Supercomputing

Our services and methods – data analysis, study design, statistical text mining etc. – make high performance computing techniques indispensable. The accelerating development of biotechnology makes traditional approaches unable to cope with the growing data volumes from text mining, new generation sequencing etc.

Based on the above mentioned needs the GenaGrid Consortium purchased an SGI Altix ICE (http://www.genagrid.hu/altixice/) grid supercomputer which is one of the most powerful machines in Hungary.

• Running independent calculations parallel: primarily this is the case at parameter study calculations. The distribution of the tasks is done by our own job distribution system that was built on Condor (http://www.cs.wisc.edu/condor/).
• Code level parallelization: we use two approaches which perform differently in the following cases:
  • When communication between the blades is necessary (the threads do not have common memory space) we use Message Passing Library from MPI (http://www.open-mpi.org/)
  • In case of SMP (symmetric multiprocessing) applications (the parallel threads reach the same memory space) we use OpenMP (http://openmp.org/) because of its ease of usage and small communication overhead

With this Altix ICE supercomputer the biomedical and bioinformatics researchers achieved a high performance computing infrastructure that was not available in Hungary before. This not only allows us to accelerate our research projects, but opens new horizons as well.