In the case where multiple kernels fit in a single AI Engine-ML, communication between two or more consecutive kernels can be established. Use a common buffer in the local data memory of the AI Engine-ML. Alternatively, in any of the three neighbor memories to which the AI Engine-ML has direct access. This case requires only a single buffer because the kernels execute one after another in a round-robin.
For cases where the kernels are in separate but neighboring AI Engine-MLs, the communication can be carried out through the data memory module shared among the two neighboring AI Engine-ML tile that use ping-pong buffers. These buffers can be on separate memory banks so access conflicts are avoided. Locks provide synchronization. Buffer locks ensure that the input and output buffers for the AI Engine kernel are ready. This type of communication save routing resources. It also eliminates data transfer latency because it does not require DMA and AXI4-Stream interconnect.