'Reversible' Computers More Energy Efficient

James Clark writes "As Congress continues work on a federal energy bill, a group of University of Florida researchers is working to implement a radical idea for making computers more energy efficient -- as well as smaller and faster." Reversible computing rears its head again.

Comments

Vaporware?

[Carnildo]

Has anyone ever built even a very simple reversible computer? Or is this like quantum computers: all theory, no practice?

Re:Vaporware?

[nestler]

This is more practical than quantum computers because it is much easier to build and can be used for general purpose things other than search and factoring.

The idea is to (down at the gate level) keep everything reversible. For example, current OR gates are not reversible (given a true output you can't definitively tell what either input was individually). If you have two outputs on the gate instead of one, you make the gate reversible. However, since you are just using it for OR, you are free to ignore the second bit you added on to make it reversible.

The bit doesn't help your computation in the sense of the answer you are looking for, but it can make things more energy efficient at the gate level.

Re:Vaporware?

[stoolpigeon]

The answers to your questions are in the article - here:

Frank, who first worked on reversible computing as a doctoral student at the Massachusetts Institute of Technology, heads UF's Reversible & Quantum Computing Research Group. Among other recent publications and presentations, he presented three papers dealing with topics related to reversible computing this summer, including "Reversible Computing: Quantum Computing's Practical Cousin" at a conference in Stony Brook, N.Y.

and here:

Frank currently is trying to persuade major chipmakers to direct more of their research-and-development resources toward reversible technologies.

Re:Vaporware?

[randyest]

I think you completely misunderstood the article, though in your defense it didn't do a very good job of explaining. The idea is not to be able to reverse logical operations -- that is of little value to anyone. Rather, they're trying to make the electrical changes (the energy transfer) reversible. That's a fundamentally differeent thing. A decent analogy, mentioned in the article is:

The concept is somewhat analogous to hybrid cars now on the market that take the energy generated during braking and recycle it into electricity used to power the car.

So, the logical realm is no different here. Physically, and electrically, there is a big difference from existing computers. Now, when a bit changes from 1->0, the voltage (accumulated charge) is simply shorted to ground (via resistive path that dissipates heat). That energy is lost. In a reversible computer, that charge would be stored, in the electrical equivalent of a spring or flywheel in a mechanical system. So, next time it needs to go 0->1, the energy is sitting there, ready to be re-used(stored in the spring's compression or flywheel's rotation).

I assume these electrical "springs or flywheels" need to be phsycally close to the transistors they're storing energy for. If all transistor's storage were common, the heat loss (and time delay) to get the energy back to where it's needed would defeat the entire purpose.

In the article, they mention that current prototypes use oscillators to store the energy (which are more like a flywheel than a spring, to continue the mechanical analogy), but the efficiency is not quite good enough to be called "reversible". Too much energy is lost in storing and un-storing the energy. The current work is focused on improving the efficiency of storing and un-storing energy from state changes.

However, as a chip designer, I know that oscillators are usually (1) much much bigger than simple logic gates and (2) much more difficult to design with (it's analog design stuff, really). So, my concerns are (1) how much bigger will dice need to be to use this system (linear increase in die size equals exponential increase in manufacturing cost) and (2) how much longer is it going to take to close a design with all those little analog cells all over the place.

I don't even want to think about the implications for STA (static timing analysis) or LVS (layout versus schematic verification) -- it makes my head hurt. :)

Re:Vaporware?

[Chris Burke]

The bit doesn't help your computation in the sense of the answer you are looking for, but it can make things more energy efficient at the gate level.

If you know how it is more energy efficient, please share. The article doesn't contain anything that actually suggests "power savings" to me.

So you have a reversible OR gate. So then you can make an Un-OR gate. Okay, so you can recover what you had before the OR. Big deal. How have you saved power? You've just done another computation through a gate that uses power.

This line from the article: "For example, when a computer 'erases' something, what it does physically is ground one part of a circuit that holds a charge, in effect converting the stored energy -- and the information it represents -- into heat, Frank said." is a red herring. Every CMOS combinational logic circuit involves shorting its output to one of the voltage rails. You're always "erasing", by forcing the voltage of a wire to a certain level.

The heat that a computer generates comes from current passing through resistive loads. If you "undo" an operation, and the voltage level of the undo result is different than the existing voltage on the wire, then you have current and thus power consumed.

So there really isn't anything to be gained just by having reversible gates. Okay, I can imagine some microarchitecture which had some power advantage due to being able to reverse some things, but that's not exactly revolutionary.

They hint in the article, however, at something more than just reversible logic. Something about "oscilators" which are "spring-like" which says to me "capacitor". But they claim to be able to capture the energy consumed in an operation and re-use it. I don't know why you'd call that "reversible" computing; sounds like "recyclable" computing to me. Still, there's basically no detail at all on what that even would mean. Are they claiming to be able to capture the heat dissipated? Or just re-use the charge stored on one wire to compute another? That's just pass gate logic. Woopdy-doo.

I'll just say I'm skeptical. Especially when I read this: "Reversible computing is absolutely the only possible way to beat this limit," he said.

Uh-huh. That doesn't sound like a scientist. That sounds like a scientist fishing for grant money.

If you've got more information, I'm interested. This article, though, convinced me not at all.

Re:Vaporware?

[olip]

I don't even want to think about the implications for STA (static timing analysis) or LVS (layout versus schematic verification) -- it makes my head hurt. :)

or heat maybe ?

Re:Vaporware?

[CXI]

Thanks for the good description. I'm still having trouble with one part of the concept. Doesn't the heat come from the flow of electricity? You are not reducing the flow with this scheme, so wouldn't it generate the same heat?

For example, consider the concept of attaching a flywheel or superconductor such that the ground of your entire circuit assembly, instead of dumping to ground, instead served to power up your storage device. This power is then fed back into the system. To me, this is the exact same concept as is presented in the article, but the current flow is still there, so the heat would still be there in your circuit. How can moving the storage close to the gate change that? Power must still flow through resistive wires and therefore generate heat!

Re:Vaporware?

[malfunct]

There would still be heat generated, but I think the key is that by storing the energy after it passed through a gate so that it may be reused we can get more decisions made with the same amount of energy.

Right now it takes one "pulse" of electricity to make one decision, with reversable logic it seems that you could get more, maybe 2 "pulses" for ever 3 decisions because you could save 1/2 the energy from each of the first two decisions which would allow you to pay for the third.

Heat would still be genererated in the process but instead of blowing 3 "pulses" worth of electricity off into heat you would only lose 2 to make the same number of decisions.

Re:Vaporware?

[cgb8176]

I think you completely misunderstood the article, though in your defense it didn't do a very good job of explaining. The idea is not to be able to reverse logical operations -- that is of little value to anyone. Rather, they're trying to make the electrical changes (the energy transfer) reversible. That's a fundamentally differeent thing.

Actually, you are wrong, in that the two things are very intimately related. I will assume that, as a chip designer, you are aware of what AND, OR, and NOT gates are, and that NAND is an example of a universal gate. NAND, however, is not reversible; you cannot in general determine the inputs by looking at the output.

The Fredkin Gate is an example of a reversible gate. As it happens, it is impossible to do reversible computing with two input gates. The Fredkin Gate (a controlled swap; two inputs, two outputs, and a control wire that passes through) has the property that it is

reversible (Fredkin inverts Fredkin), and

it has the same number of non-zero outputs as it does non-zero inputs.

To achieve reversible computing, you need reversible gates. Furthermore, with reversible gates, you can perform any computation with an arbitrarily small amount of energy; the catch is that you need more time (see adiabatic circuits, Carnot engines).

Re:Vaporware?

[Bingo Foo]

Sorry, but reversible computing is about having N distinct ouputs for N distinct inputs in any logical operation. Think thermodynamics and statistical mechanics, where reversibility is intimately coupled with "no production of entropy" which means "no loss of information."

It is at the information theory and logic level of description where reversible computing must be implemented.

Re:Vaporware?

[Chris Burke]

Thanks. You helped make some sense of the article.

Now, when a bit changes from 1->0, the voltage (accumulated charge) is simply shorted to ground (via resistive path that dissipates heat). That energy is lost. In a reversible computer, that charge would be stored, in the electrical equivalent of a spring or flywheel in a mechanical system. So, next time it needs to go 0->1, the energy is sitting there, ready to be re-used(stored in the spring's compression or flywheel's rotation).

So it is a capacitor, basically. Instead of using the power and ground rails to drive a signal, you either sink the stored charge of the wire onto a big capacitor for the 1->0 transition, and you use that stored charge to drive the next 0->1 transition. Of course it still needs an active power source, since capacitors leak and wires have resistance. I'm starting to see how it could save power, though.

If all transistor's storage were common, the heat loss (and time delay) to get the energy back to where it's needed would defeat the entire purpose.

Well, plus you have the problem that not every circuit is going to be making the same transition.
So you'd pretty much need one per signal.

In the article, they mention that current prototypes use oscillators to store the energy

Perhaps you know more of analog circuitry than I (not a tall feat), but I can't fathom what this could be that would make sense. The very concept of "oscilation" implies "movement" which when applied to charge means current and thus power loss. So unless the "oscilation" is just the charge going from the wire to the storage unit and back, it doesn't make sense to me. That's why I guessed capacitors. :)

how much bigger will dice need to be to use this system

Well, just take any design and replace every Vdd/Gnd connection in your CMOS gates with a big capacitor or oscillator or whatever. Yeah, I think you have reason for concern.

Re:Vaporware?

[Analogy Man]

Conceptually a better analogy may be balancing a circuit panel in your home. One doing this you want to put appliances and other circuits on opposite phases. This can be generalized to multi-phased systems as well. In principle the load from one phase "balances" the load on another phase so the current to neutral (which goes to ground at the main panel) is minimal.

Do a google on phase balancing or look at this URL that has a diagram: http://www.tpub.com/ceb/74.htm

If ignorance is bliss why are so many people in such a bad mood?

Re:Vaporware?

[PD]

Ideal computers aren't as wasteful as the ones we have now. Most of the heat generated is from resistive loads, as you mentioned. But some of that heat is from the actual computation done. There is a theoretical minimum energy of computation, and when you approach that the ability to reverse your calculation becomes useful. When you add 1+1, that takes some energy to get 2. But, when you reverse that computation you can reclaim the energy. The result is that less energy is wasted to ultimately become heat.

This scientist isn't the first person to think of these things. I don't know for sure, but perhaps Claude Shannon first thought of this.

Re:Vaporware?

[Weaselmancer]

Ah, thanks. I skimmed the article and didn't get it either. And I'm a BSEE. =)

That being said, it sounds like a pretty decent idea. But since you're actually in the know, any idea why they'd want to use oscillators in the first place?

My first thought would be to store the static charge in some kind of switched capacitor type thing - like maybe a single bit of DRAM.

Weaselmancer

Re:Vaporware?

[Chris Burke]

Furthermore, with reversible gates, you can perform any computation with an arbitrarily small amount of energy; the catch is that you need more time (see adiabatic circuits, Carnot engines).

Hey, thanks for the keywords. Google turned up lots of nice stuff.

Though that catch is rather a big one. According to the links, as E->0, T->infinity, which I don't like one bit. Arbitrarily low power, but arbitrarily lengthy computation.

So I've now got my own low-power logic idea. I call it the "apathetic circuit", and it works by not doing the computation at all. Same zero energy/infinite time tradeoff, but with the advantage that the basic "meh" gate can be arbitrarily small even to the point of zero area! :)

Re:Vaporware?

[cgb8176]

That's why it's Toffoli-Fredkin :)

Re:Vaporware?

[Chris Burke]

Okay, I'll admit information theory isn't my strong suit.

But I can't say I can agree that the "actual computation" is consuming energy beyond what is converted into heat by the resistive load. The computation is currents passing through loads. The only reason it's a "computation" is because we find meaning in the output.

I thought that the theoretical minimal energy of computation was based on quantum physics and the minimum energy to induce a state change in something. "Information" isn't a medium unto itself; it is carried by physical mediums, and those mediums have a minumum energy to change. Granted, I could be quite wrong.

The energy reclamation part is interesting, but really has nothing to do with reversible computing as far as I can tell. A power supply that can recycle any energy fed back into it may be necessary for reversible computing, but is a pretty great idea regardless.

Re:Vaporware?

[tho 1234]

Are you really a chip designer? 1- In CMOS technology (or any other logic type used in the last 20 years) there is absolutely no resistive path to ground. (except for gate leakage) Two complementary (the C in cmos) PMOS and NMOS transistors are used to eliminate the need for any resistive branch. 2-voltage is not "accumulated charge", its a difference in potential energy. Changing voltage levels in itself does not cause any power to be lost, and a logic level 0 certainly isn't produced by shoring vcc to ground. Power is consumed in a logic circuit precisely because it stores energy- the capacitance of the transistor causes charge to be stored and later released to ground whenever the voltage level changes. Power consumtion in chips has been reduced over the last 30 years by reducing the amount of stored energy- by making transistors smaller and reducing capacitance. Technically, you can create an oscillator by adding an inductor to the circuit, but that would increase complexity/cost with little benefit in itself. I am not familiar with reversable computing, but i would expect they would need a substantial change in logic structure to extract the stored energy. Also, without a change in logic structure, it would simply be a process improvement, not an entirely new branch of computing. Yes, analog design is more involved than digital design, but the layout and composition of transistors, capacitors, interconnects, etc on any digital logic circuit are analog in themselves. THe only reason you don't have to deal with them is they are generated automatically with CAD tools. If resonators were found to be a viable way to decrease power consumtion, they could be easily added to your CAD tools and make their design just as simple as what you're used to. However, i'm sure its not as simple as adding a resonator to the circuit, and most likely requires an entire new method of computing to extract the stored energy.

Re:Vaporware?

[Dashing Leech]

...oscillators to store the energy (which are more like a flywheel than a spring, to continue the mechanical analogy)

Technically, it's both. In system modeling (e.g., Bond graphs, which is an excellent way to compare different domains), an inductor is analogous to a mass (stores energy in motion/flow) and a capacitor is analogous to a spring (stores energy statically). (Note that gravity acts like a non-linear softening spring.) Oscillators tend to make use of both.

Re:Vaporware?

[PD]

First paragraph: OK, information theory isn't my strong suit either. I know a little about it.

Second paragraph: It's OK not to agree, but you'd be wrong. :-) Information has a definition that is not the same as the definition of "datum". One is a physical/mathematical thing with real existence and a property called entropy, and another one called energy, and some other properties too, no doubt. The other (datum) is what you described.

Third paragraph: Now you're talking! What will really bake your noodle is when you read about what astronomers believe about the information state of a black hole. i.e. no clue, many interesting guesses.

Fourth paragraph: Don't think electricity here. We're could be talking about molecular computers, so the power supply could BE heat in all liklihood.

And that just about covers the extent of my knowlege of the subject. I'd like to learn more, but franky my current computer just isn't warm enough for me to bother right now.

Re:Vaporware?

[randyest]

The purpose is not reversible logic. The purpose is reversible electrical charge movement and re-use. The fact that most current efforts to make the energy transfer reversible also happen use reversible logic gates does not mean it's the only way to do it.

