[ / / / / / / / / / / / / / ] [ dir / animu / ausneets / hypno / kc / kukuku / leftpol / sonyeon / vg ]

/tech/ - Technology

Winner of the 67rd Attention-Hungry Games
/ara/ - Mothercon

January 2019 - 8chan Transparency Report
Comment *
* = required field[▶ Show post options & limits]
Confused? See the FAQ.
Show oekaki applet
(replaces files and can be used instead)
Password (For file and post deletion.)

Allowed file types:jpg, jpeg, gif, png, webm, mp4, pdf
Max filesize is 16 MB.
Max image dimensions are 15000 x 15000.
You may upload 3 per post.

File: 41c94c5169a3227⋯.jpg (37.93 KB, 678x678, 1:1, IBM_PowerPC601_PPC601FD-08….jpg)


How difficult would it be?


Also curious.


File: abbf8bca570e9d4⋯.png (357.04 KB, 670x376, 335:188, ClipboardImage.png)

>How difficult would it be?

Ahmed Mohamed could do it.



And by hand with a soldering iron too! He put those amateurs at Intel with their "chemistry" shit to shame.


You'd need millions, no, billions of dollars and a team of 1000 high energy chinks on call 24/7 working for you. Keep them warm with the heat generated by the bitcoin miners in the warehouse downstairs.

To do it alone? You would need to be Terry A. Davis Himself.


You mean designing one?


File: 6cf5e4b7d73a6f7⋯.jpg (206.97 KB, 1440x824, 180:103, arch.jpg)

Designing a CPU is time consuming, actually building the CPU is extremely time consuming without automation.

You'll never build one in your garage powerful enough to compete with AMD/Intel due to their precise nano sized circuitry and 500+ million transistors, nevermind the security woes involved if you intend to distribute them. Depending how skilled you are in soldering creating a single prototype could take months to do. Actual products would take years to build by hand, and if you're not retarded hopefully its a low power design.

At best you could replicate a low end early 90s computer in processing power, build other computer parts and install linux on it.



If you mean designing one you could implement with a FPGA, then it's not harder than making kernel. The difficulty will rely on how complex you want the CPU to be.



>Actual products would take years to build by hand

You simply can't do it by hand, you require machinery.



You know what I mean, without industrial instruments it will be very difficult.



Babbage did it or should I say Ada Lovelace



That thing can do some sort of computation, but can you really call it CPU?


You can apparently make a 4-bit CPU from simple 7400-series logic gates.


But I guess you probably wanted to use something like FPGA?


u guys are idiots. a computer is basicaly something that tells you yes or no. 1 or 0

with todays technology it would be incredibly easy. why make it complicated?

plus you can by one for a dollar. why



It would be easy to do with vacuum tubes



why the fuck did they ever get rid of those. they were comfy as fuck



Hot light bulbs that break when bugs fly into them. Yes, why indeed?


File: ab34b1f9da47359⋯.png (31.56 KB, 300x100, 3:1, tmp_13910-tech820486439.png)


Banner was present over this thread.


Depending on how complex you're aiming.

You could make a 4 or 8 bit cpu yourself in fe months.

A 16bit would be hard.

As for modern cpu forget it, it take Townsend of people to design one


This. Take a look at https://en.wikipedia.org/wiki/MicroBlaze OP, you can even even target that instruction set with GCC.

With some understanding of digital circuits, and FPGA programming, you could realistically create your own Soft-CPU.





It's possible if you have equipment


PIC is pretty simple and documented, I think.



Mechanical computers are doable with appropriate levels of autism


File: 0bf20de892663c2⋯.png (515.65 KB, 741x351, 19:9, 20171010-060829-crazy-boar….png)

File: 6ab9ae81880fada⋯.jpg (364.78 KB, 1200x800, 3:2, 20171105-202740-csc-lcd1.jpg)

Fucking studying what a computer is.

How do they work.



Sure but how are you going to solder minecraft blocks onto your mobo?







>A 16bit would be hard

It's the save difficulty. You just change the verilog to work on 16 bits instead.





a towelhead made a cpu?



>How difficult would it be?


>designing a CPU

Doable, would take some time and effort to make it decent.

>producing a real CPU chip that people will want to use

I hope you have a few million dollars laying around. Those semiconductor fabs (all ~3-5 of them in the entire world) ain't gonna run themselves for charity.



With redstone, obviously.



That's just a matter of overclocking your minecraft world simulation.



He soldered it himself and implemented a nice suitcase clock with it :^)



> You just change the verilog to work on 16 bits instead.

i'm talking about hardware CPUs, not memes you run on FPGA



Read "CODE" by Charles Petzold (if you haven't already, but then again you probably wouln't ask) to get a rough idea.


not very difficult if it was made of cardboard, but it would require your imagination to work


File: dfec8b934362579⋯.png (174.89 KB, 553x502, 553:502, untitled.PNG)


Or out of stones.



Surprised no one has mentioned this yet. This should give you an idea of what you're up against.





If we don't go back to the beginning, it seems impossible for one person unless it is their life's work. Emulating one sounds much more fun, and you'd be mad not to emulate your design before implementing it.


