Institute Cluster II


The SCC operated a computer system called InstitutsCluster II, which was funded by the DFG (http://www.dfg.de) and jointly procured with several KIT institutes.
It was available to the institutes participating in the funding proposal in accordance with their shares specified therein.
The system was expanded to the configuration shown below thanks to additional own funds from other institutes. These institutes were able to use the cluster to the extent of their financial participation.

Photo: Holger Obermaier

Photo: Holger Obermaier




Configuration of the Institute Cluster II (IC2)

The InstitutsCluster II contained

  • 2 login nodes, each with 16 cores with a theoretical peak performance of 332.8 GFLOPS and 64 GB main memory per node,
  • 480 "thin" computing nodes, each with 16 cores with a theoretical peak performance of 332.8 GFLOPS and 64 GB of main memory per node,
  • 5 "fat" nodes, each with 32 cores with a theoretical peak performance of 340.4 GFLOPS and 512 GB of main memory per node
  • and an InfiniBand 4X QDR interconnect network.





The Institute Cluster II was a massively parallel parallel computer with a total of 487 nodes (excluding service nodes). 482 nodes and the 10 service nodes have a clock frequency of 2.6 GHz. 5 so-called fat nodes had a clock frequency of 2.67 GHz. All nodes had local memory, local disks and network adapters. A single computing node had a theoretical peak performance of 332.8 or 340.5 GFLOPS, resulting in a theoretical peak performance of approx. 162 TFLOPS for the entire system. The main memory across all computing nodes amounted to 32.6 TB. All nodes were interconnected via an InfiniBand QDR interconnect.

The base operating system on each node was a Suse Linux Enterprise (SLES) 11. KITE served as the management software for the cluster; KITE is an open environment for the operation of heterogeneous computing clusters.

The scalable, parallel file system Lustre was connected via an InfiniBand network as the global file system. Lustre achieves both high scalability and redundancy in the event of individual server failures. 469 TB of disk space was available, which other high-performance computers continue to access. In addition, each node of the cluster was equipped with at least 2 local disks for temporary data.

Detailed brief description of the nodes and the connection network:
2 16-way (login) nodes, each with 2 octa-core Intel Xeon E5-2670 processors with a clock frequency of 2.6 GHz, 64 GB of main memory and 5x1 TB of local disk space; 480 16-way (compute) nodes, each with 2 octa-core Intel Xeon E5-2670 processors (Sandy Bridge) with a clock frequency of 2.6 GHz, 64 GB of main memoryand 2x1 TB of local disk space, 5 32-way (compute) nodes, each with 4 octa-core Intel Xeon E7-8837 processors (Westmere) with a clock frequency of 2.67 GHz, 512 GB of main memory and 4x1 TB of local disk space and 10 16-way service nodes, each with 2 octa-core Intel Xeon E5-2670 processors with a clock frequency of 2.6 GHz and 64 GB of main memory. A single octa-core processor (Sandy Bridge) has 20 MB L3 cache and operates the system bus at 1600 MHz, whereby each individual core of the Sandy Bridge processor has 64 KB L1 cache and 256 KB L2 cache.
42 InfiniBand 4x QDR switches with 36 ports each served as the connection network.