Please note that JavaScript and style sheet are used in this website,
Due to unadaptability of the style sheet with the browser used in your computer, pages may not look as original.
Even in such a case, however, the contents can be used safely.
Data intensive research and development always go hand in hand with high performance computing (HPC) clusters. NEC Deutschland GmbH delivers made to measure solutions for scientific research and industrial development such as those needed in the automotive sector. Now, the successful NEC LX Series cluster solution is available with a range of new innovations centering on improved performance, scalability and availability combined with reduced power consumption and size. With the new NEC HPC 144Rc-1 server, the computing nodes consist of very high performance components because two motherboards are now integrated into a single rack unit (1U). This results in a theoretical maximum performance of 29 teraflops for a top end configuration. Integration of the new fault tolerant server NEC Express5800 FT simultaneously provides for much improved availability. These state-of-the-art technologies offers unbeatable performance and data security.
The NEC LX Series' chain of successes is founded on NEC's Stuttgart Competence Center which for many years has not only been working closely with vendors such as Intel and AMD, but has also demonstrated intense commitment in the framework of European and worldwide research projects for further developing software and middleware for Linux clusters. NEC is an active member of the Open Source Cluster Applications Resource (OSCAR), for example, where it shapes the freely accessible administration software together with other industrial and research partners.
For its computing nodes, NEC is now offering the new NEC HPC 144Rc-1 server with 16 single cores in just one 1U casing. The heart of the server comprises two motherboards, each with 2 Harpertown Series (Quad Core Intel Xeon Series 5400) processors with a 1600 MHz FSB. This is the first Xeon processor using the latest 45 nm Penryn architecture. The top end specification provides for configuration of 2 x 32 GB RAM. ScaleMP technology allows these 64 GB to be used in parallel as shared memory by each of the 16 single cores, resulting in an extremely high performance SMP server with a theoretical top performance of over 200 gigaflops.
Another innovation is the integration of a fault tolerant NEC server into the HPC cluster to serve as the master node. The server's operating availability benchmark exceeds 99.999%. With its latest fault tolerant servers, NEC has effectively left the principle of "high availability" behind it to herald in the era of "continual availability". In practice, the extremely high availability of these servers means users can expect unscheduled downtime of less than 5 minutes a year, including repair, upgrade and general maintenance times. With each component duplicated and Lockstep technology providing for continuous synchronization, the system literally runs nonstop. Damaged components can be exchanged as hot plug-ins. With this, NEC is the only vendor to offer a fault tolerant server for the master node, ensuring the greatest possible availability.
In addition to mainstream Linux distributions, the LX Series is now also configured with Microsoft HPC Server 2008 at many customers' request. NEC sees benefits here in the area of application servers e.g. for applications in the automotive industry such as LS Dyna, for which NEC offers fully configured clusters together with its partners. As a highly scalable solution, this is an interesting alternative to the Linux based systems when simple integration into Windows based work processes is involved.
But it is not only the computing nodes, switches and software that are decisive for high performance, reliable cluster computing, but also a suitable, parallel file system. Here, NEC offers its D Series storage product that serves as the basis for a Lustre technology-based file system. The LX Series also alternatively integrates other vendors' products - such as Isilon or Panasas - especially for customers placing tough demands on simple handling. NEC additionally offers application migration and optimization, as well as its NEC OSCAR PRO management software, for third party products. Professional service and support from NEC Deutschland GmbH, as a fixed element of each cluster solution, round off the portfolio.
With NEC staff participating in the framework of European and worldwide research projects for further developing software and middleware for Linux clusters, it remains optimally positioned for the future as well. NEC is an active member of the Open Source Cluster Applications Resource (OSCAR), for example, where it shapes the freely accessible administration software together with other industrial and research partners. With its product NEC OSCAR Pro, NEC is able to contribute is experience from its collaboration with the OSCAR community. A further big advantage is NEC OSCAR Pro's recent certification as Intel Cluster Ready, allowing it to be offered in connection with other vendors' HPC clusters.
A PDF file is opened in another window.Express 5800/120Rj-2 (239 KB)
A PDF file is opened in another window.Express 5800/140Rf-4 (225 KB)
A PDF file is opened in another window.HPC144 RC-1 Server (114 KB)
A PDF file is opened in another window.Intel Cluster Ready Certificate (40.5 KB)
