Something I find quite interesting is how the transistors (and area) of the Megaprocessor splits among the different kind of activity. In particular we think of processors as being all about doing calculations but the ALU forms only a small part of this processor.
Looking at the transistors for the processor (leaving out the RAM) the counts are:

Activity	Transistor Count	Proportion (%)
Multiplexing	5174	34
Control Logic	4031	27
Adders, Logic, Shifters	2280	15
Cosmetic (module bus input display)	1904	13
Registers	1808	12

The biggest proportion of transistor usage, a third, is steering the data around . We have also about a third of the transistors dedicated to working out what we're supposed to be doing. Only about a sixth of the processor is actually doing productive work.

I don't quite know where to go with the observation, but I find it interesting....

One thing I certainly should have done (hindsight is great) is given the PC its own dedicated adder. It did have early on but the adder migrated into the ALU (it became ADDER2) so I could share it with other operations. This was a mistake. Adders are big, but so are the multiplexors. And there's also the yards and yards of ribbon cables I needed to move data from the PC to the ALU and then bring the answer back.