Superconducting CPU?

For the discussion of the sciences. Physics problems, chemistry equations, biology weirdness, it all goes here.

Moderators: gmalivuk, Moderators General, Prelates

p1t1o
Posts: 771
Joined: Wed Nov 10, 2010 4:32 pm UTC
Location: London, UK

Superconducting CPU?

Postby p1t1o » Fri May 26, 2017 2:28 pm UTC

Could there be such a thing as a superconducting CPU, which drew no power and produced no heat?

>-)
Posts: 512
Joined: Tue Apr 24, 2012 1:10 am UTC

Re: Superconducting CPU?

Postby >-) » Fri May 26, 2017 2:52 pm UTC

i don't see any reason why, here's the wikipedia page on it
https://en.wikipedia.org/wiki/Superconducting_computing

quantropy
Posts: 191
Joined: Tue Apr 01, 2008 6:55 pm UTC
Contact:

Re: Superconducting CPU?

Postby quantropy » Fri May 26, 2017 3:25 pm UTC

According to Landauer's principle computers have to dissipate a certain amount of energy when they 'forget' a bit of information. This is a very small amount of energy, about 2.7e-21 Joules per bit at room temperature. At superconducting temperatures it would be considerably less, but I think that the need for refrigeration apparatus to get rid of this energy would bring the minimum back up to the room temperature amount.

User avatar
Flumble
Yes Man
Posts: 1951
Joined: Sun Aug 05, 2012 9:35 pm UTC

Re: Superconducting CPU?

Postby Flumble » Fri May 26, 2017 3:31 pm UTC

[pseudoedit]ugh, always those ninjas[/pseudoedit]

Landauer's principle argues there's a minimum amount of energy required to do any (irreversible) computation. On the other hand, that minimum goes to zero as the temperature goes to zero, the principle says nothing about reversible computing and it is being challenged.

p1t1o
Posts: 771
Joined: Wed Nov 10, 2010 4:32 pm UTC
Location: London, UK

Re: Superconducting CPU?

Postby p1t1o » Fri May 26, 2017 3:44 pm UTC

quantropy wrote:According to Landauer's principle computers have to dissipate a certain amount of energy when they 'forget' a bit of information. This is a very small amount of energy, about 2.7e-21 Joules per bit at room temperature. At superconducting temperatures it would be considerably less, but I think that the need for refrigeration apparatus to get rid of this energy would bring the minimum back up to the room temperature amount.




I was wondering if there would be any minimum energy requirement for information handling...

The question was inspired by me wondering: my computer has an 800W power supply. Suppose I make it "compute" hard enough to draw all 800W.

Is it putting out exactly 800W of heat? (say for the moment that all heat produced is removed into a simple heat sink, negating having to consider the power consumption/heat generation of cooling motors, refrigeration systems etc.)

quantropy
Posts: 191
Joined: Tue Apr 01, 2008 6:55 pm UTC
Contact:

Re: Superconducting CPU?

Postby quantropy » Sat May 27, 2017 6:25 am UTC

p1t1o wrote:The question was inspired by me wondering: my computer has an 800W power supply. Suppose I make it "compute" hard enough to draw all 800W.
Is it putting out exactly 800W of heat?

Yes. Any energy fed into the computer as electricity is emitted as heat. It doesn't 'use up' any in doing the computation. What it does is to degrade the energy from useful electricity to much less useful heat, and hopefully it produces some useful computation in the process

Tub
Posts: 323
Joined: Wed Jul 27, 2011 3:13 pm UTC

Re: Superconducting CPU?

Postby Tub » Sun May 28, 2017 4:13 pm UTC

p1t1o wrote:Suppose I make it "compute" hard enough to draw all 800W.

Congratulations, your computer is on fire.

p1t1o wrote:Is it putting out exactly 800W of heat?

The energy contained in the computer when it isn't running doesn't change much, but it changes. For example, a magnetized bit on a hard drive has a slightly different potential energy than an unmagnetized bit. But those differences are small. So energy conservation tells you that if you put in 800W and the computer cannot store large amounts of energy, then it'll have to output 800W on average.

