This week Germany’s Jülich Supercomputing Centre made what its calling “a crucial step towards Exascale” with the deployment of a Nvidia Tesla 20-series GPU cluster. The newly operational cluster is called JUDGE (Juelich Dedicated GPU Environment) and consists of 54 IBM iDataPlex nodes and supports a total of 108 NVIDIA Tesla M2050 GPU’s. Each compute node has 96 GB of main memory making JUDGE particularly well suited to CPU-intensive problems in the arenas of biology, medicine and environmental research.
Only with these technological advances can the peak performance of supercomputers, currently set at a few Petaflops, be increased to Exascale (1000 fold increase) by the end of this decade”, says Prof. Thomas Lippert, director of the Jülich Supercomputing Centre.
The cluster was deployed by the ParTec Cluster Competence Center GmbH using the company’s ClusterTools suite for installation configuration management, update management, and batch system configuration.