Primeur magazine

Edition: live - Issue: 2011-06-20

Blog

The special TOP500 session at ISC11 started with an opportunity to ask questions to Yui Oniaga, designer of the K computer and Tadashi Watanabe designer of the Earth Simulator. According to Oniaga, the K system does not have GPUs, because they really would not add to the performance of the system. The most important part is the direct network connection. Which currently is based on Infiniband. The individual speed of the nodes is not that important, as long as their connection is. Read further...

Hardware

In the Tuesday morning session on "Sustained Performance on Petascale Systems" Michael Resch from the High Performance Computing Center in Stuttgart elaborated on the practical use of Petaflop/s systems. Why do we need Petaflop/s, he asked the audience. The mission of supercomputing is to support research. This is important for society, for economy and for politics. We need simulation to solve different questions in traffic, turbulence, energy, and so on. More computing power is needed to develop solutions for difficult issues. Read further...

Applications

As Russian scientists increasingly deploy GPU-enabled supercomputers to tackle scientific challenges, Moscow State University is upgrading its Lomonosov system with NVIDIA Tesla GPUs to be one of the world's fastest supercomputers. The upgraded system couples 1,554 NVIDIA Tesla X2070 GPUs with an identical number of quad-core CPUs, delivering an expected 1.3 petaflops of peak performance, placing at number 1 in Russia and amongst the fastest systems in the world. Read further...
hpc-ch is the Swiss HPC Service Provider Community. Their mission is to promote knowledge exchange between providers of HPC services in Swiss universities and research centres and to promote HPC in Switzerland. In on-line and in regular meetings the system specialists of hpc-ch are discussing best practices and developments in HPC. Social gatherings of hpc-ch's HPC specialists are building up a team spirit across Switzerland. Despite of its size, Switzerland is an important global player in research, notably in science and engineering. This implies that Switzerland is at the leading edge on HPC and HPC supported science. At the ISC'11 in Hamburg hpc-ch presents on its booth an overview of the HPC ecosystem in Switzerland. The focus is on th hpc-ch member organisations which are currently entities hold by the states (cantons) or the confederation. The relationship between the organisations located in different parts of the country is characterized by a friendly co-operation as well as a fruitful competition. Read further...

TOP500

The TOP500 rankings have once again awarded the title of 'Europe's most powerful supercomputer' to Tera 100. Having already been named as number one in Europe in the previous listing published in the USA, in November 2010, this confirms the leadership position enjoyed by Tera 100's technology - developed by Bull along with the teams from CEA-DAM - the Military Applications Directorate of the French Alternative Energies and Atomic Energy Commission - in an ultra-competitive marketplace. Read further...

The Grid

In the Tuesday morning session on "Cloud Computing and HPC", Tom-Michael Thamm from mental images described the company's RealityServer Service in the Cloud. mental images is a rendering company which has been on the market for 25 years. Their focus, based on an OEM model, is aimed at rendering for industrial companies interested in simulation. Since 2007, mental images is wholly owned by NVIDIA. mental ray is the company's major product, together with 3D visualization component software and mental mill, a shading application. mental images is working for companies such as Autodesk, Dassault and PTC. Read further...
Last speaker in the "Cloud Computing and HPC" session, chaired by Wolfgang Gentzsch, was Josh Simons from VMware. He talked about the converging concerns of HPC, Cloud and the enterprise. VMware is an expert in virtualization and has 250.000 customers. Today, 85% of the workloads are virtualized and virtualization will be a 241 billion dollar market in the coming years. Read further...

