In an embedded design, the primary purpose of the datapath is to capture the dataflow of interest in the system application. Vision processing systems can include a dataflow where image sensors, such as cameras and LiDAR, send incoming data. This data is stored in on-chip and off-chip system memory, like BRAM, UltraRAM, and DRAM, for inline processing. For networking systems, the dataflow can be a protocol bridge between two standards like Ethernet and Interlaken.
The system maps the PL accelerators datapath (ingress and egress) to AXI4-Stream interfaces. It also maps the control interfaces to memory-mapped AXI interconnect. The software stack controls these accelerators through the control interface.
Similarly, the system maps theAI Engine accelerators datapath to the AXI4-Stream interfaces. The AI Engine kernels have a run-time parameter (RTP) interface that enables run-time updates. You can access the RTP feature through the flexible runtime (XRT) APIs. The AI Engine also supports global memory access to DDR memory for application-specific buffers.