Some of that energy is useful information, traveling via your monitor cable and maybe network cable or wifi. Some of that energy is the light from your power LED. There's a lot of sound (especially when running at 800W), and the fans move a lot of air. Heat is still the dominant source of energy loss, but not the only one.

As for the CPU itself, energy has a lot less options to leave. There's no sound (except maybe some high-pitched hum from the capacitors), there's no light, there's no airflow, so anything that doesn't leave via the data buses is emitted as heat and heat-related effects (like thermal radiation).

morriswalters
Posts: 6949
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Superconducting CPU?

Postby morriswalters » Mon May 29, 2017 12:43 am UTC

p1t1o wrote:The question was inspired by me wondering: my computer has an 800W power supply. Suppose I make it "compute" hard enough to draw all 800W.

Is it putting out exactly 800W of heat? (say for the moment that all heat produced is removed into a simple heat sink, negating having to consider the power consumption/heat generation of cooling motors, refrigeration systems etc.)
You do of course realize, that from a power supplies point of view, a load is a load. The calculation is somewhat easier since computers use DC. So the question you're asking is, in effect, can I find a resistor that that will radiate at 800 Watts, given sufficient current from a 12 volt DC power supply? To which the answer is, of course you can.

User avatar
ucim
Posts: 5640
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Superconducting CPU?

Postby ucim » Mon May 29, 2017 3:23 am UTC

morriswalters wrote:So the question you're asking is, in effect, can I find a resistor that that will radiate at 800 Watts, given sufficient current from a 12 volt DC power supply? To which the answer is, of course you can.
No, the question is more like: If I find such a resistor, can I make it do useful work (like computing), and still get 800 watts of heat out of it?". Because if you can, you've gotten free energy.

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Please help addams if you can. She needs all of us.

User avatar
Copper Bezel
Posts: 2416
Joined: Wed Oct 12, 2011 6:35 am UTC
Location: Web exclusive!

Re: Superconducting CPU?

Postby Copper Bezel » Mon May 29, 2017 7:28 am UTC

Using the phrase "useful work" in the everyday sense of those terms in a physics topic seems potentially hazardous.

Edit: And very possibly the "free" in "free energy"....
So much depends upon a red wheel barrow (>= XXII) but it is not going to be installed.

she / her / her

speising
Posts: 2079
Joined: Mon Sep 03, 2012 4:54 pm UTC
Location: wien

Re: Superconducting CPU?

Postby speising » Mon May 29, 2017 9:20 am UTC

ucim wrote:
morriswalters wrote:So the question you're asking is, in effect, can I find a resistor that that will radiate at 800 Watts, given sufficient current from a 12 volt DC power supply? To which the answer is, of course you can.
No, the question is more like: If I find such a resistor, can I make it do useful work (like computing), and still get 800 watts of heat out of it?". Because if you can, you've gotten free energy.

Jose

You don't get free energy. The cost lies in converting higher order electrical energy to lower order heat energy. Your computer increases entropy and thus furthers the end of the universe. Well done.

User avatar
Flumble
Yes Man
Posts: 1951
Joined: Sun Aug 05, 2012 9:35 pm UTC

Re: Superconducting CPU?

Postby Flumble » Mon May 29, 2017 9:24 am UTC

speising wrote:You don't get free energy. The cost lies in converting higher order electrical energy to lower order heat energy. Your computer increases entropy and thus furthers the end of the universe. Well done.

Hey now, it did sort my list. Surely that counts for something?

Frenetic Pony
Posts: 177
Joined: Fri Jan 27, 2012 6:31 am UTC

Re: Superconducting CPU?

Postby Frenetic Pony » Mon May 29, 2017 9:36 am UTC

Copper Bezel wrote:Using the phrase "useful work" in the everyday sense of those terms in a physics topic seems potentially hazardous.

Edit: And very possibly the "free" in "free energy"....


