Bull extends and updates its bullx family of supercomputers, to offer the most comprehensive and powerful range of solutions available for Extreme Computing
- With the new 'supernodes' designed and developed by its teams, Bull is now offering the latest-generation nodes as part of its bullx family, to meet the demands of even the most memory-hungry applications and reduce the complexity of large-scale computing infrastructures.
Bull boosts the speed of its blade systems, by equipping them with the latest Intel® Xeon® 5600 processors.
- A year after the successful launch of the first bullx systems - named in 2009 as the best High-Performance Computing technology in the world - unlimited innovation with no compromise has never been a more real possibility for the business and research communities.
31, 2010 -
Following on from the success of the bullx blade system - named as the world's best supercomputer by HPCwire, the benchmark American journal for the world of High-Performance Computing (HPC) - Bull is extending its family of supercomputers with the launch of its bullx supernodes. The key advantages for business? Power, memory and environmentally-efficient technologies that deliver innovation on demand, with no compromises.
Philippe Miltin, Vice-President Products and Systems of Bull Group comments: "Bull is now a technology leader capable of offering its customers new ideas and tailored solutions that effectively combine innovation and environmental responsibility. In 2009, we successfully opened up the way to Extreme Computing, with the bullx blade system; the new bullx supernodes are a further step forward. Our message to businesses and research establishments is: refuse to compromise, and open up new horizons for your organization, while at the same time minimizing the risks and reducing your carbon footprint. For Bull, innovation has to be unlimited and unconstrained: our bullx supernodes are a concrete example of this principle."
bullx supernodes: high-end compute nodes effectively combining power and memory
Having successfully launched both rack-mounted servers (R-series) and then blade servers (B-series), designed specifically for Extreme Computing, Bull is now focusing its attention on 'high-end' nodes, with the launch of its bullx supernodes (S-series). These four-socket SMP (Symmetric Multi-Processing) servers - which can be extended to 8, 12 or even 16 processors thanks to a new technology exclusively designed by Bull - offer an extremely powerful compute node in their maximum configuration, with up to 128 cores and 2TB of shared memory.
Designed and developed entirely by Bull's expert teams, in close co-operation with the French Atomic Energy Authority (the CEA), the new S-series machines offer extremely powerful 'computing units'. They capitalize on the very latest generation Intel® Xeon® 7500 series processors (codename: Nehalem-EX), designed for the most intensive applications and offering enhanced reliability, combined with exceptional scalability. Using these processors, Bull has designed an extendable SMP server capable of meeting the needs of even the most memory-hungry applications.
SMP servers enable a real concentration of computing power combined with an extremely large memory. Using bullx supernodes as part of a cluster:
- Simplifies the often complex infrastructures associated with large-scale supercomputers, by reducing the number of compute nodes by a factor of five compared with commodity servers, making it easier both to implement the cluster in the first place and to administer it day to day
- Simultaneously delivers both the very high levels of processing power and the very large amounts of memory needed by some extremely resource-hungry applications. In particular, these kinds of applications are found in:
* Pre-processing, for all mesh-generation operations (in climatology, aeronautics and automotive engineering, crash simulation, oilfield simulations¿)
* Processing, for the mechanics of complex structures, for example, calculating the fracture-points of metallic elements, or in the chemistry or physics of materials, to understand how the constituent elements behave at a molecular level.
It was these advantages that led the CEA to choose bullx supernodes for its future Petaflops-scale system, TERA 100.
Two models will be available in the S-series:
- bullx S6010: ultra-compact high-end compute nodes
The exclusive format of bullx S6010 nodes - L-shaped 1.5U drawers which fit together, with one upside-down on top of the other, to form a 3U drawer - enables a level of density in the 8 or 16-processor configuration (3 or 6U) never before available in this market. Specifically designed to act as compute nodes, S6010 servers feature the best available components for this function, and only those that are essential for processing.
- bullx S6030: service nodes, with connectivity capabilities and enhanced storage in 3U format
bullx S6030 nodes offer advanced connectivity functions, a redundant power supply and extended storage options. This means they are especially well suited to act as management or I/O nodes that are just as effective as they are reliable.
