Slashdot Mirror
Can We Surpass Moore's Law With Reversible Computing? (ieee.org)
"It's not about an undo button," writes Slashdot reader marcle, sharing an article by a senior member of the technical staff at Sandia National Laboratories who's studying advanced technologies for computation. "Just reading this story bends my mind." From IEEE Spectrum:
[F]or several decades now, we have known that it's possible in principle to carry out any desired computation without losing information -- that is, in such a way that the computation could always be reversed to recover its earlier state. This idea of reversible computing goes to the very heart of thermodynamics and information theory, and indeed it is the only possible way within the laws of physics that we might be able to keep improving the cost and energy efficiency of general-purpose computing far into the future...
Today's computers rely on erasing information all the time -- so much so that every single active logic gate in conventional designs destructively overwrites its previous output on every clock cycle, wasting the associated energy. A conventional computer is, essentially, an expensive electric heater that happens to perform a small amount of computation as a side effect...
[I]t's really hard to engineer a system that does something computationally interesting without inadvertently incurring a significant amount of entropy increase with each operation. But technology has improved, and the need to minimize energy use is now acute... In 2004 Krishna Natarajan (a student I was advising at the University of Florida) and I showed in detailed simulations that a new and simplified family of circuits for reversible computing called two-level adiabatic logic, or 2LAL, could dissipate as little as 1 eV of energy per transistor per cycle -- about 0.001 percent of the energy normally used by logic signals in that generation of CMOS. Still, a practical reversible computer has yet to be built using this or other approaches.
The article predicts "if we decide to blaze this new trail of reversible computing, we may continue to find ways to keep improving computation far into the future. Physics knows no upper limit on the amount of reversible computation that can be performed using a fixed amount of energy."
But it also predicts that "conventional semiconductor technology could grind to a halt soon. And if it does, the industry could stagnate... Even a quantum-computing breakthrough would only help to significantly speed up a few highly specialized classes of computations, not computing in general."
[I]t's really hard to engineer a system that does something computationally interesting without inadvertently incurring a significant amount of entropy increase with each operation. But technology has improved, and the need to minimize energy use is now acute... In 2004 Krishna Natarajan (a student I was advising at the University of Florida) and I showed in detailed simulations that a new and simplified family of circuits for reversible computing called two-level adiabatic logic, or 2LAL, could dissipate as little as 1 eV of energy per transistor per cycle -- about 0.001 percent of the energy normally used by logic signals in that generation of CMOS. Still, a practical reversible computer has yet to be built using this or other approaches.
The article predicts "if we decide to blaze this new trail of reversible computing, we may continue to find ways to keep improving computation far into the future. Physics knows no upper limit on the amount of reversible computation that can be performed using a fixed amount of energy."
But it also predicts that "conventional semiconductor technology could grind to a halt soon. And if it does, the industry could stagnate... Even a quantum-computing breakthrough would only help to significantly speed up a few highly specialized classes of computations, not computing in general."
Moore's Law is about device sizes and economics, not about energy use.
Can We Surpass Moore's Law With Reversible Computing? NO. This does nothing to address Moore's law and shows an ignorance of what Moore's law is by posing the question.
My favorite Slashdot stories are the ones that I absolutely do not understand. Honestly. I'm a lot more likely to actually read TFA when the summary means absolutely nothing to me.
You are welcome on my lawn.
Moore's Law is very much about energy use. In fact, the ability to decrease transistor size is directly tied to the ability to control the energy these transistors consume.
When transistors get smaller they naturally consume less energy. But that's not enough. Significant effort is requires to ensure that they consume even less than that, especially when we're dealing with 22 nm and especially 14 nm processes.
Why is that? Electromagnetic interference.
When you're dealing at extraordinarily small scales like nanometers, the electromagnetic interfere between transistors and other components within a processor becomes a huge issue. Managing it becomes extraordinarily important for the successful operation of a processor.
If the cross-talk interference between adjacent transistors isn't strictly controlled then digital errors will be readily introduced, which prevents the computation from being possible. Digital circuits don't work properly when one transistor's state change causes electromagnetic interference that causes unwanted and unexpected state changes in these nearby transistors.
It's asinine to suggest that Moore's Law isn't related to energy use. EVERYTHING about microprocessor technology is related to energy use!
therefore, no
Pain is merely failure leaving the body
"...wasting the associated energy."
Surely the energy is not "wasted", it has been used to create the output.
This sig left unintentionally blank.
