For the processing of large geodata stocks, which require high computing power and storage capacity, we as a company increasingly use procedures from the field of High Performance Computing (HPC). HPC forms the basis for applications in the fields of AI, machine learning and Deep Learning. Such calculations can no longer be sensibly executed on simple workstation computers. Hardware optimizations and parallel computing approaches must be used. In addition to multi-core and distributed computing, we use graphic processor-supported calculation methods to solve difficult and complex tasks economically.
Bereits seit 2007 gilt CUDA als der De-Facto-Standard für die Grafikkartenprozessierung (GPU) und wird laufend weiterentwickelt. Mit CUDA können Deep-Learning-Algorithmen und die Nutzung Neuronaler Netze wesentlich beschleunigt werden.
High performance computing will become increasingly important. Numerous industries benefit from the enormous computing power of these approaches. Often companies, research institutions and public authorities do not need supercomputers in data centers. With economical hardware adaptations and parallel programming routines in Software development , numerous problems can be implemented efficiently on smaller computers.