Contents
Editorial
Guest contributor
Hot topics
Business cases
Experts voice
Solutions
At a glance
Quick poll
Events
PDF version
 

Subscribe to
Bull Direct:


 

Archives
n°25  |  April   2008
Hot topics
High-Performance Computing: a major challenge for society, and a strategic commitment from Bull

Bull has become an indisputable player in the High-Performance Computing field in Europe, with exceptional growth over the past three years, major contracts, numerous records broken and significant investments in R&D

Following proposal by the French President, Nicolas Sarkozy in the presence of the German Chancellor, Angela Merkel during the opening of CeBIT, concerning a joint Franco-German supercomputer program, Bull has issued the following statement.

High-Performance Computing (HPC): a major challenge for society...

HPC has become absolutely essential in industry and in university research activities. In aeronautical engineering, energy, climatology, biosciences, financial services, sport, video games... computer modeling and simulation have permeated almost every domain. Computer simulation in particular has helped to reduce product development costs, and ensure that more reliable and innovative products, that consume less energy, get to market more quickly. HPC has become a vital investigative tool for researchers, a major competitive differentiator for companies, and even a fundamental component of State sovereignty.

...and a strategic commitment from Bull

Bull’s strategy in scientific computing is inextricably bound up with its historic expertise in complex networks and large-scale IT systems, and with its positioning as ‘Architect of an Open World™’. But this strategy really comes into its own in the current market, which is undergoing radical change with the pre-eminence of standard processors and Open Source software. It’s a strategy that has led Bull to become one of the pioneers in the design of large-scale supercomputers based on open, standard components and on networked ‘cluster’ architectures, enabling its customers to take full advantage of high performance open systems that are also highly scalable, and highly competitive compared with the leading proprietary systems.

Exceptional growth

Bull has won worldwide recognition thanks to TERA-10 , the first large-scale supercomputer designed and developed by Bull for the CEA (the French Atomic Energy Authority): ranked in June 2006 as number one in Europe and number five in the world in terms of computing power. Since then, Bull has made a significant contribution to re-energizing the HPC market; by making high-performance computing solutions much more widely available to research and industry alike. With over 100 customers in 15 countries across three continents, Bull gathered significant momentum in HPC during 2007. But the aim is to go much further, and much faster: this year we want to double our revenues from this segment.

Major contracts

The spread of countries and industry sectors covered, as well as the sheer diversity of solutions that Bull has sold, illustrates the reputation that the company now enjoys. From the first major supercomputer installed at the CEA to the latest machine – a system delivering in excess of 25 teraflops, purchased in February 2008 by Cardiff University ( Professor Sir Martin Evans FRS being awarded the Nobel Prize for Medicine in 2007 ) – with numerous supercomputers delivered to many other higher education establishments in Brazil , France , Spain , Germany ... along the way. In industry, prestigious customers including Alcan , Pininfarina , Dassault-Aviation and Alenia have chosen Bull solutions. And Miracle Machines in Singapore recently implemented a Bull supercomputer to study and help predict tsunamis.

Numerous records broken

Alongside this commercial success, the breaking of a number of world records highlights Bull’s expertise in the design and integration of the most advanced technologies. Bull HPC technologies have achieved some major performance records, particularly for ultra-large file systems, image searches in very large-scale databases (the engines of future research), and the search for new prime numbers. Recently, these systems have also been used to carry out the most extensive simulation ever of the formation of the structures of the Universe.

Significant investments in R&D

Bull has put in place all the human and physical resources needed to create and support an offering that meets the high demands of centers of excellence in scientific and industrial computing.

In order to further develop and grow its HPC business, Bull has recruited new talents: over 100 engineers with expertise in HPC technology have joined Bull in the past two years, to boost the existing teams. This recruitment drive has covered every area, from HPC development and benchmarking, to marketing, sales, maintenance and management. These technical teams now make up Europe ’s largest group of experts in this sector amongst any of the companies involved in HPC.

Bull’s acquisition of Serviware in October 2007 demonstrates Bull’s desire to be involved not only in academic projects, but also in projects of all sizes in industry, the area in which Serviware specializes. This move will widen Bull’s commercial footprint, and broaden the range of expertise that the Group can now offer its customers and prospects.

Numerous partnerships: from platform design to applications

Bull designs its HPC platforms in collaboration with leading partners in their specialist areas; most notably, Intel for processors, Quadrics and Voltaire (Infiniband) for network infrastructures, Data Direct Networks in storage, and CFS for the Lustre file management system. In the field of applications for industry, Bull has formed partnerships with the main software publishers including Fluent/Ansys, CD-Adapco and ESI. Bull is also developing an operating system specifically designed for open HPC systems – based on the Red Hat Linux – whose qualities are recognized by the development community.