I'm familiar with the infomation = energy theories which can be interepreted to mean this is the case, but I still don't buy that it's the only way. And I'm not alone. It'll probably be developed further using reverisble gates first, but then we'll find a way to avoid that annoying port bloat and time delay. Which means something else besides reversible logic will be used, because reversible gates must have 1:1 in:out port (which sucks) and the time to compute goes toinfinity as the power drops to 0 (which sucks even more). Not a good trend for something that has the goal of lowering power in circuits that are expected to get faster, or at least not slower.

Please note that many posters were highly confused about why reversing logic would have any effect whatsoever on power usage. In most cases this confusion arose from a lack of understanding of the theorized relationship between information and energy. In my case, however, I'm familiar with the idea, but I just don't believe the theories are completely sound, or at least not complete.

Re:Vaporware?

[randyest]

I realize that many people believe that. I don't. And there's no proof, theoretical or empirical, that this is the case. Universities are focusing on this approach, and they are making some decent progress lately. The industry, however, is looking for alternate implementations that do not require (1) a one-to-one ratio of inputs to outputs or (2) computation time that approaches infinity as the power consumption goes to zero. The more the uni's learn, the more it helps us all, but I'm pretty confident that commercial implementations will get 60-80% of the power savings without the hassle of extra ports and long computation delays.

If you have any proof whatsever that "near-zero-power computing" requires reversible logic gates I'd love to see it. AFAIK none exists.

Re:Vaporware?

[randyest]

0) Yes, for the last 10 years. 1) There is leakage through the substrate, and that path is resistive -- though you're correct that the majority of the current flow is not really to "ground" (which, BTW, is a relative thing), rather use in charging/discharging parasitic capacitors -- the relevent fact is that switching dissipates power in the form of heat. Sometimes simplifications are necessary to get a point across. If you want to be a pedant, we can note that there are no 1's or 0's, whip out our detailed SPICE models to explain an AND gate, and worry that there are no "holes", just electrons, though we assume there are in our models, etc. Were this "EEdot", I might have been more precise, but probably not, since it doesn't matter for the understanding of this issue. 2) You can't have a potential difference without some charge accumulation -- chicken: meet egg. This is semantic.

Bah, I can't keep up that formatting. :). As for inductors: they're really hard to make out of silicon. I'd love to hear about how you'd implement that without using way more wiring tracks than it's worth.

You're close on that last point -- the reason I don't have to deal with analog design (too much) is not that my cad tools do it, rather that we can (safely) simplify our models of logic gates down from the complex analog circuits they are and treat them as digital logic. Accounting for crosstalk, transition times, wire delays, signal integrity, etc. outside of the primary design flow allows us to maintain this gross simplification throughout most of the design flow. EDA tools that can handle a new idea usually follow initial real-world implementations by at least a few years. Check out the hierarchical design tools available over the last 5 years, then look at the hierarchical tapeouts for the last 5 years. The tools were (and are, largely) crap. But the chips work (each of mine included).

One of my favorite design tools is Perl. :)

Re:Vaporware?

[Goldsmith]

You're arguing apples and oranges.

Physics does not care what we define as output, or information.

If you have one output, and one "energy sink" which would store the energy (as the parent says the real catch is to do that without changing it at all), you essentially have two outputs. In essense, the question is, what do you do with your outputs, put it to use, or let it go?

As you point out, the engineering of such logic devices might be a bit difficult.

Re:Vaporware?

[onomatomania]

1- In CMOS technology (or any other logic type used in the last 20 years) there is absolutely no resistive path to ground. (except for gate leakage) Two complementary (the C in cmos) PMOS and NMOS transistors are used to eliminate the need for any resistive branch.

Yeah, no shit sherlock. Just because there are no explicit resistors drawn in the circuit doesn't mean that the stored charge isn't dumped to ground through a resistive path. When the NMOS gates turn on, they're effectively shorting the stored charge in the load capacitance to ground through the ON resistance of the gate. And similarly, when the PMOS gates turn on, they charge the load capacitance through the supply rails in an analogous manner.

So just because there aren't explicit resistors (thanks to complimentary logic) doesn't mean that the charge isn't effectively being just supplied to a temporary store and then dumped to ground though resistive paths, which is what the original poster was saying.

Re:Vaporware?

[Chris Burke]

Second paragraph: It's OK not to agree, but you'd be wrong. :-) Information has a definition that is not the same as the definition of "datum". One is a physical/mathematical thing with real existence and a property called entropy, and another one called energy, and some other properties too, no doubt. The other (datum) is what you described.

As Data the Sailor Droid once said: "Fas-kinatin'!"

Re:Vaporware?

[igny]

Now, when a bit changes from 1->0, the voltage (accumulated charge) is simply shorted to ground (via resistive path that dissipates heat).

Can't someone modify those resistive paths so that less heat is released into environment and some useful work is done, like charging the batteries, running motors, etc? Another silly idea. Can't someone take those resistive paths out of the CPU box so that the heat won't hurt the CPU as much?

Re:Vaporware?

[Bingo Foo]

You don't have to believe it, but it works that way. Reversible computing is an information theoretic problem. You are constrained by the second law of thermodynamics and Shannon's Law of per-bit efficiency.

Of course there will be advances to be made that do not require full reversibility, and I'm sure some trade off of reversibility and entropy will be found which is optimal from an engineering standpoint. It will involve less discarding of gate states, because it will have to. The second law of thermodynamics is a tough nut to crack.

Re:Vaporware?

[jafac]

So the obvious solution is to account for the increasing entropy (fundamental laws of thermodynamics are pretty important, I guess) - and make the gate so it can function on a wide range of charge levels, and when the stored charge reaches a certain low level (less than the threshold required for a reliable 0->1 switch), then dump that charge and ask for a new one.

I suppose THAT'S the real trick.

Re:Vaporware?

[ralphclark]

The problem with reversible computers is that they are liable to suddenly do a U-turn in the middle of a computation and instead start sucking all the intelligence out of your head, leaving you feeling completely dumb.

Actually if my wife is to be believed, this is how our computer works.

Re:Vaporware?

[randyest]

That seems awfully credulous of you, unless you have access to some proof that I haven't yet seen (and I've been looking -- it's my industry too, BTW). Assuming you're not just blindly repeating something someone told you, can you point me to a rigorous proof of your claims, or provide a basic outline of the proof so that I may investigate further?

Thanks in advance.

Re:Vaporware?

[Bingo Foo]

see this and related documents from here.

Re:Vaporware?

[randyest]

Thanks for the links (first one especially). But, I guess you've read it yourself and know that no proof is contained therein (nor in any of its references, nor in any paper on the subject I found at citeseer) -- this paper is a "brief review of the approach, techniques, and results [for establishing a theory of thermodynamics of computation]." The full paper will be published "elsewhere" (do you have it or a link?).

As much as I'd love to read the full paper, I do not believe it will support (much less prove) your assertions that (1) very low (~kT) power computation requires logically reversible computing, (2) "reversible computing is an information theoretic problem ... constrained by the second law of thermodynamics and Shannon's Law of per-bit efficiency," or (3) "the information theory and logic level of description [is] where reversible computing must be implemented."

(1) is what you initially disagred with me on. (2) is sort of a tautology -- everything, technically, is constrained by those two well-known theorems, but if (1) is not shown to be true, then (2) is tautological at best, irrelevant at worst. And (3) is sort of a re-statement of (1). I am still unconvinced, as you have yet to provide the necessary proof. In fact, after considerable research, I now confidently assert that no such proof exists.

Note that the paper you linked to attempts an axiomatic development of this theory. (As you know, that means "taken for granted without formal proof" or "derived from axioms, or theorems that are believed to be true but remain unproven"). This is an important point that should not be ignored.

Moreover, one of the axioms on which the paper relies (heavily) is Landauer's 1960 conclusion that "it is only logically ireversible operations that must dissipate energy." Note that this conclusion does say (but not prove) that

"if it must dissipate energy then it must be a logically irreversible operation."

However, and this is key, the converse is not necessarily true (and is certainly not proven in any way, by any one)! Even if you take Landauer's conclusion as axiomatically valid, as the paper's authors seem to, you still may not conclude that

"if it is a logically irreversible computation then it must dissipate energy."

...which was your original claim.

QED

Not too bad an idea, BUT...

[ghettoboy22]

I know of more than one building who relys on the heat produced by the computers as the sole source of heating. But it is good for OC'ing!

Re:Not too bad an idea, BUT...

[EmbeddedJanitor]

They'd do better to use heat pumps.

Reversed!

[dolo666]

Imagine all the cool things you could do with the heat from your computer, instead of directing back to the system... *gasp-heartfailure*

Reversing entropy?

[oGMo]

IANAP, but this sounds like trying to reverse entropy as much as possible to me. Won't it take more energy to do a reverse computation than you'll save? Where does the lost energy from that go?

Re:Reversing entropy?

[Carnildo]

There are two theories about energy usage during computation. One is that moving and transforming data requires energy. The other is that the only operation that requires energy is destroying data. Reversible computing subscribes to the second theory, so a reversible computer would not actually use energy to do computations (apart from the inevitable inefficiencies). Since there is no net energy usage, the net entropy neither increases nor decreases, and the second law of thermodynamics doesn't apply.

Re:Reversing entropy?

[jc42]

I've long been curious about this idea of destroying data inside a computer. It seems to me that before and after any opcode, your typical computer contains exactly the same amount of data. A few bits have changed from 0 to 1 or vice-versa. But the number of bits is the same, and each contains only one bit of data, so the amount of data is the same.

So far, when I've questioned people about this, I always get a response that amounts to saying "Boy, you must be a real idiot if you don't understand this." This may be true, but it's not an explanation, it's just a way of refusing to explain anything.

So whatever does it mean to talk about a computer destroying data?

Re:Reversing entropy?

[cavemanf16]

I imagine this system would simply involve designing less wasteful logic gates than currently exist, in which AND, OR, and XOR gates send excess voltage inputs back through the circuit, allowing the power supply to use the energy again rather than throw it off as heat.

Remember, in the most simple OR gate, if you have two input voltages of +5V each, the output will only be +5V, leaving that additional +5V to be thrown off as heat. Sounds like this story is talking about some kind of logic gate that "recirculates" those 'extra' +5V signals back to the power supply system.

But don't take my word for it... I've only had some of the more basic circuit analysis and design courses to date. ;)

Re:Reversing entropy?

[srn_test]

I'm not an expert, but when one talks about entropy, they mean randomness.

A computer freshly turned on has essentially random "data" in its memory; when you write a specific value into the memory it reduces the randomness of the system; this decreases the entropy.

I guess, after memory has been used and freed, it has (to anything not the original using program) some approximation of "random" bit patterns, so the next program which uses it makes it less random again. In real operating systems, the kernel itself clears out the memory before giving it out again, so the kernel is the one reducing entropy.

Of course, I could be way off :)

Re:Reversing entropy?

[jason0000042]

IANAnElectrical Engineer but I think in the context of Reversible computing they are talking about physically, on the lowest hardware level, setting a 1 to a 0. A physical bit that is set to 1 has 5 volts potential across it (or 3 in some cases). To set it to 0, that voltage has to go somewhere. Currently, the circuit gets shorted out and phzzzt, the voltage gets converted to heat and we're back to the 0.

Of course, in other contexts, 'destroying data' can mean other things. Like magnetizing a hard drive, or computer controlled paper shredders. <grin>

Re:Reversing entropy?

[Hatta]

It may have the same number of bits, but are those bits actually telling you anything? For instance, you can take 700 MB worth of audio data off of a CD and compress it to about 400 MB without losing any information. In that sense you really only had 400 MB of information in the first place, you were just expressing it inefficiently. Interestingly, ordered data can be well compressed, but disordered data cannot be compressed well. So more entropy actually is more information.

Disclaimer: IANACS

Re:Reversing entropy?

[jc42]

If this is what people mean, then the implementation is obvious: For every bit, you have a second bit that's its complement. You put them right next to each other, so that flipping their contents will be fast. (In fact, what you probably do is combine them into a single thingy, and call it a "bit". ;-)

But somehow, doubling the amount of memory on a computer doesn't really seem like the way to decrease power consumption.

Anyone know about this?

Re:Reversing entropy?

[plastik55]

It's not strictly "data" as you're accutomed to thinking about it as a bunch of ones and zeroes, but "information" in the sense of information theory--more or less another word for entropy.

The actual measurement of entropy has to do with counting the possible states that a system could be in. A computer containing a list of numbers in its memory could be in any of a large number of states depending on what you know about the list of numbers and the contents of the rest of its memory. If you instruct it to go replace the list of numbers with its sum, the number of states it could be in afterwards is decreased. So its entropy has gone down.

Thermodynamics says that the only way to reduce a system's entropy is to expend energy on it. So if you worked out a way to juggle the bits in the list of numbers around so that you would get the sum, but in such a way that you could back out the operation afterwards to recover the original list, and do it without overwriting any of the other information on the system, then you could do the operation without reducing the entropy of the computer, and wouldn't be forced to expend energy.

Now, a lot of confusion comes from the fact that Shannon decided to call mutual entropy "information" when he was working out coding theory. The concept has a lot of parallels to how we ordinarily think of information, but the correspondence isn't exact, and trying to think of mutual entropy as "information" informally will lead you to a lot of wrong conclusions. It's one of those all-too-common instances where picking a commonly understood name to stand for a subtle concept can do more harm than good.

Re:Reversing entropy?

[nusuth]

You are looking at the wrong level. Here, this device A is doing computation C and when doing so it creates k possible outputs from n possible inputs. If k is less than n it must be destroying data while operating (pigeon holes etc) on some or all of its inputs. It is the computation C that loses data, not the device A. But device A uses energy because it implements a data losing computation C.

Re:Reversing entropy?

[Dr.Enormous]

You most certainly can get an energy boost from a system like this. Think of it this way: there's some amount of energy available for use after entropy takes its cut (free energy, to be exact). Now, an efficient system pumps 100% of that into doing work (you still can't get perpetual motion because this is energy AFTER you've lost some to entropy). An inefficient system uses 50% of that, and radiates the other 50% away. This is just an attempt to recapture some of that 50% that you're just throwing away.

Entropy doesn't come into it, except to say that you can't get all the energy you lost to heat back, as you'll lose some of that to entropy as well, but it's still better than nothing.

Re:Reversing entropy?

[kfg]

You are not a real idiot. You have asked a deep question. Pity those around you choose to cover their own ignorance by being arrogant.

Me, I like to display my ignorance, so here I go:

The computer is just a big abacus really, a physical model of the data. When you shake up an abacus it still has just as many beads on it, only their state has changed, and the now random ordering of the beads can still be read as representing a number. What has been lost is meaning.

Is "2" data? (bearing in mind that we're talking about the logical number here and not the numeral that serves as its physical model)

No. It's just a number. Unless the number relates to a logical model (the number of quarts of milk in my refridgerator) it isn't data.

So the state of your all shook up (uh uh huh, uh uh huh, yeah, yeah)abacus is still a physical reprentation of a number, but it isn't data.

When we run code we keep just as many "beads on our abacus," only their state changes, but data is the physical state and what they mean in the logical model. So if we lose meaning we have lost data or if we lose the proper state to reflect that meaning we have lost data.

So to "destroy" data means to dispose of the physical model and/or its meaning in the logical model.

When we change the state of a computer we certainly don't change the total amount of data, but we certainly change its state and that state's relationship to the logical model.

