advertisement
Visit our sponsors

May 2000

Primeur is a monthly Virtual Magazine on High Performance Computing and Networking in Europe. It is produced by an editorial team composed of professionals in publishing and HPCN. You can put the editorial team to work as well. Read about our services and advertising opportunities and find out about the friendly rates .

advertisement
Visit our sponsors

Breaking news - just a click away:


Advertisement
Special advertising opportunities for PrimeurLive! from Europe's premier HPCN events
  • Supercomputer 2000, Mannheim, June, Mannheim, Germany
and more events later this year.
details and rate

Choose a section or scroll down to the articles

The Grid
SAVE Special
Media and visualisation
Linux
HPCN industry
Cluster computing
Applications
United Kingdom
The Netherlands
Sweden
Spain
Italy
Israel
Germany
France

Logicon wins $46 Million U.S. Navy contract
A team led by Logicon Inc., a Northrop Grumman Corporation company, has been selected by the U.S. Navy to support development of advanced Command, Control, Communications, Computers, Intelligence, Surveillance and Reconnaissance (C4ISR) systems. One of the contractors is the San Diego Supercomputer Center.


France

Collaboration between French research organisation CNRS and InforMax on bioinformatics
InforMax,a United States developer of bioinformatics software for accelerated drug discovery, entered a strategic collaboration with The French National Centre for Scientific Research (CNRS) in the field of computational genomics. With support from InforMax, CNRS will establish a European Center of Excellence as part of the CNRS Research Unit on Molecular Genetics and Developmental Biology in Villejuif, France.

Germany

Final programme Supercomputer 2000 in Mannheim available
The final programme of the Mannheim Supercomputer 2000 seminar is available on the web. On-line registration is possible.

Stodden new CEO of Fujitsu Siemens
Paul Stodden will be the new President and Chief Executive Officer (CEO) of Fujitsu Siemens Computers. In his previous role 52-year-old Stodden restructured the Siemens IT Service GmbH. At the same time, as President and CEO of IT Services he also worked with the Fujitsu group to provide services to customers world-wide as part of the global co-operation.

Hitachi SR8000 F1 at LRZ operational - over 1 Tflop/s of raw LinPack performance
Hitachi has successfully installed the Hitachi SR8000 F1 at the Leibniz Rechnenzentrum in Munich Germany. Currently the machine is undergoing its final stability test during 30 days of operation. In a trhoughput benchmark, the machine showed a sustained performance of 424 Gflop/s. LRZ expects that in the next TOP500 list that will be published in June at the Mannheim Supercomputer Conference, the machine will end close to the top 3, after the big US national centres.

Experts from Europe and the US at RCI conference in Mannheim
Among the confirmed speakers for the RCI Aerospace - Automotive Special Interest Group (SIG) Executive Conference to be held in Mannheim in June are David Chang, General Motors; Michael Mott, Boeing; Terry Knibb, BAE Systems; Carl Murphy, Accord Solutions and Michael Faden, DLR. The conference will cover integrated science and engineering models and representations that are becoming a strategic "need" for the R&D and design world.

Israel

SGI to deliver Real-World Visual System for Israeli Air Force F-15 flight simulator
Lockheed Martin has awarded SGI a contract to develop and install a turnkey visual system solution for the Israeli Air Force F-15 Flight and System Trainer.The system will be comprised of the commercial, off-the-shelf Silicon Graphics Onyx2 graphics supercomputer with the next-generation InfiniteReality3graphics subsystem from SGI, MultiGen-Paradigm's Vega runtime software and the SEOS PRODAS HiView display.

Italy

Advanced Research Workshop on HPC in Cetraro
an Advanced Research Workshop on HPC will be held in Cetraro, Italy - June 12 - 15, 2000. The aim of the Workshop is to contribute to assess the current and future developments in the HPC architecture and software technologies, and to discuss promising applications of HPC to new problems of science, industry and business. Work shop web page: http://www.parcolab.unical.it/hpc2000/index.html

Spain

The ACCESS project: visiting Spanish supercomputing centres through EC IHP programme
CESCA and CEPBA, the Spanish supercomputing LSFs (Large Scale Facilities) can be visited by European researchers thanks to the project ACCESS (Access to CESCA and CEPBA largE Scale facilitieS), developing since March 1st 2000.