To develop the systems of the future, Bull is a founder or member of important consortia including TALOS (which brings together Intel, the CEA, Quadrics and Stuttgart University Computing Center ) and ParMA , which forms part of ITEA2 and brings together a large number of European research centers. Finally, Bull is a founder member of the POPS consortium – under the auspices of the SYSTEM@TIC competitiveness cluster based in ‘Ile de France’ region, which is developing tomorrow’s petaflop-scale systems.

An undisputed player in the IT industry in Europe

Europe has long lagged behind in HPC. This position seems be changing. Many research centers have equipped themselves with supercomputers over recent years, and are now ranked among the pacesetters in HPC. And Bull’s momentum is absolutely in line with this movement.

The CeBIT's announcement strengthens European high-performance computing, enabling its return to the forefront. Bull is ready to face the European challenge, and act as the key industrial component – whether alone or in partnership – in meeting that challenge. Bull is very proud of the fact that, a number of years ago, it identified and anticipated the coming evolutions in computer simulation technologies. It is also something that the French and European IT industry can be proud of: going from a virtually non-existent presence in HPC a few years ago, to the point where today it now ranks among the best in the world, alongside the very biggest players in the field.

Now, more than ever, Bull is ready to be the industrial center of this new movement; bringing its experience to bear and with the support of its industrial and technological partners.

SEND TO A FRIEND
 
Latest advance in artificial intelligence: computer wins a game against a Go Master

During the Go Tournament in Paris, staged between 22 and 24 March 2008 by the French Go Federation (FFG), the MoGo artificial intelligence (IA) engine developed by INRIA – the French National Institute for Research in Computer Science and Control – running on a Bull NovaScale supercomputer, won a 9x9 game of Go against professional 5 th DAN Catalin Taranu. This was the first ever officially sanctioned ‘non blitz’ victory of a ‘machine’ over a Go Master.

More complex than chess, more possible combinations than the number of particles in the Universe

In 1997, a computer beat the then World Chess Champion, Garry Kasparov, for the first time. But the game of Go remained the exclusive preserve of Man. More complex than chess, with 10600 different possibilities – more than the number of particles in the known Universe – the game of Go is a remarkable school for strategy. So this latest officially-endorsed performance by INRIA and Bull represents a real achievement.

The victory was won as part of the ‘IA-GO Challenge’, in a so-called 9x9 game (played over nine lines and nine columns). Although Catalin Taranu beat the computer in a 19x19 configuration with a nine-stone handicap, the Go Master nevertheless rated the IA system as ‘approaching Dan standard’ in a performance that promises some formidable battles to come between man and machine.

The software used in this victory – the result of a collaboration between INRIA, the CNRS1, LRI2 and CMAP3 – is based on innovative technologies that can be used in numerous different areas, particularly in the conservation of resources which is such a vital issue when it comes to tackling environmental problems,” explained Olivier Teytaud , Researcher at INRIA, MoGo team leader. “ We are particularly delighted that, during the tournament, we were able to welcome researchers working in these areas who are interested in this technology, and we would like to thank the FFG and the tournament organizers for having given us the opportunity to learn more about these advances stemming from fundamental research .”

“Over and above the intellectual and emotional challenge that is involved in designing a computer capable of facing up to the Grand Masters of Go, our collaboration with the MoGo team has helped us to make further advances in parallelization techniques. The users of our supercomputers will reap the benefits of this work in their own environments,” affirmed Eric Monchalin , Director of High-Performance Computing R&D at Bull.

To find out more
Monte-Carlo method and Go on computers
Machine learning and optimization

1 CNRS: the French National Center for Scientific Research
2 LRI: Computing Research Laboratory
3 CMAP: the Center for Applied Mathematics at the Ecole Polytechnique

About INRIA

The National Institute for Research in Computer Science and Control is the only French public institute entirely dedicated to research in information and communication science and technology (ICST). It has a staff of 3,700 - 2,900 of whom are scientists - working in eight research centres across France . INRIA has an annual budget of 162 million euros, excluding VAT, 20% of which comes from its own research contracts and licences. INRIA has huge influence in the following fields: "networks, telecoms and multimedia", "complex systems and software" and "modelling, simulation and visualisation". It develops cooperative projects with the business world through strategic industrial partnerships and by setting up open consortiums and companies (80 start-ups in 20 years) - particularly through its subsidiary INRIA-Transfert, promoter of four start-up funds.

For more information

SEND TO A FRIEND
 
Contact  |  Site map  |  Legal  |  Privacy