Sign up for our newsletter and get the latest HPC news and analysis.

Earth Simulator to be upgraded to 131TF

NEC has announced that they have been awarded a contract to upgrade one of the world’s best known supercomputers to 131 TFLOPS (peak). No details are given, but it can probably be assumed that NEC’s SX9 vector systems will be used to perform the upgrade.

Bull to supply 300TF hybrid supercomputer for GENCI and CEA

GENCI, CEA and Bull have announced that Bull has been selected to supply a 300TF supercomputer in 2009. The system will comprise 1,068 Intel based nodes totalling 103TF coupled with 48 specialist GPU based nodes providing an additional 192TF. Details on the Bull website here.

CRL's Top5 supercomputer to be used in Yahoo! collaboration

The 4th fastest supercomputer on the Top500 list, based at the Computational Research Laboratories (CRL, www.crlindia.com), a wholly owned subsidiary of Tata Sons Limited, will be used by CRL in a collaboration with Yahoo! on cloud computing research. See here for the Yahoo! press release. This effort is the first of its kind in terms […]

Intel to release grand HPC vision

According to Pat Gelsinger, as reported by The Register, Intel will release its grand HPC vision targeting Petascale at the Intel Developer Forum in San Francisco later this year. Whilst Intel chips (mostly Xeon family) dominate the Top500 list, the most demanding supercomputers still tend to be custom designs, and in many cases using non-Intel […]

New PRACE website goes live

The PRACE project has significantly updated its website, http://prace-project.eu/, now adding more details about the work packages involved in this preparatory phase of the PRACE initiative, including management, governance and outreach activities, together with preparatory technical studies covering Distributed system management, Deployment of prototype systems, Software enabling for Petaflop/s systems, , Petaflop/s Systems for 2009/2010, […]

Intel talks Nehalem, Larrabee, Dunnington, Tukwila

Intel has issued a press briefing giving some details of its future multi-core products. The quad core Tukwila will be the first QuickPath Interconnect (QPI) enabled Itanium and will support a huge 30MB of cache. Dunnington, slated for second half of 2008, will be a six core Xeon MP processor. Nehalem is the new microarchitecture, […]

A hero for HPC

Dan Reed has a blog posting that doesn’t mention high performance computing until the end, but is a good read nonethless.

NERSC and CSCS sign MoU

The Swiss National Computing Center, CSCS, an autonomous unit of The Swiss Federal Institute of Technology (ETH Zurich), has signed a memorandum of understanding with the National Energy Research Scienctic Computing Center (NERSC) in the United States. The MOU formalises the growing exchange of expertise and experience between the two organisations. Under the agreement, the […]

Sandia Red Storm to grow to 284 TFLOPS

Sandia National Laboratory and Cray have signed an agreement to upgrade the Red Storm supercomputer, currently 124TFLOPS to 284TFLOPS. The Red Storm supercomputer is the joint development effort between Sandia and Cray that led to Cray’s succesful XT MPP supercomputers. More details here.

PRACE starts with kick-off in Julich

Over 60 representatives from 14 European Countries held the kick-off meeting of the PRACE project at the Research Centre Jülich on January 29 and 30, 2008. PRACE lays the foundations for a future European supercomputer infrastructure. At the end of 2007 the project received a grant from the European Commission towards a total budget of […]