Here is what they mean. Imagine logic elements are people passing notes. Except when you pass a note, the next person reads it, throws it in the trash, then rewrites a new one to pass on. Big waste, right? They are proposing logic gates that simply pass the note along based on its content. Much more efficient, right?.
The bad news? Good luck doing that at today's speeds. We lose more energy simply biasing the transistors heavily to make them switch faster than we ever do by erasing states. We have heat limitations due to this much more than charge lost every state transition. It might give incremental improvement in density, but it's not some silver bullet.
The trick is to undo the operations that produced the intermediate results. This would allow any temporary memory to be reused for subsequent computations without ever having to erase or overwrite it.
... which results in thermal dissipation... which results in increased entropy... which is exactly the thing that you were trying to avoid in the first place. Yet Another Free Lunch.
The only way I can see this working is if you use very low temperature super conducting grids... like they already do in quantum computers. I just can't see any improvement here without material science being involved.
I don't see the logic in your proposal. If your proposal was implemented it would require that people respond to trolls before the troll could be moderated. Responding to trolls, flamebait and off topic comments are exactly the wrong thing to do. I'm not completely sure I'm not violating this principle with this response. If you feel so strongly about this why do you feel the need to post anonymously? Besides, moderation isn't directly censorship. If you don't want to read "censored" responses just set your browsing level to -1. People choose to self censor based on how much they choose to trust the members who moderate.
Honestly, I think we'll have quantum dot cellular automata before we get reversable computing. In doing so, it would eliminate our power consumption issues in regard to computing. As always, the real problems lie with the manufacturing of these devices.
Anons need not reply. Questions end with a question mark.
This is something that calls for a proof of concept in the form of linear programming. Go ahead, show me the machine tree and its related Karnaugh maps and show this bi-directional computation performing several classic computing staples like stacks, sorts and finding primes in a manner that involves fewer steps.
Information has its limits, too, and laws somewhat similar to thermodynamics appear to govern these limits. If you have some linear function g(c(b(a))) that doesn't necessarily mean you can complete it as g(a,b,c) if c is dependent on b is dependent on a.
For instance, there are bi-directional programming languages but you still are forced to rethink your problem to be solved in a way that work toward the solution is still being done in reverse, and frankly I doubt that all real-world problems have a solution where time=t can be decremented. For starters if you need more than one output for a given input, you're kind of screwed for any linear task.
I have to agree with those who see this as a gag to win more funding. It's the equivalent of bringing again e.g. bidirectional programming over to the the hardware level, and go ahead and find me all the amazing examples of what you can do with bidirectional programming languages (go ahead, there are several and some are a number of years old).
"Stratigraphically the origin of agriculture and thermonuclear destruction will appear essentially simultaneous" -- Lee
It is interesting, as a pure mathematician, to read:
"[F]or several decades now, we have known that it's possible in principle to carry out any desired computation without losing information -- that is, in such a way that the computation could always be reversed to recover its earlier state."
Now this 'can get back to earlier state' thing is basically the 'existence of inverses' axiom of group theory. A semigroup is a structure with a well-defined associative operator, but not necessarily an identity, nor existence of inverses. Now going from one computation state to the next, as a CPU does, is essentially a semigroup operation. Or at least something like that.
Reversible computing is effectively the faithful transformation of an abstract structure (e.g. rotating an icosahedron) on which the possible transformations form a group. Such a condition means that an unbounded number of operations can be chained without loss. This means the transformation must take zero energy. Thus, in fact, no change takes place. That means that what you think is a computation is, in fact, a fixed point that you're somehow conjuring into what appears to be a non-fixed computation. Interestingly, to me this stuff isn't new, nor even recent. What the ancient mystics, yogis and others obsessed over was this sort of aspect of reality.
Getting back to a less abstract point of view, the problem I see is that if these guys (and girls) insist on reinventing group theory the hard way, they won't even be able to catch up with where group theory was middle of last century. Indeed there is a dire need to more thoroughly think through what computation itself _actually is_. The 'Turing Machine+ChurchTuringThesis' thing is a half-decent first stab, but nothing more. The infinite tape, like the successor and infinity axioms of Peano Arithmetic and ZF Set Theory also, is akin to a naive C programmer assuming that malloc() will never fail. When you're knocking up a quick prototype, and you're not bothered if a malloc() failure crashes the program, fine. On the other hand, Linux kernel module authors seem to understand the need to use malloc() when it works, but never to trust it for critical duties, whether explicitly, or implicitly (via e.g. printf).
