Chip designs that could impact Bitcoin: 3D stacking

[check_status]

So the idea is to take components that are often separate on a system – processors, main memory, networking and other peripherals – and etch them all onto chips and then stack them up in a 3D array, rather than solder them onto a motherboard and wire them together with metal stripes in a 2D array. By going 3D, the wires between components can not only be shortened - cutting the time those components need to exchange signals and potentially requiring less energy to send a signal.

"Today's chips, including those containing 3D transistors, are in fact 2D chips that are still very flat structures," explained Bernie Meyerson, a vice president of IBM Research, in a statement announcing the partnership between Big Blue and 3M.

"Our scientists are aiming to develop materials that will allow us to package tremendous amounts of computing power into a new form factor – a silicon skyscraper. We believe we can advance the state-of-art in packaging, and create a new class of semiconductors that offer more speed and capabilities while they keep power usage low – key requirements for many manufacturers, especially for makers of tablets and smartphones."

The initial plan is to come up with a way to stack up as many as 100 chips into a tower of computing power. Over the long haul IBM wants to be able to bond stacks of complete wafers together, bonding hundreds of processors at a time.

The plan, says the IBM source, is to get it into production around the end of 2013.

http://www.theregister.co.uk/2011/09/07/ibm_3m_3d_chip/

For now this is just secret sauce for IBM and 3M, but applied to GPU's or ASIC's could increase hashing power while keeping power consumption low.

[1713885647]

1713885647

[1713885647]

1713885647

[1713885647]

1713885647

[makomk]

I get the impression that chip stacking doesn't really make sense unless you need lots of high-speed interconnectivity or a really small form factor. Bitcoin mining is an embarassingly parallel problem that doesn't require much bandwidth into and out of the individual hashing cores - some of the current FPGA designs actually use standard serial links running at standard serial speeds - so there's no real reason to use stacked dies. The main problem at the moment is powering and cooling the chips, and using stacking just makes that harder to manage.

[check_status]

That's why they teamed up with 3M, to develop thermal conductive compounds.

What if they used a different medium, do you remember the solder iron sold on TV infomercials, battery operated, instant heat, melts solder?
The tip was made out of a novel type of silicon that multiplies temperature. If you put that between the layers of the chips and apply cold temps, which would be multiplied by 10, you could get an effective thermoregulator.

[DeathAndTaxes]

What if they used a different medium, do you remember the solder iron sold on TV infomercials, battery operated, instant heat, melts solder?
The tip was made out of a novel type of silicon that multiplies temperature. If you put that between the layers of the chips and apply cold temps, which would be multiplied by 10, you could get an effective thermoregulator.

Law of conservation of energy says "NO".  temp is merely a measure of thermal energy.  You can't multiply energy.  That would imply you created energy from nothing.

[check_status]

What if they used a different medium, do you remember the solder iron sold on TV infomercials, battery operated, instant heat, melts solder?
The tip was made out of a novel type of silicon that multiplies temperature. If you put that between the layers of the chips and apply cold temps, which would be multiplied by 10, you could get an effective thermoregulator.

Law of conservation of energy says "NO".  temp is merely a measure of thermal energy.  You can't multiply energy.  That would imply you created energy from nothing.

To calculate your thermal dynamic opinion, we would need to define the boundry of the closed system, have an advanced scientific calculator, and possess electron microscope data about the molecular structure and it's effects in stratum. (I don't think your clamp meter will be useful.)

Aluminium is an energy multiplier, they don't use it in spaceships because of this.

[rjk]

"Multiplier" implies an output greater than the sum of the inputs, which is impossible. Perhaps it is just a misnomer.

[BkkCoins]

IIRC The chip used for the Raspberry PI was built using a similar idea. It doesn't sound as advanced but generally similar. Broadcom mounted the dies for the GPU and RAM on top of the CPU in the same SoC package.

