The handling of wires (communication mechanisms) between attached processors has been generalized and greatly simplified.
You no longer declare wires between attached processors; the configurer invents them for you, automatically and invisibly. The only wires that you should declare are those that use explicit cables (or carrier-board equivalents).
Wire statements between attached processors in existing applications must be removed before those applications can be rebuilt with new versions of Diamond.
The configurer creates new bitstreams for your FPGAs when necessary to implement the connections you have used.
The root processor is treated as a special case; comport 3, CP:3, is created for you as it is used to boot the system from the host. This is not done for stand-alone applications.
One immediate benefit of this new way of doing things is that the internal connections between attached processors are now flexible and can be changed easily. With previous versions of Diamond it would have been necessary to create dummy FPGA tasks if you wanted to create new routing between the DSPs on a multiprocessor module such as the SMT374. Now you simply use the links you need and everything else is done for you.
You can find more information about the new way in which WIRE statements are processed under WIRE statement.