EnterTheGrid - Primeur Monthly

EnterTheGrid - Primeur is the premier Grid and Supercomputing information source in the world. With Primeur Monthly we provide you a free update with news and in-depth stories.

>Primeur Magazine
Primeur Monthly - issue September 2003
>INCITE Programme to allocate major Department of Energy Office of Science computing resources to key scientific projects
>Charm++ 2003 Workshop issues Call for Papers
>Workshop on Modelling, Analysis and Simulation of Mobile Systems calls for participation
>I-SPAN 2004 issues Call for Papers
>Workshop on Grid and Cooperative Computing issues Call for Papers
>SC2003 seeking innovative visualisations for opening video
>Second Workshop on Application Specific Processors issues Call for Papers
  >HPCN industry
>MySQL and SGI deliver MySQL on the SGI Altix 3000 Supercluster
>Cray announces record financial results
>HP delivers faster ProLiant blade servers and high-performance networking switches
>Registration open for SC2003 Conference promising to "ignite innovation"
>Former IBM executive Peter Ungaro joins Cray
>AMD introduces Opteron processor Model 246
>New 1000-processor metacluster due at the University of Utah
>Tarari delivers 20-times algorithmic acceleration for selected high-performance comouting applications
>Tarari signs Racor Systems to sell reconfigurable high performance computing content processor
>Sun completes acquisition of CenterRun
>Fujitsu Siemens Computers launches new rack-optimised PRIMERGY RX800 and RX200 servers
>ISR and Cray settle patent lawsuit
>TeraBurst Networks and SGI deliver interactive visualisation environments over optical networks
>Singapore's Nanyang Technological University unveils new Reality Theater
>Discreet and SGI deliver SAN for film mastering and video production
  >The Grid
>Avaki releases Data Grid 4.0
>Animation Company uses HP Utility Data Center for speeding up rendering
>SDSC releases version 1.0 of SKIDLkit data mining toolkit
>DataSynapse expands European presence by partnering with PCIB to serve German market
>Tenax consultant addresses legal issues at Grid Computing Conference
>Sun supports BEA WebLogic Enterprise Platform 8.1
>Inventory of European Grid projects proposes new Grid Road Map
>Architecture of the future Grid
>Key areas for future Grid development
>UNICOREpro released
>United Devices and Optive Research to expand drug discovery technology
>Grid technology from SAS powers Environmental Health Organization
>TGen, ASU, IGC Collaborative selects Altair's workload management software
>New Grid and tools deployed across NPACI partnership
>Australian Grid Computing Workshop issues Call for Participation
>CESGA and CESCA join their supercomputing resources Globus and MPICH-G2
>On-demand Grid computing is not an economical business model
>SUN and UK-Science programme release Transfer-queue over Globus (TOG) version 1.0
>European funding helps Merseyside Grid research
>French Data Grid Explorer project approved
>ZapThink and ADT announce service-oriented architecture implementation poster
>Organon implements TurboWorx’ PowerCloud to accelerate drug discovery
>e-Toile: French Grid star starts shining
>Dot Hill helps Diveo Mexico boost data center capacity
> Access Grid Toolkit 2.1.1 released
>IBM's Storage Virtualisation Technology will handle one PetaByte of data at CERN
>SGI and ABAQUS announce availability of ABAQUS on SGI Altix 3000 Servers
>Washington University Biology, Chemistry, Astrophysics researchers benefit from new SGI Altix 3000 Supercluster
>Axiom SL chooses Powerllel to enhance parallel processing
>MTU Friedrichshafen implements MSC.ADAMS to improve engine development process
>University of Queensland buys 208 processor Altix 3000 supercomputer for research into earthquake phenomena
>MTU Aero Engines invests in MSC.Enterprise Mvision to improve material data exchange and collaboration
>Cornell Theory Center joins The MathWorks Connections Programme
>Xilinx Virtex series reaches two billion dollar revenue milestone
>Evernham Motorsports uses Product Lifecycle Management from IBM and Dassault Systèmes
>Multigrid Course - Introduction to Standard Methods
>NCSA to install 17.7 Tflop/s Dell cluster
>Sandia Red Storm selects Etnus TotalView debugger
>University of Kentucky supercomputer breaks the US$100 per Gflop/s barrier
>Linux Clusters Institute organises workshops
>IBM and SuSE Linux earn first security certification of Linux
>Terra Soft provides HPC system to the US navy through Lockheed
>Sun and SuSE Linux enter into global alliance
>Los Alamos National Laboratory selects 11 Tflop/s ASCI computer system
>Los Alamos National Laboratory selects Linux Networx to build 256-node InfiniBand cluster
>Sistina integrates GFS with HP high performance clusters for Linux
>Siemens Business Services selects SuSE Linux to drive payroll accounting and HR management systems
>SGI to deploy SuSE Linux on SGI Altix 3000 Servers
>New IBM Linux database clusters
>Japan National Institute of Advanced Industrial Science and Technology orders large IBM Linux supercomputer
>ICT Kenniscongres 2003 presents business card for ICT innovation in The Netherlands
>Inauguration of 70 TByte Disk Space for GridKa in Karlsruhe
>The Grid Computing Centre at Forschungszentrum Karlsruhe
>First Water-cooled PC Cluster in Production at GridKa
>Research Centre Jülich starts operation of 1,3 TeraFlop/s system
>Mellanox surpasses 100,000 InfiniBand port sales milestone
>University of California uses Adaptec network accelerator to speed genome research
>Internet speed mark with Cern participation in Guinness World Records Book
>19th ACM Symposium on Applied Computing issues Call for Papers
>Conference on Complex Systems in e-Business issues Call for Participation
University of Kentucky supercomputer breaks the US$100 per Gflop/s barrier
Lexington 22 August 2003 Researchers at the University of Kentucky have constructed and demonstrated an innovative new, scalable, parallel supercomputer that achieves application performance of more than 1 Gflop/s for every US$100 spent on building the machine. The approach used to design and build this machine makes it cost-effective for solving a wide range of problems, from drug design using computational chemistry to design of quieter printers using computational fluid dynamics (CFD).

