Megaprocessor

It was wonderful to spend a couple of hours with James Newman and his megaprocessor.org today at the Computing Museum.

I learned a huge amount of valuable information from James and Chris Monk for which I'm very grateful.  I'll try to summarise here:

  • Lots of people should see the megaprocessor (or the nanoCPU).  Lots means at least a good fraction of the GCSE-level computer science students, but it's also important for all ages and backgrounds.
  • To get a good idea of what schools need I can look at the GCSE curriculum.  I can also go to teaching conferences and exhibit to get feedback.
  • I can pester the staff at the Computing Museum (and at Bletchley) for feedback on educational aspects before finalising the design.
  • The megaprocessor is big, really big.   It's a great educational tool, but it needs to be seen and there's only one megaprocessor.   To get to the widest audience then nanoCPU needs to be small, easy to build and transportable.   Ideally others will copy and improve and there will be lots around the world.
  • Custom PCB is expensive, but wiring on veroboard may be near impossible so it may be a necessity.   If 100 were made then the costs would come down.
  • Showing the signal routing is important - another reason to keep it simple.
  • Multi line digital scopes are expensive but speeds up the debugging massively.
  • Spice is only good for a dozen components, it won't scale to a whole CPU.  I need to use Verilog or VHDL to wire all the logic together.
  • The FPGA speeds up the debugging massively.
  • Even if there isn't an LED on every transistor, some bits should be explained in detail.   For example, one of the memory cells and one bit slice of the ALU.
  • 3mm LEDs at 10mA are quite bight enough.   In fact, it's hard to photograph something with lots of LEDs on as the LED light dominates the reflected light from everything else.