Our PhD student Christie Louis Alappat just took first place in the ACM Student Research Competition at SC18. His work revolved around the “Recursive Algebraic Coloring Engine,” which is a new method and library for hardware-efficient graph coloring. This means that he will advance to the grand finale next year. Congratulations!
Christie’s project is part of the activities in ESSEX-II, a project funded by the German Science Foundation (DFG) within the priority programme SPPEXA.
As part of the competition, Christie has prepared a video with details about his work:
The HPC group at RRZE and the professorship for High Performance Computing now have a joint homepage: https://hpc.fau.de. This is where all our teaching and research activities can be found. We also provide some basic system descriptions, but the main part of the RRZE clusters’ documentation is still (and will be, for the foreseeable future) at https://www.anleitungen.rrze.fau.de/hpc.
Intel Threading Building Blocks (TBB 2.0) on RRZE systems
On Woody and Altix the current version of Intel’s TBB library is now installed. TBB is a C++ framework for shared-memory parallel programming. If you are fed up with marrying C++ and OpenMP, TBB might be the right choice. Use it by loading the
tbb/2.0-debug modules, respectively. On Cluster32 you can access the modules from Woody by typing, e.g.,
module use /apps/modules/modulefiles/libraries, but remember that we have only provided a 64 bit version. The library was compiled with the Intel compiler and we don’t guarantee that it will work with anything else. Documentation can be found on the TBB website:
We are very much interested in sharing experiences with all users of the library. So please if you have comments or questions, contact hpc@rrze.
As of this week we are proud owners of an UltraSPARC T2 system (1 socket @ 8 cores, 32 GB, 1.2 GHz). This one is a little slower than the 1.4 GHz machine we had access to at RWTH Aachen, but the characteristics are quite the same. After some intensive benchmarking, this machine will be made available in our test cluster for all HPC users to evaluate.