Sweden

National Swedish Supercomputer Centre buys 96 Gflop/s SGI-SN
The National Swedish Supercomputer Centre (NSC) has completed the procurement of a new HPC system. The system, an SN-MIPS from SGI, will be installed during the fourth quarter of 2000. The machine will have 96 processors, each with 1 Gflop/s peak and 1 Gbyte of memory. The system will replace a current Cray C90 system.

The Netherlands

TIMA and IRISA on ACE tools
TIMA and IRISA both announced research projects using ACE Associated Compiler Experts CoSy compiler development system. These two key research institutes are exploring compiler generation with industrial implications.

Workshop New National Dutch Supercomputer
On 9 May 2000 in Amsterdam NCF, who was responsible for the acquisition of the new Dutch national research supercomputer, organises a workshop, as an associated event of HPCN Europe 2000 event. Presentations include "To acquire a new national supercomputer" and "Benchmarking for a large scale computing facility".

United Kingdom

NEC invests in UK semi conductor plant
NEC Corporation will invest 60 million in its subsidiary NEC Semiconductors (UK) Ltd. in the current fiscal year, raising total investment in the company since its establishment in 1982 to over GBP1 billion.

First customer for Systolix math processor array
Analog Devices Inc. will be first to use the "PulseDSP," a customizable math processor array devised by startup Systolix Ltd. (Liverpool, England). Introduced at the DSP World Spring Conference, the array consists of 144 separate multiply/accumulate cells whose function is user-configurable. It will provide real-time digital filtering for the A/D converter Analog Devices plans to introduce later this year.

Edinburgh EPCC calls for TRACS participants
Edinburgh Parallel Computing Centre (EPCC) is coordinating an EC-funded project to bring European researchers for short visits to associated departments in Edinburgh to collaborate on projects involving High Performance Computing (HPC). Next application dead line: 26th April 2000.

Weapons Experts Present Experiences of ASCI Programme
3-4 April 2000; Keble College, Oxford, UK and the natural history museum next to it, were the venues for an International Conference: "Exploiting Leading-Edge High Performance Computing", hosted by the UK Atomic Weapons Establishment (AWE), under new management from 1st April. This two day event attracted 135 people, more than half employees of Weapons Federal Laboratories in the USA and their counterpart establishments in the UK and France. The presentations from the USA revolved around their achievements and pains during the implementation of the ASCI programme, while the Europeans looked with envy and talked about their efforts with much more modest computer resources. Presented side by side it became patently obvious that the capability gap, between the USA and Europe, in using computer simulation techniques for the stewardship of nuclear weapons, is in reality an enormous chasm. The French, at least, have a clear path in following the USA by ordering a 5 Teraflop/s system from Compaq, while in the UK, no specific plans have emerged. It was alluded however, that a similar path will be followed in the near future.

Limits and expectations of HPC systems
Tera's MTA designer, Burton Smith, said at the HPC conference in Oxford, that he foresees two distinct lines of supercomputers that will survive indefinitely. The scalar MPPs with little communication capabilities compared to their processing and the vector PVPs which are better characterized by memory bandwidth than theoretically infinite scalable peak performance. Will this really be the future? Let us look at the last decade to see whether we can foretell what is likely to happen.


Applications

IBM delivers DB2 for NUMA-Q servers
IBM announced the general availability of DB2 Universal Database for its Intel processor-based NUMA-Q servers. This combination creates a foundation for e-business and business intelligence solutions that require very high performance for multi-terabyte databases.

Primerica finacial service on Shark
US financial services provider, Primerica Financial Services, has chosen the IBM 'Shark' Enterprise Storage Server, the company's most advanced disk storage system, to manage its increasing customer service operations and protect its data. At the Atlanta data centre, Primerica's IT staff manages two Enterprise Storage Servers at 1.6 TByte each.

