The Future of CPU cooling?

[finners]

Just came across this on the BBC news website-

http://news.bbc.co.uk/2/hi/science/nature/4955398.stm.

Probably a long time away yet if at all.

 

[Bobthelost]

short version? No

Many chips have cooling circuits nowadays, but these require pumps to drive the coolant, which in turn generate even more heat.

You site the pump externally, and stick on a larger radiator, simpler, cheaper and a world more easily implemented than this.

 

[BrownTown]

sounds comletely useless, and just another neat science expiriment that has little if any real world application. The heat produced by a pump is produced externally and is no concern except for in price, and this type of new tech is unlikely to win in price, so it has little use.

 

[d3lt4]

It is more likely that chips will use so little watts that only a small heatsink is needed instead of better cooling. Sounds interesting, but I doubt it will ever be used.

 

[suszterpatt]

Two words: quantum computing

 

[Bobthelost]

Originally posted by: suszterpatt
Two words: quantum computing

Ninteen words (well 17 and one repeated twice): Not bloody practical with today's hardware, still science fiction rather than science fact, utterly irrelevant to the topic at hand.

 

[suszterpatt]

Originally posted by: Bobthelost

Ninteen words (well 17 and one repeated twice): Not bloody practical with today's hardware, still science fiction rather than science fact, utterly irrelevant to the topic at hand.

It's not that far away, and it will most likely solve any cooling problems.

 

[TrevorRC]

As we shrink and shrink and shrink with die shrinks, chip heat will become less and less of a problem [supposedly.]

Maybe we can run back to passive cooling, eh?

45nm is on the horizon.

9nm will be around by 2015, if Intel's roadmaps are to be believed.
--T

 

[suszterpatt]

Originally posted by: TrevorRC
As we shrink and shrink and shrink with die shrinks, chip heat will become less and less of a problem [supposedly.]

Maybe we can run back to passive cooling, eh?

45nm is on the horizon.

9nm will be around by 2015, if Intel's roadmaps are to be believed.
--T

Shrinking the transistor size does reduce heat, but we also continue packing more and more transistors on a chip. Even more so with dual core, not to mention when quad- and octa-core roll around (which I'm guessing won't be that far).

 

[imported_Imp]

If it gets to the point where water cooling becomes a necesity, I sure hope they start finding new people to work towards a 'softer' path (efficiency rather than shoving more and more power through something).

 

[suszterpatt]

That would be quantum computing.

 

[A5]

Originally posted by: suszterpatt

Originally posted by: Bobthelost

Ninteen words (well 17 and one repeated twice): Not bloody practical with today's hardware, still science fiction rather than science fact, utterly irrelevant to the topic at hand.

It's not that far away, and it will most likely solve any cooling problems.

If by not far away, you mean "doesn't work yet and has no current practical means of contstruction", then yes, you'd be right

 

[blackllotus]

Originally posted by: suszterpatt
That would be quantum computing.

Its not even clear that its possible to make a general purpose quantum computer

 

[TrevorRC]

Originally posted by: suszterpatt
Shrinking the transistor size does reduce heat, but we also continue packing more and more transistors on a chip. Even more so with dual core, not to mention when quad- and octa-core roll around (which I'm guessing won't be that far).

Check out Conroe's thermal envelope

The new chips will be made using 65nm process technology and will have thermal design power (TDP) of 65W or lower, something which is expected to deliver on Intel?s promise to release chips featuring leading performance per watt ratio.

Yes, new process; I'll admit that.

I simply don't see something the size of a pinky-nail-clipping [9nm] consuming ~200W of power.

Maybe it's just me.

--Trevor