Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Do superconductor computers produce waste heat? Will they always?
31 views
Skip to first unread message
therealst...@gmail.com
Mar 21, 2013, 12:37:23 AM3/21/13
to
So superconductors don't produce waste heat when they have a current running through them. They have no resistance, so you can spin up a current in a circuit and it'll just keep going, around and around, forever (ok: for a long, long time). MRI machines use this.
But how about superconducting computers? Could a superconducting transistor exist that has (functionally) no resistance and (functionally) no energy loss? If you can have a circuit that runs a current without needing to keep adding energy, could you put a computer in that circuit and keep it going? Could you have a computer circuit that does computations without energy input?
Or does this break some inviolate law? Is there anything in the laws of thermodynamics / the laws of physics about computation? Does computation necessarily cost energy in and of itself? Is there some 'minimum perfect-efficiency cost' of thought? I can't find anything on this, but I might just be looking for the wrong terms . . .
To be clear, I'm not necessarily asking if this is possible with current technology. I'm asking if this is established and/or strongly-believed to be straight-up *impossible* by our current understanding of physics. Alternatively, if it is possible, what would make this difficult, and/or why haven't we done it (or have we?)?
Now, on earth you'd need energy to keep it cool, but out beyond the heliopause and into interstellar space, you could cool yourself down to superconductor-level temperatures with passive radiators (no?). Then you can just sit and think deep thoughts until entropy eventually gets around to rusting something important . . .
But how about superconducting computers? Could a superconducting transistor exist that has (functionally) no resistance and (functionally) no energy loss? If you can have a circuit that runs a current without needing to keep adding energy, could you put a computer in that circuit and keep it going? Could you have a computer circuit that does computations without energy input?
Or does this break some inviolate law? Is there anything in the laws of thermodynamics / the laws of physics about computation? Does computation necessarily cost energy in and of itself? Is there some 'minimum perfect-efficiency cost' of thought? I can't find anything on this, but I might just be looking for the wrong terms . . .
To be clear, I'm not necessarily asking if this is possible with current technology. I'm asking if this is established and/or strongly-believed to be straight-up *impossible* by our current understanding of physics. Alternatively, if it is possible, what would make this difficult, and/or why haven't we done it (or have we?)?
Now, on earth you'd need energy to keep it cool, but out beyond the heliopause and into interstellar space, you could cool yourself down to superconductor-level temperatures with passive radiators (no?). Then you can just sit and think deep thoughts until entropy eventually gets around to rusting something important . . .
Erik Max Francis
Mar 21, 2013, 12:40:34 AM3/21/13
to
On 03/20/2013 09:37 PM, therealst...@gmail.com wrote:
> So superconductors don't produce waste heat when they have a current running through them. They have no resistance, so you can spin up a current in a circuit and it'll just keep going, around and around, forever (ok: for a long, long time). MRI machines use this.
>
> But how about superconducting computers? Could a superconducting transistor exist that has (functionally) no resistance and (functionally) no energy loss? If you can have a circuit that runs a current without needing to keep adding energy, could you put a computer in that circuit and keep it going? Could you have a computer circuit that does computations without energy input?
No.
> So superconductors don't produce waste heat when they have a current running through them. They have no resistance, so you can spin up a current in a circuit and it'll just keep going, around and around, forever (ok: for a long, long time). MRI machines use this.
>
> But how about superconducting computers? Could a superconducting transistor exist that has (functionally) no resistance and (functionally) no energy loss? If you can have a circuit that runs a current without needing to keep adding energy, could you put a computer in that circuit and keep it going? Could you have a computer circuit that does computations without energy input?
> Or does this break some inviolate law?
> Is there anything in the laws of thermodynamics / the laws of physics about computation? Does computation necessarily cost energy in and of itself?
--
Erik Max Francis && m...@alcyone.com && http://www.alcyone.com/max/
San Jose, CA, USA && 37 18 N 121 57 W && AIM/Y!M/Jabber erikmaxfrancis
The cosmos is a magnet. Once you have been there, all you can think
of is how to get back. -- Yuri Romanenko
Wayne Throop
Mar 21, 2013, 1:15:10 AM3/21/13
to
: therealst...@gmail.com
: But how about superconducting computers? Could a superconducting
: transistor= exist that has (functionally) no resistance and
: (functionally) no energy l= oss? If you can have a circuit that runs
: a current without needing to keep = adding energy, could you put a
: computer in that circuit and keep it going? = Could you have a
: computer circuit that does computations without energy inp= ut?
Actually a complicated question, but mostly... no.
Computations must be embodied in physical systems, and physical
systems are subject to thermodynamics (even if they have superconductors).
So, there's an energy (actually, and entropy) cost to slinging bits.
: Or does this break some inviolate law?
Sort of. http://en.wikipedia.org/wiki/Von_Neumann-Landauer_limit
On the other hand, there is great interest is finding loopholes
in this, hence http://en.wikipedia.org/wiki/Reversible_computing
But there are limits to what you can compute if it's entirely reversible.
Further, if you read out your answer, it makes reversibility difficult.
Bottom line, the issue isn't absotively posolutely settled, but
superconductors don't help. If you have a superconducting gate,
it takes energy to flip its state. And if you never flip states
of gates... not much computing going on.
: But how about superconducting computers? Could a superconducting
: transistor= exist that has (functionally) no resistance and
: (functionally) no energy l= oss? If you can have a circuit that runs
: a current without needing to keep = adding energy, could you put a
: computer in that circuit and keep it going? = Could you have a
: computer circuit that does computations without energy inp= ut?
Actually a complicated question, but mostly... no.
Computations must be embodied in physical systems, and physical
systems are subject to thermodynamics (even if they have superconductors).
So, there's an energy (actually, and entropy) cost to slinging bits.
: Or does this break some inviolate law?
Sort of. http://en.wikipedia.org/wiki/Von_Neumann-Landauer_limit
On the other hand, there is great interest is finding loopholes
in this, hence http://en.wikipedia.org/wiki/Reversible_computing
But there are limits to what you can compute if it's entirely reversible.
Further, if you read out your answer, it makes reversibility difficult.
Bottom line, the issue isn't absotively posolutely settled, but
superconductors don't help. If you have a superconducting gate,
it takes energy to flip its state. And if you never flip states
of gates... not much computing going on.
Stephen Hartley
Mar 21, 2013, 9:16:46 AM3/21/13
to
Interesting. This is pretty much what I was looking for.
Man. That limit is really, really, *really* low. My napkin-math says that if you run at the CMBR temperature, you can run a trillion operations from one fissioning atom of U-235.
Divide that by a couple hundred to account for inefficiencies and overhead, it still looks insane. Are there extant gates that even vaguely approach this level of energy efficiency? Are there theoretical designs that could (theoretically) do so?
Man. That limit is really, really, *really* low. My napkin-math says that if you run at the CMBR temperature, you can run a trillion operations from one fissioning atom of U-235.
Divide that by a couple hundred to account for inefficiencies and overhead, it still looks insane. Are there extant gates that even vaguely approach this level of energy efficiency? Are there theoretical designs that could (theoretically) do so?
gracefool
May 7, 2013, 8:35:02 AM5/7/13
to
You'll probably find interesting the paper "Ultimate physical limits to computation": http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.251.4494&rep=rep1&type=pdf
0 new messages