Model Technology and Platform Computing increase simulation capacity for server farms
Model Technology, a Mentor Graphics company and a supplier of HDL simulation tools, and Platform Computing announced a joint solution that dramatically increases simulation capacity in distributed computing environments. The integration of Platform's Load Sharing Facility (LSF) software and Model Technology's ModelSim simulator allows customers to maximize the capacity and utilization of their simulation farms - enabling both more simulation cycles per day and a better return on investment. By coupling ModelSim with LSF's job management software, designers can now easily construct and manage simulation farms with any number of workstations across multiple locations.

The next frontier in aerospace/automotive application of HPC
RCI organises a special event on June 7, 2000 called "Integrated Science/Engineering modeling: The next frontier in aerospace/automotive application of high-end computing" in conjunction with the Mannheim Supercomputer 2000 Seminar. The programme is now available with key speakers from automotive and aerospace companies in Europe and the USA.

NAG Parallel Library 3.0 with new features
The Numerical Algorithms Group (NAG), announced a new release of the NAG Parallel Library. Release 3.0, allows the user to make use of the performance of parallel machines or networks of workstations and PCs behaving as if they were a single parallel machine. The library offers greater speed of execution over conventional sequential numerical software, particularly on networks of workstations and PCs.

New release of NAG SMP library
The Numerical Algorithms Group (NAG), has announced a new major release of the NAG Fortran SMP Library specifically optimised for use on Symmetric Multi-Processor computers.

Western Resources Relies on IBM S/390 Data Center Servers for New Business Growth
Western Resources has reorganised its data center applications team. An early choice the IT team made was to strengthen their existing use of IBM S/390 enterprise servers as the core of its data center. Wstern Resources serves millions of customers, both in th US and in Europe.

Next gen Electromagnetic Effects Simulation Software from ESI
ESI announced PAM-CEM 2000, its 3D Computational Electro-Magnetics (CEM) simulation software. PAM-CEM, provides realistic virtual testing of complex electromagnetic phenomena arising from the tremendous increase of electronic devices. Relying on physics-based simulation models created from CAD or CAE data, PAM-CEM analyses electromagnetic compatibility and interference phenomena. PAM-CEM enables easy visualization of the invisible electromagnetic world.

Cluster computing

Compaq to ship new servers and upgraded version of Tru64 Unix
Compaq Computer Corporation announced major enhancements to its high availability AlphaServer systems, and introduced NonStop eBusiness solution infrastructure capabilities with Oracle on Compaq Tru64 UNIX and TruCluster Server. Announced were major enhancements to the AlphaStation and AlphaServer DS20E and ES40; a new, pre-configured AlphaServer ES40 Cluster Base System for Tru64; and enhanced versions of Tru64 UNIX V5.0A and TruCluster Server V5.0A.

HPCN industry

New technique boosts chip performance
IBM has developed a new method for building microchips that can deliver up to a 30 percent boost in computing speed and performance. IBM's new manufacturing technique uses a material known as a "low-k dielectric" to meticulously shield millions of individual copper circuits on a chip, reducing electrical "crosstalk" between wires that can hinder chip performance and waste power.

Nullstone supports ACE' DSP-C language extensions in its Automated Compiler Performance Analysis Tool
ACE Associated Compiler Experts announced that Nullstone Corporation is now supporting ACE's DSP-C language extensions in the latest release of their compiler performance analysis product. Nullstone's new release will be available later this month. ACE's DSP-C is provided standard with the DSP release of its CoSy compiler development system.

Intel buys Kuck & Associates
Intel Corporation today announced it has entered into a definitive agreement to acquire privately held Kuck & Associates, Inc. (KAI) for an undisclosed cash amount. It is also not clear which direction Intel will take with the newly acquire KAP multiprocesssing tools.

HP Unix servers show good growth results
Hewlett-Packard Company closed its fourth quarter of 1999 with substantial growth across its entire line of Unix system servers, according to International Data Corporation's (IDC) Unix Servers Quarterly Tracker market-share report. The report shows a increase of 33 percent in factory revenues for entry-level, midrange and high-end HP 9000 Enterprise Servers between the fourth quarter (Q4) 1998 and Q4 1999.

