The following figure shows an example system block diagrams on the Alveo U200 card which is inserted into the PCIe® slot of the host server. Two DPUCADF8H instances at most can be implemented on the two SLRs and one or two SLRs can be reserved for customers to implement customized functions. The system can be used to perform deep learning inference tasks such as image classification, object detection, and semantic segmentation.
Figure 1. Example System with Integrated DPU