Cloud-based service to provide intelligent backbone to customer queries
IBM is working on a cloud service based of Watson, its natural language processing and information retrieval system, for use by other developers. The service will allow developers to pay for time on the supercomputer, allowing for apps and services to use its "cognitive computing" technology for commercial purposes, such as answering queries within smartphone apps.
Shutdown of AMD, Cell-based supercomputer despite #22 ranking
The Los Alamos National Laboratory is shutting down its Roadrunner supercomputer, after deeming it too energy inefficient. Constructed by IBM in 2008, the five-year-old supercomputer was at one point the fastest in the world after achieving a speed exceeding one petaflop: one million billion calculations per second.
Intel Xeon processor use allows for over 1 million cores
Cray has launched its new supercomputing creation, the Cray XC30, codenamed Cascade. The new system combines Intel Xeon processors with the Aries interconnect, new cooling and power technologies and Cray's integrated software environment, to create a supercomputer capable of workloads higher than 100 petaflops.
GPU-based supercomputer to help biological, physical modelling
The world's fastest open-science supercomputer has finished its NVIDIA Kepler-based upgrade. The newly-rebranded Titan at the US Department of Energy's Oak Ridge National Laboratory in Tennessee now contains 18,688 Tesla K20 GPU accelerators that helps it achieve a peak performance of more than 20 petaflops.
NYC to earn share of crime system profits
The New York Police Department and Microsoft have constructed a supercomputer to help fight crime. Costing between $30 and $40 million to create, the Domain Awareness System (DAS) uses a combination of CCTV, license plate readers, police reports, and live incident reports to track criminals and help protect the city from potential threats such as terrorist attacks.
IBM system uses 98,000 nodes, 1.5 million cores
A supercomputer in the United States tops the list of the world's top 500 for the first time since November 2009. Sequoia, an IBM Blue Gene/Q system installed by the Lawrence Livermore National Laboratory for the Department of Energy, clocked in at 16.32 sustained petaflops during tests, 50 percent more powerful than the second place “K Computer” in Japan.
Removal may force university IT, researchers to find new solutions
OS X Mountain Lion was officially unveiled on Monday, and is shipping in July -- but conspicuously absent from the list of features is Apple's ad-hoc supercomputer tool Xgrid. Apple's Xgrid provided network administrators a means of constructing a Beowulf cluster, allowing them to use computer idle-time to drive calculations that can be divided into smaller operations. Xgrid client has been installed in all computers running OS X 10.4 or greater by default, so installation and configuration of a massively parallel grid was completed with little time or money outlay. The job controller is included in Mac OS X Server up to OS X Lion and as a free download.
Supercomputers based on Tegra 3 chips
NVIDIA has introduced a new technology for workstations, known as Maximus, and what it claims to be the first supercomputer with ARM-based CPU/GPU hybrid hardware. Maximus enables computers to simultaneously handle interactive graphics and complex computations required to the resulting renderings.
System centers on 16-core processors
Fujitsu has introduced a new supercomputer, known as the PrimeHPC FX10, that is said to be capable of theoretical processing performance up to 23.2 petaflops. The system relies on the company's new VIIIfx processor, which serves as the successor to the existing IXfx chips. The 16-core water-cooled design supports standalone performance up to 236.5 gigaflops, with performance-per-watt reaching 2 gigaflops.
Sunway BlueLight MPP uses 8,700 nation-built CPUs
China has announced that it has created its first supercomputer built upon its own microprocessor chips. The computer, the Sunway BlueLight MPP, is rated at one petaflop per second (1,000 trillion calculations per second) and is comprised of 8,700 ShenWei SW1600 microprocessors. The computer has been up and running in the country's National Supercomputer Center in Jinan, a city in the Shandong Province in eastern China, since September.
Tesla GPUs could deliver up to 20 petaflops
The Oak Ridge National Laboratory will use the parallel processing capability of the NVIDIA Tesla M2090 GPU in a new supercomputer, a Cray XK6 nicknamed "Titan." The new supercomputer could deliver over 20 petaflops, making it more than twice as fast as the most powerful supercomputer today, the Fujitsu K supercomputer. Titan will also be three times more energy efficient.
Takes top spot from Chinese supercomputer
A computer developed by Fujitsu and the RIKEN Advanced Institute for Computational Science has claimed the title of the world's most powerful supercomputer on the Top500 list. The Fujitsu K Computer is a series of 672 computer racks with a total of 68,544 CPUs. The K Computer handles 8.162 petaflop/s (quadrillion floating point operations per second) as measured by the LINPACK benchmark. Researchers expect the final configuration of the K Computer to exceed 10 petaflop/s. The supercomputer it bumped from the top spot, the Tianhe-1A supercomputer at the National Supercomputing Center in Tianjin, China, performs 2.6 petaflop/s.
Configurations available with up to 256 sockets
SGI has introduced the Altix UV, a new series of supercomputers geared for large databases or analysis operations. The computers are based on Intel's Nehalem platform, with support for configurations utilizing quad-, six, or eight-core Xeon CPUs. Devices can be scaled to integrate up to 2,048 cores, with architectural provisioning for up to 262,144 cores.
VT installs Mac cluster
Virginia Tech has recently installed a new Mac-based supercomputing cluster, according to an announcement. The cluster is being used by the university's Center for High-End Computing Systems, and consists of 324 Mac Pros operating as servers, with eight cores apiece, for a total of 2,592 cores. The computers are networked together via a 40Gbps copper LAN connection.
IBM today announced it will commission a $100 million hybrid supercomputer, the world's fastest, codenamed Roadrunner, to the US Department of Energy's National Nuclear Security Administration. The supercomputer is claimed to be able to perform 1,000 trillion operations per second, or one petaflop. According to IBM, the array has the same computing power as 100,000 of today’s fastest laptop computers combined. The hybrid computer combines AMD's x86 processors with Cell Broadband Engine chips originally developed for video game platforms, such as the ones used in Sony's PlayStation3.