Cores keep shrinking, how low can they go?

N

nigerian_businessman

[H]ard|Gawd
Joined
Sep 3, 2004
Messages
1,535
I guess this question will be best answered by an electronics engineer, but for the sake of promoting discussion and feeding my (and hopefully others) curiosity, what happens when they run out of space?

I'm not sure if they really can run out of space, but in laymans terms, what I mean is -- how small can they keep shrinking these processors before we have to find a whole other way to create them? We've seen big (or small, depending on your POV) drops in nm, from 130 to 65 in a matter of a few years. From what I understand, Intel is lining up to push out 45nm processors probably by the end of 2007 or early 2008. What happens when they get down to 16nm, which is the lowest standard defined by the ITRS ? Can we go lower than that?
 
From what I've heard. silicon can get down to about 10-12nm before they become useless because the insulator become too thin to block electon flow between circuit pahts.
 
ryan_975 said:
From what I've heard. silicon can get down to about 10-12nm before they become useless...
Click to expand...
The big quetion is, what happens after that?
wink.gif


There aren’t many options as far as I can tell:
1) Design more efficient architectures.
2) Find something better than Silicon to make microprocessors out of.
3) Start adding more cores and get Reverse Hyper Threading working.

With 1 and 2 you do get more speed, but you’ll eventually hit a wall again. Option 3 is more expandable, but you can only put so many cores on one processor before they get so large that latency becomes an issue between the cores.

Something is going to need to change soon, otherwise we’ll be exercising all of these options at once and we’ll still be running head first into another performance wall.
 
note235 said:
i heard new material was in the makes.
Click to expand...
I've heard good things about carbon nanotubes...that would allow things to contenue to get faster for a while, but there's still a poaint at which you'll nolonger be able to gain more performance with them.
 
The whole reason for getting smaller component sizes is to reduce heat and power consumption while at the same time be able to increase efficiency and speed. So a new technology doesn't necessarily have to be smaller to be more effiecient, cooler, or faster. They just need to find something that works better than silicon, but it doesn't HAVEto be smaller.
 
A Silicon Atom is about 0.22 nm across so no smaller then that I bet, though I think we will hit a limit much earlier.
 
arman01 said:
i think the new material is Indium antimonide.
Click to expand...
It's Indium Antimony but same difference. Similar idea to Gallium-Arsenic, take a group 3 and a group 5 element to make an "apparent" group 4 device.

Anyway, I recently read an article that researchers are looking into using electron spin to store data.
 
drizzt81 said:
It's Indium Antimony but same difference. Similar idea to Gallium-Arsenic, take a group 3 and a group 5 element to make an "apparent" group 4 device.

Anyway, I recently read an article that researchers are looking into using electron spin to store data.
Click to expand...

The atoms are actually bonded together, so the "ide" is proper. These are not alloys. They are lattice site substitutions.

http://en.wikipedia.org/wiki/Indium_antimonide
http://en.wikipedia.org/wiki/Gallium_arsenide

The wiki you linked even calls InSb Indium Antimonide. :p
 
A physics professor I had a discussion with at university of Iowa a couple years ago was researching a way in which they could sprinkle a certain substance on a silicon die, and found that the atoms had an ordered way of staking themselves, tending to be roughly 7 atoms tall about as many wide. He mentioned that with further refinement this could potentially be used in processors. This would put transistor sizes in the angstroms, if properly refined.
 
I can't wait until they figure out how to make processor that use photons instead of electrons. They already have photonic transistors and logic gates.
 
ryan_975 said:
I can't wait until they figure out how to make processor that use photons instead of electrons. They already have photonic transistors and logic gates.
Click to expand...

That's a bit harder since photons aren't real. By real, I mean in the sense that we can observe and monitor their output, let alone be able to store information on them.

drizzt81 said:
Anyway, I recently read an article that researchers are looking into using electron spin to store data.
Click to expand...

This is probably the next or next-next step in making procs smaller. Nanotubing is useful, but not in the practical terms of conduction of information. Nanotubes made out of carbon though can prove to be very very good insulators at the atomic level for electrons.

...Now if we can only get those tricky anti-matter particles under our control :D
 
crazjayz said:
That's a bit harder since photons aren't real. By real, I mean in the sense that we can observe and monitor their output, let alone be able to store information on them.



This is probably the next or next-next step in making procs smaller. Nanotubing is useful, but not in the practical terms of conduction of information. Nanotubes made out of carbon though can prove to be very very good insulators at the atomic level for electrons.
Click to expand...

