NVidia’s 16-nm FinFET-based Tesla P100 GPU targets deep learning applications, incorporating stacked, High Bandwidth Memory 2 (HBM2) and Chip-on-Wafer-on-Substrate (CoWoS) technology.