Ivy bridge build in 1986 process would be a 529 meters^2 die

[Hugo Drax]

Quick math. looking at the 80386 1.5 micron process in 1986 (275K transistors)

And todays Ivy bridge 1.4 billion transistor count. Building an Ivy bridge die using 1986 1.5 micron process would result in a big ass chip.

The die would be 529 meters^2 total area. compared to 160mm^2 using todays 22nm process

That would be big as hell. And astounding in terms of how technology has progressed. You could do one of those Carl Sagan walk on the die 529meter die while holding the Ivy bridge chip in your hand.

 

[rudeguy]

We need to get you laid man

 

[Markbnj]

And it would take a direct feed from the local power station to run it.

 

[lupi]

I remember going into a department store in the late 90's and they had a math coproccesser in a bin a miscellaneous computer stuff, was soo tempted to buy it then.

 

[Eug]

What's even scarier is the fact that the chip in the new iPhone is also over a billion transistors. In a frackin' phone.

 

[Bateluer]

Fascinating!

How much power would that 530m chip take to power?

 

[mmntech]

A while ago I got the idea to find out how large a modern tube computer would be. I had found a large triode on the internet with a shipping weight of 150 grams. It appeared similar to the ones used in ENIAC.

Assuming at a modern transistor and a triode are the same thing, that Ivy Bridge would come it at a whopping 220,000 metric tonnes. Or about two Nimitz class aircraft carriers. Of course that's not including any of the mounting hardware, capacitors, relays, or wiring.

Assuming it scaled up, it would take roughly 12 jigawatts to operate!

 

[Eug]

My first computer had 48 KB of RAM. However, I upgraded it to increase it by 33% to 64 KB!

 

[mjrpes3]

wut?

160mm area = 12.65mm
1.5μm = 1500nm / 22nm = 68x
12.65mm x 68 = 860mm = .74m area

 

[yhelothar]

My first computer had 48 KB of RAM. However, I upgraded it to increase it by 33% to 64 KB!

I got curious what one of those dinosaurs looked like. 48KB of RAM was huge!

www.ebay.com/itm/Siemens-6ES5350-3KA21-6ES5350-3KA21-RAM-Memory-Module-48-KB-/380676651773

 

[veri745]

wut?

160mm area = 12.65mm
1.5μm = 1500nm / 22nm = 68x
12.65mm x 68 = 860mm = .74m area

I think his math might be off.

If you use the transistor density of the 80386 and scale it up to 1.4B transistors:
http://en.wikipedia.org/wiki/80386#i386DX

275000 transistor / 104mm^2 = 2644 transistors/mm^2

1.4B transistors * 2644 trans / mm^2 = 529500 mm^2

This is 0.529 m^2, not 529m^2, since you have to divide by 1000^2, not 1000

This is why we have rockets crashing into mars.

 

[dank69]

I think his math might be off.

If you use the transistor density of the 80386 and scale it up to 1.4B transistors:
http://en.wikipedia.org/wiki/80386#i386DX

275000 transistor / 104mm^2 = 2644 transistors/mm^2

1.4B transistors * 2644 trans / mm^2 = 529500 mm^2

This is 0.529 m^2, not 529m^2, since you have to divide by 1000^2, not 1000

This is why we have rockets crashing into mars.

 

[Puppies04]

Learn to ²

 

[alkemyst]

learn to ²

accomplishment achieved

level 99 nerd!

 

[ElFenix]

I think his math might be off.

If you use the transistor density of the 80386 and scale it up to 1.4B transistors:
http://en.wikipedia.org/wiki/80386#i386DX

275000 transistor / 104mm^2 = 2644 transistors/mm^2

1.4B transistors * 2644 trans / mm^2 = 529500 mm^2

This is 0.529 m^2, not 529m^2, since you have to divide by 1000^2, not 1000

This is why we have rockets crashing into mars.

i'm also going to guess ivy has a higher percentage of its xtors devoted to cache than the 386 did, which means the 386 is comparatively less xtor-dense.