In the CAE Symposium Dave Turek, IBM Vice President Deep Computing, gave a broad view on Linux, IBM's hardware strategy, application areas, the HPC on demand facility and compared Blue Gene with the Earth Simulator. Then he presented the Deep Computing activities and offerings.

The Lehrstuhl für Rechnertechnik und Rechnerorganisation (LLR-TUM), Professor Arndt Bode, at Technical University Munich got a cluster donation from Intel. It is based on the new Intanium2 processor architecture. This grows the existing cluster to 40 processors. (Uwe Harms)

The new Cray X1 supercomputer is running challenging applications up to 25 times faster than previously achieved by the Department of Energy's Center for Computational Sciences (CCS) at the Oak Ridge National Laboratory (ORNL).

SC2003, the annual conference on high-performance computing and networking, concluded lasy Friday with record numbers for overall attendance (over 7500), exhibit space sold, technical program registration and tutorials attendance.

At the SC2003 conference last week, the winners of the Gordon Bell Prizes, the HPC Challenge, and the best research papers and poster were announced. The conference itself gives awards for Best Paper, Best Student Paper, Best Poster, and the HPC Challenge and Bandwidth Challenge. In addition, SC2003 serves as the venue for presenting the Gordon Bell Prizes, which reward practical uses of high-performance computers, including best performance of an application and best achievement in cost-performance. Additionally, two special awards are presented by the Institute for Electrical and Electronic Engineers (IEEE) to recognize longtime innovators in high-performance computing.

HP has launched new HP Integrity and ProLiant servers and Linux clustered systems. Announcement includes the HP XC3000 and XC6000 Linux clusters for high-performance technical computing.

HP solidified its position as the leading global server vendor with 29.7 percent of total unit shipments worldwide, according to preliminary results for the third calendar quarter of 2003 released by analyst research firm Gartner Dataquest, HP said.

Winners of the 2003 Gordon Bell Prizes, one of high performance computing's most prestigious honours, will be announced at the SC2003 conference on high performance computing and networking. The Gordon Bell Prizes are awarded each year at the annual SC conference to recognize outstanding achievement in the field of high performance computing.

On October 29, the first Belgian Grid workshop was held, organised by the Belgium reserach network organisation Belnet. At the workshop, Rosette Vandenbroucke, Grid co-ordinator at Belnet, introduced BEGrid, the Belgium academic research Grid initative. BEGrid will start on a small scale and initially will use the EDG software. The Flemish government, one of the regional governments in Belgium, said they will support the initiative.

At the BEGrid workshop in Brussels, Guy Wormser explained the EGEE project that will start early next year as a follow-on to the DataGrid project. EGEE will install a production quality Grid infrastructure across Europe. According to Guy Wormser, the negotiations with the Europan Commission have been finished, so it is almost certain the project will start in April 2004. The whole EGEE infrastructure and support structure will cost tens of millions of euros.

Platform has introduced an adapter for its integrated Symphony Suite, which makes Microsoft Excel Grid-enabled. Big, complex Excel sheets can be partitioned and sent to different servers. This shortens the computing time, improves and fastens the financial trading operations.

The Singapore National Grid Office has been established to help steer Singapore towards a Grid-enabled economy. The broad objective is a Grid which allows suppliers and users, public and private, to plug into a high-speed network easily and securely. The Singapore government must progressively allow market forces to allocate resources and incentivize innovation in the Grid. This has to be done in a gradual and experimental way, as stated by Minister for Trade and Industry George Yeo at the launch of the National Grid Pilot Platform in Singapore.

As part of TeraGyroid, supercomputers in the UK and the US have been linked to carry out an interactive scientific experiment. The "TeraGyroid" experiment was jointly funded by the UK's Engineering and Physical Sciences Research Council (EPSRC) and the National Science Foundation, USA (NSF). TeraGyroid is based on the e-Science pilot project RealityGrid.

