Earlier sections review the baseline MATLAB model of the SAR BP algorithm, identify the specific compute workloads required to perform the algorithm, formulate a specific set of AI Engine kernels to tackle these workloads, and consider various algorithm adaptations that yield attractive AI Engine implementations of those kernels. Early prototyping work validates the performance of these algorithm variants in the context of the MATLAB system model using Vitis Functional Simulation. These results may now be summarized in the functional block diagram of an AI Engine based computation engine for SAR BP shown below. The diagram captures ten different functional blocks and identifies the required data flow between them. Configuration input data is shown in red text and the I/O data path is shown in gold text. These are identified from the baseline MATLAB model. The legend indicates an early identification of implementations of these blocks based on the analysis above including library blocks, customized library blocks or fully custom blocks. The next step is to conduct system partitioning to identify AI Engine solutions for all blocks, define the data flow between them, and quantify the throughput performance and resource estimates for the design proposal.