Numerical processes in science often require very high computing power. This computing power is required to perform calculations in large numbers independently of each other (capacity computing) to use a large number of processors in parallel, or very large main memory (capability computing or to combine both methods). The HPC service provides users with computing capacity and basic software environment to perform such calculations. It should be noted that the HPC environment is different from a normal desktop system, in particular, there is no graphical interface. Shareholder will have the opportunity to process thousands of computing jobs per month and use several hundred computing cores in parallel. For the latter, program code is necessary, which was written for parallel processing and on appropriate programming access libraries (e.g. MPI, OpenMP).
The Scientific compute clusters service consists of two computing clusters which are located in Zurich (Leonhard) and Lugano (Euler). In total around 45 000 processor cores are available for scientific calculations. The two clusters have high-speed networks with storage and parallel file systems. Calculations (jobs) must be started using the login nodes from the batch system (LSF). Frequently used scientific software packages and compilers are centrally installed.