ExaGrid Systems, an emerging company providing reliable data recovery through Grid Protected Storage, has launched the company and its approach to addressing the cost and complexity associated with ineffective storage, backup and disaster recovery.

Research staff from the San Diego Supercomputing Center (SDSC) and the National Center for Supercomputing Applications (NCSA) in partnership with IBM will demonstrate the use of IBM's General Parallel File System (GPFS) in a large-scale Grid environment spanning several sites and long distances. The demonstration will be presented in the SDSC booth at SC2003, the annual high-performance computing and networking conference, to be held November 15-21 in Phoenix. The demonstration will take place Tuesday, November 18 and Thursday, November 20.

The National Center for Supercomputing Applications (NCSA) has joined with Intel Corporation, Oracle, and HP to demonstrate Grid-based bioinformatics tools using an Oracle 10g database on Intel Itanium 2-based systems at SC2003.

Cybermation, a developer of enterprise job scheduling and software change management solutions, announces the immediate availability of ESP System Agent Release 6. This release offers added functionality to strengthen the performance of Cybermation's enterprise job scheduling solution across UNIX and Windows platforms.

Stuttgart University has chosen Bull NovaScale to allow academic users and industrial partners to test their High Performance Computing (HPC) applications in a Windows Server 2003 environment. Beyond its traditional missions, Stuttgart University hosts the High Performance Computing Center (HLRS) for Research and Industry, with particular emphasis on the automobile and aeronautic sectors. Among its customers are DaimlerChrysler and Porsche.

IBM has become a reseller of Scali's platform independent Message Passing Interface (MPI). As part of the agreement, Scali MPI Connect has been tested and approved for distribution on custom built systems and the IBM eServer 1350 Linux cluster.

The Protein Data Bank (PDB), an international resource for biomedical research with facilities at Rutgers, The State University of New Jersey, became the focus of an international collaborative agreement today. The agreement establishes the Worldwide Protein Data Bank (wwPDB) and, in doing so, ensures a single, uniform and enduring archive of 3-dimensional models of biological structures, proteins specifically. An announcement of the agreement appears in the December issue of Nature Structural Biology.

ModViz Inc., an expert in parallel rendering and visualization cluster software, has made available Renderizer Visualization Cluster Software for Microsoft Windows. Renderizer Visualization Cluster Software from ModViz offers real-time data distribution, synchronization and load balancing capabilities, and helps to enable high-end visualization and scalable parallel rendering on workstation clusters. Renderizer Software Development Kit (SDK) hides the complexity of a cluster and provides developers an efficient way to port single system applications to clusters. No parallelization of existing application code is required.

The new International Supercomputing Conference "ISC Award" has been created to motivate innovative supercomputing researchers to contribute outstanding HPC findings. Starting in 2004, it will be held yearly to reward and promote those people who have distinguished themselves through state of the art projects. Closing date for submissions is February 16th 2004.

The seventh fastest supercomputer in the world according to the 22nd TOP500 supercomputing list. issued in November 2003, at Lawrence Livermore National Laboratory (LLNL) has been in General Availability Production since October 1, 2003.

According to the 22nd Edition of the "TOP500" list of the world's fastest supercomputers released this week, Intel processors continue to gain acceptance with high-performance computing (HPC) solutions. Nearly 38 percent of systems on the list use Intel processors, 15 percentage points more than the closest competing architecture and overtaking RISC-based systems for the first time.

A Cray X1 system with 252 applications processors achieved a speed of 11.5 Gflop/s per processor, or 2.9 Tflop/s in total. This was the highest per-processor speed of any currently available supercomputer in the rankings. Cray has installed 256-processor Cray X1 systems at multiple customer sites in the United States.

A Dell supercomputing cluster at The University of Illinois' National Center for Supercomputing Applications (NCSA) was recognized as the world's fourth fastest supercomputer, according to the bi-annual TOP500 list. The TOP500 list has emerged as the defacto source to determine the fastest supercomputers based on a benchmark that measures a system's sustained performance. The total number of Dell entries on the list has increased to 18, reflecting the growing trend toward standards-based systems configured in high-performance computing clusters (HPCC) to address sophisticated computing needs in academia, laboratories and commercial organisations.