SC2000 proposals due by April 28
Submission of proposals for SC2000, the conference of high performance networking and computing for technical papers, tutorials, education program, panels, and Gordon Bell Prize nominations must be made via the conference web site. SC2000 will be held Nov. 4-10, 2000, in the Dallas Convention Center, USA. Submissions in these categories must be made by Friday, April 28. Notification of acceptance for technical papers, tutorials, education program, and panels will be made by Monday, June 26.

Starfire leads high-end
SUN says that according to International Data Corp.'s Q4CY99 Server Tracker report, Sun shipped 428 units of its Starfire servers for that quarter. Combined, IBM and HP managed to ship just 148 high-end Unix server units.

SGI Appoints Larry Sheffield Senior Vice President, the Americas
SGI appointmed Larry Sheffield, 50, as senior vice president of the Americas. Immediately prior to joining SGI, Sheffield was vice president of sales for the Enterprise Systems group at IBM for the Western United States.

Bioinformatics workshop in Amsterdam
The Dutch Platform HPCN will organise a workshop on datamining, focusing on the relation with bioinformatics, on May 9th during the International HPCN Europe conference in Amsterdam. Several projects of the Platform HPCN will be presented. Further, a "Virtual Laboratory" project shows emerging new research environments. Gigaport presents the broad band network possibilities in the Netherlands and the Dutch ministry of Economic affairs will explain the "Actieplan Life Sciences" policy plans.

Tera completes Cray acquisition
Tera has closed its acquisition of the Cray vector supercomputer business unit, and the Cray brand name from SGI. The acquisition was made in exchange for $15 million in cash, one million shares of unregistered common stock and a nine-month promissory note in the amount of $36.3 million subject to adjustment upon final audit.

Linux

NEC Soft Launches NetVault From BakBone Software for Linux Storage
BakBone announced that NEC Soft . (NES), the systems-integration arm of NEC, has selected BakBone's NetVault for its Linux Enterprise Oracle Server Solution. BakBone Software's NetVault is a storage-management software application that supports department to enterprise systems, including all major Unix Windows NT and Linux platforms.

Linux company into mainstream IDC rankings
International Data Corporation (IDC) announced that it will re-categorize and track VA Linux Systems as a ``brand-name'' vendor in its Commercial Systems and Servers program. With the re-categorization, VA Linux Systems ranks as one of the top five vendors in IDC's statistics on Linux server shipments for Q4 1999.

VA Linux to Open Offices in France, UK, Germany, The Netherlands and Switzerland
VA Linux Systems has named William Cobert as managing director of European operations. The company also opens offices in several European countries.

Linux Cluster supercomputer at NOAA unveiled
The $17 million Linux cluster called JET at NOAA, the US National Oceanic and Atmospheric Administration, has been unveiled this week. Cluster top speed is 5 Tflop/s. JET was developed by High Performance Technologies Inc.

(Fujitsu)-Siemens partner with SuSE
Fujitsu Siemens Computers, Siemens Business Services (SBS) and Siemens IT Service (ITS) have extended their Linux offering by signing a Memorandum of Understanding with SuSE Linux AG. The agreement is part of a major drive to bring the benefits of Linux as an Enterprise IT platform, and combines the strengths of all four organizations to provide European organizations with a powerful partner for all their Enterprise Linux requirements.

Media and visualisation

Visual Numerics Enters Open Market via ComponentSource
Visual Numerics and ComponentSource inked an agreement whereby ComponentSource will electronically distribute Visual Numerics' entire product line for the first time to the global open market.

Manchester Visualisation Centre installs 1 million euro VR facility
The Manchester Visualisation Centre ordered a 1 million euro VR facility with Virtual Presence, a subsidiary of MUSE Technologies. The order includes setting up high-performance visual computing platforms for researchers, students and commercial users to project real time graphics in a three-screen theatre. The presentation of data in a theatre setting using extremely high-resolution graphics improves interaction with the data so that large groups of users from different disciplines can collaborate and reach conclusions as a team. Virtual Presence has previously collaborated with the University of Manchester on other projects, including the development of MIST (Minimally Invasive Sur>

Transfer interrupted!

based system designed to teach and assess minimally invasive surgical skills. More than 125 MIST systems are in use at 70 medical institutions worldwide.

