Why don't they just slap a heatsink on this and make a round

Well they would but 80% of the plate would be defective and it'd cost $100,000

>>55730462
Eventually they will, since that'll be the only way to increase cpu power.

>LGA360
not bad

>>55730462
You know I never really understood the point of circular wafers for rectangular CPUs. Don't they end up with a ton of waste for all the cut off dies around the edges?

>>55730738
I think that they have this stuff that they squish flat and it forms a circle.

>>55730738
They do
The reason is because they forge cylinders of silicon and then cut in slices to work with it

>>55730738
see >>55732581

Silicon is forged as cylinders

>>55732581
>>55732638

I think it's also easier to coat circular wafers evenly.

>>55730462
It would have a TDP of 10 kilowatts

>>55730462
Did they finally finish your mom's wafer?
Or something.

File: silicon-2.jpg (90KB, 220x588px) Image search: [Google]

90KB, 220x588px

medal benis :D :D

>>55733001
*metal, *penis
And it's not metal, it's silicone you idiot.

>>55733515
Its a doggie dog world out there, man.

>>55733515
Lurk more

I think they say ingot or crystal not cylinder.

>>55730462
the yields would be something like 1%

>>55733515
>silicone

>>55730462
Literally what the fuck is that thing ?

>>55733515
fucking newfag kys yourself

>>55733001
It surprises me how there isn't a porn with this as insertion.

>>55730462
There is a limit to how large a CPU can be due to the fact that electricity actually needs time to move place to place, which could lead to desync, meaning that data in one place of the chip could be processed too quickly or slowly, and no longer match up with the rest of the CPU. Multi-CPU systems rely on a really slow (in relative terms) bus to connect them, so sync issues are unlikely, since the CPU has to wait for the bus anyway.

>>55733853
A wafer from which chips are cut out, each group of rectangles on its surface is potentially a CPU die.

>>55733929
I own a cylinder of silicon, and this stuff is insanely dense. A chunk a bit larger than your fist weighs at least two kilograms. If you wanna insert it somewhere, you have to be able to lift it first.

>>55734058
Girls could just sit on it, no?

File: Trilogy_WSI_Wafer.png (3MB, 1280x1280px) Image search: [Google]

3MB, 1280x1280px

>>55730462
https://en.wikipedia.org/wiki/Wafer-scale_integration

>Many companies attempted to develop WSI production systems in the 1970s and 80s, but all failed. TI and ITT both saw it as a way to develop complex pipelined microprocessors and re-enter a market where they were losing ground, but neither released any products.

>Gene Amdahl also attempted to develop WSI as a method of making a supercomputer, starting Trilogy Systems in 1980[1][2][3] and garnering investments from Groupe Bull, Sperry Rand and Digital Equipment Corporation, who (along with others) provided an estimated $230 million in financing. The design called for a 2.5" square chip with 1200 pins on the bottom.

>The effort was plagued by a series of disasters, including floods which delayed the construction of the plant and later ruined the clean-room interior. After burning through about 1/3 of the capital with nothing to show for it, Amdahl eventually declared the idea would only work with a 99.99% yield, which wouldn't happen for 100 years. He used Trilogy's remaining seed capital to buy Elxsi, a maker of VAX-compatible machines, in 1985. The Trilogy efforts were eventually ended and "became" Elxsi.

The main problem is that any defect in the wafer might make it unusable as a whole.

>>55730475

Yields were better than that I thought.

>>55733515
MEDAL PEBINS

>>55734094

All wafers have defects, its just getting them super small enough and uncommon enough to generate good yield.

>>55733515
newfag

>>55733515
Go back to rebbit you fucking retard

>>55730738
>>55731984
>>55732581
>>55732638
>>55732722
I figured it was because the lenses used for the lithography were round.

>>55733515
>silicon is no the same as rubber
the more you know

>>55733854
Retard of the hour goes to- (you)

>there are people on /g/ right now who don't know how CPU's are made

File: si_einkrist_inset.jpg (115KB, 607x809px) Image search: [Google]

115KB, 607x809px

>>55734077
better call your mother i got something for her

>>55737955
What am i looking at?

>>55738123
https://en.wikipedia.org/wiki/Czochralski_process

File: Silicon_grown_by_Czochralski_process_1956.jpg (211KB, 840x1085px) Image search: [Google]

211KB, 840x1085px

>>55738176
Damn those are some nice tits

>>55738123
mono-crystalline silicon ingot
this gets sliced into wafers

is everyone on /g/ a software weenie? from this thread it looks like nobody's building anything physical.

File: 1263852205541.png (25KB, 400x400px) Image search: [Google]

25KB, 400x400px

>>55738176
>The process is named after Polish scientist Jan Czochralski,[1] who invented the method in 1916 while investigating the crystallization rates of metals.[2] He made this discovery by accident, while studying the crystallization rate of metals when, instead of dipping his pen into the ink, he did so in molten tin and drew a tin filament, that later proved to be a single crystal.

I bet he was drunk.

>>55740159
Because building a CPU from logic chips is a pain in the ass and DIY lithography was quite expensive to get into last I checked.

>>55740238
Building a CPU from TTL isn't really that hard if you have the tools to put it together, making it do something really useful is another story however.

The Magic-1 is probably the greatest piece of computing autism I have ever witnessed.

>>55734094
maybe they could do better nowadays, by designing chips to have non-working bits fused off. I know Intel, AMD, and Nvidia all do this.

>>55740505
Thats like saying its easy to manufacture a Ferrari 458 as long as you have a factory and everything that comes with it.

>>55738211
I masturbated to this picture

>>55740505
>Building a CPU from TTL isn't really that hard if you have the tools to put it together
Sure, that part's relatively easy. Designing it on the other hand isn't. I don't see many people taking the time to build a CPU out of TTL chips if they didn't design it. Most of the people who want something more hands on/in depth than just assembling a modern computer will build something around an older CPU that's still available such as the Zilog Z80 where you get the same end results with less work and better documentation.

>>55740812
Not really, the Magic-1 that he's talking about didn't involve any DIY lithography, just a metric fuckton of wire wrapping/soldering. Was also built by someone who works as a botnet engineer for Google.

>>55738211
hot nips do