This page contains a Flash digital edition of a book.
news and products


HIGH-PERFORMANCE COMPUTING


Products in brief


DataRunner


Panasas has announced DataRunner, file transfer software designed to overcome the data movement challenges found in distributed scale- out storage environments. Averting performance bottlenecks associated with multi-site content distribution and large-scale file sharing, DataRunner assures transfer speeds up to 16 times faster than traditional network transfer protocols, the company says. www.panasas.com


3DBoxx 4150 Xtreme Boxx Technologies has introduced the 3DBoxx 4150 Xtreme, a workstation featuring an overclocked fourth-generation Intel Core i7 processor. The latest Intel processors offer a significant performance increase over previous Intel technology while providing better efficiency. Boxx states that it is the only workstation manufacturer featuring the new micro architecture safely overclocked and backed by a three-year warranty. www.boxxtech.com


CarnotJet container Green Revolution Cooling has announced its first container-based CarnotJet submersion cooling data centre. GRC’s submersion cooling racks are installed in a lightly modified 40-foot ISO shipping module. According to GRC, the end result is a drag-and-drop solution that enables premium performance at the lowest cost in the industry. www.grcooling.com


For more products, please visit www.scientific-computing.com/products


Way 2) supercomputer, developed by the NUDT, and with a performance of 33.86 petaflops on the Linpack benchmark (and a peak performance of 54.9 petaflops). Not only did this greatly exceed the performance of the previous leader on the Top500 list – Titan, a Cray XK7 system installed at the US Department of Energy’s (DOE) Oak Ridge National Laboratory, which achieved 17.59 petaflops – it almost doubled it. Furthermore, Tianhe-2 will be deployed at the National Supercomputer Center in Guangzho, China, by the end of 2013 – two years ahead of schedule. According to Dr Yutong Lu, director of the System Software Laboratory at the National University of Defence Technology (NUDT), the country plans to build a 100-petaflop machine by the end of the current five-year plan in 2015. It was, she conceded, a tough target and the pressure was on to achieve it. But her demeanour gave no hint that she thought the goal unattainable.


A


The Tianhe-2 has a combined total of 3,120,000 computing cores, comprising 16,000 nodes, each with two Intel Xeon IvyBridge processors and three Xeon Phi processors. Some observers expressed surprise that the Chinese had gone all-out for performance using the first- generation of the Intel Phi, sometimes known as ‘Knights Corner’ rather than waiting for the second phase, ‘Knights Landing’. These processors will be fabricated at the 14-nanometre


16 SCIENTIFIC COMPUTING WORLD


t the opening of ISC’13 it was announced that the fastest system in the world is now the Tianhe-2 (Milky


scale and will be able to act as CPUs in their own right. They can be stand- alone CPU or PCIe coprocessors, so they will not be bound by ‘offloading’ bottlenecks and they will have integrated on-package memory. However, China itself had developed the interconnects, the operating system, front-end processors and software for Milky Way 2, so most of the features of the system were Chinese and the Intel components were used for the main compute part. (The Tianhe-1A system, which was the world’s fastest in 2010 and is now 10th in the ranking, uses Nvidia GPUs to accelerate computation.) According to Peter Beckman of the US Argonne National Laboratory, China had a rational strategy for achieving leadership in


For regular news updates, please visit www.scientific-computing.com/news


China to dominate HPC landscape for years?


delivering 2,143 Mflops/W. Sequoia, an IBM BlueGene/Q system installed at DOE’s Lawrence Livermore National Laboratory, also went down one place and is now claiming the third spot. Sequoia was first delivered in 2011 and has achieved 17.17 petaflops on the Linpack benchmark using 1,572,864 cores. Sequoia is also one of the most energy-efficient systems listed, consuming a total of 7.84 MW and delivering 2,031.6 Mflops/W. Installed at the RIKEN Advanced Institute for Computational Science (AICS) in Kobe, Japan, Fujitsu’s K computer has slipped down one place to fourth position, with a performance of 10.51 Pflops on the Linpack benchmark using 705,024 SPARC64 processing cores. With 8.59 petaflops on the Linpack benchmark using 786,432 cores, Argonne National


SOME OBSERVERS EXPRESSED SURPRISE THAT THE CHINESE HAD GONE ALL-OUT FOR PERFORMANCE


supercomputing. ‘China has invested in the people,’ he said. ‘Over the past six years, they have invested in the software and interconnects that are home-made. It is software and interconnects that make a supercomputer a supercomputer.’ He said that processor chips entirely developed and made in China are waiting ‘off in the wings’. Dropping one position to number two on the Top500, ORNL’s Titan has 261,632 Nvidia K20x accelerator cores, and is one of the most energy-efficient systems on the list, consuming a total of 8.21 MW and


Laboratory’s Mira – a BlueGene/Q system – is at number five. Completing the Top 10 are the upgraded Stampede at the Texas Advanced Computing Center of the University of Texas, Austin; Juqueen at the Forschungszentrum Juelich in Germany (and the most powerful system in Europe); an IBM BlueGene/Q system at Lawrence Livermore National Laboratory; SuperMUC, an IBM iDataplex system installed at Leibniz Rechenzentrum in Germany; and Tianhe-1A at the National Supercomputing Center in Tianjin, China.


@scwmagazine l www.scientific-computing.com


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32