Fabio Gallo, Vice-President and Director of Extreme Computing Solutions at Bull Group comments: "In Extreme Computing, the needs of businesses and the research community are extremely diverse. Our customers are looking for solutions that enable them to run a wide spectrum of applications within the same system. Which is why Bull has devoted itself to designing a range of supercomputers well-suited to the architectural models used in the world of HPC: 'scale-up' and 'scale-out', as well as accelerators. So our customers can continue to innovate with no real limitations."
A comprehensive family of supercomputers, for customized, unlimited innovation, with no compromises
With the launch of its bullx supernodes, Bull is continuing the roll out of its family of supercomputer nodes designed specifically for Extreme Computing. HPC users have extremely diverse requirements, and are increasingly searching for hybrid solutions to cover the widest possible spread of applications. Bull has recognized this, and has been gradually building an exhaustive range of solutions that includes 'thin nodes', 'fat nodes' and accelerators, that can be combined with each other in a customized way, in a single bullx cluster, and administered and managed as a single system using the bullx cluster management software suite.
With the addition of the S-series 'fat nodes' launched today, the bullx family effectively enables very diverse requirements to be met, with:
- The R-series models (rack-mounted servers) offering a high degree of flexibility, with a wide choice of options, including coupling with GPUs
- The B-series (blade servers) delivering extremely high density, with 18 bi-processor compute nodes in just 7U. They also offer an exclusive solution featuring GPU accelerators integrated directly into a blade. Capitalizing, as always, on the very latest technologies, the bullx blade system has also just been updated: the B500 blades now feature Intel® Xeon® 5600 series (Westmere-EP) processors, boosting their power by 50% and increasing the maximum power of a bullx rack full to capacity with blade chassis to 15.2 Teraflops.
Much more than just a complete hardware offering: a wide range of solutions
As well as its comprehensive hardware offering for Extreme Computing, Bull offers its customers a wide range of solutions:
- The bullx cluster suite, developed by Bull's Extreme Computing center of excellence, is a robust and effective solution which provides all the functions needed to manage a server cluster at every stage in its lifecycle: installation and software implementation, surveillance and error processing, optimization and extension of the cluster. It can be used to manage all bullx systems, including those based on GPUs
- The mobull¿container, a true mobile Data Center that enables a new IT facility to be established or an existing one to be extended in the shortest possible timescales, by freeing the organization from the constraints associated with building a permanent Data Center. All that is needed is an electricity supply, a cold-water supply and a small amount of space in a car park. A mobull¿container is capable of hosting the equivalent of up to 227.8 Teraflops (CPU)
- Bull also provides customized hosting services for HPC systems and sells computing power on demand, for organizations who want to focus on their core business rather than on managing a Data Center. This is also a valuable solution when it comes to meeting peaks in business demand.
Environmentally friendly technologies for 'Green Power'
Using environmentally friendly technologies, bullx offers genuine 'Green Power' to overcome the issues related to the high electricity consumption of high-power supercomputers:
- An ultra-capacitor module improves the efficiency of the electrical power supply by 10-15% compared with traditional power supplies: a super-condenser protects the server against micro-outages of up to 300ms. In areas with good quality electricity supply networks, this does away with the need for a UPS and enables savings of up to 15% on electricity consumption.
- A cooling door consumes 75% less that standard air-conditioning
- A processor control system optimizes electricity consumption
- Bull's expertise in infrastructure design optimizes the Flops/m² ratio.
bullx supernodes are available now; drawer interconnection features will be available in the second half of 2010.
Bull is a leader in secure mission-critical digital systems. The Group is dedicated to developing and implementing solutions where computing power and security serve to optimize its customers' information systems, to support their business.
Bull operates in high added-value markets including computer simulation, Cloud computing and 'computing power plants', outsourcing and security.
Currently Bull employs around 9,000 people across more than 50 countries, with over 700 staff totally focused on R&D. In 2011, Bull recorded revenues of 1.3 billion.
For more information visit: http://www.bull.com