Actually it's a useful term here. Most heat from any computer today is going to come from leakage. The actual "useful" work here is just the charge and discharge of the gates/running a charge through whatever memory type you have to check resistance/etc. That computers can use so much power really is, in this case, due to inefficiency. Graphene, as an example, has incredibly little leakage (little gets accidentally turned into heat) which means vs silicon graphene based semiconductors, and thus a graphene CPU would do much more "useful" computational work per unit of energy than silicon.

Though ultimately disorder would rule the day as "output" of course. Reading information from memory, any memory, or etc. still causes disorder (Maxwell's demon and all that). But useful work in this case is doing that vs just sending heat off into your room before doing any computational work.

morriswalters
Posts: 6949
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Superconducting CPU?

Postby morriswalters » Mon May 29, 2017 2:18 pm UTC

The whole point of my comment was to inform the poster that power has little to do with the act of computation. The question,
Suppose I make it "compute" hard enough to draw all 800W.
Isn't meaningful. The OP should look at the definition for electrical power. If he wants to know if computation breaks the Second Law, good luck to him.

mfb
Posts: 944
Joined: Thu Jan 08, 2009 7:48 pm UTC

Re: Superconducting CPU?

Postby mfb » Mon May 29, 2017 2:27 pm UTC

Current transistors need some amount of energy to switch. It is part of the design, replacing them by superconductors would not work. During the switching process, it can also happen that there is a short circuit (for a very short time), that would draw even more power with superconductors (although the difference would be small).

Transistors always have some leakage current, that needs some power.

Superconducting wires would eliminate losses in the wires, but the CPU would still draw significant power.

User avatar
ucim
Posts: 5640
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Superconducting CPU?

Postby ucim » Mon May 29, 2017 2:32 pm UTC

speising wrote:Your computer increases entropy and thus furthers the end of the universe. Well done.
Hey... I don't run Windows. Friends don't let friends run Windows! :)

morriswalters wrote:The whole point of my comment was to inform the poster that power has little to do with the act of computation.
...but it doesn't have nothing to do with the act of computation, and that was p1t1o's question's point. Computers are inefficient, sure, but more computing still draws more power. Perhaps the question would make more sense in reverse, which is what I attempted to do.

There is a certain minimum amount of energy that is theoretically needed to change the state of internal components in order to generate the answer; that answer will have lower entropy than any other value, simply by virtue of its being the answer. This requires entropy to be increased somewhere else; to wit: waste heat. But waste heat costs energy as well as entropy; that energy is not available to flip bits. This is the root of the inefficiency in the first place.

So, the useful work uses some energy, the waste work uses some energy, and the total energy is some given amount (800W, in this case).

The waste work is what's radiated (as heat and noise and such), but the useful work is not. So, there must be less than 800W of heat generated.

If this were not true, then we could sell 800W heaters that also, as a side effect, calculated pi for free. And while I like free desserts as much as the next guy, it ain't gonna happen.

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Please help addams if you can. She needs all of us.

User avatar
Copper Bezel
Posts: 2416
Joined: Wed Oct 12, 2011 6:35 am UTC
Location: Web exclusive!

Re: Superconducting CPU?

Postby Copper Bezel » Mon May 29, 2017 3:04 pm UTC

But we absolutely can. Remote the display and wireless communication so you're not losing anything to radiation that might get out of the room, and a computer is a 100% efficient heater, as all electrical devices are if you follow the chain long enough. The useful computation absolutely is spending entropy, not energy.

You can't even appeal to the hard drive as storing energy in one magnetic state or another, first because its state before and after is a normal number and therefore the proportion of 0s to 1s has not changed, second because it's bloody 2017 and why are you using a magnetic drive as your boot device it's unconscionable
So much depends upon a red wheel barrow (>= XXII) but it is not going to be installed.

she / her / her

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 25823
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Superconducting CPU?

Postby gmalivuk » Mon May 29, 2017 3:25 pm UTC

ucim wrote:that answer will have lower entropy than any other value, simply by virtue of its being the answer.
But whether a configuration is "the" answer depends on what question was asked. Do all answers to possible questions you can ask a computer have equal minimum entropy?
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
ucim
Posts: 5640
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Superconducting CPU?

