I realized I have yet another problem - I can't do a RAM->RAM transfer. At least not in a single clock cycle. I can only address one location in memory at a time. I suppose I can do a RAM read in one instruction, and then a RAM write in a second instruction. But I'm not sure what to do if the software tries to perform a RAM->RAM transfer. I could try to set it up so that the destination address takes priority. So it will write back the same data that's in there. At least I think that's what it would do. Maybe at some point I could have exceptions, but my chip count is already getting uncomfortably high. I do have to build this thing at some point.