Next gen graphics for SGI Onyx2
SGI launched the InfiniteReality (IR3) graphics engine for the Silicon Graphics Onyx2 visualisation workstation. Extending the industry-leading InfiniteReality graphics architecture, InfiniteReality3 provides the most advanced capabilities for image quality and realism. IR3 is suited for applications in the modelling and simulation, oil and gas, manufacturing and entertainment markets. For scientific visualisation, researchers can now interact with medical and seismic data sets with greater visual accuracy due to performance increase on volumetric data sets up to 512x512x512 in size.

Intergraph announces new Zx10 ViZual Engine
Intergraph Computer Systems (Nasdaq:INGR) today announced its new Zx ViZual Workstation, the Zx10 VE (ViZual Engine), a high-performance 5U rackmount and Ultra-Tower workstation based on Intergraph's Wahoo Technology, which delivers unmatched system throughput and I/O bandwidth. The combination of the Zx10 VE with Intense3D's latest Wildcat 4210 3D Graphics, achieves several of the industry's highest benchmark performance scores.

SGI to provide high-performance visual systems for U.S. Army AVCATT-A
SGI will provide its advanced high-performance visual computing equipment and servers to power the U.S. Army's new Aviation Combined Arms Tactical Trainer-Aviation Reconfigurable Manned Simulator (AVCATT-A).

SGI Names Gary Geissler Vice President of Engineering
Gary Geissler, 47, has been named SGI vice president of Engineering based in Chippewa Falls, Wisconsin. A Chippewa Falls native, Geissler will be responsible for scalable server development, as well as leading and managing SGI's Engineering Technology Services and Interconnect Technology organizations.

SAVE Special

12th Conference of the Fujitsu Siemens SAVE AK "Scientific Computing"
On March 23 and 24, the SAVE (Siemens Users Group) working group "Scientific Computing" had its spring meeting in Paderborn. About 70 participants from all over Europe came together, to discuss their experiences in using Fujitsu Siemens computers. Summaries of the most interesting presentations are given in separate articles. Head of the AK Scientific Computing is Dr. Klaus Brühl from RWTH Aachen, a technical university.

Fujitsu Siemens' hpcLine takes off
In the status report of hpcLine, which was announced end of February 1999, Fujitsu Siemens reported about 11 new customers in academia and industrial research. This system is an integrated cluster of SMP (Symmetric Multiprocessor) nodes, each containing two off-the-shelf Intel Pentium III processors. They are connected via a high-speed system area network, mostly SCI (Scalable Coherent Interface) from Scali Norway. The operating system is optional, Linux or Solaris. For academia and research Fujitsu Siemens has a special offer, the Starter Kit: 8 nodes with 2 processors each - 16 Pentium III (650 MHz), 8x20 GByte IDE disks, 8x512 MByte SDRAM (ECC) main memory, SCI network, ScaMPI library, ScaCONF tools, Fast Ethernet adapters plus other hard- and software components for DM 130.000 (66.468 euro plus VAT.

GP7000F - the new workhorse from Fujitsu Siemens
Although VPP supercomputers and hpcLine parallel machines catch a lot of attention, even in HPC, much work is done on workstations and small servers. The GP7000F is Fujitsu Siemens' work horse for these cases. The company offers a broad spectrum, starting with the models 200 (275 MHz, 4 MB 2nd level cache, up to 2 CPUs, up to 4 GB memory), 400 (300 MHz, 8 MB 2nd level cache, up to 4 CPUs, up to 8 GB) and 600 (the same as model 400 but up to 8 CPUs). The models 1000 and 2000 use 330 MHz SPARC64-GP chips, 8 MB 2nd level cache and up to 64 GB memory (64-way interleaved) with 6 to 32 processors (mod. 1000) and 8 to 64 CPUs in up to two cabinets (mod. 2000). The highend systems have internal disk space up to 1 TB and external up to 700 TB. The memory crossbar has a bandwidth of 40.96 GB/s (51.2 GB/s for the coming 128 processor system). The operating system is SUN Microsystem's Solaris, identically running on SUN SPARC-servers.