[niko]

IBM Reaearch in recent years serves mostly PR function. A typical example of science hype is their "centipede" memory vaporware.
Stacking wafers certainly increases packaging density compared to PCBs. Their press release misleads readers that this somehow menas lower power, whereas in reality all it means is more cooling problems, and thereby the need for lower power. So, they word the problem in a positive way.
Stacking is not a new idea. Lots of research and real devices exist. Microfluidic cooling - evaporative or even with liquid metal alloys (eGaIn) - is not a new idea, lots of published research on that topic.
IBM simply announced in a bombastic tone that they will try and build upon this. Chances are, nothing will come out of it except for some stock price manipulation. Once in a while something real actually comes out of it, but in this case it wouldn't be anything revolutionary.

[check_status]

"Multiplier" implies an output greater than the sum of the inputs, which is impossible. Perhaps it is just a misnomer.

Nope, not a misnomer. Stellar radiation, upon striking aluminium, radiates 2-3x more than input. Nasa had an engine designed, for satellites and interplanetary observatories, that used a radiation multiplier propulsion engine (which produced 8x output), through stages I believe.

[BkkCoins]

"Multiplier" implies an output greater than the sum of the inputs, which is impossible. Perhaps it is just a misnomer.

Nope, not a misnomer. Stellar radiation, upon striking aluminium, radiates 2-3x more than input. Nasa had an engine designed, for satellites and interplanetary observatories, that used a radiation multiplier propulsion engine (which produced 8x output), through stages I believe.

That's interesting. But what happens to the aluminium? Does it degrade over time as I expect the energy must come somewhere, ie. the chemical bonds or something.

[AmpEater]

"Multiplier" implies an output greater than the sum of the inputs, which is impossible. Perhaps it is just a misnomer.

Nope, not a misnomer. Stellar radiation, upon striking aluminium, radiates 2-3x more than input. Nasa had an engine designed, for satellites and interplanetary observatories, that used a radiation multiplier propulsion engine (which produced 8x output), through stages I believe.

That's interesting. But what happens to the aluminium? Does it degrade over time as I expect the energy must come somewhere, ie. the chemical bonds or something.

It's not interesting, it's false.

[Littleshop]

"Multiplier" implies an output greater than the sum of the inputs, which is impossible. Perhaps it is just a misnomer.

Nope, not a misnomer. Stellar radiation, upon striking aluminium, radiates 2-3x more than input. Nasa had an engine designed, for satellites and interplanetary observatories, that used a radiation multiplier propulsion engine (which produced 8x output), through stages I believe.

Ha.  Link please.  Maybe, just maybe, you were reading about solar sails and got most of the details wrong. 

[tvbcof]

"Multiplier" implies an output greater than the sum of the inputs, which is impossible. Perhaps it is just a misnomer.

Nope, not a misnomer. Stellar radiation, upon striking aluminium, radiates 2-3x more than input. Nasa had an engine designed, for satellites and interplanetary observatories, that used a radiation multiplier propulsion engine (which produced 8x output), through stages I believe.

That's interesting. But what happens to the aluminium? Does it degrade over time as I expect the energy must come somewhere, ie. the chemical bonds or something.

It's not interesting, it's false.

"Oh, ye seekers after perpetual motion, how many vain chimeras have you pursued? Go and take your place with the alchemists."

  - Leonardo da Vinci (1452-1519)

[rjk]

Yeah, it sounds like you might be misinterpreting something about an ion thruster maybe. Those weren't efficient enough to use, and they definitely required energy to use.

[AmpEater]

"Oh, ye seekers after perpetual motion, how many vain chimeras have you pursued? Go and take your place with the alchemists."

  - Leonardo da Vinci (1452-1519)

What a fantastic quote.  

"Aluminium is an energy multiplier, they don't use it in spaceships because of this." - What the actual fuck man? The space shuttle was largely made of aluminum.

