An octree is a space-adaptive data structure. This thesis shall demonstrate the integration
of a new octree-based particle container for the AutoPas simulation library. The new particle
container is intended to serve as an additional option to let AutoPas’s special tuner unit
choose from in order to find a better, truly optimal simulation configuration. Detailed insight
regarding the implementation of the octree container is given as well. Furthermore, the
performance of the new container is compared to an existing linked cells implementation.
It could be shown that the octree container is working as intended, but it falls behind the
existing, optimized containers by a factor of 3 in terms of speed. Tools that were developed
during the implementation are presented as well. The outlook provides concrete ideas and
inspiration on how the existing implementation can be improved and further analysis can be
conducted.