Description

AI workloads must efficiently distribute CPU-bound tasks (like video decoding or text tokenization) and GPU-bound tasks (like deep learning model training).