fermi), 22GB of memory, 1.7TB of storage, and a 10Gb/s Ethernet connection, Amazon's Jeff Barr said.
"With Amazon Cluster GPU Instances, our customers now have the power of high performance computing, the efficiency and speed of GPUs, and the highly scalable and affordable cloud environment our customers have come to expect fro Amazon Web Services (AWS)," said Peter De Santis, GM of Amazon EC2. "We're excited to help our customers access the raw power of GPU technology and look forward to the innovation this will enable."
Using all that GPU power can be tricky, but Nvidia says hundreds of applications have already been ported to the Nvidia CUDA architecture, making it easy for programmers to dive right in.
Intel, the world's largest chip maker, and Nvidia can lay claim as the heart and soul of the fastest supercomputer on this or any planet that we know of. That's because the Tianhe-1A supercomputer located at the National Supercomputing Center in Tianjin, China contains 14,396 Intel Xeon 5600 series processors and over 7,000 Tesla M2050 GPUs.
All that hardware translates into 2.57 petaflops of computing power, making Tianhe-1A the top dog on the 36th edition of the Top500 supercomputer list.
"Our Xeon processor roadmap continues to deliver hugely powerful supercomputers that are helping solve mankind's greatest challenges," said Rajeeb Hazra, general manager of Intel's High Performance Computing organization. "Securing the top position on the Top500 is a source of great pride for Intel, and demonstrates the tremendous leaps in performance and versatility that our processors are delivering across a range of compelling workloads."
Just another notch on Intel's supercomputer belt, really. The Santa Clara chip maker says it's responsible for powering some 80 percent of the systems on the Top500 list, including four of the top six, and more than 90 percent of new systems on the Top500 list.
The supercomputing crown now belongs to China and the country's Tianhe-1A machine, which was able to crank out 2.507 petaflops with more than a little help from Nvidia and Intel, according to a Nvidia press release.
The Tianhe-1A supercomputer houses 7,168 Nvidia Tesla M2050 graphics chips and 14,336 Intel Xeon processors. It consumes 4.04 megawatts, and according to Nvidia, it's three times more power efficient than a CPU-only supercomputer. It would take more than 50,000 CPUs and twice as much floor space to achieve the same level of performance if using CPUs alone, the GPU maker claims.
"The performance and efficiency of Tianhe-1A was simply not possible without GPUs," said Guangming Liu, chief of National Supercomputer Center in Tianjin. "The scientific research that is now possible with a system of this scale is almost without limits; we could not be more pleased with the results."
Nvidia says the Tianhe-1A system will be operated as an open access system for large scale scientific computations.
The next time someone asks about that bulge coming from your khakis, you can confidently respond, "Why yes, yes that is a supercomputer in my pocket, and I am happy to see you."
You'll be telling the truth if you install MIT's new software designed to perform complex simulations on mobile phones. The software simulates physical phenomena, such has how cracks form in building materials and how fluids flow in irregular channels, crunching the numbers in seconds that would typically take a supercomputer hours to calculate.
As the eggheads at MIT explain it, the software is designed for situations where the general form of a problem is already known in advance, just not the details.
"This is a very relevant situation," says David Knezevic, a postdoc in the department who helped lead the project. "Often in engineering contexts, you know a priori that your problem is parametrized, but you don't know until you get into the field what parameters you're interested in."
MIT News has a whole bunch more on the topic right here.
And the award for the world's most energy efficient supercomputer goes to... Grape-DR, a supercomputer taking residence in the Department of Information Science at the University of Tokyo.
Grape-DR led all others in the Little Green500 List of energy efficient supercomputers in June, putting out 815.43 MFLOPS/W. That was enough to top IBM's second place machine in Germany, which puts out 773.38 MFLOPS/W.
Helping the Grape-DR top the list are 64 Intel Core i7 920 processors, plus an accelerator chip that's able to pump out 200 gigaflops with the same amount of power it takes to run a lightbulb.
For all of its power and efficiency, however, the Grape-DR could use a little TCL in the cable management department, don't you think?
During a recent press meeting, the Tokyo Institute of Technology talked up details of its Tsubame 2.0 project, a next-gen supercomputer slated to start crunching numbers in the fall of 2010.
"It will be the first petaflops computer in Japan," said Satoshi Matsuoka, professor at the Global Scientific Information and Computer Center (GSIC) of the University. "And it will be the world-class supercomputer system for our university."
At full bore, Tsubame 2.0 will be capable of computing 2.39 PFLOPS, making it the second most powerful supercomputer in the world. It will also be one of the greenest supercomputers on the planet, helped in large part by 173.9TB of SSD storage.
"By using them to input and output local data (that are not shared by other nodes), the performance of the entire system can be enhanced," Matsuoka added.
According to the latest list of the world's 500 fastest supercomputers, China is home to two of the top ten machines, including Nebulae, the second fastest supercomputer on the planet.
"China's ambitions to enter the supercomputing arena have become obvious," said the compilers of the list.
In terms of theoretical peak performance, Nebulae leads the pack with 2.98 petaflops, the highest ever on the Top 500 list. Nebulae hasn't actually achieved that level of performance, however, so far topping out at 1.271 PFlop/s. Jaguar, the world's fastest supercomputer, resides in Tennessee and holds the record with 1.75 PFlop/s, though its theoretical peak performance is 2.3 PFlop/s.
China is now home to 24 of the top 500 supercomputers in the world.
Novell has perused the most recent list of the top 500 supercomputers in the world, and nine of the top ten run on Linux. Not only that, but 85% of the entire list run Linux. It’s not hard to guess why Novell might take an interest in this; Novell’s SuSE Linux is the distro of choice in six of the top ten supercomputers.
After tabulations were complete, Novell wasted no time patting themselves on the back. “Supercomputers are helping to push the boundary of science and knowledge around the world, and SUSE Linux Enterprise Server from Novell has been chosen as the optimal operating system to power many of these HPC environments for good reasons,” said Novell VP of Business Development Holger Dryoff.
The current king of the supercomputer hill is the Jaguar computer at the US Department of Energy Oak Ridge lab. It's quite a monster of a machine, capable of 2.3 petaflops. It runs Linux, but not Novell's SuSE. Rest assured, as soon as one of these supercomputers develops sentience and proceeds to wipe out humanity, odds are it will be running on open source. There will be no blue screen to save mankind.