Postby ucim » Mon May 29, 2017 5:24 pm UTC

gmalivuk wrote:But whether a configuration is "the" answer depends on what question was asked. Do all answers to possible questions you can ask a computer have equal minimum entropy?
Is there anything special about round numbers? The number line doesn't care - a point is a point is a point. Entropy is a measure of the relationship between the set of sequences that are possible vs the set of sequences that "look like" the sequence in question. But just because something "looks random" doesn't mean it is. (And doesn't mean it makes a good password). Nonetheless, whatever the answer is, it's the only thing that is the answer. And if "almost" counts, then "almost zero" also counts, and there are many ways to get "almost zero".

One thing special about "the" answer, is that it tells the computer when to stop. If it didn't, then the computer would be incapable of finding the answer, because recognizing it is part of finding it.

Copper Bezel wrote:it's bloody 2017 and why are you using a magnetic drive as your boot device it's unconscionable
Get off my lawn! I still use floppy drives sometimes.

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Please help addams if you can. She needs all of us.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 25823
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Superconducting CPU?

Postby gmalivuk » Mon May 29, 2017 6:00 pm UTC

ucim wrote:
gmalivuk wrote:But whether a configuration is "the" answer depends on what question was asked. Do all answers to possible questions you can ask a computer have equal minimum entropy?
Is there anything special about round numbers? The number line doesn't care - a point is a point is a point. Entropy is a measure of the relationship between the set of sequences that are possible vs the set of sequences that "look like" the sequence in question. But just because something "looks random" doesn't mean it is. (And doesn't mean it makes a good password). Nonetheless, whatever the answer is, it's the only thing that is the answer. And if "almost" counts, then "almost zero" also counts, and there are many ways to get "almost zero".

One thing special about "the" answer, is that it tells the computer when to stop. If it didn't, then the computer would be incapable of finding the answer, because recognizing it is part of finding it.
Can you clarify what any of this has to do with my question?
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

morriswalters
Posts: 6949
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Superconducting CPU?

Postby morriswalters » Mon May 29, 2017 8:25 pm UTC

ucim wrote:Computers are inefficient, sure, but more computing still draws more power. Perhaps the question would make more sense in reverse, which is what I attempted to do.
No. And I don't understand why this is so difficult. The whole point of the all the power used is to produce computation. How much power you use to do computation is a function of the efficiency of the device doing the computation. There may be some minimum imposed by the second law in terms of how efficient that device can be. And I don't know the answer to that question. But there is no dividing the power into that producing heat and that producing computation. All the heat is a product of computation.

User avatar
ucim
Posts: 5640
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Superconducting CPU?

Postby ucim » Tue May 30, 2017 1:16 am UTC

gmalivuk wrote:Can you clarify what any of this has to do with my question?
Sure, but I can't guarantee that it answers it.

Your question sounds like: "what has greater entropy, a shuffled deck or a brand new card deck in numeric and suit order?", to which the obvious answer is the shuffled deck. I'm questioning this answer after asking it badly. To ask it better, "which has greater entropy, the brand new deck, or one that is ordered by the amount of ink used?" The latter case "looks like" a shuffled deck, but is in fact a specific order; the unique order that answers the conditions (or question). So long as the answer is unique, all answers to all possible (such) questions should have the same entropy, because there is only one way for the answer to be the answer..

morriswalters wrote:All the heat is a product of computation.
Yes, but if one were designing a heater, could one get more heat per input watt by not computing? IOW, does doing useful work (such as computing) reduce the amount of heat produced? Certainly doing useful work such as raising a mass reduces the amount of heat produced per watt expended, because some of those watts got turned into potential energy, from which we could get heat later (or never).

It may be that the design of the computer is such that 800 watts can never be drawn (and the power supply is just oversized). But there is some theoretical maximum power draw, and from that point, "computing less" will draw less power. This is the basis of downspeeding laptops to extend battery life.

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Please help addams if you can. She needs all of us.

Frenetic Pony
Posts: 177
Joined: Fri Jan 27, 2012 6:31 am UTC

Re: Superconducting CPU?