We don't store information or monitor output on electrons eithers. We store information in base on whether or not they're present.
 
ryan_975 said:
We don't store information or monitor output on electrons eithers. We store information in base on whether or not they're present.
Click to expand...

You're right, information is either a yes or no sort of "storing". I figure though it's much easier to play with something that actually has physical properties, than with something that's massless and moves at the speed of light.
 
crazjayz said:
You're right, information is either a yes or no sort of "storing". I figure though it's much easier to play with something that actually has physical properties, than with something that's massless and moves at the speed of light.
Click to expand...

Light isn't totally without mass, and it has physical properties that can be measrured. Plus unlike electrons, you can mix different wavelengths of light without corrupting what's stored in it.
 
ryan_975 said:
Light isn't totally without mass, and it has physical properties that can be measrured. Plus unlike electrons, you can mix different wavelengths of light without corrupting what's stored in it.
Click to expand...

If it isn't without mass, what is the mass of a photon?
 
ryan_975 said:
Light isn't totally without mass, and it has physical properties that can be measured.
Click to expand...
Right you are, and there's even practical applications for it. By using a "solar sail" you can harness the pressure exerted by light alone to pull spacecraft around (at least away from the sun ;))

ryan_975 said:
Plus unlike electrons, you can mix different wavelengths of light without corrupting what's stored in it.
Click to expand...
Good point you have there. Even if you only used wavelengths of light from the visible spectrum, there are enough distinguishable bands there to send and receive massive amounts of data.
 
Unknown-One said:
Good point you have there. Even if you only used wavelengths of light from the visible spectrum, there are enough distinguishable bands there to send and receive massive amounts of data.
Click to expand...

That is a good point. Actually, that's a great point. But going back again to the origial question, what are the physical properties of a photon?
 
After going through the Wikipedia page on Photons, and their page on Wave-Partical Duality, I now know more than I ever wanted to about how light works.

The artical on Photons first states Photon's are generally accepted to be massless, but then gives a conservative upper limit for the mass of the photon, 6×10−17 eV...not sure how that works.

Theres also a bit that talks about how the momentum carried by a photon can be transferred when it interacts with matter. They even use the same example I did for a practical aplication for radiation pressure, a solar sail.
 
Unknown-One said:
After going through the Wikipedia page on Photons, and their page on Wave-Partical Duality, I now know more than I ever wanted to about how light works.

The artical on Photons first states Photon's are generally accepted to be massless, but then gives a conservative upper limit for the mass of the photon, 6×10−17 eV...not sure how that works.

Theres also a bit that talks about how the momentum carried by a photon can be transferred when it interacts with matter. They even use the same example I did for a practical aplication for radiation pressure, a solar sail.
Click to expand...


Wikipedia is not an academic source. Even in the academic community, this is not the more widely accepted theory. Solars sails are not a reality yet. If they do become real, there are plenty of particles flying around with mass to drive them. The sun emits a plethora of particles.

I'm just pointing this out. Not that I believe one way or another. When I was in school we were strictly taught that photons do not have mass. The way I understand it, the definition of "mass" has been changed, if only to accomodate this new theory that photons do have "mass" - and the reason the theory exists is some scientists are unable to reconcile conservation of enery-mass with the current particle-wave theory. To me, anything will work. There's no way to measure the mass of light because when you stop it from moving, it is no longer light. So, the meaning has been changed by some to regard mass by the way it impacts things, basically it's energy-effect.
 
Bao01 said:
Wikipedia is not an academic source. Even in the academic community, this is not the more widely accepted theory. Solars sails are not a reality yet. If they do become real, there are plenty of particles flying around with mass to drive them. The sun emits a plethora of particles.

I'm just pointing this out. Not that I believe one way or another. When I was in school we were strictly taught that photons do not have mass. The way I understand it, the definition of "mass" has been changed, if only to accomodate this new theory that photons do have "mass" - and the reason the theory exists is some scientists are unable to reconcile conservation of enery-mass with the current particle-wave theory. To me, anything will work. There's no way to measure the mass of light because when you stop it from moving, it is no longer light. So, the meaning has been changed by some to regard mass by the way it impacts things, basically it's energy-effect.
Click to expand...

There was a device my physics teacher had in high school that she used to prove the point that light has at leasta non neglibile mass. It was a evacuated bulb with a little thing in it that had three metal sail looking things. One side of each sail was painted white, while the opposite sides were painted black. When it was in a lit room it would spin. the brighter the light shineing on it, the faster it would spin.
 
