Atos installs its second Bull supercomputer at the German Federal Waterways Engineering and Research Institute (BAW)


September 9, 2015

The first Bull supercomputer has been successfully running simulations for BAW since 2012

Bull, the Atos brand for technology products and software, will provide a second High Performance Computing (HPC) cluster to the German Waterways Engineering and Research Institute (BAW). Following the installation of the first computer in 2012, the Federal Agency selected aBull supercomputer for one of their HPC clusters dedicated to complex simulations. The new supercomputer will be operational in October and under the new contract Bull will also provide maintenance services for five years.

BAW supports the Federal Administration responsible for managaging navigable waterways and navigation (WSV) by enabling them to ensure that all its equipment and facilities are compliant with legal and safety requirements. BAW uses high-performance computers to perform complex calculations, based on two and three dimensional numerical models, so as to simulate water flows and transport processes in rivers and canals, as well as flows around engineering works.

”Our existing Bull HPC cluster has proved its worth in the last three years. Its flexibility, its availability and its compute performance, as well as the support provided by Bull, have convinced us. During the tender process for our new cluster, Atos again came up with the most compelling proposal that included a complete solution based on Bull technology, s+c software and support services at an outstanding price/performance ratio,” said Prof. Dr.-Ing. Christoph Heinzelmann, director of BAW.

”We are delighted that BAW chose a complete high performance computing solution from Atos for their simulations at their Karlsruhe location”said Winfried Holz, CEO of Atos in Germany. ”Our close and longstanding collaboration underpins our expertise to ensure we provide research institutions with the tailor-made hardware they need.”

The new bullx B520 cluster is composed of five racks, which contain a total of 250 compute nodes, equipped with 5,000 processor cores. The cluster has a peak performance of 208 Tflop/s (208 Tflop/s = 208 * 1012 floating point operations per second). This is equivalent to approximately the capacity of 2,000 current desktop PCs.

Users will have access to 600 terabytes of data storage over a Lustre parallel file system. Like the existing Bull cluster, the new supercomputer will be managed through the scVENUS HPC management tool developed by the Atos affiliate – science + computing ag – , a tool which automates software installation and distribution, job scheduling and configuration management to considerably simplify cluster administration. The cluster also operates at an excellent energy efficiency level. The heat generated by the water-cooled racks will be re-used to heat the BAW buildings, as is already the case with the Bull cluster installed in 2012.

 

Agenda