How realistic is designing your own hardware and ordering it from a manufacturer? no need to reinvent the wheel...

i think you are underestimating the amount of work that goes into designing a CPU
making a motherboard is realistic if you wanna make you own ARM board. you could probably learn how in about 2 years if you have at least some electrical engineering background. plus maybe another 1 or 2 for all the failed attempts until you have one. ordering one from china would cost about 500-800 at low quantities

I'll make the logo.

>how difficult would it be to design a CPU as fast as an i5-4200 and a supporting motherboard
Very difficult and extremely expensive. Modern microprocessors are too complex with an incredible amount of optimization, such as:
- branch prediction
- speculative execution
- register renaming
- out-of-order execution
- transactional memory

You might want to go with something like RISC-V, which has a very open design that can be copied and adapted, but still it will be a lot of very specialized work. Last I heard people could get 100 RISC-V chips made for around $14,000.

not a fucking CPU but there's a lot of companies that you can send PCB schematics to to get a board made if you ever need it.

Take a look at a few cores that have open licenses. riscv.org/risc-v-cores/

Imagine that this is as complex and expensive to design as (probably more than) an operating system. You don't see new operating systems popping out everywhere.

>designing your own hardware
If you want a Z80 clone, not that hard if you use an FPGA.
>ordering it from a manufacturer
Only feasible if you want to make thousands of them.
>to design a CPU as fast as an i5-4200 and a supporting motherboard?
Forget about it.

>You don't see new operating systems popping out everywhere.
Actually you do, there's a huge hobbyist scene. A guy wrote an OS with a GUI entirely in assembly.

>how realistic is designing your own hardware and ordering it from a manufacturer?
assuming by your picture, you mean ICs
custom ICs are expensive, and only something you do if you want millions of them
for most basic uses, there's microcontrollers
for something in-between, there's FPGAs
>how difficult would it be to design a CPU as fast as an i5-4200
far, far beyond one person, you're literally wasting your time even thinking about that

If you want to find out first-hand, learn how to use an FPGA and create a working implementation of a RISC architecture, like MIPS. Run some benchmarks and compare it with your i5-4200. Then create a VLSI version of your design, and see how much your theoretical clock rate improves over the FPGA. Then look at the x86 ISA, and you should realize how massive the task you laid out is.

It would be far easier to do budget research on equivalent VM solutions; these are equally scalable as physical architectures; but your profit point is scalable as well. VM is the future...

You can make a processor in minecraft. If you give me your server IP I can help

Attached: Language.png (931x561, 54K)

It's just a chunk of silicon, how hard do you think it can be?

>i think you are underestimating the amount of work that goes into designing a CPU
What about RISC-V, weren't those copy-pastable?

I dont know which is worse, the shit posting twat of an OP, or the idiots discussing his post.

>how realistic is designing your own hardware and ordering it from a manufacturer?
Pretty realistic. You can order any IC you want.
Deigning IC is realistic too.
> no need to reinvent the wheel, just copy well tested design principles without doing anything fancy.
CBAКC… Кoгдa вы зaбaтитe дoвoльнo вopoвaть нacтoящий лyчший
In case of modern CPUs it is near impossible to see the structure on die. You would need like same equipment they use to project stuff on Si-wafer, but with in other direction.
> for example how difficult would it be to design a CPU as fast as an i5-4200 and a supporting motherboard?
Well, around 1000 engineers would do this in couple of years. Keep in mind, those engineers won't work for less than $150 000/yr each.
Not big deal.

Attached: 2010.07.11.jpg (490x290, 62K)

Just diy your own op-amp on silicon and you'll understand why you can't do a full CPU. Custom op-amps are at least in the realm of diy though it usually takes someone skilled in electronics with the proper test equipment and design software and tools to pull it off.

Attached: LM741.png (1200x772, 2.16M)

Doesn't look too complicated.
I see couple capacitors in silicone, resistors...
But I kind don't see transistors, I might be blind, or retarded.
Not op though.

>2 companies
dont forget IBM, it still makes processors, and you can license them to build them yourself if you could.
based power9

you could use opensparc on a FPGA. other than that you won't be able to do much

In function no, it's not too terribly complicated. If you know your electronics fundamentals it's not hard to understand how it works. But understanding how it works is only half the battle.

Designing the die and getting all the nuances of the design right and then more importantly for diy, the manufacture, that's the killer. This is where it goes from being easy to being really hard.

Attached: 741+op-amp+internal+structure.png (780x480, 77K)

How many years have people been talking and talking and talking about RISC-V and it barely even has SIMD extensions, let alone anything remotely competitive to desktops from 20 years ago?

C'mon man, it's fun.

>how difficult would it be to design a CPU as fast as an i5-4200 and a supporting motherboard?

Attached: MNrzdvS.jpg (600x656, 51K)

>how realistic is it to design a highly technical and specialized product in your garage that rivals that of a 50-year-old company that spends 10 BILLION dollars annually in R&D
What do you think?

idk, do you have tens of thousands to piss away on even the oldest of technology nodes? Assuming of course that you could come up with a decent netlist and actually synthesize it which are huge problems on their own.

I guess it is like laying out PCBs.
Some people use 9000 unnecessary jumpers on 4 layer board, while others can do it on singe layer with zero jumpers.
I'm autistic, and I like laying out everything on one (1) layer, because I'm too poor for proper PCBs.

now fan out an FPGA on one layer without resorting to retardedly small traces biiiiiiiitch

>CAD software is proprietary
>The design libs are proprietary
>It costs more than your kidney
Possible in theory OP, but the design and testing tools cost so much you'll never be able to pay the price, and the design process is so long and complex you'll never finish.
t. former VLSI