The Helmholtz-Zentrum Dresden-Rossendorf is a member of the Helmholtz Association of German Research Centres pursuing new insights that will allow us to maintain and improve all of our lives. That’s why the HZDR conducts research in the sectors health, energy, and matter in Dresden and at three other locations. Three of our five large-scale facilities are also available to external guests from around the world to help answer the decisive questions of our society.
Institute of Ion Beam Physics and Materials Research
The Institute of Ion Beam Physics and Materials Research conducts materials research for future applications, e.g., in information technology and for energy conversion. To this end we make use of the various possibilities offered by our Ion Beam Center for synthesis, modification, and analysis of thin films and nanostructures. The analyzed materials range from semiconductors and oxides to metals and magnetic materials. They are investigated with the goal to optimize their electronic, magnetic, optical as well as structural functionality. This research is embedded in the Helmholtz Association’s program “From Matter to Materials and Life”.
Scientific Projects with GPU Background
GPUs are used in two field of our current research.
The main activity is related to programming and application of Lattice Monte Carlo simulation on GPUs which has been performed in the group of Dr. Karl-Heinz Heinig.
The evolution of nanostructures in systems in thermal equilibrium as well in systems driven by ion irradiation is studied employing the atomistic 3D Kinetic Lattice Monte Carlo (KLMC) method. The KLMC method models the on-lattice kinetics of particles in solids on a stochastic basis, enabling the treatment of systems containing billions of particles. Compared to running on a single CPU, the available CUDA implementation speeds up such large scale simulations from months to days.
Another type of Lattice Monte Carlo simulations is used to model surface growth as described by the Kardar-Parisi-Zang stochastic differential equation. In cooperation with the group of Dr. Géza Ódor at the Research Centre for Natural Sciences, Institute for Technical Physics and Materials Science of the Hungarian Academy of Sciences surface growth was studied in 2+1 dimensions. Here GPGPU helped to produce a precise estimate of the growth exponent by enabling large scale simulations at roughly 250 times the speed of optimized (single) CPU code.
A second activity is related to the application of the well-known Molecular Dynamics code LAMMPS on GPUs. This code is used in the group of Dr. Matthias Posselt for several purposes ranging from investigations of basic processes of friction to studies of ion irradiation, defect formation, defect evolution, and ion beam mixing.