StealthyFish said:
-ide is attached to a cation that is monoatomic.
Click to expand...
wouldn't it be mono-molecular or mono-crystalline? Mono-atomic = 1 atom. Any "thing" with > 1 nucleus is not mono-atomic.
 
soory for the double post:
http://www.lbl.gov/Science-Articles/Archive/sabl/2005/March/03-polarons.html

some stuff on what is called "spintronics" which is what I was referring to earlier. The spin of an electron (up or down iirc) will be a 0 or 1 state. If it's possible to keep the sping "stable" and read it 'easily' it should be possible to store multiple bits/ atom.. very neat.

As to the light issue:
EDU article on it.

And here is a paper on it too :)
 
ryan_975 said:
There was a device my physics teacher had in high school that she used to prove the point that light has at leasta non neglibile mass. It was a evacuated bulb with a little thing in it that had three metal sail looking things. One side of each sail was painted white, while the opposite sides were painted black. When it was in a lit room it would spin. the brighter the light shineing on it, the faster it would spin.
Click to expand...

Was she sure the energy from the absorption of photons did not result in the release of a few ions or created an electromagetic effect especially since the sails were made of metal? There must be a little bit more than a high school physics teacher to explain this stuff. I mean my last college physics class with post doctorates acting as assistant professors and doctorate candidates as teaching assistants was like 12-13 years ago? man. I'm sure they must be teaching this stuff at accredited colleges but, I'm not aware that it's a mainstream theory. Well, it's been a long time, but I really doubt it's proof.
 
drizzt81 said:
soory for the double post:
http://www.lbl.gov/Science-Articles/Archive/sabl/2005/March/03-polarons.html

some stuff on what is called "spintronics" which is what I was referring to earlier. The spin of an electron (up or down iirc) will be a 0 or 1 state. If it's possible to keep the sping "stable" and read it 'easily' it should be possible to store multiple bits/ atom.. very neat.

As to the light issue:
EDU article on it.

And here is a paper on it too :)
Click to expand...


The EDU article is from a usenet. It's interesting, though.
 
Bao01 said:
Was she sure the energy from the absorption of photons did not result in the release of a few ions or created an electromagetic effect especially since the sails were made of metal? There must be a little bit more than a high school physics teacher to explain this stuff. I mean my last college physics class with post doctorates acting as assistant professors and doctorate candidates as teaching assistants was like 12-13 years ago? man. I'm sure they must be teaching this stuff at accredited colleges but, I'm not aware that it's a mainstream theory. Well, it's been a long time, but I really doubt it's proof.
Click to expand...

I could be very wrong about what it was made of. But I just remember we were talking about the non negligible mass of light and this thing was used to show it. The more direct light that hit it the faster it spun. We took it out into the sun that it went crazy fast.

And 12-13 years ago I wasn't in high school ter.
 
ryan_975 said:
There was a device my physics teacher had in high school that she used to prove the point that light has at leasta non neglibile mass. It was a evacuated bulb with a little thing in it that had three metal sail looking things. One side of each sail was painted white, while the opposite sides were painted black. When it was in a lit room it would spin. the brighter the light shineing on it, the faster it would spin.
Click to expand...
Sounds more like an experiment with EM fields than light mass. Light IS an electromagnetic field, but all a photon does in that sense is communicate forces (it's called a "messenger" particle for that reason). It sounds to me that the interaction between the light and the plates caused disturbances in the electromagnetic field that caused it to spin, and not actual photon momentum.

If proving that light was not massless were that simple, it wouldn't be such a contested issue.

Another thing I'm curious of...was the orientation of the bulb important in relation to the light? Or would it spin no matter the location or angle relative to the light? If orientation were not important, then it definitely is just magnetic field trickery going on.
 
WHOAguitarninja said:
Sounds more like an experiment with EM fields than light mass. Light IS an electromagnetic field, but all a photon does in that sense is communicate forces (it's called a "messenger" particle for that reason). It sounds to me that the interaction between the light and the plates caused disturbances in the electromagnetic field that caused it to spin, and not actual photon momentum.

If proving that light was not massless were that simple, it wouldn't be such a contested issue.

Another thing I'm curious of...was the orientation of the bulb important in relation to the light? Or would it spin no matter the location or angle relative to the light? If orientation were not important, then it definitely is just magnetic field trickery going on.
Click to expand...

I don't recall the details. It was just a random thing I happened to remember the she did while we were talking about light.
 
Can someone translate this entire thread to become idiot friendly. I have no idea what's going on with the ion flux particle accelerators and the anti matter particle laser beams.
 
You must log in or register to reply here.
Back
Top