Designing Hardware

28 Feb 2010

Dave and I both did the NAND-to-Tetris course a while back. Actually, I believe Dave did the whole course, but I stopped around the “compiler” assignment because it was too close to what I was doing at work at the time to be entertaining.

The course was geared towards very beginner CS students which is great. It gives a nice vertical slice of all the bits and pieces that go into getting your average piece of code to run, and did it all in a way that was simple to follow along with. The assignments and test cases given were excellent in the way they walked you down the garden path very clearly, but you still felt like you were “figuring it out” yourself.

I work firmly in the realm of software, but that course was enough to catch a bit of the hardware bug and have me wanting to venture below machine code.

I found a few nifty/slightly crazy attempts to recreate pre-Z80-era micros (like Magic-1 and Mark 1 FORTH). Those two are particularly crazy in that they don’t use any microprocessor. The largest IC they use is a 4-bit adder, and everything is built up from TTL 74xx series chips (basically just and/or/flip-flops/etc.). I was all amped up to try to build something similar, but looking at some wire wrap pictures and then watching Bill Buzbee’s bring-up attempts of his CPU pretty much scared me off of that idea. They look extremely cool, and I’m duly impressed, but that’s a whole lot of tedium I’m not quite prepared for yet.

So, I started looking into FPGAs. Doing a CPU on FPGA isn’t nearly as cool as having your own wirewrapped CPU. Mostly because it’ll just look like some generic standard board, so when the victorious moment arrives when it finally prints “Hello, World” or the answer to fib(7) over a serial cable, onlookers may not be as impressed as they ought to be.

However, it does turn hardware into a (relatively) non-tedious/not-too-expensive project, so I think I’ll go that way. My other excuses are that I don’t have a serial port on my computer, and TTL chips are hard to find nowadays. This FPGA beginner dev board has a bunch of nifty connectors, and isn’t too expensive. They have others that are cheaper too, with less connectivity.

Next time, my take on trying to come up with an interesting processor that isn’t just a generic RISC clone.