is aircooling gonna cut it with new CPU´s?

[vidar]

wont there be a point in CPU development where air cooling will fail to keep Cpu´s cold enough? i mean, the heatsink and fan for a 3-4 gigahertz pentium chip will have to be huge! am i right?(or horribly wrong?)

 

[papasmurf]

I dont believe so, with the shrinking of cpu die sizes and other enhancements cpu's are beggining to get cooler slowly but surely. I HAVE A DREAM, A DREAM WHERE AN OVERCLOCKER IS NOT BOUND BY THE COLOR OF HIS HEAT SINK, WHERE ALL CPUS ARE EQUAL, WITHOUT ANY HEATSINK WHAT SO EVER! I HAVE A DREAM!

Introducing Tapeworms! The new big thing for weight loss!

 

[10GHZ]

ye, the maufactureing process getting smaller and smaller, eg, a Xp 2000+ uses 0.18 micron manufactureing process and a Tbred will use a 0.13 micron process, when the tbred hits the 2000+ mark, it runs significantly cooler than the XP beacuse of its manufactureing process. so as the CPu speed gets faster, the processes shrinks even more from 0.13 to 0.09... and so on.

 

[CALV]

I cant imagine it either, just imagine if every pc HAD to be watercooled !!


If they squeeze olives to get olive oil, how do they get baby oil?

 

[10GHZ]

if one day, they can no longer manufacturer CPU that can be cooled sufficiently with a conventional hsf, then i think water cooling kit will replace hsf and becoming a standard.

when the speed gets to 10GHZ maybe? ;p

 

[WingDing]

Would be a lot quieter too. Bring it on.....

Sex is great but it's no substitute for the real thing.....

 

[bum_jcrules]

What if computers went optical. No electrical internal resistance and no heat. They could drop the amps on existing systems to. Less Wattage. There are a lot of options for systems in the future.



<b>"Sometimes you can't hear me because I'm talking in parenthesis" - Steven Wright</b> :lol:

 

[Guest]

It only seems logical to convert as much as possible to optical systems. Problem at the moment is trying to manufacture affordable yet decent optically based circuitry.

Heat is not the only limitation electronics designers are battleing with. We are quickly approaching several insurmountable walls with respect to electronic design. I mean the fastest circuits I've ever heard of are on the order of hundreds of Gigahertz, but that is all analog, and they have been stuck in around the same frequency range for many years. We just don't have any science to make going any faster cost effective. Digital stuff will always be somewhat behind speedwise, at best.

 

[papasmurf]

hmm another thing to ponder, cpu manufacturers believe that they will hit the limit with silicon at 0.001 microns, witch should take us into the next decade, but it will not be very cost effective, it gets more and more expensive. but there is still hope for the future, IBM has had success with nanno tubes to conduct electricity, immagine nanno tube CPU's, with chips soooo much smaller think .00000000001 micron! also this thechnology has much more potential for speed and should halve the cost of manufacturing. food for thought.

Introducing Tapeworms! The new big thing for weight loss!

 

[10GHZ]

hummm... thats not gonna happen in my lifetime

 

[papasmurf]

well optical would have the benefit of no heat and light travels far faster than electricity can, however I believe this will be a long time in coming, my reasons are simple: remember the first electronic calculator, it filled a space of an entire home and was basicaly the first computer. was it practical or cost effective for a person to buy one of these, no, it would cost millions and they couldnt just cary it around in their back pockets, besides why not just do the math in your head? in terms of computer chips like cpu's, think about how complex they are, now think about how hard it would be to make light travel through tunnels so small and numerous and make it work faster than what is currently available for the same price and we are talking a miracle. perhaps in a century or so but I think we will just stick to electricty for now.

Introducing Tapeworms! The new big thing for weight loss!

 

[cakecake]

Using light to store data is one thing, but the most recent issue of Scientific American talks about Spintronics, a technology currently used partly in hard drives. You know how hard drives have the dilemma of how to cram more and more data into smaller spaces? So does the microprocessor. With spintronics what they hope to do is be able to use the natural 'spin' of electrons themselves. Electrons naturally follow a 'magnetic' pattern, and if this technology worked we could see infinitesimally small microchips.

<A HREF="http://www.scientificamerican.com/2002/0602issue/0602awschalom.html" target="_new">http://www.scientificamerican.com/2002/0602issue/0602awschalom.html</A>

Of course this is the very scientific and jargon-filled and 'impossible-to-read' description. Basically this will be better known as magnetic semiconductors. Everyone already knows about MRAM and they talk about it here too.

 

[bum_jcrules]

Nice idea, but controlling ellectrons and their orbit? Man that is really science fiction. I am not saying it can't be done but man that is out there. (Anyone heard of Quantum theorums vs. Partical theorums?) No scientist can fully define what gravity is let alone give me a cup of it and paint it red while their at it.

I know that MRAM already exists in R&D lab over at IBM but to make it a viable solution right now is a little out thereas well. Talk about your EMP, Hysteresis, EMI, etc.

I do like the spin idea however. Those aren't shots at you. I think optical and molecular computers are going to be a reality before we see the need for controlling the spin of electron on the molecular computers. Molecular alone would be 3-D vs. 2-D of today. Now take that 3-D to a new level. You are talking a 5th dimention. (Length x Width x Depth x Time x Spin) Ouch!!! My brain hurts.

Back to you...

<b>"Sometimes you can't hear me because I'm talking in parenthesis" - Steven Wright</b> :lol:

 

[vidar]

man, this post got pretty out there!