JuRoPA cluster at Jülich Supercomputing Center - 274,8TF @ 91,6% efficiency
During recent HPL (LinPack) benchmarking tests with ParaStationMPI on the JuRoPA machine, an impressive 274.8 Trillion floating point operations/second was recorded over a sustained 11 hour period. This was achieved using 3221 compute nodes each containing 8 cores and a Nehalem, dual socket, serverboard architecture. This LinPack performance constitutes 91.6% of peak making JuRoPA the global leader in parallel efficiency for supercomputer class clusters.
The name JuRoPA is itself an acronym which reflects the broader goals of the project of which this machine is just a part. JuRoPA stands for Jülich Research on Petaflop Architectures. The machine will be made available to over 200 research groups across Europe and will be used for data-intensive applications.
JuRoPA's architecture was developed by HPC experts from the Jülich Supercomputing Center. Partner companies Bull, Sun, Intel, Mellanox and ParTec were responsible for the realization of the machine which consists of 3288 compute nodes and a total computing power of 308 Teraflops peak.
During recent HPL (LinPack) benchmarking tests with ParaStationMPI on the JuRoPA machine, an impressive 274.8 Trillion floating point operations/second was recorded over a sustained 11 hour period. This was achieved using 3221 compute nodes each containing 8 cores and a Nehalem, dual socket, serverboard architecture. This LinPack performance constitutes 91.6% of peak ? making JuRoPA the global leader in parallel efficiency for supercomputer class clusters.
The JuRoPA cluster is an aggregate of two smaller clusters, known individually as JuRoPA-JSC and HPC-FF. Both machines share a common interconnect fabric (QDR Infiniband) in addition to a common management network. ParaStationV5, ParTec's current release of it's cluster operating system, enabled both clusters to be integrated to form a single heterogeneous cluster entity capable of solving the most challenging problems facing researchers today.
The JuRoPA-JSC machine is the larger of the two components forming the JuRoPA cluster. This machine, a supercomputer class cluster of its own right, is based on the 6048 Blade system from Sun Microsystems. JuRoPA-JSC alone consists of 2208 compute nodes with a total computing power of 207 Teraflops peak.
The HPC-FF machine (High Performance Computing for Fusion) was the result of a cross-vendor collaboration between Bull, Intel, Mellanox and ParTec. HPC-FF consists of 1080 compute nodes with a total computing power of 101 Teraflops peak.
The performance of the JuRoPA cluster ranks it among the top 10 of the world's most powerful supercomputers. This is an impressive achievement for a true general purpose machine with a modest footprint and power consumption.
ParTec's ParaStationV5 cluster operating system combined with ParaStationMPI delivered an integrated, easy to use and reliable compute cluster environment which made a significant contribution to the success of this project. ParaStationMPI delivered proven scalability running more than 25,000 MPI tasks with parallel efficiencies in excess of 91,6%. The Linux operating system of the whole system is Suse Linux Enterprise Server (SLES11).
The JuRoPA cluster is seen as a model for the next generation of general purpose supercomputers designed using open standards and commodity hardware on an Intel platform.
Complete JuRoPA System (HPC-FF & JUROPA-JSC combined):
- 3288 compute nodes (26304 cores)
- 308 Teraflops peak performance
- QDR Infiniband
- ParaStationV5 Operating and Management
- ParaStationMPI
- GridMonitor system monitoring
- Novell SLES11 Linux
- Intel Compiler and Tools
- MOAB Scheduler
- Lustre Filesystem