mspangler wrote:
OK, I'm baffled by this whole thread. How can a collection of NANDgates with blow-upable inputs act as a general purpose CPU?
Well, it ain't '70s no more... We are not talking "blow-upable inputs".FPGAs are infinitely programmable. Meaning that you can load new hardware functionality as many times as you can run a different program on your PC.
As a logic
device replacer (MMU/IOU/74ls154 ect) it makes sense. But given you don't know what instructions are coming in what sequence, and there is microcode in the CPU that has to execute in a strict time sequence, I'm baffled.
What do you think a real 6502 is made of? Of a Penitum 4 for that matter... There are just gates arranges in a particular way. There was a project somewhere on the 'net - a CPU made of 74LSxx chips. In fact there were CPUs made of discrete transistors and vacuum tubes before that. A transistor or a tube is just a gate. The FPGA I used for the FPGApple happens to have 200,000 gates. The ones I use for *real work* have a few million gates.
Just read a few weeks ago somewhere: the navigation computer on an Appolo spaceship was made of 4000+ DTL chips. That is what was available in the 60's before TTL.
Is there is nice simple link that explains once over lightly how this works?
I'm partial to Verilog over VHDL, so I suggest this as a good place to start:
http://members.aol.com/SGalaxyPub/useful_links_verilog.htm
confusedly yours; Mike
-Alex.