Take Linux in a Nutshell off your shelf (a physical representation of data) and burn it. Have you not destroyed data? Now fill the same space on the shelf with a gardening book.

You are now still in possession of the same amount of data as you had before, but both its state and meaning have changed. You aren't going to be able to use that book to look up a vi command because the gardening book doesn't contain that meaning.

You destroyed that data.

KFG

Re:Reversing entropy?

[srn_test]

No, not really. That's enthalpy.

Entropy often results in heat (not the loss of heat), but that's simply because heat is essentially disordered states.

When I say I'm not an expert, I did do a few years of University Chemistry and Physics.

"Reversible" a bad name?

[cant_get_a_good_nick]

Wouldn't "regenerative", like regenerative braking on most electrics/hybrids been a better term?

Re:"Reversible" a bad name?

[Anonymous Coward]

This is an area where information theory and physics meet. To minimize heat you must minimize entropy production and that means you must in fact make your computations reversible, as much as possible.

Re:"Reversible" a bad name?

[commodoresloat]

The name is a holdover from the original research project, which took place at Moscow State University in SOVIET RUSSIA.

Re:"Reversible" a bad name?

[mindriot]

No, it should in fact be "reversible". Energy is used mostly when information is destroyed, so to speak. If you do not destroy information, you re-use the energy (yes, that's a little oversimplified). Currently, every time you calculate a XOR b for two bits a and b, you produce exactly one result bit, destroying one bit worth of information. Now, if you instead calculate, for example, (c, d) <- (a, a XOR b) (also known as a CNOT gate), you can reverse the operation (by simply applying CNOT twice to get (a, a XOR (a XOR b)) = (a, b) ). If information is reused as much as possible, you save energy.

Reversible computing also has a lot to do with quantum computation, since all QC operations (except for measurements, of course) are described with unitary matrices, i.e. reversible.

Re:"Reversible" a bad name?

[Hard_Code]

Yeah, but it doesn't say WHICH state you have to reverse to. For instance, any operation could simply shut down the system entirely. That would effectively "reverse" the state to before it was turned on. :)

In my layman's understanding, I don't think it really matters WHAT state you reverse to as long as it has lower entropy, and hopefully, is a state which you want to be in some time in the future. E.g. if my program recomputed a lot of values, it would be nice to be constantly "reverting" to those precomputed values. For that matter, it would be nice to "revert" instruction cache also.

Sounds good, but...

[Tin Foil Hat]

I have to admit that I'm no chip designer, but I have to wonder why this hasn't been done before? What are the problems with this technic?

It sounds good, but what's the catch?

Re:Sounds good, but...

[pclminion]

Because it seemed totally pointless. It was a theoretical curiosity.

People started looking at reversibility in earnest when quantum computing came on the scene. A quantum computer HAS to be reversible in order to function. That made it a very important field of study.

We only recently realized that reversible circuits are also more energy efficient. So basically, we didn't do it before because we didn't know. There is no "catch."

Re:Sounds good, but...

[mikeee]

Although it should have been obvious all along, and probably was if anyone cared. It follows directly from thermodynamics, although the result is a little odd; in essence, there's no theoretical lower bound on how much energy it takes to compute; it's forgetting that takes energy. Ergo, in theory a computer that never loses any data (is reversable) doesn't necessarily use any energy.

re: sounds good, but...

[ed.han]

stupid question perhaps but doesn't this in essence require something like a database in order to store, arrange and recall the desired results? am i missing something here?

b/c if that's true, then does this mean some day in a few decadees, someone's gonna introduce a relational--

[gets killed by angry mob of unemployed DBAs]

ed

Re:Sounds good, but...

[AArmadillo]

Reversible computing has also become more of an issue now that significant evidence points to Moore's Law failing within the next ten to twenty years. CPUs in the near future will become very heat-limited -- in fact, Dr. Michael Frank (the researcher in the article) is also very invovled in researching the physical limits of current CPUs and computer architectures. Reversible computing would reduce the heat produced by CPUs, effectively expanding their physical limits. Reversible computers are essentially a stepping stone, allowing computing speeds to continue to increase until quantum computing is at a mature and usable level.

Re:Sounds good, but...

[mikeee]

You can't save the charge unless you save the result. Thermodynamics occasionally sounds like magic, but it will bite you in the end somehow.

But, no, you don't ever lose data. You can just back out the operations sequentially.

How exactly am I supposed to ...

[burgburgburg]

fry eggs if this sort of thing becomes the norm?

Insensitive clod!

Re:How exactly am I supposed to ...

[pmz]

Yeah, and I'll miss making muffins for my co-workers in my Easy Bake PC!

Reversible Computing

[JamesD_UK]

Before I read the article, I saw the title and assumed someone had designed a computer that gave me answers before I'd even decided what program to run. :-) Useful for those moments where I find myself sitting down at the computer, wondering what it was I was meant to be doing. Current solution to my problem: load up slashdot and wait a little while.

Re:Reversible Computing

[MarcQuadra]

Sometime last year I lost all by bookmarks and now all I do is compulsively read Slashdot. Really.

I wish I remembered the cool sites I used to go to, I'm slowly building a new list of things to do on the net besides Slashdot, but it's still no good.

The worst thing is that the more often you refresh the main page, the less amazed you are. I don't think I've been 'engaged' by the main page in at least three months. Back when I only got to read /. twice a week I really liked it.

What terrible addictive behavior!

sort of like recycling....

[smd4985]

your computer could spit out: "these CPU cycles made of 75% post-CPU-consumed waste" :)

Re:sort of like recycling....

[pmz]

your computer could spit out: "these CPU cycles made of 75% post-CPU-consumed waste" :)

Unsuprisingly, Windows will win the post-CPU-consumed waste category, too :(

Something to worry about...

[Nom du Keyboard]

In fact, unless reversible computing is achieved, computer chips are expected to reach their maximum performance capabilities within the next three decades

Boy, that's something to worry about today. I'll just have to find a spot to insert it on my Worry List. Maybe I can drop Global Warming to make space.

Re:Something to worry about...

[Carnildo]

Assuming Moore's law continues to apply until then, that means that those computers will be a million times faster than the ones available now.

So, what do you plan to do with your 3 exahertz Pentium 17?

Re:Something to worry about...

[pizen]

So, what do you plan to do with your 3 exahertz Pentium 17?

Heat my house.

Re:Something to worry about...

[avgjoe62]

So, what do you plan to do with your 3 exahertz Pentium 17?

Why, play DukeNuke'em Forever, of course!

Re:Something to worry about...

[mariox19]

Boy, that's something to worry about today. I'll just have to find a spot to insert it on my Worry List.

Let me guess, you're a COBOL programmer, retired since 1969, who coded all dates using two digits, right?

Re:Something to worry about...

[BiggerIsBetter]

I don't think you'll have to worry about Global Warming - millions of mulit-KW ExaHertz PCs around the planet should take care of that just fine.

Re:Something to worry about...

[HuguesT]

Well, the climate seems to be warming (although the cause is disputed, large ice shelves have begun to break up in the last few years), there is definitely a hole in the ozone layer over Antarctica, don't your read your newspapers?, and no one is putting oil back into the ground, so how exactly are they lies?

Re:Something to worry about...

[Kynde]

Wrong. The climate is not warming.

Tell me about it. Mid November way up north in Finland and the excessive heat produced by hardware these days is the best thing since the invention of internal heating using alcoholic liquids (which is probably the reason why our ancestors wound up here anyway).

I'll curse the cool reversible computers, embrace the P4 and drink 'till I'm sun bathing in the Caribbean.

I have one of those jackets ...

[didjit]

Its cool cause sometimes I feel like wearing blue and sometimes black. They work best on an every other day cycle. People think I can afford two jackets.

Sigh... I knew I shouldn't RTFA

[winkydink]

Here I thought it was an Intel box that, when turned inside out, became a Mac.

Sigh.

Re:Sigh... I knew I shouldn't RTFA

[shish]

Intels are beautiful on the inside? Macs aren't?

Cool

[pclminion]

Yet more evidence that information is in fact a quantifiable property. We're starting to see hints that information and energy are flip sides of the same coin.

I'm not just spewing. There are serious theoretical problems associated with how information "disappears" when it falls into a black hole. Fortunately, you get the information back again from Hawking radiation, as the hole converts mass into energy. From a theoretical standpoint it's really starting to look like "information == energy," or to put it more precisely, there is a specific equivalence between information and energy like the equivalence between matter and energy.

We've already got space == time, matter == energy, why not also information == energy? There are starting parallels between Shannon's information theory, and the theory of thermodynamics. There is some mysterious shit going on here.

Another boost to my pet theory of the universe: everything is equal to everything else, and we delude ourselves into perceiving imaginary distinctions between things.

Re:Cool

[gowen]

We've already got space == time, matter == energy, why not also information == energy? There are starting parallels between Shannon's information theory, and the theory of thermodynamics. There is some mysterious shit going on here.

I'm not a theoretical physicist, but I thought the equivalence was more along the lines of

Information = order = !entropy

and it was entropy constraints that gave rise to Hawking Radiation.

Re:Cool

[gordyf]

everything is equal to everything else, and we delude ourselves into perceiving imaginary distinctions between things.

Sounds like a really good hit of acid, to me...

Re:Cool

[Have Blue]

Energy is equivalent to WORK. Information is a static property of a configuration of elements.

What's going on here is a circuit implementation detail. In a normal chip, when you have a bit set to 1 and a bit set to 0 and you flip them both, the bit set to 0 is charged with fresh energy from the power supply and the energy in the bit set to 1 is converted to heat. In this proposed system, the charges would be moved from the 1 to the 0 with no loss and no additional draw on the power supply. Less work, same informational content.

Re:Cool

[hchaos]

Yet more evidence that information is in fact a quantifiable property. We're starting to see hints that information and energy are flip sides of the same coin.

I'm not just spewing. There are serious theoretical problems associated with how information "disappears" when it falls into a black hole. Fortunately, you get the information back again from Hawking radiation, as the hole converts mass into energy. From a theoretical standpoint it's really starting to look like "information == energy," or to put it more precisely, there is a specific equivalence between information and energy like the equivalence between matter and energy.

Actually, you are just spewing, at least kind of. As long as the Second Law of Thermodynamics holds true, there is no "conservation of information" law in this universe.

What's really happening here is a lot more simple. In a digital computer, information is stored as a series of energy states. A bit is either 1 or 0, with a 1 meaning that a circuit is energized, while a 0 means that the circuit is not energized. The important thing is that both energized and non-energized circuits hold exactly the same amount of information.

The only thing that this article is talking about is storing the energy from the energized bits in an "energy cache" once the 1 has been switched back to 0, so it can then be used to power other bits. It's really not a very radical idea at all. The only semi-radical thought here is that it would be worthwhile to recover this energy, and that chip manufacturers would benefit from investing in this research.

Re:Cool

[pclminion]

I won't make any claim that it's "insightful" or that nobody else has had similar ideas.

I studied CS, not philosophy. At least I can claim to have rediscovered an idea without prompting from somebody else, right? How about a few inches of slack?

Re:Cool

[Rostin]

"Another boost to my pet theory of the universe: everything is equal to everything else, and we delude ourselves into perceiving imaginary distinctions between things." E may equal m*c^2, but it is hard to understand how you can make the leap from that to the conclusion that matter and energy can't be distinguished at all.

Re:Cool

[Jedi1USA]

Well, I have read a lot of stuff and said to my self....What a load of crap. So to finish out that thought... Information=mass, and since Mass=energy...You're right Information=energy!! ;^)

Re:Cool

[pmz]

Energy is equivalent to WORK.

This must explain why all the lazy bums out there read slashdot all day.

Re:Cool

[RobinH]

What's going on here is a circuit implementation detail. In a normal chip, when you have a bit set to 1 and a bit set to 0 and you flip them both, the bit set to 0 is charged with fresh energy from the power supply and the energy in the bit set to 1 is converted to heat. In this proposed system, the charges would be moved from the 1 to the 0 with no loss and no additional draw on the power supply. Less work, same informational content.

Well, actually you're only talking about a register. A bi-polar type register will still consume power while it's sitting there holding its state, on or off. CMOS is different (see below). What you're describing is DRAM (dynamic random access memory) which uses charges to store values. However, even DRAM consumes power because the charges leak out relatively quickly, and have to be refreshed every 2 microseconds (I think).

Remember, however, that a lot of the energy in a CPU is used by the logic gates (AND, OR, NOT, etc.). If you're talking CMOS technology, or something similar, then you're right... you have to "charge" the gate of the transistors to get them to switch, and "discharge" to ground again. I suppose that reversible would mean that every transistor in the whole chip has an opposite pair, and if you turn one transistor on, you have to turn the other one off (to borrow the charge) rather than using power from the power supply. Some energy will still be used during the switch, due to losses.

However, I think the article is talking about something even more radical. It's talking about MEMS, which makes me think that inductors are involved, and therefore rather than moving charge between two capacitors, you're moving them between capacitor and inductor, which can be done rather efficiently at the right frequency.

That's about as far as I can follow the article.

Re:Cool

[PDAllen]

Hmm.. no.
Information =/= energy; this is easy to prove. Just observe that you can consider any (finite) string of bits to represent any (finite) message, as you're allowed to choose the encoding, and you see that the information content (the ordering of the bits) has the same energy content as the string of bits given by sending as many 1s as in your message, followed by as many zeroes.
However, information processing does require energy, and that is where your parallel is worth considering.
As far as Hawking radiation goes: firstly the entropy situation was never a real theoretical problem, more an inconvenience (the second law of thermodynamics is a statistical law which holds in Newtonian dynamics and which would be nice to have in all cases). Secondly, the information is not returned by the Hawking radiation. The point is that the probability of there being more order in the Hawking radiation than in the swallowed matter approaches zero as time (and matter swallowed and re-emitted as Hawking radiation) approaches infinity. Thus instead of matter ending up in a black hole somehow removing entropy from the universe, you can consider all mass in a black hole to 'have no order', hence any mass falling into a black hole causes total entropy to increase.
Next thing to think about: if you can encode information in any way you choose, what exactly does it mean for something to be disordered?
Answer: this is hard. On one hand, you can examine the probability of certain events (such as all the 1s being in a block, not spread around). This attempts to give a measure of disorder to matter at a given instant. On the other hand, you can examine the processes affecting the matter as time goes on. Here you find that certain processes output energy, and others require energy input. The processes that output energy tend to cause the disorder of the matter to increase in any fixed metric for order (result in more-probable states), and those that require energy tend to decrease that disorder (result in less-probable states). From this point of view, entropy is likely to increase when a sequence of processes occurs (second law of thermodynamics). The sequence of processes is approximately equivalent to performing a sequence of calculations on the data represented by the matter; entropy increase is due to the fact that highly-ordered states tend to look dissimilar, while disordered states are all fairly similar in any given metric. Then if you want to look at parallels with Shannon's theorem, observe that you can consider a calculation to be the same as encoding your data in a different way, possibly throwing away some of the output. Here it becomes obvious that a reversible calculation tends to preserve entropy, while irreversible ones increase it, so that transmitting accurate information through a channel requires energy input, more energy for better transmission (in the sense of more expected accurate data).

Patent office would love this

[bugnuts]

A Perpetual Computing Machine!