Postby Frenetic Pony » Tue May 30, 2017 2:01 am UTC

morriswalters wrote:
ucim wrote:Computers are inefficient, sure, but more computing still draws more power. Perhaps the question would make more sense in reverse, which is what I attempted to do.
No. And I don't understand why this is so difficult. The whole point of the all the power used is to produce computation.


No, no it's not. Most isn't, it's just leaked off as heat due to inefficiency. I just said this. Shrinking gates (Moore's law) lowers the inefficiency of chips, this is why a 400 watt computer from 12 years ago fits inside a phone that draws 5 watts today. A computer is perfectly efficient at turning electricity to heat (minus power output to like, your monitor or USB charging device). It is NOT perfectly efficient at computation. "Useful work" vs not useful work is perfectly applicable, all power drawn "is used for computation" is like saying all power from a gas engine is used to make a car go forward. That a bunch of heat and stuff is also produced over the given short time period T is just part of the car going forward!

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 25823
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Superconducting CPU?

Postby gmalivuk » Tue May 30, 2017 3:05 am UTC

ucim wrote:
gmalivuk wrote:Can you clarify what any of this has to do with my question?
Sure, but I can't guarantee that it answers it.

Your question sounds like: "what has greater entropy, a shuffled deck or a brand new card deck in numeric and suit order?", to which the obvious answer is the shuffled deck. I'm questioning this answer after asking it badly. To ask it better, "which has greater entropy, the brand new deck, or one that is ordered by the amount of ink used?" The latter case "looks like" a shuffled deck, but is in fact a specific order; the unique order that answers the conditions (or question). So long as the answer is unique, all answers to all possible (such) questions should have the same entropy, because there is only one way for the answer to be the answer..
Are there orderings of a deck of cards that aren't answers to such questions?
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
Copper Bezel
Posts: 2416
Joined: Wed Oct 12, 2011 6:35 am UTC
Location: Web exclusive!

Re: Superconducting CPU?

Postby Copper Bezel » Tue May 30, 2017 4:21 am UTC

ucim wrote:Yes, but if one were designing a heater, could one get more heat per input watt by not computing?

No.

All electric heaters are equally 100% efficient. If there's no energy being transmitted away, and no energy stored, then all energy consumed must be released as heat, because that is the only other place it can go. The only way to build a less efficient heater is by sinking the energy somewhere.

Frenetic Pony wrote:No, no it's not. Most isn't, it's just leaked off as heat due to inefficiency. I just said this. Shrinking gates (Moore's law) lowers the inefficiency of chips, this is why a 400 watt computer from 12 years ago fits inside a phone that draws 5 watts today. A computer is perfectly efficient at turning electricity to heat (minus power output to like, your monitor or USB charging device). It is NOT perfectly efficient at computation. "Useful work" vs not useful work is perfectly applicable, all power drawn "is used for computation" is like saying all power from a gas engine is used to make a car go forward. That a bunch of heat and stuff is also produced over the given short time period T is just part of the car going forward!

The efficiency of an engine is the proportion of energy it manages to transform from one form into another, as opposed to the portion that is lost to heat. It's a ratio that is less than, but can be arbitrarily close to, one. Computation is not a form of energy, and there is no meaningful definition of a perfectly efficient computing device. Whether your computer is ENIAC, an Intel Xeon workstation, or a completed copy of Babbages' analytical engine, electrons or switches or gears move around inside in patterns and produce nothing physical but waste heat, and the efficiency is expressed purely in computations per watt. There's no proportion of the energy put in that "gets through" to transform into some other thing, your "useful work", exactly all of it is used in exactly the same way, no matter how "efficient" or "inefficient" the machine.
So much depends upon a red wheel barrow (>= XXII) but it is not going to be installed.

she / her / her

morriswalters
Posts: 6949
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Superconducting CPU?