Several European Fujitsu Siemens VPP vector users upgraded their systems
The Fujitsu Siemens vector computer users reported their experiences during the SAVE users group meeting in Paderborn. There are several Fujitsu Siemens vector supers in Europe. There was specific interest for the first experiences with the new machine VPP5000. Research centre Karlsruhe, ECMWF, AUDI AG, Meteo France, Electricite de France and CEA (Commissariat Energie Atomic) presented their centres.

VPP5000 at AUDI AG shows good performance on CFD and crash code
During the SAVE meeting, Karl Huber, head of CAE support at AUDI AG, described the usage of vector supercomputers and discussed the reasons for acquiring the Fujitsu VPP5000 in addition to the still existing VPP300 with 16 processors. Because of shortening the product development cycle at Audi, an enlargement of computer simulation is necessary. He also compared the performance of a RISC processor and a VPP5000 processor for typical Audi applications.

The Grid

EuroPVM/MPI conference
The 7th European PVM/MPI User's Group Meeting will be a forum for the users and developers of PVM, MPI and other message-passing programming environments. Final Date for Submissions is 26th April 2000. You can get more information on the following web page: http://www.lpds.sztaki.hu/EuroPVM_MPI2000

Metacomputing Workshop focuses on distributed simulation
The High Performance Computing Center Stuttgart (HLRS) will bring together experts in the field of metacomputing to discuss relevant issues for distributed simulations on high performance computing systems. HLRS has invited experts from the US, Japan and Europe for a small and focussed workshop in June.

CSP Inc. introduces Highly Scalable Cluster Computing System
CSP introduced its new highly scalable clustering system: FastCluste. This new system is the first product of its kind to run the Linux operating system. FastCluster was designed to meet the high performance computing requirements of engineers in the scientific and technical computing fields, as well as Internet applications like high performance search engines and Web database servers.

Legion Spring Workshop
A joint University of Virginia-Applied Metacomputing spring Legion workshop will be held May 16-17, 2000, at the university's Computer Science Department. Day one will cover introductory information and a hands-on lab for running applications. Day two will involve more advanced system administration and applications, as well as information on customizing and troubleshooting systems.

Workshop on OpenMP Applications and Tools
Deadline for abstract submisson for the Workshop on OpenMP Applications and Tools (WOMPAT) at the San Diego SuperComputer Center, San Diego, California July 6th-7th 2000, is May 5. The workshop It is dedicated to the OpenMP language and its use. WOMPAT provides an opportunity for users and developers of OpenMP to meet, share ideas and experiences, and to discuss the latest developments. Web site: http://www.cs.uh.edu/wompat2000/

375 Gflop/s Linux cluster with 512 processors called LosLobos in Albuquerque
The National Computational Science Alliance in the USA will take delivery of a 512-processor Linux supercluster within the next month. The new supercluster, called LosLobos, will be located at the University of New Mexico's (UNM) Albuquerque High Performance Computing Center (AHPCC), one of the Alliance Partners for Advanced Computational Resources sites.


MCI WorldCom operates NSF's High-speed back bone for another three years
As expected, the National Science Foundation and MCI WorldCom have agreed to a three-year, no-cost extension that will keep the very high performance Backbone Network System (vBNS) operating until at least 2003. The network connects 94 U.S. universities to other research institutions at speeds up to 2.4 Gbit/s.

Adaptec's Ultra160 SCSI in Celcius
Fujitsu Siemens Computers will begin including Adaptec's Ultra160 SCSI technology in its Celcius workstations.

DANTE and Infonet interconnect European research net TEN-155 to the European commercial Internet
Infonet Services Corporation and DANTE, the company organising advanced data networking services for the European research community, signed a contract which awards the provision of connectivity between the TEN-155 network and the European commercial Internet to Infonet. The first interconnection between TEN-155 and the European commercial Internet under this new contract was made in Amsterdam in September 1999. The interconnections in Geneva, London and Stockholm were made in October, November and January respectively.