HLRS Hornet to deliver four petaflops peak performance
The supercomputing infrastructure of GCS centre HLRS (High Performance Computing Center Stuttgart) at the University of Stuttgart will soon enter the next stage of its HPC systems installation phase. Code-named Hornet, the future HLRS supercomputer will gradually be installed at the Stuttgart HPC facility, so that the vast user community of HLRS has a smooth transition from current supercomputer Hermit to Cray’s next generation high-end HPC system. In its final configuration, Hornet will deliver a peak performance of four petaflops, outperforming Hermit’s maximum performance by a rough factor of four.
The installation of Hornet, a Cray XC30, is being carried out according to the earlier agreed HPC systems roadmap of HLRS, which defined Hermit as the initial installation step. With Hornet’s implementation to be gradually conducted over the coming 18 months, the next installation phase will be completed by the second half of 2014. The new supercomputer will provide 500TB of main memory and about 6PB of disc space. It will be equipped with 100,000 computing cores and will feature Intel’s next generation of microprocessors which, according to the manufacturer, are specifically designed to optimise power savings and promise significant performance enhancements.
‘A large percentage of our user community comes from the field of scientific engineering where highly memory demanding applications are typical,’ explained Professor Resch of HLRS. ‘Especially in the automotive and aerospace research and industries, HPC users are depending on systems that feature leading-edge supercomputing technology and at the same time high sustained performance. HPC systems are an indispensable tool in the researchers and scientists’ pursuit to achieve breakthrough discoveries and innovations, and they ultimately help to strengthen the world-renown reputation of German Engineering as a synonym for innovation and quality of the highest degree.’
As with Hermit, the system expansion at HLRS is funded through project PetaGCS with support of the Federal Ministry of Education and Research and the Ministry of Higher Education, Research and Arts Baden-Württemberg.