Data processing engine (DPE) array routing

A method for compiling, the method comprising:

• by a processor-based system:
• modeling a communication network among switches interconnected in an array of data processing engines (DPEs), each of the DPEs including a hardened processor core, a memory module, and one or more of the switches, each of the switches comprising an input port or output port that is capable of being shared by multiple routes, one or more ports of each of the switches being modeled as one or more respective nodes;
• generating routes for an application on the modeled communication network, wherein generating the routes includes using an A* algorithm that includes a congestion costing function based on a capacity of respective nodes in the modeled communication network and a cumulative demand for the respective nodes;
• and
• translating the routes to a file.

A design system comprising:

• a processor;
• and
• a memory coupled to the processor, the memory storing instruction code, the processor being configured to execute the instruction code to:
• model a communication network among switches interconnected in an array of data processing engines (DPEs), each of the DPEs including a hardened processor core, a memory module, and one or more of the switches, each of the switches comprising an input port or output port that is capable of being shared by multiple routes, one or more ports of each of the switches being modeled as one or more respective nodes;
• generate routes for an application on the modeled communication network, each of the routes corresponding to a net of the application, wherein the processor being configured to execute the instruction code to generate the routes is further configured to:
• generate routes using an algorithm that is based on a capacity of respective nodes in the modeled communication network, a cumulative demand for the respective nodes, and a history of over-utilization of the respective nodes;
• and
• if a node of the routes has a cumulative demand that exceeds a capacity of the respective node:
• iteratively process the algorithm until no node has a cumulative demand that exceeds capacity of the respective node:
• increase the history of over-utilization of each node that has a cumulative demand that exceeds a capacity of the respective node, nets that correspond to a respective route that includes any node for which the history of over-utilization is increased forming a group of nets;
• and
• iteratively process the algorithm until no node has a cumulative demand that exceeds capacity of the respective node or until the nets of the group of nets have been rerouted:
• remove a route corresponding to a net of the group of nets;
• and
• generate a route for the net of the group of nets using the algorithm;
• and
• translate the routes to a file.