When the static region accesses an NSU in the dynamic region, it is essential for ownership to remain within the static region. In this NoC topology, NMUs are programmed with the destination IDs of the NSUs that the NMUs need to communicate with. Therefore, if an NMU is positioned within the static region, reprogramming the NMU during dynamic reconfiguration is not allowed. This means that you must determine the maximum number of NSUs within the dynamic region that the static NMU needs to communicate with, and you must make connections to the static NMU during the initial implementation.
The following figure shows an example of an NMU in the static region that needs to access three NSUs in the dynamic region, each with different bandwidth requirements. However, not all reconfigurable modules (RMs) require the use of all three NSUs. Because the path is owned by the static region, all three networks (Network-1a, Network-1b, and Network-1c) are established and configured with their corresponding Quality of Service (QoS) requirements during the initial implementation. These configurations are preserved and followed in all child implementations.
In each child implementation, you can choose which of the three NSUs (NSU1a, NSU1b, or NSU1c) to use based on the specific QoS requirements of the corresponding RM. You can opt to use all or a subset of NSUs based on the RM requirements. In this example, RM2 variant (child implementation-1) uses NSU1a and leaves NSU1b and NSU1c unused. Conversely, RM3 variant (child implementation-2) uses NSU1c and keeps NSU1a and NSU1b unused.