Postby morriswalters » Tue May 30, 2017 4:58 am UTC

Consider for a moment the abacus. All the power used in computation is used to physically move the beads. The data is agreement on what the positions at some stage represent. The computation correlates with the power used to arrive at the resultant. But the energy was actually consumed by moving the beads(causation). Anyway I don't have anything useful to add past this.
Copper Bezel wrote:All electric heaters are equally 100% efficient. If there's no energy being transmitted away, and no energy stored, then all energy consumed must be released as heat, because that is the only other place it can go. The only way to build a less efficient heater is by sinking the energy somewhere.
Thank you.

User avatar
Eebster the Great
Posts: 2770
Joined: Mon Nov 10, 2008 12:58 am UTC

Re: Superconducting CPU?

Postby Eebster the Great » Tue May 30, 2017 5:56 am UTC

ucim wrote:There is a certain minimum amount of energy that is theoretically needed to change the state of internal components in order to generate the answer

That may be true, but the energy doesn't disappear. The "energy that is needed" is just heat. Total energy is conserved.

speising
Posts: 2079
Joined: Mon Sep 03, 2012 4:54 pm UTC
Location: wien

Re: Superconducting CPU?

Postby speising » Tue May 30, 2017 6:38 am UTC

A computer uses energy the same way as a waterwheel uses water.

The wheel is driven by the flow of water, and performs useful work. But it doesn't use up any water, the volume is just the same down- as upstream.
What it does use is kinetic energy, it transforms the flow from fast to slower, as the computer uses entropy (~), and transforms the energy from ordered to unordered.

quantropy
Posts: 191
Joined: Tue Apr 01, 2008 6:55 pm UTC
Contact:

Re: Superconducting CPU?

Postby quantropy » Tue May 30, 2017 11:18 am UTC

I wonder sometimes whether nVidia or the like might make a bunch of last year's GPU's into a computer and sell it as a heater. My idea would be no interface for the owner, rather it just connects to the internet and takes part in a distributed computing project when it's turned on.

User avatar
Flumble
Yes Man
Posts: 1951
Joined: Sun Aug 05, 2012 9:35 pm UTC

Re: Superconducting CPU?

Postby Flumble » Tue May 30, 2017 11:27 am UTC

Sounds like a great idea. You can even market it as an IoT device. (You probably need to, because most people don't just buy a really expensive heater, even if it promises to do science.)

User avatar
Copper Bezel
Posts: 2416
Joined: Wed Oct 12, 2011 6:35 am UTC
Location: Web exclusive!

Re: Superconducting CPU?

Postby Copper Bezel » Tue May 30, 2017 11:30 am UTC

And you'd have to keep it expensive, else your free electricity scheme turns into people junking your "heaters" for their juicy GPU innards.

Plus, you need to connect to the user's WiFi, so you're not losing everything you're making in cellular data bills. Easiest way to do that is to ask and provide an interface.
So much depends upon a red wheel barrow (>= XXII) but it is not going to be installed.

she / her / her

morriswalters
Posts: 6949
Joined: Thu Jun 03, 2010 12:21 am UTC

Re: Superconducting CPU?

Postby morriswalters » Tue May 30, 2017 1:34 pm UTC

quantropy wrote:I wonder sometimes whether nVidia or the like might make a bunch of last year's GPU's into a computer and sell it as a heater. My idea would be no interface for the owner, rather it just connects to the internet and takes part in a distributed computing project when it's turned on.
You've already bought the heater, why use last years parts? And you can help me with a distributed computing problem by allowing me to use your all your computing devices to mine bitcoin. You get warm and I get rich. Thank you.

User avatar
ucim
Posts: 5640
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Superconducting CPU?

Postby ucim » Tue May 30, 2017 1:37 pm UTC

Copper Bezel wrote:All electric heaters are equally 100% efficient. If there's no energy being transmitted away, and no energy stored, then all energy consumed must be released as heat, because that is the only other place it can go. The only way to build a less efficient heater is by sinking the energy somewhere.
Ok.

gmalivuk wrote:Are there orderings of a deck of cards that aren't answers to such questions?
Probably not. Is a new deck in a lower entropy than a shuffled deck?

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Please help addams if you can. She needs all of us.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 25823
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Superconducting CPU?