Turn it on and it generates cycles from microscopic springs and pulleys, we call "Springons" that can recover the computing power expended, sending the cpu "Wheel" into another revolution.

--

funny how all these machines require a battery or plug.... :-)

WTF is reversable computing?

[EvilTwinSkippy]

Energy efficient, got you. But as an Electrical Engineer, and a seasoned network engineer, I've never heard the term before. And I'm pretty damn well read.

Re:WTF is reversable computing?

[pclminion]

It means you can take the outputs and recover the inputs.

If I tell you that (x && y) == 0, can you tell me what x and y are? No: it could be (0, 0) or (0, 1) or (1, 0). Therefore, the operation AND is not reversible.

A reversible computer always performs operations that can be uncomputed. Given the outputs, you can reconstruct the inputs. This means, for one thing, that a reversible computer has no concept of boolean AND. Or OR, for that matter. NOT is reversible, though.

Re:WTF is reversable computing?

[Sebastopol]

Unfortunately, only NOT is reversible. (Name one other gate that is? ... buffers don't count ;-)

So that gets us nowhere.

I believe the article is hinting at reusing the charge required to change state. Although I have no idea how it is stored, or how it is re-routed.

This article is entirely fluff.

Re:WTF is reversable computing?

[pclminion]

Well, AND and OR aren't reversible, but that doesn't mean there are no binary operations that aren't. It just means that a reversible computer is going to work a lot differently than you are used to.

In particular, most reversible logical operations have more than one output. The example of NOT is one of the few than has only a single output.

Re:WTF is reversable computing?

[sukotto]

Hmm... Given the output, you can figure out what the original input was.

I think that's illegal under the DMCA now.

Better luck next time!

Re:WTF is reversable computing?

[merlin_jim]

This means, for one thing, that a reversible computer has no concept of boolean AND. Or OR, for that matter

Actually, you can add one additional output to any binary logic gate in order to make it reverseable; most reverseable computing designs focus on that and the logic circuits themselves ignore the secondary output...

Re:WTF is reversable computing?

[anandrajan]

Here's a link that could clear up things. On the other hand, it may not.

Re:WTF is reversable computing?

[hesiod]

> Unfortunately, only NOT is reversible. (Name one other gate that is?
> So that gets us nowhere.

Very true. That is why they are not using your standard gates, they are making reversible gates so it overcomes your limitations.

> Although I have no idea how it is stored, or how it is re-routed.

I believe this is where the problem in your thinking crops up.

As you know, most gates have two inputs and one output. Therefore, if the power from both inputs exceeds what needs to be sent on the output pin, the extra energy is converted to heat.

Now these reversible gates come in. The fact that they are reversible is not important in this context. The important bit here is that the excess energy can be sent on a second output pin to be stored (in a capacitor, whatever) instead of being automatically wasted to heat.

However, I may be misunderstanding this technology, as I only grokked this much 2 minutes ago in a burst of insight.

Wrong tagline...

[Anonymous Coward]

Should be "Reversible computing rears its butt again"

This is perfect for Microsoft...

[ConceptJunkie]

You can start with the Blue Screen of Death and acheive any functionality.

Re:This is perfect for Microsoft...

[nmg]

least funny comment on slashdot ever

Re:This is perfect for Microsoft...

[Artifakt]

But only if the OS version counts down - 94 - 93 - 92... each time you turn it on.

Re:This is perfect for Microsoft...

[DarkSarin]

interesting sig, but generally see the orginal quote attributed to Isaac Asimov, not Arthur C. Clarke.

Can you provide a source? I can--look up the book "Magic" by Asimov on amazon, and read that. Also the Foundation books that have been written in his honor by (I think) Greg Bear and others also reference this quote.

Re:This is perfect for Microsoft...

[StenD]

Here are "Clarke's Three Laws", including "Any sufficiently advanced technology is indistinguishable from magic.". In "Magic", Asmiov was speculating on whether the reverse was true, but Asimov was hardly the first to reverse Clarke's quote.

Re:This is perfect for Microsoft...

[rcjhawk]

Clarke's three laws, curtesy of Wikipedia ( http://en2.wikipedia.org/wiki/Clarkes_Three_Laws)

• Clarke's First Law When a distinguished but elderly scientist states that something is possible he is almost certainly right. When he states that something is impossible, he is very probably wrong.
• Clarke's Second Law: The only way of discovering the limits of the possible is by venturing a little way past them into the impossible.
• Clarke's Third Law: Any sufficiently advanced technology is indistinguishable from magic.

And yes, Clarke came up with number 3. In the 1950's or 1960's, I believe. http://www.lsi.usp.br/~rbianchi/clarke/ACC.Laws.ht ml says they're listed in his Profiles of the Future, which I read in High School, a long, long time ago.

mjm

Re:This is perfect for Microsoft...

[dillon_rinker]

It's Clarke's Third Law. Look here. The page includes the book it was published in. Google for "clarke's law" for more.

I prefer the the contrapositive form of his third law:
Any technology distinguishable from magic is insufficiently advanced.

Re:This is perfect for Microsoft...

[Artifakt]

I appreciate several posters providing references to the original question so I didn't have to dig for page and paragraph numbers. Thanks to all. Let me add that Clarke and Asimov had a "mutual non-agression treaty" while Isaac was still alive. In part, when anyone asked who was the greatest living SF author or such questions, they always said the other one was, and if there was a fannish arguement about who had written first on a given idea or similar point, either of them would usually say something diplomatic rather than risk giving facts that might enflame the arguement.

Heated my first apartment

[BWJones]

.....it also could boost their speed, because these chips are becoming so fast that the heat they generate limits the speed at which they can operate without overheating and malfunctioning.

Bah, this idea is nothing new. From the two SGI's with two 20in displays, two macs and five displays attached to them, my tiny little first apartment had more than enough heat production to warm things up. :-)

This seems really confusing!

[MooseGuy529]

Correct me if I'm wrong, but does this just take the energy from the ground of a component and pull it back into a power source? Couldn't you just do this by putting a capacitor on the ground and switching to it occasionally instead of... (train of thought derailed) So it basically takes electricity and pushes it between a capacitor and a circuit? That's clever, but I thought the heat just came from all the electricity "moving around" so much--won't this just make it move back in the same direction instead of to ground? And, if done right, couldn't this make really energy-efficient processors? And if you made a multi-processor system where one processor runs off positive voltage and the other runs of negative voltage, could you wire this up to be *really* efficient?

Just some ideas, flame/reply away...

Theory

[cristofer8]

The idea here is that when you use any 2->1 gate, such as an and gate, you lose one bit of information. Since information is actually just energy, you have to dissapate that energy somewhere, usually as heat. If instead of a 2->1 gate you used 2->2 gates, where one bit is the and product and the other is enough information to reverse the operation, you aren't discarding any information and thus, aren't dissapating any heat.

Re:Theory

[TeknoHog]

Entropy S = k ln W
(k = Boltzmann's constant, W = number of states)
Information (in bits) I = log_2 W = ln W / ln 2
Hence S = kI/ln 2 or roughly S = kI.

Heat Q = ST = kTI.

Let's say we destroy 100 gigabits of information at a temperature of 300 K. Since k = 1.38E-23 J/K, this means a heat of about 4E-10 Joules. Which is not very much, and does not really contribute to the heat produced by CPUs.

In fact, I think this is the way to find a theoretical minimum for the heat produced from information processing. We can try and make more efficient processors to get closer, just like we can increase the efficiency of engines to get closer to the thermodynamic limit.

Re:Theory

[owlstead]

Dunno, I just burned some of my CD collection, and I am sure there was a lot more heat than that.

Maybe 1's burns more brightly than 0's cause most of them contained data.

Re:Theory

[Repran]

Conservatively it is save to assume that in a processor running at 3 Ghz, every cycle 1 Mbit of information gets destroyed. That are 3145728 Gigabits a second right there.

Re:Theory

[Tim C]

Surely a 32 bit processor consumes at most 32 bits of information per cycle? (Well, 32 bits per instruction pipeline, but there's only a couple of them).

Around 100 Gbps sounds about right to me, although I'm more than prepared to be proved wrong.

No more breakfast

[AtariAmarok]

I was so used to cooking breakfast on the top surface of my Apple Cube. I'll miss this if the energy gets recylcled elsewhere, and I'll likely have to go buy a Foreman grill to make up for the loss of this nifty cooking appliance.

From the article...

[TamMan2000]

While he was at MIT, Frank worked on a team that built several simple prototypes of reversible chips.

It has at least gotten to the chip level so far...

Re:From the article...

[FrostedWheat]

It has at least gotten to the chip level so far...

I have a prototype of a Holly Hop Drive. It dosen't quite work yet, but still ... it's a prototype :)

All I need now is funding...

At long last! Reversable computing!

[bugnuts]

Finally a hardware device that can decrypt backwards writing.

Now if only they'll invent Transposed Computing that can hardware decrypt Rot13.

Shit, I thought my computer could run backwards

[Saint Stephen]

I wish my computer was reversible in the sense that I could press the rewind button and everything I did would happen in reverse.

I'm telling at as a joke, but I've always wondered why no chip designer ever wrote this. It should be possible to log every instruction that passes through the CPU and play them in reverse order. Imagine how cool that would be!

Re:Shit, I thought my computer could run backwards

[wed128]

I'm telling at as a joke, but I've always wondered why no chip designer ever wrote this. It should be possible to log every instruction that passes through the CPU and play them in reverse order. Imagine how cool that would be!

What would be the practical application of this?

Re:Shit, I thought my computer could run backwards

[SmallFurryCreature]

How would you undo the logging? Keep a log of the log? Then how do undo that? etc etc.

To explain the log would write to something. Thereby overwriting what was already there. This would need to be undone as well.

End logic error.

Re:Shit, I thought my computer could run backwards

[br0ck]

What would be the practical application of this?

Am I the only one that wakes from an eight hour daze every day wishing there was some way to figure out what the hell I've been doing all day?

Re:Shit, I thought my computer could run backwards

[Jeremi]

No need to wait for a hardware implementation... start hacking your favorite open source JVM today!

Of course, where it breaks down is in the I/O ... good luck trying to get your TCP socket to unsend data ;^)

Re:Shit, I thought my computer could run backwards

[hesiod]

> good luck trying to get your TCP socket to unsend data

If you melt down every wire and piece of hardware that packet was on, philosophically speaking, did it ever exist to begin with?

I'm no Plato...

Reversable memory?

[autopr0n]

It seems to me that what they are saying is that most of the heat comes from memory elements being discharged. What about energy spent by fliping a CMOS gate? Isn't that where most of the power is lost? I mean, even CPUs that only have very small on-die caches still generate a lot of heat.

Re:Reversable memory?

[Dachannien]

That's what I thought as well - that during the switching of a CMOS gate from one state to the other, the positive supply and ground are momentarily shorted (well, there's still a lot of resistance there, but compared to normal, quite a lot of current flows).

OK Homer quote

[LittleBigScript]

Homer: I know. And this perpetual-motion machine she made today is a joke! It just keeps going faster and faster. In this house, we obey the laws of thermodynamics!

Thermodynamics 101

[majid]

You get the most energy efficiency from a machine when it works in a thermodynamically reversible way, for instance the most efficient thermal motor possible is one that uses a Carnot cycle. Most real-world engines use different, less efficient cycles like the Otto or Stirling cycle because they yield higher speeds or torque.

Losing the ability to reverse computations means increasing entropy and thus lower efficiency. Interestingly, there is a whole class of functional programming methods that is intrinsically reversible (because evaluating expressions without side effects is reversible).

The best explanations of the issues involved is in Richard Feynman's "Lectures on Computation", that show how thermodynamics constrain what is ultimately possible with a computer.

Re:Thermodynamics 101

[iggymanz]

funny this topic has popped up again after I saw it featured in Scientific American over 10 years ago....the real problem is that no one wants to halve the number of useful gates on a chp in order to bulid all the extra circuitry required to reduce (of course, not eliminate, entropy still will increase though at lessened rate) the thermodynamic cost of "forgetting" data.

I attack instead the basic premise, that there is a shortage of energy, or that we must accept lower standard of life or lower capability in our machinery. What we DO need to do is get smarter about where we get our energy - instead of adding to net heat budget and pollution budget of earth getting really serious about solar energy (which might just mean making hydrocarbon fuel out of plant & suitable waste materials)

Re:Thermodynamics 101

[Theaetetus]

I attack instead the basic premise, that there is a shortage of energy, or that we must accept lower standard of life or lower capability in our machinery. What we DO need to do is get smarter about where we get our energy - instead of adding to net heat budget and pollution budget of earth getting really serious about solar energy (which might just mean making hydrocarbon fuel out of plant & suitable waste materials)

Um, that's not the basic premise. The basic premise is that with each bit of information lost, that bit is converted to heat. More bits lost = more heat = limit on how fast a processor can be due to temperature-caused failures. Removing that problem results in much faster theoretical limits on processors.

-T

Re:Thermodynamics 101

[Zardoz44]

As I understood from the article, they weren't so much concerned with how much energy computers used, but rather how this lost energy is limiting advances in CPU speed.

They mentioned that 10% of the energy used in the US can be attributed to computer usage, but the only reason this was a Bad Thing is that most of that energy is lost as heat. Too fast of a CPU == too much heat == stuff melting.

On the downside, I might have to buy a real whitenoise generator when/if computer fans are no longer needed.

Re:Thermodynamics 101

[greg_barton]

I attack instead the basic premise, that there is a shortage of energy, or that we must accept lower standard of life or lower capability in our machinery.

Then you are attacking a straw man, at least with respect to the article. A reversible chip would be no less capable. In fact, in the long run, it would be more capable. Less energy heat dissapated means we can continue to use the same materials far into the future with faster and faster chips. As it stands silicon will become ununsable once the heat dissapation reaches the melting point of silicon. (Far sooner, actually.)

Now, I'd love to see chips made from artificial diamond, but I think reversible chips will come sooner.

Re:Thermodynamics 101

[onomatomania]

Silicon melts at approx 1400C (2600F)

Aluminum melts at approx 660C (1200F)

Copper melts at approx 1100C (2000F)

Silicon's melting point is NOT the problem. The problem is the metal interconnects, and the increased rate of diffusion under high temperature. Microchips depend on there being gradients of impurities to create P and N regions, and as the temperature increases the natural diffusion of these impurities causes them to move, to redistribute themselves.

Aside from that, you have to remember that modern chips these days are like huge tall deli sandwiches, with a stack of 15 or more layers of various compositions. All of them have to maintain their form, shape, and alignment for things to work. As you heat things they expand and contract at different rates, and the thermal stresses involved will cause havok. This is why the processing steps involved have a "thermal bugdet". You cannot perform high temperature operations on the chip in its later stages of life on the production line because it becomes more and more susceptible to these thermal gradients. You have to schedule the steps so that you do all the high-temp things at the beginning of the processing. Remember that the various materials in use for microchips are all extremely brittle, and when they're all bonded together it doesn't take much sheer force at all to cause fracture.

Not exactly a "radical idea" ...

