HPC (supercomputer)
Rőczei Gábor
<>
NIIF Intézet
Hornos Tamás dr.
<>
NIIF Intézet
One of the most important area of state-of-the-art scientific work is scientific computing. In scientific computing solve mathematical models of physical problems by computer simulation. Simulations are a cheaper and more flexible approach than physical experiments. In science, this approach is called "in silico" research. Most simulations are resource intensive and demand a supercomputer or High Performance Computing (HPC) environment.
The main topics of the HPC tutorial are the following:
- HPC in Hungary
- PRACE : The European HPC initiative
- The SMP Supercomputer at Pécs (guided tour during the break)
- Supercomputer clusters: Debrecen, Szeged, Budapest
- Job schedulers: Sun Grid Engine, SLURM
- Parallel programming paradigms: MPI, OpenMP
- Remote visualization: VirtualGL, TurboVNC
- Grid with HPC: ARC middleware
- Scalasca parallel profiler
- Compiler optimizations: gcc, Intel compilers, mathematical libraries
- GPU resources