18

CMOS greatly reduces the current draw of ICs because one of the complementary FETs is always in the non-conducting mode, so there is only a flow of current during the transition between states, which is just the amount of charge on the gate's equivalent capacitance and maybe some leakage when both gates are open momentarily.

Is it theoretically possible to make a logic gate that has zero leakage while changing states (using any realistic technology), and the signal is just passed through the circuit as changes in voltage causing other changes in voltage? If not, what's the theoretical minimum?

tyblu
  • 8,167
  • 6
  • 40
  • 70
endolith
  • 28,494
  • 23
  • 117
  • 181

5 Answers5

10

Yes. You can make a gate that switches with zero current if you don't mind waiting an infinite amount of time ;) Since current is change in charge over change in time, as the change in time goes to infinity the current goes to zero. Run your logic as slow as you can while meeting your other system specifications.

Your homework assignment for tonight is to read the "Thermodynamics of Computing" chapter from "Feynman's Lectures on Computation" ;)

tyblu
  • 8,167
  • 6
  • 40
  • 70
jluciani
  • 11,646
  • 1
  • 34
  • 54
9

It is not possible to make an electronic logic gate that functions even when its current is always zero.

However, it is possible to arrange CMOS electronic logic gates in such a way that the energy capacitively stored on the transistor gates is later returned to the power supply, so it is using almost zero net power. Once the system is powered up and all the bypass capacitors are fully charged, those logic gates can do an arbitrarily large amount of computation while pulling nearly zero current from the battery. Such arrangements are often called non-destructive computing.

Also, there are many ways to build logically equivalent computational structures without any electronic devices. Such non-electronic logic gates naturally use zero current, although nearly all of them require much more power to operate than their logically equivalent electronic logic gate.

non-electronic computing

Some non-electronic logic gates are listed in the article "Ten weirdest computers".

A few more non-electronic logic gates that are apparently not quite weird enough to make that article:

David Cary has designed a CPU to be built entirely out of spool valves, and is still pondering whether to power the thing with traditional hydraulic oil pressure, water pressure, or air pressure.

The fluidic logic gates have no moving parts, if you don't count the fluid moving through them as a "part".

(Is there an article on Wikipedia or some other wiki with a list of ways to implement the abstract concept of a "logic gate" ?)

non-destructive computing

Non-destructive computing, also called reversible computing, Charge Recovery Logic, or Adiabatic Logic, involves gates that use almost zero power.

When a computational system erases a bit of information, it must dissipate a theoretical minimum energy of kT ln(2) -- the von Neumann-Landauer limit -- where k is Boltzmann's constant and T is the temperature.

Most logic gates erase a bit of information for every logic operation. However, there are a few logic gates that preserve every bit. In theory these non-destructive logic gates could use far less power than the theoretical minimum power of bit-destructive logic gates.

"Reversible Logic" by Ralph C. Merkle at Zyvex

RevComp - The Reversible and Quantum Computing Research Group has some nice photos of their reversible CPU.

