Supercomputers at GCS

The Gauss Centre can boast of a state-of-the-art high-performance computing and networking infrastructure:
  • the JSC will provide a 5 PetaFlop/s IBM Blue Gene/Q JUQUEEN by November 2012,
  • LRZ offers its recently installed IBM IDataplex System SuperMUC with more than 3 PetaFlop/s peak performance, and
  • HLRS offers its recently installed CRAY XE6 system with more than 1PetaFlop/s peak performance.
The architectures of these machines are different yet complementary. Each one favours special types of applications. The SuperMUC (and its migration system) can be used as a shared-memory machine. The Cray XE6 Hermit at HLRS with its exceptional high-speed network is in particular suitable for scalable applications with demanding communication needs. The Blue Gene is adequate for applications which scale to extreme processor numbers.



FZJ Logo Jülich Supercomputing Centre

IBM Blue Gene/Q

Technical data of the IBM Blue Gene/Q JUQUEEN:

Peak Performance 1.6 PFlops
Nodes 8192
Cores 131072
Memory 131 TB

 Computers at JSC




LRZ Logo   Leibniz Supercomputing Centre in Garching

SuperMUC

Technical data of the SuperMUC (under construction):

Peak Performance approx. 3 PFlops
Cores > 100,000
Memory > 300 TB
Storage capacity > 10 PB
Fat Node Islands 1
Thin Node Islands 18
Nodes > 9,000


 Computers at LRZ



HLRS Logo  High Performance Computing Center Stuttgart

NEC SX-8
Technical data of the CRAY XE6 Hermit:

Peak Performance 1.045 PFlops
Number of compute nodes 3552
Number of compute cores per node 2 sockets with 16 cores each: 113 664 in total
Processor compute nodes Dual Socket AMD Interlagos @ 2.3GHz 16 cores each
Memory/node 32 GB and 64 GB
Disk capacity 2.7 PB
Node-node interconnect CRAY Gemini (low-latency high-speed network)

 Computers at HLRS