HLRS

Home > Calls > Partnering > HLRS

The High Performance Computing Centre (HLRS) is a research and service institution affiliated to the University of Stuttgart (USTUTT). It has been the first and is now one of the three national supercomputing centres, which are also members of the GAUSS centre for supercomputing in Germany. HLRS offers its HPC resources to users from academia and is one of the few HPC centres offering those services to industry as well. Besides access to computing resources, HLRS also provides consultancy services and training for industry and academia to develop large-scale applications and to adapt existing algorithms for performing the experiments.

Over the last 20 years HLRS has built up world-leading expertise in supporting end-users, from the engineering domain at large, in parallel programming (for example inside the Teraflop Workbench). It has worked on the development of productivity tools for parallel programming, including integrated development environments. Finally, HLRS has significant expertise in applying software engineering methods to the HPC domain.

HLRS participates in a number of European collaborative supercomputing initiatives, such as Partnership for Advanced Supercomputing in Europe (PRACE), and Distributed European Infrastructure for Supercomputing Applications (DEISA) and their forthcoming successors as well as other infrastructure projects such as DORII, BonFIRE and HPC-Europa. Additionally it is involved in more than a dozen projects investigating the challenges of future hardware architectures in order to tackle the Exascale challenge.

The available supercomputing facilities for Fortissimo and Fortissimo 2 are the following:

The CRAY XC40 system, Hazel Hen, equipped with 185.088 Intel Haswell @ 2.5 GHz processing cores and 128 GB memory per node, offers nearly 8 PFLOPS of peak performance to its users. The system is currently the 9th fastest system all over the world and the fastest HPC system in Europe.

The NEC Clusters, LAKI and LAKI2, are two independent, heterogeneous computing clusters installed in two different buildings. The 751 computing nodes are equipped with different processor types, different memory sizes (up to 1 TB RAM) and accelerators in order to support all kinds of applications and programming models.

Both systems are integrated with the existing High Performance Storage System (HPSS), which has been developed to manage Petabytes of data stored on disks and in tape libraries. As a consequence, HPSS at HLRS manages 512 TB of disk cache storage and more than 10 Petabytes tape storage. Both HPC systems make use of the Batch system torque/moab and support various development tools, libraries, and applications, for instance:

  • Development tools: make, CMAKE, Git, SVN, Eclipse, …
  • MPI libraries: Open MPI, MVAPICH2, MPICH2, …
  • Compilers: GNU, Intel, PGI …
  • Libraries: BLAS, LAPACK, FFTW, PETSc, MPT...
  • CFD (Fire, Fluent, OpenFOAM), CSM (ABAQUS), Computational Chemistry (GAMESS-US, Gaussian), amongst others

For further information, especially for the system configurations and environments, please visit www.hlrs.de or contact Michael Gienger directly.