Hardware acceleration for AI (GPU/TPU) plays a particularly central role in the fields of artificial intelligence, Big Data and Smart Data, and digital transformation. This involves using special computer chips – known as GPUs (Graphics Processing Units) and TPUs (Tensor Processing Units) – to make AI applications much faster and more efficient than with conventional processors.
Normal computer processors (CPUs) often reach their limits when processing large amounts of data or training artificial intelligence. GPUs and TPUs are specialised for performing many computational tasks simultaneously. This enormously accelerates the training and utilisation of AI programmes.
A simple example: when recognising faces in photos, billions of pixels can be analysed. While a CPU might take hours to do this, a GPU or TPU can often achieve it in minutes or even seconds.
This makes hardware acceleration for AI (GPU/TPU) a crucial factor for rapidly implementing complex AI processes. For example, this is relevant for real-time analysis of large amounts of data, in medicine, or for intelligent voice assistants.
Deutsch 
English (UK)