Application of small massively parallel systems to deep learning“? The systems to be evaluated should be the Tegra X1 system (or similar systems) and it should be investigated, if high level software development in the context of deep learning/machine learning and based on, e.g., Python, is possible.