[Dlugar]

The article mentions itself that this idea "date[s] back to the early 1960s" ... Feynman gave a lecture at some conference in Japan, I believe, in which he gave detailed explanations of his ideas for reversible logic gates, and the theoretical uses and limits for such machines. You can find it in the book The Pleasure of Finding Things Out, which I have, or apparently also in Feynman Lectures on Computation, which I don't have. I don't have either book on-hand to check out the exact date, but it was quite a while ago.

However, while it may not be particularly new or radical, it is quite interesting.

Dlugar

Read the Feynman book

[rarose]

http://www.amazon.com/exec/obidos/tg/detail/-/0738 202967/103-2222180-5559862?v=glance

He has a great deal of info about how reversable computers work and why they save energy.

Re:Read the Feynman book

[MillionthMonkey]

I second this. Feynman's lectures on computation are at a very fundamental level, so they are impractical for day to day use, but the theory is solid. Thousands of years from now, computers will undoubtedly have changed a lot, but the principles in this book will still apply to them since they merely describe how the laws of physics affect any computational system.

Here is an interesting excerpt on pages 149-150 that explains Maxwell's demon in terms of reversible computing:

The demon has a very simple task. Set into the partition is a flap, which he can open and shut at will. He looks in one half of the box (say, the left) and waits until he sees a fast-moving molecule approaching the flap. When he does, he opens the flap momentarily, letting the molecule through into the right side, and then shuts the flap again. Similarly, if the demon sees a slow-moving molecule approaching from the right side of the flap, he lets that through into the side the fast one came from. After a period of such activity, our little friend will have separated the fast- and slow-moving molecules into the two compartments. In other words, he will have separated the hot from the cold, and hence created a temperature difference between the two sides of the box. This means that the entropy of the system has decreased, in clear violation of the Second Law!

This seeming paradox, as I have said, caused tremendous controversy among physicists. The Second Law of Thermodynamics is a well-established principle in physics, and if Maxwell's demon appears to be able to violate it, there is probably something fishy about him. Since Maxwell came up with his idea in 1867, many people have tried to spot the flaw in his argument. Somehow, somewhere, in the process of looking for molecules of a given type and letting them through the flap, there had to be some entropy generated.

Until recently, it was generally accepted that this entropy arose as a result of the demon's measurement of the position of the molecules. This did not seem unreasonable. For example, one way in which the demon could detect fast-moving molecules would be to shine a demonic torch at them; but such a process would involve dispersing at least one photon, which would cost energy. More generally, before looking at a particular molecule, the demon could not know whether it was moving left or right. Upon observing it, however this was done, his uncertainty, and hence entropy, would have reduced by half, surely accompanied by the corresponding generation of entropy in the environment.

In fact, and surprisingly, Bennett has shown that Maxwell's demon can actually make its measurements with zero energy expenditure, providing it follows certain rules for recording and erasing whatever information it obtains. The demon must be in a standard state of some kind before measurement, which we will call S: this is the state of uncertainty. After it measures the direction of motion of a molecule, it enters one of two other states- say L for "left-moving", or R for "right-moving". It overwrites the S with whichever is appropriate. Bennett has demonstrated that this procedure can be performed for no energy cost. The cost comes in the next step, which is the erasure of the L or R to reset the demon in the S state in preparation for the next measurement. This realization, that it is the erasure of information, and not measurement, that is the source of entropy generation in the computational process, was a major breakthrough in the study of reversible computation.

Future Energy Star Notice:

[farrellj]

This computer uses 100% recycled electrons. No electrons were destoryed or harmed by this computer.

Hallelujah!

[mblase]

I, for one, am all for this. Providing machines with a way to reclaim their own heat energy for power is the only way to ensure we don't all end up jacked into a computer-generated Matrix so they can suck up ours.

'Splain somthin' to me... (aka. IANAEE, but...)

[OrbNobz]

From the article:
"In theory, these oscillators could recapture most of the energy expended in a calculation and reuse it other calculations."
What the hell does this mean?
4(0100) + 3(0011) = 7(0111)
Ok, now, let's take that 0111 and use the bits for the answer to 7+8.
Is that really what they're saying?

"The concept is somewhat analogous to hybrid cars now on the market that take the energy generated during braking and recycle it into electricity used to power the car."
Ummm, no. The car analogy would work if we captured the waste heat thrown off, and converted it back to electricity. The concept here is that we don't waste the heat to begin with. This would be analogous to driving back to point A in reverse and reclaiming the fuel.

How could this possibly work?

- OrbNobz
I, for one, welcome our new reversible friends!!sdneirf elbisrever wen ruo emoclew ,eno rof ,I
The preceding statement took absolutely no energy to write.

Re:'Splain somthin' to me... (aka. IANAEE, but...)

[Theaetetus]

"The concept is somewhat analogous to hybrid cars now on the market that take the energy generated during braking and recycle it into electricity used to power the car."
Ummm, no. The car analogy would work if we captured the waste heat thrown off, and converted it back to electricity. The concept here is that we don't waste the heat to begin with. This would be analogous to driving back to point A in reverse and reclaiming the fuel.
How could this possibly work?

It's a little esoteric, but stay with me...

Bits of information are made up of energy - make that assumption, and the rest makes sense, but it's okay to make, since it's held with experiments since the 1960s when Feynman was talking about it.

So, you've got an OR gate in your computer - two bits enter, one bit leaves! [/thunderdome flashback]. Well, since that extra bit can't just stay inside the gate, it has to go 'somewhere' - and that somewhere is out, as heat. As more and more operations are performed, more heat is generated - thus limiting the speed of the computer due to temperature problems.

The solution is to make a 'reversable gate' - two inputs, two outputs, one of which is the calculated output from the gate, and the other of which is a checksum output or something, which you can ignore. This way, no information has left the system, and therefore no heat has left. The only time you remove information is at the output of the processor, resulting in much less heat loss.

Why this is called 'reversable' is that from the output of your new gate, using both of those bits - the output, and the checksum - you can tell exactly what the input bits were (which you can't do with an OR or an AND gate).

-T

Re:'Splain somthin' to me... (aka. IANAEE, but...)

[JPelorat]

How can you just ignore the second one? Don't you have to do something with it, or it dissipates and becomes heat? Or do you end up with a lot of spare bits just cycling around inside the system?

Now it'd be cool if you could have a pressure release valve like outside your home, and every now and then it opens up and spits a bright blue arc across the yard as all that pent-up information is released. Dog next door barking all night? No problem, just leave SETI running; that'll take care of it.

Re:'Splain somthin' to me... (aka. IANAEE, but...)

[hchaos]

From the article: "In theory, these oscillators could recapture most of the energy expended in a calculation and reuse it other calculations." What the hell does this mean? 4(0100) + 3(0011) = 7(0111) Ok, now, let's take that 0111 and use the bits for the answer to 7+8. Is that really what they're saying? "The concept is somewhat analogous to hybrid cars now on the market that take the energy generated during braking and recycle it into electricity used to power the car." Ummm, no. The car analogy would work if we captured the waste heat thrown off, and converted it back to electricity. The concept here is that we don't waste the heat to begin with. This would be analogous to driving back to point A in reverse and reclaiming the fuel.

I'm now going to get technical on your ass.

Actually, what happens in a standard x86 processor is that if I want to add 4 + 3, and then add 8 to the result, I'm already reusing the bits from the 7 in the process.

Here's how it works: I want to add 3 to 4. I put 0100 in my processor's A register, and I put 0011 in the B register. I then send an "add" command to the processor. The result is that I now have 0111 in the A register. Now let's say I want to add 1 to this. I put 0001 in the B register, which switches one bit from a 1 to a 0. There was energy in that 1, which, in a normal processor, was "lost" (converted to waste heat) by grounding that bit. I now give the processor another add command, which changes the A register to 1000. The processor grounds out three bits, and takes energy from the power supply to energize the fourth bit. In this process (which takes a few nanoseconds) I have lost the energy to power 4 bits (assuming my registers initially were 0000).

What a "reversible computer" would do is, instead of grounding out those bits, it would move the energy into a reservoir, and then use the energy from that reservoit to power other bits. It would also take the energy included in the bits that I sent to the processor as the "add" commands and put it back into the reservoir. If 50% of the bit energy could be reclaimed (which is a very conservative estimate), the processor would generate a lot less heat, requiring less cooling (which means less energy wasted by your CPU fan as well). This is almost exactly what a hybrid car does during braking, as a hybrid car does not capture the waste heat thrown off by the brakes, but instead uses magnet-assisted braking, which directly converts some of the car's kinetic energy to electricity.

Re:How much is that?

[z4ce]

I would imagine about as much as a 100watt light bulb... All of the 100 watts has to end up as heat in your room some how.. :)

Ian

Sounds like

[karlandtanya]

..collaboration with this guy would be productive.

Re:How much is that?

[Theaetetus]

According to the article, current CPUs put out about 100 Watts as heat. How much is that? Can anyone give me a decent comparison to an ordinary houshold object that would allow me to fathom how much power that is?

Sure! A 100 watt lightbulb. Glad I could help. ;)

-T

Re:Horrible article

[Crimson Midget]

I refer you to a decent article on Kuroshin:
http://www.kuro5hin.org/story/2003/9/8/ 14125/70302

I also suggest MIT's reversible computing page:
http://www.ai.mit.edu/~cvieri/reversible.ht ml

and there's:
http://www.supercomputingonline.com/arti cle.php?si d=4894
http://www.spie.org/conferences/calls/01/a m/confs/ AM429.html

Is PC power use really a big issue at this time?

[Anonymous Coward]

I don't know about AMD, but at least Via and Transmeta and I think Intel are already producing processors that can handle most any PC application including playing high resolution videos while running on around ten watts of power. That's not a terribly significant amount of power even compared to flourescent lighting.

I'm thinking these will all be made in Russia....

[DeionXxX]

These reversable computers must be Russian because I've heard on /. that:

In Soviet Russia computers use you.

-- D3X
NeoX3.com: The One site you'll ever need for XXX.

Re:What about cars?

[Anonymous Coward]

Internal combustion engines used in mass market automobiles have an efficiency of closer to 18%. Pretty horrible, really.

The energy lost is through exhaust heat, water heating, and friction.

HOW does it make it more efficent?

[autopr0n]

I mean, say you have a CMOS OR gate. If both of the inputs are high, then the NMOS transistors will close and the PMOS transistors will open. Energy is lost only when electrons 'leak through' when the gate changes (and of course, electrons that leak through but don't affect the computation, which I guess happens all the time). How would reversing the computation affect this? Maybe if you were using plain PMOS or something...

Re:HOW does it make it more efficent?

[mikeee]

Well, there isn't exactly a how. Thermodynamics guarentees us that non-reversable gates will use energy. It doesn't guarentee that it's possible to build a reversable one that doesn't; it does guarentee that any gate which doesn't lose energy is reversable.

Re:HOW does it make it more efficent?

[randyest]

Enegy is lost always (leakage current) because the gate is not a perfect insulator. The smaller the gates, the more ther leakage. This is called static power.

Energy is also lost during switching, as the charge needed to switch is moved around. This is called dynamic power.

Reversible computing endeavors to reduce/eliminate dynamic power. It does nothing for static power. A long time ago, dynamic power was dominant and static power was negligble. Now, gates are so small, static power is approaching the same order of magnitude as dynamic.

So, even though they're only addressing about 1/2 of the problem, it would be great to have the magnitude of that big problem halved.

Re:HOW does it make it more efficent?

[TigerNut]

Static power might be an increasingly large issue for Pentium-class (0.1 micron, low core voltage) CPUs and graphics cores, but for all other CMOS circuits, the dynamic dissipation will still be the lion's share of the power used.

One technique that I imagine you might be able to use is to tie a small inductor to each gate (making a tiny oscillator), and then you run the system such that you somehow drive all of the oscillators into a synchronous mode of operation. The energy of the system is continuously being changed from electric to magnetic. While it's in the magnetic state, you could update the electric connections (again, "somehow") and then your computational results would be propagated by energy redistribution as opposed to continuously pulling more charge from the supply rails and using it to charge and discharge caps.

Re:HOW does it make it more efficent?

[randyest]

Static power might be an increasingly large issue for Pentium-class (0.1 micron, low core voltage) CPUs and graphics cores, but for all other CMOS circuits, the dynamic dissipation will still be the lion's share of the power used.

Static power becomes significant (within 1/10th) compared to dynamic power at about 0.15um. According to the Peddie Report, that includes about 75% of all design starts from 2002 on.

Even if the "lion's share" were 90% (it's not, see above), I'm still worried about static power because (1) it happens all the time, on every gate, (2) I can't stop it by stopping my clocks and (3) I personally don't do anything in technologies older than 0.13um :)

Oh BTW, neat idea, but it's really hard to make inductors out of silicon.

Re:HOW does it make it more efficent?

[CustomDesigned]

Think of the memory of a reversible system as initialized so that you have your program and data, plus a store of bits - some (half maybe) initialized to 1 and some initialized to 0. Imagine that every gate is perfect and loses no energy. However, no energy can be added to the system either. You are forever stuck with the same number of 1 bits and 0 bits. All you can do is shuffle them around. A real reversible computer is like this, except that the gates aren't perfect and lose some energy to heat during operation - but less than a conventional computer.

All your logic operations accomplish the equivalent of rearranging bits. For instance, there is no 'copy' or 'move' instruction. Instead you have an 'exchange' instruction. If you want to make a variable have the value of zero, you can't copy 0's to it. But you can exchange it with one of your known 0 words.

You carefully write your program so that it keeps track of values exchanged for known 0 and 1 bits so that it can reuse them later. But this becomes more and more difficult as the computation becomes more complex. Eventually, your supply of known 0 and known 1 bits dwindles and you have an increasing supply of "garbage" bits with essentially random values. They are really only pseudo-random, because all logic operations are deterministic (not to mention reversible), but their value is not known easily enough to be useful. When you run completely out of known bits, your computation has reached maximum entropy, or 'heat death'. To continue, you can clear some of the garbage bits to 0 (draining any charge in them to ground and dissipating heat in the process), and set other garbage bits to 1 (adding energy to charge any 0 bits). Having discharged its entropy into the host environment, your reversible computer can continue its computation.

Don't think you can get known 0 or known 1 bits for free by 'testing' them. Your 'test' instruction is reversible too. It must be possible to reconstruct the original memory state by simply running the logic in reverse. If this isn't obvious, remember that 'testing' a bit drains its energy, and the energy must be put somewhere. So the 'test' instruction operates by exchanging the tested bit with another bit - which can then be used to reverse the operation. Think Uncertainly Principle.

Once you get the hang of thermodynamic programming, you have a handle on understanding entropy and quantum mechanics. It makes you wonder: are entropy and the uncertainty principle necessary properties of any possible world? Or does the similarly between our mechanical model and our universe indicate that our universe is actually a gigantic reversible computation?

A practical device would probably not grind to a halt to set and clear a vast store of known bits. Instead, it would reuse bits as much as practical, but provide for a continuous supply of known 0s and known 1's (at the expense of adding energy and dissipating heat to do so).

Finally, when modelling physics, all computations should be reversible - because physical processes are reversible. An egg is difficult to unscramble without running time backward for the same reason that the memory of your reversible computer fills up with "garbage" bits. Enforcing reversibility provides a check for the correctness of the simulation.

Re:HOW does it make it more efficent?

[tjb]

Oh BTW, neat idea, but it's really hard to make inductors out of silicon.