But I'm just curious and don't really know anything. I'm amazed nobody's ever mentioned any of this in my life. Like there's a mental blackout for this device I use all day everyday.



are you a racist to doubt it?



>hi im new holds up spork



>Learn assembly.

>Print out someone's code.

>Start reading it, jotting down the state of the registers/memory as you go.

Boom, YOU are the CPU.



I know how to build a cpu, just take a bunch of small cpu's and solder them together!



>YOU are the CPU.

Too bad it'd be a sub-MHz one.



It's not practical to build a CPU without integrated circuits.



>to build a cpu, just take a bunch of small cpu's

So it's CPUs all the way down?







>74ls161 TTL 4-bit counters are CPUs

you are this new



>verilog is only used for FPGAs.

FPGAs let you iterate over a design before you make it in silicon, aka a "hardware CPU"



I don't think OP has the financial resources to make his own wafer


File: 6ac5ed92da49dba⋯.png (85.71 KB, 640x468, 160:117, soc_top_v5.png)


Depend what you mean by "creating a CPU".

If your main concern is avoiding any possibility of botnet, there are open-source RISC-V cores available that can be implemented on FPGAs.


Sun also released a UltraSPARC T1 and T2 cores fully open, but I'm not sure if there are FPGAs big enough to hold them yet.


File: 053e1d651a4c51a⋯.png (331.08 KB, 1312x544, 41:17, Digilent.png)


I've been looking in to this some more, there's been quite a lot of progress on open hardware recently.

The lowrisc project have got to the point of a functional RISC-V SOC design that can be implemented on a mid-range ($320) FPGA board and boots a functional Linux system with basic text-mode VGA and ethernet.



This is still a long way from something powerful enough to run a desktop, but it does raise the possibility of other uses. Perhaps we could have completely open-source hardware home routers in the not too distant future. The FPGA development is ongoing while they work towards the goal of taping out ASICs for a fabrication run. I guess there will be some Pi-type boards made eventually.

I'm tempted to invest in one and have a play.



1.) Get an FPGA

2.) Design a prototype in a hardware design language

3.) publish the spec based on the prototype

4.) Contact a semiconductor fab and have them print a chip based on the spec

It will run you several thousand bucks or more but there's no reason you can't really


Is there any efficiant way to etch silicon in your garage? How did Commadore and Atari do it?



Didn't Commodore use Motorola chips?



They fabricated their own chips, but used existing MOS designs.



>with appropriate levels of autism

This is the end boss of autism.





ask VIA how successful it is at competing with the big guys


File: 9f2642cfdc8f588⋯.jpg (11.84 KB, 182x268, 91:134, MV5BMTkyOTcwMjY1NV5BMl5Ban….jpg)


Just build one the size of a room, and miniaturize it slowly over time.



File: 7535d1e51dba2aa⋯.gif (2.9 MB, 500x540, 25:27, datamind.gif)

1.) Learn a Hardware Design Language https://en.wikipedia.org/wiki/Hardware_description_language

2.) Run code in an FPGA simulator

3.) Purchase FPGA and run real-world field tests, perform bug checks.etc

4.) Draft specification based on FPGA

5.) Contact a semiconductor fab like TSMC, get rates for custom fab jobs.etc spoiler alert; it'll likely run you into the thousands of dollars to produce a single sample because wafers aren't cheap and you wont have the advantage of scale of economics

congratulations! You have now made compooter!



That's a nice board, I picked one up recently. If you go with that, be sure to check out these free resources

Digital design (basic)


Vivado tutorials (VHDL or Verilog)


I also recommend Ashenden's book on VHDL, and his Digital design book is a comfy read too.



difficult if you have to ask


Seize the means of production



>gommie spodded :-DDDDDD


File: a006c8a445b0a43⋯.mp4 (5.76 MB, 1920x1080, 16:9, Refused-The Apollo Program….mp4)


The destruction of everything is the beginning of something new

Your new world is on fire and soon you'll be too

Throw a rock in the machine

Sabotage will set us free

Throw a rock in the machine




The botnet goes back further than we thought



Shut the fuck up stupid fucking shithead



File: 10219f1bffe35ab⋯.png (1.25 MB, 677x846, 677:846, dis buk.png)

read this and you will know the most things you have to know



>wasting so much time, energy and money to circlejerk about how you know to make CPU out of discrete transistors




i always trust mr. money's cpu designs


File: a02c6a40f3aa345⋯.jpg (84.34 KB, 465x357, 155:119, screenshot-23.jpg)

A literal underage did it so...


You can actually make a SUBLEQ CPU in Verilog super easy, and load it onto an FPGA.



You can make one with water and cardboard



There were non-general computers made with water, I think they were used even in the 70s.



Depends on what you classify as a CPU. You could buy a shit ton of relays and have at it if you want.



He did it by refurbishing an old 70s one. Apparently the people who used to work on the machine were so impressed with him getting it back up and running that they gave him tips on creating chips.

What he did is impressive but not all of us can have old fab equipment around the corner with people who know how to use it.

[Return][Go to top][Catalog][Nerve Center][Cancer][Post a Reply]
Delete Post [ ]
[ / / / / / / / / / / / / / ] [ dir / animu / ausneets / hypno / kc / kukuku / leftpol / sonyeon / vg ]