"When the Space Shuttle was first proposed in the late 1960s, planners from the National Aeronautics and Space Administration (NASA) wanted a vehicle that would be much larger than any that had flown in space before. But the amount of high-temperature metal required to protect a large vehicle would have been very heavy and this would have affected vehicle performance. Designers chose to use conventional aluminum for the main body and to protect it with a layer of heat resistant material." - http://www.centennialofflight.gov/essay/Evolution_of_Technology/TPS/Tech41.htm

P.S.

check_status - I hate you.  Dirtying the name of a legitimate scientific entity to sell your bullshit notions. "It's true because NASA!"  It's a lie, you didn't read that somewhere trustworthy, you made it up, or you blindly parroted something another idiot made up.  And then you construct a sentence that conveys some sort of expertise in the field to convince laypersons that you're a reputable source.  Fuck.

[wabber]

"Multiplier" implies an output greater than the sum of the inputs, which is impossible. Perhaps it is just a misnomer.

Nope, not a misnomer. Stellar radiation, upon striking aluminium, radiates 2-3x more than input. Nasa had an engine designed, for satellites and interplanetary observatories, that used a radiation multiplier propulsion engine (which produced 8x output), through stages I believe.

2-3x the input you say? So we can just take a small piece of aluminium heat it up with a lighter, use that piece of aluminium to heat up 3 new pieces, use these three pieces to heat up 9 pieces of aluminum etc
Then we could have a room full of very hot aluminum and use it to boil water. The steam could then be used to generate electricity. Awesome a simple lighter and some aluminium could replace a nuclear power plant.

But there are more ways to get free energy, just check out youtube.

[unclemantis]

This is news is not new. I heard about Intel doing something like this a few years back.

Things are about to get really interesting really fast!

[kjj]

IIRC The chip used for the Raspberry PI was built using a similar idea. It doesn't sound as advanced but generally similar. Broadcom mounted the dies for the GPU and RAM on top of the CPU in the same SoC package.

This is POP packaging.  Part on Part.  It takes two normal chips, fully encased as usual, and stacks them, usually using something like BGA.  They are different dies, built at different times, and probably in different fabs, then assembled later.

3D chips are different.  A typical flat chip is made up of dozens of layers of depositation and etching.  You start with a bare wafer, and etch part of it away, then deposit a thin layer on top of that, then etch part of that away, then deposit another layer, and etch, etc, etc.  You end up with a flat 2D structure of 3D objects, made all at once.

A real 3D chip has layers of groups of layers, creating a 3D structure of 3D objects.

Right now, we have problems moving heat out of the chips we have, and not so much problems with a lack of transistors available on a die.  But that's how progress works, the best we can do is the product of a whole pile of limitations, and people are busy fighting back all of them.  In time, 3D chips will become just "chips", just like color TVs all became just "TVs".

[check_status]

@AmpEater  - Maybe you and I should spend a little more time researching our statements before posting and we can limit the abuse of techtrolls.

In interplanetary space, it is believed that thin aluminum shielding would have a negative net effect.

Spacecraft can be constructed out of hydrogen-rich plastics, rather than aluminum. Unfortunately, "Some 'galactic cosmic rays are so energetic that no reasonable amount of shielding can stop them,' cautions Frank Cucinotta, NASA's Chief Radiation Health Officer. 'All materials have this problem, including polyethylene.'"

http://en.wikipedia.org/wiki/Health_threat_from_cosmic_rays

It was a show on PBS, possibly Nova, involving space travel. It's not possible I confused it for solar sails, though it is possible, I would concede propulsion for long term electrical generation. I still claim multiplication via GCR interacting with aluminum.

There was a motor, a rotor/stator design built with permanent magnets, which doesn't produce heat while spinning, but produces cold instead. It was developed by a scientist from JPL and he was awarded a patent solely because of the novel effect it produced.

I only brought it up as speculation for cooling the 3D CPU stack, maybe a bit tongue in cheek, but now I'm defending anomolous physics claims.  

In Thermal Decay, all heat moves towards cold, so being able to cool the medium between the chips would be an advantage in removing excess heat.

[tvbcof]

...
I only brought it up as speculation for cooling the 3D CPU stack, maybe a bit tongue in cheek, but now I'm defending anomolous physics claims.  

...somewhat unconvincingly.