Company news

Cloud computing has swiftly moved from industry hype to a top IT initiative, with more CIOs building a private Cloud, evaluating service providers for public Cloud services, or employing a hybrid model that combines the best of both. Through an innovative global ecosystem of Cloud partners and a product portfolio that offers new integrated Cloud management capabilities and pre-validated Cloud infrastructure solutions, NetApp helps enterprise customers accelerate and simplify their transition to private, public, and hybrid Cloud models. Read further...
Bull and seven Founding Members have created the Bull User group for eXtreme Computing (BUX) on 20 June 2011, at ISC'11 in Hamburg. BUX is an independent worldwide group of users that will co-operate to increase the capabilities of the large-scale, parallel scientific and technical computing solutions supplied by Bull, to promote the exchange of information and understanding of these systems, and to provide guidance to Bull on the essential development and support issues for large-scale technical systems. Bull, as an Affiliate, is committed to supporting this new user group. Read further...
Bright Computing, an expert in cluster management software, is now shipping Bright Cluster Manager release 5.2. This latest version of the product includes multiple new features and a broad range of expanded capabilities for managing HPC clusters. These capabilities include enhanced multi-cluster support, full CUDA 4.0 support for NVIDIA GPUs, tightly integrated support for on-demand SMP, additional options for integrated workload management, support for the latest releases of Linux distributions, cluster reliability improvement and an enhanced web-based user portal. Read further...
T-Platforms, a global HPC company providing comprehensive supercomputing systems, software and services, will participate in a new programme jointly funded between the European Union and the Russian Ministry of Education and Science, and aimed at improving supercomputer performance. The programme is called HOPSA, which stands for Holistic Performance System Analysis, and has been funded for two years. Read further...
Just before the start of the Big Exhibition Party at ISC'11, the organizers took media representatives on an exciting tour crowded with HPC innovation in industry. At the Intel booth 12 demonstrations are running and the MIC technology is being displayed. HP has launched a new series of second generation servers for HPC with GP/GPU use. The new X9000 storage line has 1 single interface for administration and many, many cores. They also have a new data centre in a container which will be shipped at the end of this year. The cost of energy has been tremendously reduced with this solution. Read further...
Xyratex Ltd., a provider of enterprise class data storage subsystems and hard disk drive capital equipment, is exhibiting at the International Supercomputing Conference (ISC'11) in Hamburg, Germany. This year marks the first exhibitor level participation for Xyratex at ISC with a large booth on the exhibition floor, informative presentations on 'The Future of HPC Data Storage', Lustre roadmap futures from Xyratex at the HPC Advisory Council workshop, multiple sessions introducing ClusterStor 3000 and inclusion in the the world's first FDR 56Gb/s InfiniBand ISCnet demonstration along with other key HPC organisations including Microsoft, Mellanox, HP, Dell, and Fujitsu. The HPC Advisory Council, a leading organisation for HPC research and education, 56Gb/s InfiniBand demonstration will interconnect participating exhibitors' booths via the ISCnet network to demonstrate various HPC applications and products including Xyratex's new ClusterStor 3000 storage solution. Read further...
Super Micro Computer Inc., a global expert in server technology innovation and green computing, will demo its latest HPC solutions at the 2011 International Supercomputing Conference (ISC) in Hamburg, Germany. Supermicro is spotlighting its new 1U (1026GT-TRF) and 2U (2026GT-TRF) high-density GPU SuperServers supporting up to 4 and 6 GPUs respectively, the GPU SuperBlade (SBI-7126TG) providing 20 GPUs and 20 CPUs in a 7U enclosure, the TwinBlade (SBA-7222G-T2) supporting up to 3,840 Cores per rack and the 4-Way SuperBlade (SBA-7142G-T4), a high-performance (up to 60x 4-Way servers per 42U SuperRack) compute platform with QDR 40Gb/s InfiniBand or 10GbE connectivity per blade. Also on display will be their enterprise-class 8-Way, 5U SuperServer (5086B-TRF), a high-performance platform, supporting up to 80 Cores/2TB of memory and designed for mission-critical, high-availability computing environments. Read further...
Mellanox Technologies Ltd.'s recently introduced end-to-end FDR InfiniBand connectivity products have been selected by the Leibniz Supercomputing Centre (LRZ) to provide world-leading connectivity performance for its upcoming SuperMUC supercomputer. Scheduled for full deployment in 2012, the SuperMUC system will utilize the performance power of IBM servers with onboard Mellanox FDR ConnectX-3 InfiniBand mezzanine adapters and next-generation Intel Xeon processors, all connected and managed via 36-port and 648-port Mellanox FDR InfiniBand switches, software and copper and fiber cables. Read further...
High-performance computing (HPC) applications put incredible stress on an organisation's IT infrastructure as a result of the extreme amount of data and information that must be stored, managed, and processed. Addressing this challenge is no easy task and requires a storage solution that can deliver the performance, bulletproof reliability, and efficiency necessary to enable customers to succeed in today's hypercompetitive business environment. In May, NetApp unveiled the NetApp E5400 storage system, which is available to OEMs as part of the E-Series Platform, to help customers manage the massive amount of data resulting from big-bandwidth and high-performance applications. To help meet the growing needs of customers in high-performance environments, NetApp's E5400 has been benchmarked for use with the Lustre file system, an open-sourced high-performance file system used by the majority of organisations on the Supercomputer 500 list. Read further...
VMware Inc. has launched VMware vFabric 5, an integrated application platform for virtual and Cloud environments. Combining the market-leading Spring development framework for Java and the latest generation of vFabric application services, vFabric 5 will provide the core application platform for building, deploying and running modern applications. vFabric 5 introduces for the first time a flexible packaging and licensing model that will allow enterprises to purchase application infrastructure software based on virtual machines, rather than physical hardware CPUs, and to pay only for the licenses in use. Read further...