So I read the article and have a basic understanding of the technology. I can see how "reversible" applies at the low level, but it is a poor choice for a description of this process. Adiabatic computing might be better, but people who have never taken thermodynamics probably don't understand that word. I'd suggest something like "No Waste Computing" or "No Heat Computing" might be a better description (neither is strictly true, but the potential waste heat is extremely low, i.e. just saying "low heat" doesn't seem adequate). My descriptions are not necessarily a better description in terms on understanding what is going on, but a description of what the benefit of the process is, which I think is more appropriate in this case. Reply to This Share
Yeah, but, where's the example of an implementation that saves anything?
"Stratigraphically the origin of agriculture and thermonuclear destruction will appear essentially simultaneous" -- Lee
Somebody should do a text file in old BBS/Usenet style, titled "Why Your Proposal To Remove Moderation From Slashdot Isn't 'A Thing'."
"Stratigraphically the origin of agriculture and thermonuclear destruction will appear essentially simultaneous" -- Lee
While I agree some of the causes of your angst, please remember that crowd moderation will always have this kind of side effect.
Any crowd-moderated community will frown upon comments which don't agree with said community's majority point of view/beliefs/preferences. Here on Slashdot it happens to be Linux, Android, F/OSS, and the USA.
Posts disagreeing with group think will be downvoted anywhere, ad if that's not possible, they will simply be ignored with the same outcome. just deal with it or go to a forum which agrees with you :)
Note: I am generally not in agreement with group think here, but I don't mind, I'm active on a dozen other communities and I can get my kool-aid there :)
...gis sdrawkcab (usually not responding to ACs; don't bother posting as AC)
I was forced to turn up the level of skepticism since conclusions drawn from false assumptions cannot be trusted. Sounds more like somebody wants to jump on a current fad and see if they can ride it to some publicity. Probably best to forget them and let the real researchers come up with something more real.
"Still, a practical reversible computer has yet to be built using this or other approaches."
Since quantum computers of any kind have to be reversible due to the very nature of QM, every realization of quantum computation is a reversible computer.
This includes the controversial D-Wave machine as well IBM's QC chip that you can play with online.
What a dickwit. If we are all such a great union of minds here with our own understanding of what is intelligence and what is dribble, why do you even want to post your shit here? Who do you expect to read your brain farts and fall in love with your ideology, if you have already determined that nobody here wants to hear your shit? This is bordering on the clinical definition of insanity. Just close the fucking web browser, or go to some other site, where maybe a big bunch of fuckwits are desperately awaiting your arrival to share your munted view of the word with them. Cos that aint us.
Right-o!
Get rid of modpoints, shadowban the trolls.
I suggest that people who are really interested in understanding this subject read and understand the papers reprinted in "Maxwell's Demon: Entropy, Information, Computing" (first edition: ISBN: 978-0691605463; second edition: ISBN: 978-0750307598).
makw the transisters out of gas, plasma or quarks. Ever smaller!
They've discovered antispacetime.
Switching is an erasure of state
Imagine you have a container with a wall in the middle. There is a door that can be opened or closed. There is a person who can open or close that door to let individual gas particles through. Now suppose he opens or closes this door such that all the fast moving particles are moved to the left side of the container and all the slow to the right. It turns out that this person can do this without consuming any energy so long as he remembers how he did it in such a way that it can be reversed. The minimum energy required for him to forget how to reverse this opening and closing is the amount of energy a Carnot engine could produce using the hot and cold sides of the container. Basically, calculating is free, forgetting is what consumes energy.
Reading this, I thought of their essays.
This isn't the first time I've heard of reversible computing and its purported benefits, and over the years every time I looked it up I haven't seen significant advancements or implementations. This article is no exception. And I'm still not convinced there exists any design of a reversible-logic processor is practical or useful for general purpose computing, assuming that the physical hardware problems have been solved.
I would be quite happy to see a software simulation of an 8 bit processor with a simple instruction set and a small amount of RAM, all implemented in simulated reversible-logic gates. Does anyone know if such a simulation exists? Or is the answer "well, it's complicated" which means probably not?
The moderation system is designed based on the assumption that moderators are unbiased and just, which is very far from reality (they are often biased and/or malicious).
Rather than just allowing people to moderate stuff, they should also provide a short, one-sentence summary why they are moderating the post that way. This allows a 3rd-party reader to judge whether the moderation is accurate.
