Sadly, I don't have enough motivation to make a full documentation for the circuit, I'll still try to explain a bit.
there are 2160 8 bits registers put in series (so it needs 2160 clock cycles to program), each register will configure a part:
- the PLU is the part making logic algebra (combinational and sequencial) so the registers tells what operation to perform
- the PRU is a simple routing part: for each output bit on each side, you have to choose one of the 3 other side (you stay in the same byte and the same bit though)
- the PIRU is an powerfull interstage routing unit, each can output follow any input (that's why it has the biggest amount of register)