Postby gmalivuk » Tue May 30, 2017 1:57 pm UTC

If every arrangement of cards is the answer to such a question, then according to you they all have the same entropy.
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
measure
Posts: 111
Joined: Sat Apr 04, 2015 4:31 pm UTC
Location: Time-traveling kayak

Re: Superconducting CPU?

Postby measure » Tue May 30, 2017 2:49 pm UTC

gmalivuk wrote:If every arrangement of cards is the answer to such a question, then according to you they all have the same entropy.

FWIW, some such questions are considerably more complicated than "sort by suit, then by rank" (for appropriate measures of complexity).

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 25823
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Superconducting CPU?

Postby gmalivuk » Tue May 30, 2017 3:21 pm UTC

Sure, but ucim's claim didn't refer to differences in the questions.

In any case, however much entropy it has, the arrangement of bits in a computer's memory can only account for up to some constant amount of energy. Asymptotically, then, the average heat output over time equals the electrical input.

Edit: To continue the cards analogy, the energy it takes to shuffle or otherwise rearrange the deck isn't stored in the final arrangement. If you keep adding energy to keep rearranging the deck, you don't end up with any more stored as a result. All the power (energy-over-time) it takes to move cards is lost as heat.
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
Soupspoon
You have done something you shouldn't. Or are about to.
Posts: 2544
Joined: Thu Jan 28, 2016 7:00 pm UTC
Location: 53-1

Re: Superconducting CPU?

Postby Soupspoon » Tue May 30, 2017 10:14 pm UTC

As an aside, as something that just occured to me, does every possible combination of cards relate to a (say) 6-bit Linear Feedback Shift Register (or similar mechanism to generate >52 unique states1 in order to 'extract' cards, exactly once, from an arbitrary-but-consistent2 starting condition) that can be defined by <initial bits> + <logic taps>? Because, if so, every single deck combination can be defined as the answer to at leas the question "taking the starting condition of <initial bits> and the <logic taps> configuration, what resulting pack outputs?" Whatever other questions this particular sort-order (like "alphabetically sorted according to the common English name") answers.

Let's work it out. Umm, number of possible pack combinations is 52!, and we have to match that. That value is 8e67ish

Initial bit-states are 63, with one 6-bit value (all zeros for XOR, or all ones for XNOR) is an unchanging sequence of one. (There's a complication in that sequence n1>n2>n3…n63>#HALT# is effectively the same as n2>n3…n63>n1>#HALT# if n1 is a skip-over number, etc... Ignoring that for now.) This is a trivial number that hardly matters, but let's say that the remainder of our decisions must account for 1.3e66ish permutations.

"Ignorable" numbers are the all-zero (or all-one) combination by default, plus our choice of eleven-from-sixtythree numbers to skip. Again, choosing a skipable number - or, worse, a series of skippable numbers - that happen to be located at one and/or the other end of the sequence we'll define could cause duplication of results for multiple initial conditions. To solve this problem, I propose we exclude from the "ignore list" the very first number of our sequence (so that it cannot duplicate an alternate "question format" where the sequence turns out identical except for rotation of (some of) the ignorable(s) at the head of the sequence now being represented at the tail-end. So eleven-from-sixtytwo is 5e11ish, I think. We now account for 2.5e54ish combinations.

PPE:
Spoiler:
There are other 'duplicated results from differing initialities' to consider. If the raw sequence somehow contains two numbers that are numerically adjacent, and we take the rather basic assumption that an Ignorable number is skipped over both when encountered in the sequence and in our setmof unignored-numbers as mapped onto the cards, then the choice-from-eleven difference between the one number and its neighbour merely says something like "instead of skip 12, and 13 is the King Of Clubs, 12 is the King Of Clubs and we skip 13". Or whatever numbers 'work' in the sequence. Although I haven't studied LFSRs enough to confirm that sequential-adjacency can be preserved. Likely not. Thus possibly another reason (see below) why we cannot provide every single shuffling (or, indeed, unshuffling, e.g. sequentially 1>2>3>4…50>51>52 from initial condition of 000001 whilst ignoring 53…63, by some rather specific "adder-like" tap-combos) from the initial conditions. But that conclusion is yet to be arrived at. This is by way a post-facto spoiler...
(Read after finishing the rest...)