Sun said that more than 50 percent of its customer sites highlighted on this year's Top 500 Supercomputers list are powered by its Sun Fire V60x Compute Grid supercomputers, i ncluding the University of Queensland (Australia) and Saudi Aramco, and San Diego Supercomputer Center's RockStar cluster. The latest Top 500 list includes 11 Sun Fire V60x Compute Grid system entries demonstrating the rapid adoption of Sun's low-cost computing initiative in the high performance and technical computing (HPTC) market.

For the fourth consecutive time, HP ranks as the No. 1 supercomputing provider on the TOP500 Supercomputer list.

With 165 entries, representing nearly one-third of the posted sites, HP has more installations on the TOP500 list than any other vendor. With its powerful HP Integrity, Superdome and AlphaServer systems, HP increased its share on this exclusive list from its June 2003 showing of 159 entries. HP also secured two of the top five spots: No. 2 Los Alamos National Laboratory and No. 5 Pacific Northwest National Laboratory.

Dr. Karsten Gaier, Altraia, published a study on the Linux Cluster usage in the German industry coupled with a questionnaire concerning actual aspects of Linux Clusters in the MDA (Mechanical Design Automation). The development departments in the engineering arena have risen the usage of the clusters because of the price/performance reasons. The study, based on data from May 2003, looks for the distribution in the technical and scientific market. (Uwe Harms)

In a joint effort with Los Alamos National Laboratory (LANL), the Panasas ActiveScale Storage Cluster has achieved more than 10 gigabytes per second of sustained data bandwidth for highly parallel Linux cluster applications. Three additional bandwidth records were set for tightly coupled cluster applications using LANL's MPI-IO benchmark. Together, these tests represent the range of throughput demands for technical computing applications. The Panasas storage cluster is currently deployed at LANL to provide scalable I/O for a variety of supercomputing clusters used for simulation.

Linux cluster software developer PathScale will introduce a suite of cost-effective, best-of-breed technologies that will add significant value and performance enhancements to users of servers based on Advanced Micro Devices' highly-regarded Opteron microprocessors.

Mellanox Technologies has launched the InfiniScale III single chip InfiniBand switch device offering 480 Gb/sec total bandwidth. The third generation device features 24 integrated 10Gb/sec full duplex InfiniBand ports capable of non-blocking full wire speed operation. The device is the first to offer 12X InfiniBand link aggregation supporting 8-Ports at 30Gb/sec. This level of integration advances the state of the art in fabric interconnects, delivering the highest performing commercially available single-chip solution, far surpassing the port density of any other 10 Gb/sec technology.

Mellanox Technologies L introduced the MTS2400 InfiniBand switch system powered by third generation InfiniScale III silicon - the first switch of its kind to deliver 30Gb/sec links. The MTS2400 is a high performance, ultra-low latency switch, offering enterprise customers an industry leading 480 Gb/sec non-blocking aggregate switching capacity and wire-speed forwarding for all twenty-four 10 Gb/sec ports. This capability is implemented in a compact 1U platform with the option to support up to eight 30 Gb/sec links. The MTS2400 is designed for creating low cost, high performance compute clusters from industry standard servers.

Cisco Systems Inc. continues to deliver advanced Internet Protocol version 6 (IPv6) features along with multicast and quality of service capabilities, through Cisco IOS Software, that enable the real-time information delivery necessary for advanced applications like video conferencing, supercomputing and Grid computing. With such advanced features, National Research Networks (NRN) such as SURFnet and Japan Gigabit Network (JGN) can use Cisco Catalyst 6500 switching and Cisco(R) 12000 Series routing platforms to introduce new applications and network-based services that can advance ongoing academic and scientific research.