One of these days I'll probably get sucked into the whole stupid Minecraft thing and be forced to build awesome structures out of blocks. Until then, I'm happy to waste my block-building time with Freebuild. .)

Waste? If you want to play later, I would be game

That might definitely be a good idea! I need to work my way through this sudden postflood, finish and submit my homework for the evening, and then I'll probably be down. Let's not hijacked Ashbad's thread too severely, though.

Well, my friend lost interest in making this out of household items, and suddenly wants to make a 64 bit CPU in logisim x.x we were naive at first with the original idea, but it was doable with our current knowledge and some help -- but 64 bit with no knowledge? Anyways, we kinda let this thing go now, since we don't have nearly enough wire for what I planned to do and he a lost interest at my new opcode table x.x so, I'll be trying to build this in logisim as a super fast 4 bit computer

Sorry to hear that you won't be continuing work on the hardware version. 64-bit processors are no harder than 2-bit processors, assuming the number of bits refers to the number of bits in the registers of the processor. You just need to replicate the per-bit logic more times. The hard part will be for you to understand the different pieces that go into making a processor.

Yeah x.x so it's just more work then? I see . Either way, imma gonna try to make a 4 bit one first, to see how it works out the. I can go up from there.

Well, I'm oversimplifying a bit, but essentially, yes. I think you're absolutely better off starting with a 4-bit design with very few opcodes that is very very simple.

The way I have it planned is that I reads programs in nibble pairs (technically bytes, but nibbles will be the word size) so it allows for 256 different opcodes, though probably only 100 different instructions.

I'll be sure to share any questions I have here, I'm sure that besides just you kern there are also a ton of other EE experts around cemetech

Well, Benryves is an excellent EE and all-around (white-hat) hardware hacker, and we have several other experts and experts-in-training. I think 100 instructions to start is setting yourself up for disaster; start with 8 instructions or something! Perhaps the BF opcode set?

That's all really cool and everything, but if you want this to work, I would stick with 8 or maybe 16 bit word size. You can do OODLES (thats right, oodles...) of stuff with just an 8 or 16 bit word size. 2^16 already means that for JUST a simple demux you would need over 65 thousand 16 input NOR gates and 16 NOT gates!! Trust me, I know how cool it would be to have a 64-bit word size, but Logism just can't handle anything above about 16-bits... I mean, you get lag beyond imagining!

Your OTHER option is to use an FPGA. They are pretty cheap (less than 100 dollars.) and can do 64, 128, 256.... bit CPUs at amazing speeds. You would have to learn EVERYTHING about CPUs first and learn Verilog programming, but it is possible. (I hope, its what I'm busy working on every waking hour doing!)

I still think that you and your friend should go on and make a little relay computer. A plank of wood, some nails, and some wires and you could make a really cool 2-bit full adder. Just search YouTube for relay computers and you could see.

If you're set on making a 64-bit CPU, I would recommend you check out this page and the whole web-ring: "http://www.homebrewcpu.com" You can even telnet into his Magic-1 homebuilt CPU and play games on it! If you could build that, you would understand what it takes to make a 64-bit CPU. Just multiply the cost by a factor of a thousand, create black-outs throughout the entire Midwest powering the thing, and you wouldn't learn anything more than you would learn from making a simple 8-bit CPU, but it would be way funner... Anyway, if you have any questions about building CPUs, I'm your man!

@ Kerm
He means to build something called a "relay-computer" which is TOTALLY awesome. They were invented in the 19th century and used at telephone exchanges, ya know, to complete the call to the other end. Vacuum tubes were FAR faster and required a lot less power... but even today its a cool project to make a 4-bit relay adder to show how simple technology is. Just don't try to make the diode without wire... its impossible.
Also, do you think you could help me out with Verilog programming at all? I'm working on MY 64-bit CPU, and want to use an 80 dollar FPGA with 3.4 million gates in it. I have the design, just don't know SQUAT about Verilog.

I'm well-familiar with the concept of a relay computer; I don't think that's what he was intending. It didn't seem like he had any sort of switching devices like transistors, relays, etc in mind at all.

Wasn't there a topic here about papercraft logic gates?

http://www.cemetech.net/forum/viewtopic.php?t=5827&highlight=papercraft+logic+gates

Mr. Sir the adept, those are probably good ideas, but I didn't understand half of it I just started dabbling in this stuff out of interest 2-3 days ago; can you explain a few of those things to me, like why the demux has to be that big, what an FPGA is, etc. -- sorry if I sound rude :/

Adept, that's not how demuxes work. A 16-to-1 mux/demux definitely is built of a chain of a few dozen gates, not 2^16 gates. Ashbad, an FPGA, or Field-Programmable Gate Array, is sort of like a processor that you can change the gate connections of on the fly. It's a massive grid of pieces that can be logic gates or a few bytes of storage, and you can configure the FPGA to connect it all together to do many cool things. For example, you could write a z80 wiring design (a very very long and complex process, of course) to make the FPGA act as a z80. Contrary to what Adept said, FPGAs can be very expensive, in the thousands and tens of thousands of dollars, but the FPGA companies generally sell academic versions of their products for a few hundred dollars.

Wow, that's expensive x.x but I get how they work now -- so, it would be a real life version of logisim in a sense then? That sounds like it would be pretty cool to test out processors easily

But also, adept, it's not THAT laggy -- people on YouTube have made 64 bit processors in it using an upper and lower 32 bit halfword for each byte -- and the lag is noticeable, but it's not like it brings up a BSOD

A real-life version of logosim is a good way to put it. The FPGA development and testing boards usually also have tons of different ports like USB, PS/2, the various types of video in and out, IDE and SATA, sound, and all manner of other fun things that FPGAs can bitbang or use.

Wow, that sounds awesome... But before I use that I should most likely just start with logisim before I either break something or waste money on something I don't fully understand

Yes, that's definitely a good way to go. If you discover that you have a latent penchant for processor design, then it might be something good for you to consider in the future. Also, have you ever messed with microcontroller programming?

Micro controller programming? Sorry, never heard of it, or at least that term (sorry if I sound rather like a newb in the subject)

Also, I find this actually extremely fascinating. I even told my dad that I would consider being an EE after just seeing all of this, and unlike my aspirations to be a game programmer or a priest, he thought it would be a really great idea

So, Kerm, your excitement in this subject is really wearing off on me