Funded by


Subproject D.2

Flexible Many Particle Simulations on Different High Performance Computing Architectures

What challenges high performance computers have to fullfil for calculating many-particle simulations?

Simulation of molecular dynamics and particle systems allow researchers to visualize the invisibles. For realistic simulations however, an extremely high number of particles must be implemented using complex models of interaction, which pose high requirements on the computer and most often may only be accomplished using HPC Systems.

The goal of this research project is to implement new methods for HPC systems, such as GPGPU accelerators and many-core systems of the newest type available. Early implementations of methods are based on parallelization techniques using CUDA, CAPS HMPP respectively OpenACC.


Further information about this subproject


Simulation of a billard table with 200 billiard balls.