Earlier this month we said that China was poised to sit atop the supercomputing summit with its Tianhe-2, a system built by the Chinese National University of Defense Technology (NUDT). Armed with over 3 million physical processing cores and 1 petabyte of memory, there was no way this thing wouldn't shoot right to the top, and so it did, taking pole position on the TOP500 chart.
Tianhe-2 notches up 30.65 petaflops during benchmark test
China scaled the supercomputing summit in late 2010, when a 2.507-petaflop machine named Tianhe-1A (or Milkyway-1) was ranked as the fastest supercomputer by TOP500, which publishes a list of the fastest 500 supercomputers twice a year. Its stay there, though, was brief, lasting all of six months. But if you think it was just a flash in the pan, the Chinese National University of Defense Technology (NUDT), the organization which developed the Tianhe-1A, is building the Tianhe-2 in order to prove you wrong.
Roadrunner won the race to 1 petaflop five years ago.
Even supercomputers sometimes have relatively short lifespans. So it is with Roadrunner, the first supercomputer to break the petaflop barrier by posting better than 1 million billion calculations per second five years ago. Back then, it was the world's fastest supercomputer, and scientists used it to gain a better understanding of energy flow in nuclear weapons and its relation to weapon yield.
The building of a supercomputer in 90 awesome seconds.
We missed this one when it was first posted in mid-December, which is our way of saying we know this is old news, but it's also really cool. It's a 90-second time lapse video showning the construction of the Fujitsu Primergy high-performance supercomputer at the National Computational Infrastructure (NCI), which was completed late last year. It's now located at the Australian National University.
Nvidia today formally introduced its K20 GPU family, supposedly the highest performing and most efficient accelerators ever built. If you recall, these new GPUs were featured in Titan, the world's fastest supercomputer that was unveiled by the U.S. Department of Energy's (DOE) Oak Ridge National Laboratory (ORNL) a couple of weeks ago.
The U.S. Department of Energy's (DOE) Oak Ridge National Laboratory today announced the completion of Titan, a supercomputer capable of processing more than 20,000 trillion calculations per second. That's equivalent to 20 petaflops, which is 10 times more powerful than Jaguar, ORNL's previous flagship supercomputer. All that power will be put to use to research energy, climate change, efficient engines, materials, and to play Crysis. Wait, what?
November 2009 was the last time a United States supercomputer sat on top of the TOP500 list of the world's fastest supercomputers, and thanks to Sequoia, an IBM BlueGene/Q system residing at the Department of Energy's Lawrence Livermore National Laboratory, the U.S. is back out in front of the pack after it achieved 16.32 petaflop/s on the Linpack benchmark. Over a million and a half cores (1,572,864, to be exact) comprise Sequoia, which TOP500 describes as one of the most energy efficient systems on the list.
The Nvidia GTX 690 is real, and it's amazing -- both in specs and in price. But while the tech world swooned at the announcement of the dual-GPU behemoth, another new product outlined at the GTX 690's unveiling holds even more intriguing potential for the gaming world at large: the cloud-based "GeForce Experience," which promises to automatically optimize the graphics settings in games based on the components in your individual PC.
When we first heard about San Diego's Gordon supercomputer, we envisioned a spunky thirty-something with a near endless database of PC knowledge and a custom program designed for epic rants. But then we remembered that's our own Gordon Mah Ung. Surprisingly, no one has named a supercomputer after intrepid Deputy Editor (yet), but San Diego did name one after Flash Gordon, an appropriate namesake since it's the world's first supercomputer to rely entirely on flash memory for storage chores.
ARM’s built its business around power-efficient chips that are perfect for mobile applications (like tablets and smartphones), but that pedigree could transfer over to another technical arena as well, one that has traditionally been dominated by Intel and AMD: high-powered computing. In fact, Sumit Gupta, who serves as the senior manager of Nvidia’s Tesla GPU Computing HPC business, says that ARM chips are “inherently much more energy efficient than an x86 CPU” – and that fact makes Nvidia feel that the future of supercomputing lies in ARM.