Multiplexing Output Signals vs 3-State Busses
You can implement an output multiplexer by using three-state controls and
tying the pins together off-chip. This uses more pins but you do not need a
macrocell to implement the multiplexer as shown below.
You can implement a multiplexer in another macrocell and bring the data out
though a single pin. This saves pins but costs a macrocell to implement each bit
of the multiplexer as shown below.
You can register the output macrocell to shorten the clock-to-output delay as
shown below.