I'm not an LSI guy (though I do work closely with our LSI team), but IIRC from a college course a million (ok, 4:) ) years ago, there is a way to simulate an inductor using capacitors and op-amps. I'm pretty sure that analog-LSI people do this all the time when they make our uber-cool programmable A/D-D/A chips with tunable filters and adaptive hybrids.

But from my fuzzy recollection, it seemed like really convoluted deign - you'd probably waste more power trying to simulate the inductor than you would save by having your sim-inductor.

Tim

Re:What about cars?

[Jerf]

And if it is realistic, then how much more powerful are oil companies politically than electric power companies that the latter are going to just stand by and let this happen?

You mean, the power companies are going to force Intel to make their chips more wasteful, causing progress to halt and people to buy fewer Intel chips? Yeah, sure.

I mean, there's paranoia, and there's paranoia.

Come on, wake up. I won't claim that kind of thing never happens but by and large capatalism is too powerful; Intel isn't going to act against its own best interests for any mere money the power companies can throw at it, because it won't be worth it. Growth is worth more then mere money to Intel. (If you don't understand why, go learn about business; the explanation is too complicated for a Slashdot posting.)

The power company is made of people like you and me; far too busy to hover over various scientific journals and swoop around like super-villians repressing "dangerous" information.

We already have reversible robots

[AtariAmarok]

We already have reversible robots.. Why not reversible computers?

"Apple Toast-Or! From G5 Power, to nice warm toast, back to G5 Power again!"

Are there other uses for MEMs

[aashenfe]

I'm no physicist, but I was wondering, couldn't these mini electro-mechanical devices be a more efficient way to make electricity from heat.

If I remember correctly heat is manifest as molecular motion, or radiation.

So couldnt a small enough array of MEM's make electricity out of this molecular motion while cooling at the same time?

I'm thinking if this is the case, it could result in a lot of nice applications in the future.

For instance super fuel efficent electric generators, Freezers/ air conditioning that generates electricity, etc

Although this sounds like it would break some law of physics.

Re:Are there other uses for MEMs

[PDAllen]

If you try to design one, you run into problems. Let's say you try to make a ratchet-wheel sort of thing, so random motions of particles can knock the wheel round one way, but not the other because of the ratchet - and you generate energy from this nice one-way motion. But then you find that either your contraption is too stiff (the ratchet has to be lifted) to move at all, or the ratchet occasionally gets jumped by the random motions and the wheel goes backwards - and no net motion occurs. You can actually prove that there isn't any design that will output energy (let's say in the form of a turning shaft for convenience, but any 'better' form of energy than heat will do) given only energy in the form of a constant temperature background. You can of course generate energy given a heat-sink as well as a source; e.g. Carnot cycle, or a thermoelectric junction. This will always cause the cold side to heat up and the hot side to cool down, though (again provable).

Re:What about cars?

[Carnildo]

Even a perfect internal combustion engine can't be more than about 25% efficient, because of the nature of heat engines. Cars are already getting pretty close to this limit, so any improvements to fuel efficiency will come from techniques like lighter-weight vehicles, better aerodynamics, and techniques like hybrid engines that let the engine run at top efficiency all the time.

Stock Posturing

[MojoRilla]

In fact, unless reversible computing is achieved, computer chips are expected to reach their maximum performance capabilities within the next three decades, effectively halting the rapid advances in speed that have driven the information technology revolution, Frank said. "Reversible computing is absolutely the only possible way to beat this limit," he said.

Anyone who is this emphatic about his own technology, and that it is the "only possible way" is trying to pump stock prices, plain and simple.

All technologies have a serious risk of failure, and even sound technology might fail due to uncontrolable forces. And there is never just one way to accomplish anything.

More about 'Reversible' Computers

[rpiquepa]

I commented on this University of Florida news release a week ago on my blog. Not only you'll see more references and details than on the news release, but you'll also read comments by Michael Frank, the UF assistant professor behind this research effort.

a practical use

[MrLint]

Imagine a computer that ran on heat and got colder the more you used it.. then i could pay video games and have ice cold beer.

Oh thats not what they mean by reversible? Damn

Photographs of "a very simple reversible computer"

[Dlugar]

Universiteit Gent has some pictures of reversible logic gates, including a four-bit adder composed out of Feynman's "NOT, the CONTROLLED NOT, and the CONTROLLED CONTROLLED NOT" reversible logic gates, and some other circuits they've built.

They also have links to other sites about reversible logic and reversible computing, such as Ralph Merkle's Reversible Computing page (from Xerox).

Also note the bottom of the page: there's a vacancy in the research group, for all those just aching for a chance to work on reversible computing! (Looks like you'll have to speak Dutch, though.) ;-)

Dlugar

You are confusing something.

[mindstrm]

Electron flow, current flow, and energy are not the same thing.

Energy does not come from the wall and go into the ground.

A better way to look at it is:

Energy is generated at the power station, and is consumed at the house. Both the ground AND the wire are part of the same circuit; power is transferred over the entire mechanism.

Just get rid of x86!

[EmbeddedJanitor]

An ARM processor running at a few hundred MHz uses around 1W and does not even get warm to the touch -- no heatsink required. THe same goes for MIPS, SH4 etc. Most of the heat in an x86 is caused by trying to make an inherently inefficient architecture work faster. Intel, quite long ago, demoed an ARM core running at over 1GHz. Making a 2 or 3 GHz ARM is not impossible, it's just that the ARM niche is portable devices etc and they have not been put into the x86-dominated space

I remember a thread where people were slagging off at SUVs that they don't get better mileage than a Model T and "is this progress?". Well my first PC has a 120W power supply and that was overkill for the job. My latest has a 350W power supply. Now then folks "Is this progress?".

Re:Just get rid of x86!

[EmbeddedJanitor]

Yes, they'd perform like a Pentium but would only use a small fraction of the energy. An ARM switches less transistors per operation, hence uses less power. It is also register rich, so needs less RAM cycles and a smaller cache too.

The sticking point is software. Same problem is likely to be the case for any "reversable" device too.

Linux runs very nicely on ARM. WinCE does too. XP etc are nailed tp x86 for a long time though.

Basic Problems with Reversible Computing

[reaperbean]

Reversible computing is severly limited in terms of normal processor operations. This means that operations such as modular multiplication start to build up a lot of data since you need to 'remember' the two number multiplied in order to undo the operation.

Consider multiplying two numbers, a and b. So a * b = c. Now to undo the operation you only need c and either a or b. So with normal multiplication (or addition, etc) you have two inputs as such and you need to remember two outputs. This gets worse with modular multiplication (depending on the exact set up) since you may need to remember a, b, and c to undo the operation.

When you think of standerd computer operations, most of them are lossy. The problem with reversible computing is coming up with algorithms that are reversible and still useful. This is the case with quantum computers -- quantum operations are not allowed to lose info, so they are reversible. The most famous quantum algorithm, Shor's Algorithm, will factor very large integers quite easily on a quantum computer. It is actually a probabalistic algorithm, and quite complicated (and interesting). Although the entire opeartion is not reversible (and hense not all quantum), the key components are indeed reversible. Other than Shor's Algorithm, there are not a whole lot of algorithm's for quantum computers becuase they are reversible by nature, and, as such, are limiting to work with.

I agree with the author of the article that more research should be done on reversible chips, algorithims, etc. However, I feel that people should understand the limitations inherent in such a system.

Re:What about cars?

[rarose]

Don't forget that there are drivetrain losses through friction and viscous couplings... that's got to be a few percentage points there too.

And finally slowing the car down by hitting the brakes wastes all of the kinetic energy as waste through the brake pads.

I think some city was experimenting with adding storage flywheels to their buses. Energy normally dissipated through braking would instead go into spinning up a large flywheel, and once the bus started forward again the flywheel energy would help the engine accelerate the vehicle. Seems mighty dangerous though... there'd be a lot of energy in the flywheel that would be unleashed if somebody crashed into the bus.

Huh?

[autopr0n]

Thermodynamics also says that you lose non-heat energy in reversable systems as well. If you throw a ball into the air, you lose some energy from wind resistance, from converting chemical energy in your arm into mechanical energy, etc.

Re:Huh?

[Aardpig]

Thermodynamics also says that you lose non-heat energy in reversable systems as well. If you throw a ball into the air, you lose some energy from wind resistance, from converting chemical energy in your arm into mechanical energy, etc.

Sure, but mechanical losses can always be recovered and put back into a system. Heat losses can't, which is the point of the second law of thermodynamics.

Resistance == Heat

[POPE Mad Mitch]

This may be an overly simplistic view of things, but the heat dissipation of a device is caused by the energy used up in pushing against the electrical resistance of the materials used. Every conductor has a resistance, and even the insulators used arent perfect so they have a resistance, and when a current flows through this resistance some of the energy is converted to heat.

This heat generation happens no matter what the direction of flow is, so even if as this article suggests you arrange the gates and circuits so you can also run them backwards, well thats still going to be current flow which generates heat.
A standard domestic light bulb runs on AC right, where the current is flowing back and forth constantly 'undoing' itself, but it still generates a heck of a lot of heat.

You cant turn the heat back into electricity, you dont make cold by running a hot circuit backwards, and your not transforming the electrical energy into some kind of mythical information energy, you are merly controlling the path of electrons on their inevitable flow back towards a state of equillibrium.

Re:Resistance == Heat

[anubi]

Interesting...

So it looks as if we could get rid of the resistance, we could have essentially "perpetual" computing, much like we have essentially "perpetual" current flow in a superconducting ring... except this time with switches directing the electron flow amongst many parallel channels.

I think you are onto the superconducting computer.

I read the main article and was kinda confused about the use of "resonators" to store energy with much less loss. I design a lot of switching power supplies, and I use those techniques a lot to boost the efficiency, as well as reduce stress, on my power supply components. By doing resonant designs, I can use stray capacitances to my advantage, storing their energy in inductors during switching intervals, then re-introduce the stored energy back into the circuit at the proper time to make some really cool power converters.. ( pun intended ).

But here's the problem.. my frequencies are determined by the laws of physics and are either sinusoidal or sinusoidal derivatives. Data is not. I would find it hard to store energy is some sort of inductor, as the energy will bounce back at me in a given time... and if I am not prepared to route the energy in a constructive way when it comes back at me, its wasted, only thing it does then is expend its energy heating up and stressing my switch. I have looked at enough core-dump to know data is not periodic.

It doesn't look like an easy thing to do to try to recover energy from the edges of many switching lines that are all switching at asynchronous times. I would have to know a lot more about this before I could really generate a cogent comment.

Re:Resistance == Heat

[PDAllen]

The point is that power = current x voltage drop.
When you look at the power wasted due to the insulator being imperfect, you've got maximum voltage drop (+V line to ground) but very small current; power wastd flowing along a wire has relatively large current but very small voltage drop (due to small resistance of wires). However setting or resetting a bit will have the large voltage drop and the large current, hence very significant power lost. That last bit is what is being removed.
In fact, you can sometimes cool things down by running a circuit backwards - a thermocouple, for example.

For more info...

[Niten]

You can find more information about Dr. Frank's research on his homepage.

Stirling engine

[bs_02_06_02]

Isn't it just easier to use the excess heat to power a Stirling engine to recapture waste energy?
Maybe the Stirling idea is going too far.
How about a more efficient circuit? It's been awhile since college, but isn't excess heat a sign that the circuit is inefficient?
While it's not completely frivolous research, it's not the first avenue I would approach when looking at this problem. It seems more difficult and time-consuming to add in circuitry to re-use the energy to perform other actions inside of a CPU. It seems like you'd have a better chance at compounding the problem, rather than helping it.
However, make the circuit more efficient, you'll generate less heat. That would be my first goal. What kind of efficiency do they get with today's CPUs?

With this reversible thinking, I have an idea. I need a little help from the anti-SUV crowd... wouldn't all gasoline engines be better off with really big flywheels?

Re:Stirling engine

[zymano]

1. Stirling engines are too bulky . While efficient heat engines they are not practical . They have been used once in project by detroit but didn't produce much power. Power is key ,what everyone wants.

2. Circuits compared to reverse logic are probably inefficient but the heat comes from Clock speed and thinner interconnects and poor insulators(dielectrics at present sizes).

This is the reason for the push for spintronic transistors . google that.

Re:Stirling engine

[dbIII]

I need a little help from the anti-SUV crowd... wouldn't all gasoline engines be better off with really big flywheels?

It's a trade off - more weight, size etc. I've been on a bus with a large flywheel, which worked well due to frequent stops and starts. Really big flywheels give you problems due to momentum, cornering becomes difficult in one direction (eg. Concrete mixer trucks for parts of the world that drive on the left spin the other way to those that drive on the right - to stop the things falling over when drivers take tight turns off the side of the road).

In this case we are talking about electrical energy - which is why a bit more creative thought is required.

Re:Stirling engine

[Mignon]

Isn't it just easier to use the excess heat to power a Stirling engine to recapture waste energy?

Homer: Lisa, in this house we obey the laws of thermodynamics!

Re:Stirling engine

[kavau]

wouldn't all gasoline engines be better off with really big flywheels?

Or an electric motor, acting as a generator, as in a hybrid car. With the flywheel you might run into trouble: install it with its axis oriented horizontally, and it acts as a gyroscope, preventing the car from making turns. Install it with the axis oriented vertically, and your car would swerve every time you hit the brakes. Maybe two flywheels aligned vertically would do the job... but they'd make the car much heavier, of course.

Re:Stirling engine

[drinkypoo]

Flywheels store energy, so in that sense they are useful. However your engine goes through a broad range of RPMs because so does your powertrain. In the case of a diesel engine, which never turns high RPMs and thus operates in a smaller range of RPMs, a heavier flywheel makes sense - and diesels do in fact have heavier flywheels.

Having a heavier flywheel means you have to do more work with the clutch. In the case of a friction clutch, this translates into clutch wear, and in the case of a fluidic clutch (automatic transmission) this translates into more heat, which is wasteful. So the answer is no.

Actually, the reverse is true; a lighter flywheel (or rotating assembly in general) is beneficial, because the engine has less rotational mass to fight when its revolutions are being matched to those of the powertrain. However studies and experience show that most people will stall their car out all the time if the rotating mass is too little. Performance cars, whether sold that way or upgraded, tend to have lightened flywheels because they make the engines more responsive.

Re:Stirling engine

[Kynde]

wouldn't all gasoline engines be better off with really big flywheels?

Or an electric motor, acting as a generator, as in a hybrid car. With the flywheel you might run into trouble: install it with its axis oriented horizontally, and it acts as a gyroscope, preventing the car from making turns. Install it with the axis oriented vertically, and your car would swerve every time you hit the brakes. Maybe two flywheels aligned vertically would do the job... but they'd make the car much heavier, of course.

Err, the low efficiency of a combustion engine cannot really be helped by a flywheel. The compression ratio defines an upper bound for the thermal efficiency, which for common engines is IIRC somewhere around 20%. The waste energy actually comes out of the exhaust pipe.

Well, I'm not saying that a flywheel in theory couldn't help a bit, but it's help is negligible compared to the other sources of inefficiency.

Re:Stirling engine

[kavau]

Err, the low efficiency of a combustion engine cannot really be helped by a flywheel.

I was actually more thinking along the lines of storing the excess energy generated when braking in the flywheel, instead of dissipating it into heat. How much good that would do, if technically feasible, I have no idea.

