NeMeSys is a NUMA-aware graph pattern processing engine, which uses the Near Memory Processing paradigm to allow for high scalability. With modern server systems incorporating an increasing amount of main memory, we can store graphs and compute analytical graph algorithms like graph pattern matching completely in-memory. Our system blends state-of-the-art approaches from the transactional database world together with graph processing principles. We demonstrate, that graph pattern processing - standalone and workloads - can be controlled by leveraging different partitioning strategies, applying Bloom filter-based messaging optimization and, given performance constraints, can save energy by applying frequency scaling of CPU cores.