A decade ago, supercomputers cost about a million US$ per Gflop/s performance. By using standard PC parts, "Beowulf" cluster supercomputers dramatically reduce the cost, but as processors and other components have become faster and cheaper, the network needed to coordinate them has become relatively expensive. The University of Kentucky researchers made their first breakthrough in reducing network cost in May 2000, when KLAT2, Kentucky Linux Athlon Testbed 2used standard 100mb/s Fast Ethernet hardware in the world's first machine-designed asymmetric cluster network -- and achieved $640 per Gflop/s, breaking the $1,000 per Gflop/s barrier. Their newest machine, KASY0, Kentucky Asymmetric Zero, uses a more advanced type of asymmetric network design to break the $100 per Gflop/s barrier.

A well-known reference for supercomputer performance is the TOP500 500 list. Performance on the Linpack uses depends partly on the theoretical peak Gflop/s of the processors, but also on the parallel implementation and efficiency of the network that allows the processors to work together. In the current (June 2003) list, most systems use expensive, specialized, network hardware. The machines explicitly listed as using standard 100mb/s Fast Ethernet achieve an average of less than 8.5% of peak. The average for the systems listed as using Gigabit Ethernet is somewhat better, at about 30% of peak. In contrast, KASY0's 100mb/s Fast Ethernet network allows it to achieve 187.3 Gflop/s, over 35% of peak using a double-precision version of the benchmark (HPL). Using a single-precision version, the $39,454.31 KASY0 obtains over 471.5 Gflop/s, more than 44% of its theoretical peak and less than $84 per Gflop/s.

The remarkable thing about KASY0's price/performance is that, while network hardware is often the dominant cost for a system of its size (128 plus 4 spare nodes), less than 11% of the system cost went for the network hardware. The AMD Athlon XP 2600+ processors were more than 35% of the total system cost; memory was 21%. Even more significantly, the network design technology that made this possible can be applied with similar benefit to cluster supercomputers with thousands of nodes. KLAT2's network was the world's first Flat Neighborhood Network; the enhanced version used for KASY0 is the world's first Sparse Flat Neighborhood Network (SFNN). KASY0 also is the first supercomputer to have its physical node and switch placement optimized by a computer program. FNN design technology and tools have been freely available and used by various other groups; so too will the new SFNN technology be freely available.

KASY0 is not a toy or a "hack" -- it is a serious demonstration of a fundamental new advance in network design. The only other supercomputer we have seen claim close to the price/performance measured for KASY0 is this $50,000+ system built by the National Center for Supercomputing Applications (NCSA) using 70 PlayStation2 units. Not only does KASY0 have a vastly superior network and significantly higher peak floating point performance per node, but KASY0's lower price yields many more nodes and real application performance, not just high peak numbers.

For example, KASY0 also has set a new world record for rendering a complex image using the Persistence of Vision Raytracer (POV-Ray). Executing pvmpovray 3.5 on KASY0 to render the standard benchmark.pov scene yielded a time of 72 seconds. According to this site, the previous record was 107 seconds set on August 1, 2003 by a cluster costing $79,000.

The primary architect of KASY0 is Tim Mattox, a research assistant who has been developing the Sparse Flat Neighborhood Network concept for his Ph.D. thesis. As an educational experience available to anyone, the physical construction of KASY0 was done entirely by volunteers at the University of Kentucky.

From the creation of the first Linux PC cluster in February 1994 to the construction of KASY0, Hank Dietz and his students have continued to improve cluster performance by making compilers, hardware architecture, and operating system work together more efficiently. At the University of Kentucky, as Professor of Electrical and Computer Engineering and James F. Hardymon Chair in Networking, Dietz's goal is to develop and freely diseminate the new technologies that will allow scientists and engineers to solve their most important computational problems.


EnterTheGrid - Primeur

James Stewartstraat 248

1325 JN Almere

The Netherlands



© EnterTheGrid - Primeur Monthly