Here are some links. Mit's pendulam project.

[zymano]

Mit's reversible computing page and pendulam project

Another site with Reverse logic gate pictures

Re:What about cars?

[xanderwilson]

First off, I added that line as a joke. Have you ever heard of the urban legend that says that an oil company has a patent on a design for an ultra-efficient car and whenever anyone is about to come up with a design that's similar, they are able to close down its development? You mean, the power companies are going to force Intel to make their chips more wasteful, causing progress to halt and people to buy fewer Intel chips? Yeah, sure.

No, I don't think power companies will have any direct contact with Intel, but I think they would take the same position regardless.

I don't see Intel having much incentive for energy-efficient chips so long as consumers are interested in speed and power--in much the same way that US consumers are more likely to pay more for a less-efficient larger car with more power than an energy-efficient car. A focus on energy-efficiency will likely be a shift in resources for Intel, and, if so, will likely work against "growth" in speed and power in the short term.

So I'd say if there's going to be encouragment or discouragement towards more energy-efficient chips, then it will likely come from a political level (rightly or wrongly, successfully or unsuccessfully). Automobile companies have balked at and/or fought against legislation such as gradual percentage increases in fuel efficiency, and, if any political body takes a similar interest in the efficiency of microprocessors, I see Intel taking a similar position.

Alex.

Re:Ya'll got that computer on backwards...

[Smidge204]

My computer is reversable! I back it up every weekend...

=Smidge=

Asynchronous Logic will be here first.

[Cordath]

Asynchronous Logic (i.e. no clock) has many of the same benefits, as well as potentially increasing the speed of processors significantly.

A rather large portion of the heat genreated by a processor is just from the clock signal propagating to every bloody logic gate in the mess including the parts not in use. With asynchronous logic, if a part isn't in use, it gets no current. Of course, clock signals have been used for the last half century for a reason. Clock signals are used to time signals so that you don't have 3 digits of a number showing up before the rest, etc. With asynchronous logic you have to worry about path lengths down to the picometer so you don't need the clock to act like a traffic warden. The biggest holdup to asynchronous logic has been the immense design difficulty involved, but that is changing as new design tools are developed.

Anyways, the big reason why Asynchronous logic is going to arrive on the processor scene long before reversable logic is that it already has. Intel and other manufacturers are already incorporating asynchronous logic into their designs, and plan to increase the ammount used as time goes by. The different manufacturing techniques required are slowly being phased in. Reversible computing, on the other hand, has virtually no chance of showing up within the decade.

My point is that the article linked made no allowance for the increasing use of asynchronous logic. It's going to have a significant impact on heat dissipation in the neBuew years.

Re:Asynchronous Logic will be here first.

[mrtroy]

Asynchronous blows for non-trivial computation.

You are correct, the clock signal needs to get stronger/faster as speed increases.

But try designing a whole motherboard using asynchronous design...it would be VERY hard.

Hence why nobody has (that I am aware of)

Clocked is much simpler...

Another benefit of asynchronous would be speed benefits...instead of something taking 1/3 of a clock cycle having to wait, it just finishes when its done.

Re:Asynchronous Logic will be here first.

[Yobgod Ababua]

"Asynchronous blows for non-trivial computation."

Please cite your references or evidence to this statement if you wish to be taken seriously.

Several companies are currently working on complex and high-performance designs using asynchronous techniques. It's true that it is currently more difficult, predominantly because current design tools are all geared towards generating and testing "standard" clocked logic, but it is being done and it does not by any stretch "blow".

It will be quite some time before all of the components on a motherboard are asynchronous, but groups -have- designed processors, memory controllers, and other components in asynchronous.

For but the briefest of examples... check out this article or this article. No, it isn't the answer to everything... but it's much farther along than you seem to realize

Re:Asynchronous Logic will be here first.

[drinkypoo]

The solution to avoid digits of a number getting somewhere out of order is to use serial connections between functional units (possibly multiple serial connections for multiple data.) In fact, this is being done in asynchronous logic today. While this is an issue, it can be solved through intelligent design, rather than brute force approaches like making all your traces the same length. Besides, it's not that they don't have clock signals, they just don't have a master clock signal. You put the data on the lines, you send the clock, rather than putting the data in a buffer, waiting for the clock, putting the data in the send buffer, waiting for the clock. You just toggle the "clock" line (or in this case the ready to send signal) and then when the next functional unit has the data, it toggles its ready line. It's like magic! You can do it a bit at a time from and to buffers on a serial bus, or you can do it multiple bits at a time, either way it's known and solved.

You could certainly design a whole motherboard in this fashion. As the sibling to my comment points out, the tools for designing asynchronous logic are only starting to be there, which is probably the only reason it hasn't been done. Gate counts are simply too high today to just do it by hand :P

Re:Asynchronous Logic will be here first.

[mvdw]

"Anyways" is not a word.

Re:Asynchronous Logic will be here first.

[wavestream]

With asynchronous logic you have to worry about path lengths down to the picometer so you don't need the clock to act like a traffic warden.

No, the most successful class of Asynchronous circuits is wire length invariant. It's called QDI (Quasi-Delay Insensitive). The challenge is asynchronous logic is implementing efficient logic that obeys the hankshaking between concurrent processing elements.

Most of the popular press articles on async are bogus. If you really care, you should read some technical papers

Re:Asynchronous Logic will be here first.

[hughk]

Actually I have seen asynch logic used for non-trivial stuff - how about engineering design? Of course, it was in 1976 and that computer was to be retired in favour of a synchronous machines. Many early computers used asynchronous design because of their physical size. This particular one was an Atlas.

More info

[cgb8176]

For those of you who are seriously interested in this topic, I recommend the book "Minds, Machines, and the Multiverse" by Julian Brown.

Aside from dealing with the above mentioned topics of information (== -log probability), entropy, and reversible computing, it moves on to show the relation of these topics to quantum computing. It even has a good bit of history.

A very good read.

don't know exact details

[nestler]

I don't know the exact details of how it is more efficient. It was explained to me once in a quantum computation course (where among other things they were using equations to relate energy to information).

So I don't know how to explain in terms of currents and transistors, but it is similar to what mikee is saying in this thread (that thermodynamic laws say that destroying information will always consume energy).

The reason quantum computation guys tend to know about this area is because all logical operations on a quantum computer (except for the measurement at the end) are reversible operations.

Uh, been there, done that

[JCMay]

Information is already thought of that way by the people that actually do communications theory. Shannon's Theorem states that the amount of data a given channel can propagate expressed in bits per second is

C= Wlog2(1+S/N)

where W is the channel bandwidth in Hertz, S is the signal power in Watts, and N is the channel noise power in Watts. S is defined as

S= kEb/T

where k is the number of bits per symbol, Eb is the energy per bit in Joules, and T is the symbol duration time in seconds. Note that Joules/Seconds= Watts.

A little later, the paper I reference above defines the minimum bit energy required for reliable reception:

(Eb/N0)>= (lim as nmax->0){log2(((2^nmax)-1)/nmax)}

for infinite bandwidth, this becomes (Eb/N0)= ln(2)= -1.59 dB.

In any case, I hope it has become readily apparent that those that deal with communications and signalling theory have considered information to be energy for going on sixty years.

Wait a minute!!! (was Re:For more info...)

[BabyDave]

Frank ... or Frink?

Re:What about cars?

[AKAImBatman]

Actually, we have figured it out. That's the whole point of the Hybrid cars. There's no way that a hybrid could get more energy out of the system than what's going in. However, it *can* try to use power more efficiently and recover it where possible. Thus you use electricity for acceleration, the engine for cruising, and regenerative braking to get back some of the used energy.

This article is basically describing something similar for computing.

Feynman Lectures on Computation

[John Sokol]

This book by Richard Feynman is based on a series of lectures given at CalTech in the mid 1980s.

In it he discusses Reversible Computation and the Thermodynamics of Computing and quantum computing.

As usual, Feynman was way ahead of his time.

I highly recomend this book.

The basic idea is heat is only generated when information is destroyed. So don't destroy information when performing computations.

How this relates to something actualy practical is hard to say, but it didn't strike me as something that would apply to silicon very easily.

John

Re:Feynman Lectures on Computation

[Alomex]

This book by Richard Feynman is [from the ] 1980s. In it he discusses Reversible Computation and the Thermodynamics of Computing and quantum computing.

As usual, Feynman was way ahead of his time.

Reversible computing had been proposed twenty years earlier by an IBM engineer and widely recognized as an important idea, so one can hardly credit Feynman for this one.

There has been steady research on reversible computation over the last ten years or so. In fact the best paper award at one of the major CS conferences in 1993 was for a reversible computing paper.

Re:Feynman Lectures on Computation

[John Sokol]

I don't think I actualy credited Feynman as it's inventor, just as someone who lectured on it....

And now has a book printed in 1996 that seems to do a good job explaining it.

Reversible means constant entropy (of the world).

[gnalle]

However the whole idea is much more easy to understand from an energy perspective. A ram chip with data has a certain potential energy. When the data is deleted. This potential energy is converted into heat. The reversible computer is able to convert some of the potential energy into usable power. This way it produces less heat.

Size penalty

[toybuilder]

The problem with reversability is that for any given semiconductor process, it effectively doubles the number of gates that need to be built on the chip, and manufacturers are currently more interested in cramming more features into the chip; not to make them more efficient.

It might be theoretically possible to build smaller and faster chips by reducing the energy/thermal issues, but I suspect most companies are not willing to take that leap of faith.

I bet the first places we'll see reversible gates being used in a full-fledged MCU/CPU would be for a mobile/handheld processor running reversified version of an older (less gates) core using latest processes...

Re:Size penalty

[foniksonik]

What about elegant design? From reading the summary it sounds like they want to do more than just add more gates to 'reverse' the computations... they want to use new design methods such as oscillators and springs to capture and hold the energy as potential which would then be reused when needed in an alternate process.

The point is not to over-engineer for this but to intelligently engineer. it will take more R and D time but will hopefully gain enough to justify the expense.

The truth is known...

[pmz]

"When chips perform millions or billions of erasing (emphasis mine) and other operations in a short time, the total heat becomes substantial, limiting both the performance of the chip and the number of chips that can be packed together in a small space, he said."

Now we know why Arthur Anderson got caught.

Re:What about cars?

[randomencounter]

There exist carbon fiber flywheels that disintegrate to easily blocked fibers for safety. I don't know if they use them in regenerative braking systems, but they do exist.
Other methods used are: storing the energy in batteries (Prius), and storing the energy as compressed air (some busses, I think semi's have something like this too).

Good easy read about reversible computing

[snooo53]

A good book I've read that covers the subjects of reversable and quantum computing is Minds, Machines, and the Multiverse by Julian Brown. It's a couple years old, so it doesn't have the latest advances, but as for basic principles of Reversable and Quantum computing, it's an excellent introduction. Explains RSA, quantum logic, code breaking, q. teleportation, etc... I own a copy and enjoyed it.

Reversable versus Probabilistic Computation

[John.P.Jones]

If you are interested in reversable circuits read what Feynman had to say about them in his lectures on computation.

While they may be helpful for certain things, especially quantum computers (but that is a whole different story) there is a snag. They are deterministic; great CS people like Rabin have taught us the value of probabilistic turing machines and today we use them as the basis of determining what is computationally efficient (BPP, see Michael Sipser's intro to computation and complexity). Every once in a while you have to take a non-reversable step to pick a random number (as well as through away garbage you don't want to store any more) and this negates the thermodynamic advantages of reversible computing.

No Free Lunch

Didn't Cray already sort of do this?

[Anonymous Coward]

Reading the article (nice and short, it was!) reminded me of the way the Cray-1 was designed: All the logic signals had both true and complement forms. This was necessary to drive the twisted-pair interconnect if the signals went off-module, and also had the advantage that the power supply mostly saw a DC load instead of a wildly-varying load depending on what was happening in the CPU. Thus, the power-supply filtering could be a lot smaller than it otherwise would have had to be, which was good, because it drew a LOT of power!

I realize that the point here is to not draw a lot of power, but somehow the two things seem related...

Re:What about cars?

[pmz]

Even a perfect internal combustion engine can't be more than about 25% efficient

Does perfect also mean zero-mass pistons and connecting rods and perfect-insulator cyinder walls? Or is perfect the best possible for mass-production?

It always bothered me slightly that car engines have four, six, eight, etc. masses being thrown back and forth several thousand times a minute. I think electric cars have tremendous reliability potential from a mechanical standpoint...I wonder if mechanics are wondering about their job security.

The Fredkin Gate

[Anonymous Coward]

Seminal paper on conservative logic which is clear, readable and built into an interesting unification of computation and physics.

is this possible?

[Major_Small]

i don't know much about processors, but is it possible to send all the unused electricity to a capacitor somewhere nearby on the motherboard, and then draw some of the power for something else (a smaller chip?) from that capacitor?

what I'm thinking is that the CPU does billions of calculations/second, but some other chips don't run as fast and don't need as much power, so they can take what's left over from the CPU and other chips and use some outside energy.

is that possible? like i said, I don't know much about electrical engineering, so I don't even know if it's practical to map a ground pin to a capacitor...

Re:is this possible?

[aXis100]

Possible, but not practical

In order to move the charge to a capacitor, you need an electromotive force. Normally, the differences in voltages could do this, but:
1) The logic voltage is very small
2) As soon as you start charging the capacitor, the capacitor voltage rises and your potential dissapears.

You could do it with more complicated circuits like charge pumps, voltage doublers, DC-DC converters etc, but that would be very demanding and expensive.

It's just not a practical level of waste to try and deal with.

Worst Article Ever

[alexq]

After reading that article very closely, thinking "that doesn't make any sense", thinking some more, reading some other posts, and _finally_ getting the basics of this technology (as badly named as it is)...

I can only suggest that this James Clark seek a career working for SCO's legal department. With his ability to confuse an issue and their desire to do so, it's a match like peanut butter and jelly.

HUGE memory usage problem with reversible machines

[snooo53]

The HUGE problem with reversible computing (IIRC) is that it takes gobs of memory to do relatively simple operations. I believe it's called the "Garbage problem" and is the reason reversible computing hasn't taken off when it was conceived in the 60's.

Why? Because the machine has to save all the intermediate results in a computation. Conventional processors don't do this, and those intermediate results end up as heat. If that seems wierd, imagine a simple AND or OR gate... the 2 inputs become one output... then the gate has to be reset before the next clock cycle, which costs energy to do.

Favorite Quote...

[lordvdr]

In fact, unless reversible computing is achieved, computer chips are expected to reach their maximum performance capabilities within the next three decades...

Ok, lets look back at personal computing 30 years ago. Oh, we can't. How can they predict what is going to happen in 3 decades? We can't predict the state of computing in 3 YEARS. He says

reversible computing is absolutely the only possible way to beat this limit

This is absolute BS and is FUD on the scale of MS and SCO. He's just trying to bump his research grants up. ignore him.

-lv

Re:What about cars?

