h a l f b a k e r yNot so much a thought experiment as a single neuron misfire.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
It seems to me that the problems with heat dissapation in laptops might be solved if the processors were created differently. They are currently produced as squares, concentrated into the smallest physical space - which produces the problem of getting rid of the heat. What if the processor was produced
in long strips rather than squares? I understand there would be a loss of speed due to farther travel along signal paths, but with the proper engineering, couldn't this be minimized? Would the loss be that great (I am asking, I don't know). If the processor took up the same area as a square one, but was "stretched out" - wouldn't cooling be easier?
[link]
|
| |
Proper engineering has yet to alter the ultimate speed of electron flow; propagation delay is what it is. Path length dictates the total delay and is the chief reason chip architectures are reduced to ever smaller scales. |
|
| |
There are a bunch of subfactors, such as crosstalk delay, that contribute to propagation delay that can be improved through special signal encoding and routing schemes but, for now, the elemental rule is the smaller the scale the faster it can be. |
|
| |
I think the problem with this is that if a processer were stretched out it wouldn't be the same area, but larger. |
|
| |
But if you could arrange your circuitry so that it 'flowed' without everything needing to access everything else, maybe you could. |
|
| |
But, given the highly iterative flow demands of a cpu could such a design be stretched and flow optimized to prevent a bottleneck route? With heat being the horrid factor that it is would it not have been done already? |
|
| |
Would Mobius strips would make it easier to program an infinite loop? |
|
| |
If the processor is spread out, the longer distances would create more resistance. Resistance reduces clock cycles and creates heat. To overcome this resistance, more power must be used. More power means more heat. |
|
| |
using this logic, shouldn't we have spherical chips then ? |
|
| |
I'm sure they'd love to build solid spherical chips, and 3-D chips have been at least designed. The issue is being able to manufacture one. Circuits are etched onto a silicon surface using a masked light and photopolymers. If you find a way of embeding circuits in silicon, let us know. |
|
| |
"Spherical chips" Heh. Maybe they'd be called Cocoa Puffs. |
|
| |
I'm picturing a whole fried potato. |
|
| |
I guess this is pretty well covered, but there's really a very small portion of the chip which is generating most of the heat. A square die probably does as good a job (or better) as an elongated one at transfering this heat into the heatspreader. |
|
| |
Anyway, getting heat off of the chip is easy. The difficulty with laptops is transferring out of the case. |
|
| |
way off topic: I would expect a 3d chip to be sort of like a pencil stub |
|
| |