Something tells me that "well chosen" taps from the whole phase-space of possible logic tap combinations just isn't going to cut it. Even before we cut out the non-m-sequential tap combinations, we're tapping, or not, each of those 6-bits, before passing on (sequentially) to XOR with the next-most-significant-chosen-tap, and that'd be 64 (mostly useless) combinations. I don't think it's worth it trying to calculate (or just generate!) the number of globally-relatively-prime combinations, which I suspect would be a polynomially-inspired formula, but would winnow down the number to significantly less than 64 (even with the option of using XNORs across the board instead of XORs, for any given valid tap-pattern, which would double the possible tap profiles we can apply, at a stroke).

So... that seems like (maybe an order of magnitude more than!) 4e52ish additional (unmappable) pack-orderings for every variation of Question that I think I could generate, given my seemingly rather naïve postulation.

I wonder what else I could add to the question to include roughly 52 orders of magnitude of additional starting options, thus to match "a valid Question for every Answer"..?

Or maybe I've made a silly mistake, as I write this (as you will have deduced) entirely on the fly, very much being the stream of my deliberations.

(I started this wondering if the Interesting Number Paradox had its counterpart in this issue, with some Unexpected Hanging Paradox thrown in as well. Given a finite set, and some agreed-upon way of singling out set-members as being the nth most non-boring element by having n-1 elements already identified as more immediately non-boring, then all the way up to the very last of those elements it must be possible to say "and this element is identifiably the most uninteresting element, and if I told you to look for the most uninteresting element you would identify this element as it". With each element being a specific card-(dis)order.)


1 And, in this case, drop/ignore 12 of those states. Perhaps not just off the end, and thus you supply
2 Has to be consistent, however arbitrary in other regards, otherwise the answer is to have a single, static LFSR mechanism operate identically upon every possible starting conditions of the deck, to get a 1->1 translation onto every possible deck result.

quantropy
Posts: 191
Joined: Tue Apr 01, 2008 6:55 pm UTC
Contact:

Re: Superconducting CPU?

Postby quantropy » Wed May 31, 2017 11:00 am UTC

Copper Bezel wrote:And you'd have to keep it expensive, else your free electricity scheme turns into people junking your "heaters" for their juicy GPU innards. Plus, you need to connect to the user's WiFi, so you're not losing everything you're making in cellular data bills. Easiest way to do that is to ask and provide an interface.

I don't think it would be that hard to prevent users from using the GPUs for their own purposes (well maybe not users who are really dedicated, but the effort they put in would probably be better spent earning money and buying the latest GPU board). I'm assuming that the a lot of the cost of GPUs is development cost, so manufacturers could afford to sell them cheaply if they were happy that this wouldn't affect sales of the latest version. Older GPUs tend to be less energy efficient, but this is being sold as a heater. Yes, it would cost quite a bit more than a simple fan heater, but you'd be surprised what people are prepared to pay for electric heaters with a fancy name: https://www.amazon.co.uk/d/ajl/Electror ... B00FEEU9US

I would see it as allowing the user to enter their WiFi details, but have no other user interaction.

User avatar
HES
Posts: 4796
Joined: Fri May 10, 2013 7:13 pm UTC
Location: England

Re: Superconducting CPU?

Postby HES » Wed May 31, 2017 11:29 am UTC

quantropy wrote:I would see it as allowing the user to enter their WiFi details, but have no other user interaction.

You're proposing a heater where the user cannot change the heat. At least throw in some fancy smartphone temperature control to justify the markup.
He/Him/His Image

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 25823
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Superconducting CPU?

Postby gmalivuk » Wed May 31, 2017 11:55 am UTC

Yeah, what's the point of an IoT device that doesn't have gaping security holes?
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)


Return to “Science”

Who is online

Users browsing this forum: Bing [Bot] and 10 guests