davidcary
  • 17,426
  • 11
  • 66
  • 115
  • Adiabatic logic is pretty much what I was looking for. Something to improve on/get around the imperfections of CMOS. – endolith Aug 17 '10 at 23:51
  • Though I'd still like to know the lowest possible amount of energy theoretically necessary to process information. – endolith Aug 24 '10 at 21:08
  • 2
    All computer architectures alternate between storing bits in one place, piping stored bits through some combinational logic such as an ALU, and then storing the result bits elsewhere. Storing a bit requires a theoretical minimum of kT ln(2). There seems to be a debate over whether the lower limit of the combinational logic is actually zero or merely small compared to kT ln(2) -- or in other words, over whether the theoretical maximum amount of combinational logic operations one can do with a given amount of energy is infinite or finite. See http://www.cise.ufl.edu/research/revcomp/ . – davidcary Sep 10 '10 at 10:49
  • CMOS gates use much more than kT ln(2), though, don't they? That's like 10^-21 J per bit, or picowatts at 3 GHz. – endolith Jan 24 '11 at 03:55
  • 1
    Can could reversible computer be Turing-complete, being able to solve any problem with a bounded function of the amount of memory that would be required for a non-reversible Turing machine>? – supercat Mar 04 '11 at 23:53
  • 1
    @endolith: Yes, *today*, gates use much more than kT ln(2) -- both destructive and non-destructive CMOS gates. However, if [current trends continue](http://www.cise.ufl.edu/research/revcomp/RevComp-trend.jpg), the energy used by non-destructive CMOS gates will drop below kT ln(2) around 2025. Destructive gates, whether CMOS or anything else, cannot use less than the theoretical minimum of kT ln(2). – davidcary Mar 07 '11 at 13:19
  • A quote from Wikipedia: "Theoretically, room‑temperature computer memory operating at the Landauer limit could be changed at a rate of one billion bits per second with only 2.85 trillionths of a watt of power being expended" https://en.wikipedia.org/wiki/Landauer%27s_principle – endolith Apr 19 '12 at 19:22
6

No, it is not possible.

The gate capacitance is a function of the transistor geometry and the properties of the transistor materials. There will always be capacitance. In an effort to minimise capacitance there will always be trade off between transistor speed, voltage breakdown, gain and other device properties.

Not only that, but in order to use the output of the gate, the transistor must drive any output capacitance. Again, the output capacitance is a function of the wire geometry and the properties of the surrounding materials.

There are also other leakage effects. Across the drain and source of any transistor in the off state and even some leakage current into the gate. While these effects are for the most part negligible in actual silicon parts, you would come up against them sooner or later in your quest for a zero-current gate.

endolith
  • 28,494
  • 23
  • 117
  • 181
Clint Lawrence
  • 2,526
  • 17
  • 13
  • I didn't say it had to use transistors. – endolith Mar 16 '10 at 00:44
  • 5
    Maybe not, but you did mention them and tagged the question CMOS. So it seemed like a reasonable assumption. – Clint Lawrence Mar 16 '10 at 02:47
  • 1
    I mentioned it as an example of a new technology that greatly reduced current usage. I'm wondering if there's theoretically something better. – endolith Mar 22 '10 at 16:37
  • In other words, is it possible to measure the voltage on a wire without drawing any current from it in the process? Is it theoretically possible to make a switching device that has *no* input capacitance? Is it possible to measure the pressure in a pipe without letting any water out? – endolith May 05 '10 at 22:51
  • 1
    > Is it possible to measure the pressure in a pipe without letting any water out? Yes. Just install rubber membranes on both ends. But you cannot do this without moving any water and moving water loses energy to the tubing. :) – jpc May 07 '10 at 11:23
  • @jpc You could also measure pressure by measuring the amount of force you have to apply to the membrane/piston to *prevent it* from moving. – endolith May 30 '11 at 01:24
  • endolith: Good idea. Now, are there methods for measuring force without everything moving at least a tiny bit... ;) – jpc May 30 '11 at 13:35
  • @jpc Why is that necessary? I suspect Maxwell's Demon is relevant: "Szilard ... suggested that proper accounting of entropy is restored in the process of measuring the molecule position. This explanation became the standard one until 1981, when Bennett showed that the fundamentally dissipative step is surprisingly not the measurement (which can be done reversibly) but the logically irreversible erasure of demon’s memory, to make room for new measurements." http://strangepaths.com/reversible-computation/2008/01/20/en/ – endolith Jun 09 '11 at 20:09
3

au contraire:

Your posed headline question can be solved with out using current, or any type of circuit.

http://www.youtube.com/watch?v=SudixyugiX4

stuckie27
  • 397
  • 1
  • 9
  • 1
    If you don't count all the energy required from the human to reset the gates each time... – davr Aug 17 '10 at 18:54
2

If you don't have to run the program to get the result, that would seem like a step in the direction of computing something for nothing, though their apparatus must have been dissipating some power.

JustJeff
  • 19,163
  • 3
  • 48
  • 75