[Carnildo]

That means zero-mass pistons, friction-free bearings, perfect-insulator walls, infinite-diameter valves, the whole nine yards.

computers don't use 10% of energy

[danharan]

That's a canard perpetrated by companies that don't want the US government to sign on to Kyoto- "if you force us to reduce consumption, we could lose our high-tech edge". Besides, there are many things that are making our computers far more efficient. Hibernation mode, better screens, etc... There is already significant economic incentive to produce energy efficient CPUs for laptop makers, since this could boost effective battery life. Hopefully one day, we'll have digital ink displays that only consume a fraction of what they do now. With efficient chips (which also require less ventilation), we should have laptop batteries that can last more than a day :) I don't know much about all this reversible computing stuff... but anything that's sold as a solution to a non-existent problem, however elegant the technology, is intrinsically bad. Just my $.02

Re:computers don't use 10% of energy

[hesiod]

> That's a canard perpetrated by companies that don't want the US government to sign on to Kyoto

Why is there always at least one asshole who makes every story political? Your statements in no way apply to what is going on here. The idea is to reduce HEAT, not consumption.

He's trying a lil too hard

[NanoGator]

" ?Reversible computing is absolutely the only possible way to beat this limit,? he said."

I'm not sure I agree with this statement. There's another way to make a computer process faster. Run less data through it. That's how they have improved video cards. They compress texture data to fit it through the narrow pipe and get it to the screen. Who's to say that other forms of compression will make their way into processing?

I think it's funny how he promotes this as the next generation in processors. I just don't see it that way. I mean, it'll probably help quite a bit. But I mean it's not like we'll see 2x the speed here. Some operations that generate heat won't happen as often, but under real-world conditions we won't see it happening so much that they'll be able to clock it much higher than it's at now.

I'd appreciate being corrected on this if I'm mistaken. However, the way the article reads, it doesn't sound like much more than an overhyped tweak.

Re:Horrible article

[Logger]

Thanks. These links should've made up the original slashdot post. :)

GigaHertz crisis

[verloren]

"In fact, unless reversible computing is achieved, computer chips are expected to reach their maximum performance capabilities within the next three decades"

So if Moore's law holds true, by that point processors will have barely passed the 3 million GHz level. Clearly we need to get this working now!

Re:What about cars?

[pmz]

zero-mass pistons, friction-free bearings, perfect-insulator walls, infinite-diameter valves, the whole nine yards.

You know, this sounds just like a software requirements document!

Upgrade to 400 MHz and TWM !

[billstewart]

Sure, reversible computing can radically increase your performance! Upgrade from that 4 GHz Pentium 4 with Windows 2004 to a 400 MHz XScale with Linux and TWM, and you'll find things go a lot faster, quieter, and smaller! (And if that's not fast enough, upgrade to MGR....)

Slightly more seriously, though, you might try IceWM or Blackbox if you want something newer than TWM but still lightweight. Are there any other good low-overhead window managers?

Re:Upgrade to 400 MHz and TWM !

[aschlemm]

I find TWM much too Spartan. For systems lacking CPU horsepower I find that WindowMaker isn't too bad and is much nicer than TWM. At least with WindowMaker you can resize a window by dragging on the bottom corners. TWM is too much work since you have to select "Resize" from the WM menu first and click on the window to resize it. Too much work IMHO.

I've found WindowMaker is also available for the Cywin/Xfree86 port and I've found it to be a reasonable replacement for the rather expensive Hummingbird Exceed product for Windows.

I love my new muitneP processor

[Anonymous Coward]

Not only is it reversible, it cools my home.

not so hard to believe

[TubeSteak]

1707, "Papin might also claim to have been the first to operate a boat on steam power. In 1707, he built a paddle wheeler, powered by one of his engines, with which he intended to travel to England. A few kilometres from Kassel, where the boat was built, he was stopped by members of the Association of Canal Transporters. They dragged the boat onto the bank and, during the night of June 26, took it to pieces."

1830's, the horse drawn train lobby made hell for the steam locomotive companies to try and keep them out of business.

1880's-1890's The horse and buggy lobby pushed through all types of complicated laws to try and force auto manufacturers out of business. Horse and Buggy manufacters actually sued car drivers to get them to go back to using carriages.

1895, Rudolph Diesel the diesel engine and originally ran it on peanut oil. He suspsiciously died on a boat trip, and the petroleum industry was more than happy to bury the idea of biodiesel fuels and instead sell off this crap they didn't have any use for as diesel fuel.

World War II: As part of the wartime reparations, Germany was forced to give up much of its textile industry. Dupont (thanks to the U.S. gov't) got its hands on the recipe for nylon and figured out a cheap way to make paper from wood. Up till then, everything cloth that had to be durable was made from hemp. The Bill of Rights and the Declaration of Independence were written on hemp paper, canvas sails were actually made from hemp, official gov't documents... written on hemp. Anways, Dupont lobbied the Dept of Treasury then later on, the Fed Bureau of Narcotics outlawed it completely.Marihuana Tax Act of 1937
Someone's Masters Thesis on the whole boondogle

More recently 1970s~1980s, record companies sued over audiocassetes, 1976 tv companies sued over VCRs. 1982 The RIAA encouraged everyone to replace their old LPs and cassettes with CDs. 1990 first CD-R writers were made, 1992 Audio Home Recording Act passed. The MPAA encrypted DVDs 1996 to avoid the RIAA's mistake with CD copying. MPAA sued cause DeCSS was broken, 1998 DMCA was passed... the list of corporate interests trumping the public good goes on.

Automobile companies are currently touting a study showing that reducing weight increases the risk of death in an accident. Why is this important you say? Because removing weight is the cheapest way to increase fuel efficiency.

Now to stay on topic ;o) Intel does have an incentive to make smaller cooler chips (Centrino, Pentium M) but a technologicaly based drop in heat output would be a Bad Thing(tm) for coal/electric companies because it would point to a long term drop in demand for electricity (assuming we don't all have solar cells on our roofs in 20 yrs.) Intel has definitely showed their desire to move away from the GHz race and towards greater IPC

Most large industries are not driven by innovation, they're protecting whats theirs and (sometimes) searching for efficiency in the process. Sorry for the rant, but companies are swooping around suppressing information (that college kid's masters thesis where he mapped out most of the US infrastructure, Edward Felten's breakage of the RIAA mp3 watermarks) and you're right capitalism is too powerful, thats why we're running around in nylon running suits instead of wearing our parents hemp clothing or driving around in cars burning vegetable/hemp oil.

I swear i'm not a hippy.

Recharge the battery?

[foniksonik]

It's a little off-topic but why isn't there a way to use all this extra energy to recharge my laptops battery? That would be more analogous to the hybrid car example in any case and much more useful today...

And behold...

[DrewCapu]

The "Gatorchip" is born (sort of).

I'm sure a lot of sweat will go into making those.

Oh wait, that was something else :)

orders of magnitude...

[Jeff DeMaagd]

At 100 watts per CPU being fairly high end today, an automobile with 200kWatt engines (~270HP) are a FAR greater concern.

Re:What about cars? less than 1% efficient

[danharan]

That's only primary energy efficiency: how well the engine extracts energy from fuel and turns it into mechanical energy.

If you consider how much of the initial energy is actually used to move the passengers, it drops to about 1% - pretty dismal.

Just tell me

[thewiz]

where are they going to put a gear shift on a laptop?

Will it cause the hard drive on the system spin in reverse?
Will I be able to see Satanic messages when I retrieve my Word documents backwards?

My computer is already reversible,

[reboot246]

but frankly, the change from putty-colored to plaid just doesn't work for me. Maybe if I changed the other color to blaze orange?

words I understood in the above post.

[Anonymous Coward]

you
both
close
guess

Re:Energy transfer

[PDAllen]

That's not the point. No-one really cares much about the 30W of power you don't have to pay for, or the heating you do, that's going to cost/gain you not much even if your computer is permanently on.
The point is that the 30W of power does not have to be got off the chip and dissipated without letting the chip get too hot and burn up. If you could get your Pentium IV to run on only 0.5W, say, at normal speed, then you could overclock it by a factor of 60 or so, with no extra cooling equipment. (well, if the rest of the computer didn't die)

Arrogance of the scientist

[dhammabum]

From the article:

Frank said. "Reversible computing is absolutely the only possible way to beat this limit," he said."

Don't people ever learn?

Vaporware? Nope!

[WillWare]

Has anyone ever built even a very simple reversible computer?

Carlin Vieri at MIT designed and fabricated a reversible processor, described in his PhD thesis defense. It was really a fascinating gadget. It was very interesting how far-reaching the implications of reversibility were -- he needed to be able to reverse all the conditional jumps in the code, so it influenced compiler design as well as hardware design.

If by "vaporware" you mean "I can't buy one today at CompUSA", then you're correct. If you mean "it has never been built and it never will be built" then you are misinformed.

You want to save *how* much electricity???

[stienman]

You want to cut back on the 100W of heat being released by today's processors?

100W?

I piss 100W when I get up in the morning.

100W will cost you $79 [US] a year if you run it hard and constant every second 24/7/365. ($0.09 per KWH)

In the US, each average family has more power, more cheaply than some cities in other parts of the world.

Furthermore, the energy is still going to be released as heat at some point. Where else does it go??? Sure, you might be able to switch a given transister 3-4 times with the same energy, but once it drops in voltage and current, the transister no longer switches. Furthermore the chips are already being run at 1.x volts, which is barely enough to account for the voltage drop anyway. To get enough energy back after a transister you'd have to put in a greater initial voltage, wasting more heat.

Furthermore, more transisters means more complexity, more electricity, and more speed problems. I'm sure there's some savings, but once you add everything up it simply isn't worth it for mainstream desktop processors.

It may be worthwhile in battery operated, low speed, high efficiency processors, but it'll be a long time before a wall is hit that only this technology can help with.

The reality is that this guy's patent is running out, and he's shopping it around to see if he can eke anything out of it.

-Adam

Re:You want to save *how* much electricity???

[foniksonik]

100W = $79 US x N million cpus = $79,000,000 x N each year... what do you think 79 billion dollars in energy? It's not inconceivable considering business servers in addition to consumer PCs plus mobiles plus embedded....

Yeah I'd say it's worth it if it works.

79 BILLION dollars a year.. we could rebuild a war ravaged nation each year... oh yeah and there are more CPUs made every minute so just keep adding to that number.

no.

[rebelcool]

manufacturers are currently more interested in cramming more features into the chip

Completely untrue. A PDA doesn't need a P4. But it does need an energy efficient chip that won't drain the batteries in 30 minutes. There are alot more cellphones/pdas/other low-power portable devices out there than PCs. A 'manufacturer' as you put it would have to be ran by monkeys, stupid ones at that, to not cater to the millions of low-power devices demanded.

In the EE departments of every decent university these days, energy efficiency is being taught right alongside gate and speed efficiency in chip design classes.

Re:no.

[toybuilder]

Touche'. I should have known better to make such a blanket statement -- after all, I'm working with one such low-power processor (AMD/Alchemy Au1000).

But it seems to me that a good chunk of the work in improving efficiency today still comes from process improvement and selective block powering, and not from more "exotic" approaches being discussed on this topic.

YOW! (off topic)

[RoboProg]

Zippy's got nothing on this guy (the link) for off the wall fortune-fodder.

Energy recovery

[Colonel Panijk]

The concept is somewhat analogous to hybrid cars now on the market that take the energy generated during braking and recycle it into electricity used to power the car.

Somewhat off the subject of "reversible computing", but that reminds me... since CPUs put out so much heat, could you stick a thermoelectric generator (Peltier effect, IIRC) between the chip and the heat sink and generate some electricity from heat that is otherwise discarded?

I'd tell you my sig, but then I'd have to kill you.

Information == Entropy

[caveat]

...and that's the end of it. Refer to a physical chemistry text for the details.

More energy efficient than

[rofthorax]

I thought that was a great idea, learning about CMOS in school.. Just rewrite all registers by XOR 1 when the voltage rises above the potential half voltage (when all bits are high), flip polarity, high voltage is represented by low voltage, low by high, statistically making full on voltage improbable.. Simple and efficient energy and concept.. Now what could top that!!

100 watts less A/c too

[hughk]

Unless you live in Alaska, you may notice that leaving desktops running contributes a fair bit of heating to the room. In summer that means pumping that 100 watts outside, which itself is taking energy. Take that from the home to the office where there are hundreds of PCs then the heat output is extremely important.

The thing is that data is stored in a mechanism whereby charge is topped up because of leakages but never dumped. Effectively the power usage (and thus the heat) should then be minimal, being the amount needed to maintain the charge and the switching losses.

Gate count doesn't *have* to go down so much

[logpoacher]

Two problems with your comments.

One, as has been observed, is that the power issue is to do with managing the heat generated by the CPU, not the costs of providing that power.

The other is that an increase in component count doesn't imply a proportional decrease in gate count. A large amount of chip space is devoted to wiring and module interconnections, so just because the gates are twice as big doesn't mean that the die needs to be twice as big. In fact, if the extra transistors can be "snuggled in" to the existing cell designs, then you might find that the gate sizes don't increase much either.

It might even help - the power supply tracks won't need to be so meaty!

Now, it may be that reversible systems require a massive increase in hard-to-place transistors and long wires - in which case the above doesn't apply! But I guess my point is that the relationship between component count and chip area is non-linear - I remember when we first started working with CMOS instead of NMOS, and it became clear that just because there were twice as many transistors, it didn't mean that the gates were twice the size.

RTFA

[danharan]

"The research comes at a time when computers are estimated to consume as much as 10 percent of electricity in the United States"
http://www.napa.ufl.edu/2003news/efficientcomputer .htm

So I guess reading the article makes me an asshole on /., eh? D'oh!

Re:RTFA

[hesiod]

> So I guess reading the article makes me an asshole on /.,

No, using this story to emphasize your political views makes you an asshole. You brought up Kyoto, which was not even alluded to in the article.

> > > anything that's sold as a solution to a non-existent problem [...] is intrinsically bad

And reusing electricity and reducing heat aren't nonexistent problems.

Re:RTFA

[danharan]

FYI, I don't even agree with the Kyoto protocol. I didn't bring up my political views, but another group's lie that made it into the article as a rationale for this type of development.

To be clear: computers do not consume 10% of electricity. This was parroted in the article, even though the source of that assumption distorted statistics to suit their political ends.

The distinction is subtle.

On another note, you have repeatedly called me an asshole while I haven't. However if you persist in doing so, I will be forced to conclude you are one.

Re:RTFA

[hesiod]

> That's a canard perpetrated by companies that don't want the US government to sign on to Kyoto

Sorry if I misunderstood your point or something, but the statement quoted above sounds like you were saying "The outright lies from companies are the only reason we haven't signed this wonderful treaty." And that is so totally political. BTW there was only one place (not "repeatedly") I outright called you an asshole. That was only after making the witty comeback of "I guess reading the article makes me an asshole," although the thing I was arguing wasn't even IN the article. Even without my persistence, your conclusion may be correct, as I have a rather biased opinion of myself :)

My exact words were: "Why is there always at least one asshole who makes every story political?"

That means that I think whoever brought it up is an asshole. If you weren't the one who brought up Kyoto, then that means it doen't apply to you, unless it was a joke and I made a mistake, which did not appear to be the case. However, I was of the belief that you were the one who brought it up, since there was no mention of Kyoto anywhere in the article itself.

I use to program in 1's and 0's

[KJSwartz]

With the newfangle Differential Computing nowadays, people are just programming with +/-1's.