I did not read this article, so I'll just make up some numbers I find plausible.
Somehow we need to equate entropy (or information loss) with energy. The assumption of 1eV per bit is probably ok. One electron, either changing potential of 1V or not.
So by making computations reversible, we could avoid this inevitable 1ev loss if the computation is not reversible. Nice. But if we currently burn 1keV per switch, there is no point talking about this technology right now. Let's shave off another 990eV first. Then we can think about reversibility.
You have to understand that making the computation reversible (not loosing this bit of knowledge) gets you from 1000eV to 999eV. That's premature optimization. And you all know that this is the root of all evil. Though we might keep an eye on it.
reverse computation back. Honestly, or was it 1997? Unfortunately, I dumped all my old Bytes due to not enough space, but I know that somewhere in the second half of the 1990's there was already an article about reversible computing. Since in those 20 years time, there haven't been further advancements in this field, I would think that this an idea that is born dead. Also, as reversible computing can be thought of as the electronic equivalent of a weight-counterweight system, I do not see that his helps Moore's law. It would just use twice the number of logic. Reversible logic has only to do with trying to use less energy.
No it isn't, just that their bias is useful. I have very little bias so I don't get a vote.
The moderation system is designed around the law of large numbers, with one filter: the better an account's karma is, the more moderation points it receives in time.
I have been wrongly moderated a couple times, in topics where TFS, TFA or both were incorrectly bashing Microsoft and I pointed it out, but at the same time I expected the moderation to swing that way. But generally I am happy with how it works.
The trick is to set the right expectation and not care too much.
...gis sdrawkcab (usually not responding to ACs; don't bother posting as AC)
The IEEE author, Michael P. Frank, derides conventional computing as "space heaters that happen to do a bit of computation".
I would firstly point out, this article was published on the IEEE site which certainly uses conventional computers. Next, it is 99.9% likely the author wrote the article on a conventional computer. I'm writing my /. reply on a conventional computer.
Meanwhile, Mr. Frank has accomplished what with his neato idea? He advised a student in 2004. And "still, a practical reversible computer has yet to be built using this or other approaches."
Oh, I get the appeal of fundamental research. Maybe Mr. Frank is on to the Next Big Thing. We certainly could use some breakthroughs now that Moore's Law is growing weaker every year.
However let's be real for a minute. Honest-to-gosh breakthroughs that revolutionize conventional computing are rare. Maybe 1 in 10,000 ideas get any traction at all, and most of the successful ones still wind up being niche solutions at best. Zip drives, anyone?
Perhaps Mr. Frank ought to show a bit of humility when tossing off those "space heater" zingers. My space heater happens to be one of the most powerful machines mankind has ever created. It is also cheap enough that most people can afford one. And Mr. Frank advised a student 13 years ago, aren't you special!
As I understand reversible computing it's basically a recycling of data to preserve electrons before they are allowed to disappate as heat. The idea being that the more you reuse an electron the less heat a chip will create. The problem is not so much that the chips aren't designed this way today its got more to do with how fast chips lose electrons as heat due to the fabrication technology they are built with. As most hardware savy people know, the smaller the chip the less space between transistors there is for electrons to be lost as heat. So as fab tech improves it requires less energy to keep a chip running because less energy is being lost between transistors [just keeping the chip functioning without data corruption]. As fab tech improves we will get to a point where it becomes feasible to use reversible computing because there will be so little energy being lost in this way. In other words more heat will be being produced not from the current leakage of a chip but from data waste caused when a bit is no longer needed. Last I heard, that's around the 5nm mark. There is also reversible computation, the software equivilant of this which is an interesting read imo.
When I see "frictionless", I think perpetual motion machine. Which can't exists because of the laws of thermodynamics. Then there is this "fully reversible" concept here that claims it's not only allowed, but inspired by the laws of thermodynamics. So I'd like to ask : Although we are putting new energy in the processors, how can we make information go through it without it loosing any energy or experiencing any friction? Isn't it impossible? Or how do we compensate?
Now on the internet: More tears from a loser who no one wants to listen to.
If you're getting downvoted, it's for a reason. There are pro-America and pro-Microsoft posts that receive positive moderation.
Some things, like Microsoft and anti-piracy, run contrary to the free flow of information. The open source movement and Slashdot (on average) both approve of these things. In order to receive positive moderation, you need to be insightful enough to overcome readers' biases.
Can't do that? Then go cry somewhere else.