I must admit that my understanding of the term Turing complete may not itself be complete. Yet it seems to me that passive circuits cannot perform all computations. Digital computers are perhaps ever so slightly circumstantial evidence of that. (Citation needed)

Do you think the addition of the memristor to the shaker would allow all possible computations to be made by passive circuits? Provided of course there was previously a limitation on this.

I'm not sure the question even makes sense, but I would really like to know your thoughts on the subject.

 

Personally I wouldn't call memistors passive, they require current, they have memory. I'm curious where their going with them myself, besides memory that is.

 

HP finds partner to build memristors into chips

 

Will we ever see passive memristors, or are they confined to the realms of silicon?

 

Bill. A cacitor is just two pieces of metal with an insulator unless an electron differential is applied to the control nodes. It's considered passive, electrolytic caps are even more active they require polarity to maintain it's electrical state and functionality.

A resistor is just a substance of a known compositional state until a voltage differential is applied to it, which will result in a current of approximately known rate flowing through it. Keep that in mind when you think of a thermistor which varies it's resistance a known degree depending on it's temperature, which is still a passive device.

An inductor is just a coil of wire until again a voltage is impressed upon it, even straight wires have inductance.

A memresistor is the same thing, it just has different uses. It's effective resistance will change depending on the direction and amount of current that passes through it. Memresistors are currently being made with titanium dioxide compounds not silicon. If silicon is used it's just because it's an easy to use substrate, silicon itself does not mean it's an active device anyways. Many people consider non-linearity as prime requisite of a passive component, which doesn't occur in the real world in real world passive in the first place.

Any state change requires energy transfer, effectively speaking there is no such thing as a passive device. Memresistors however because their construction can be made so much smaller and have incredible potential for both computing as a whole (replacing transistors even) and static memory, it blows away the mechanical complexity of flash memory which is actually far more complex than most people think.

 

So? Memistors require power to program and read, while they are nonvolitle memory they are most definately not passive. If used as a computational device they are also not passive. What they are is extremely efficient and dense, both good things.

 

Bill, capacitors require power to program and read, according to your logic they're not passive components?

Wikipedia's entry on Memristor states that they're a passive device, the entry for passivity states that,

In control systems and circuit network theory, a passive component or circuit is one that consumes energy and cannot control the flow of electrons, but does not produce energy

That's the requirement for thermodynamic passivity, which the memresistor meets.

The voltage/current going through a memresistor does not directly effect the electron flow, it effects the bulk resistance of the materials they're made out of, that resistance change is what effects the electron flow, it's that hands off indirect portion of the memresistors operation that qualifies it as a passive device.

Diodes aren't active devices because they can't cause power gain, they can be used in circuits with bias to provide current gain but the diode is still passive while the circuit itself becomes active.

 

Bill, capacitors require power to program and read, according to your logic they're not passive components?

Wikipedia's entry on Memristor states that they're a passive device, the entry for passivity states that, That's the requirement for thermodynamic passivity, which the memresistor meets.

The voltage/current going through a memresistor does not directly effect the electron flow, it effects the bulk resistance of the materials they're made out of, that resistance change is what effects the electron flow, it's that hands off indirect portion of the memresistors operation that qualifies it as a passive device.

Diodes aren't active devices because they can't cause power gain, they can be used in circuits with bias to provide current gain but the diode is still passive while the circuit itself becomes active.

I think we are going to have to agree to disagree on this one. Memistors require a programming current, they don't come at a fixed value (though there may be a default value). A capacitor has an intrinsic value when they are finished manufacturing, as do coils and resistors. This is not the case with a memistor (BTW, your spelling it wrong). The fact they have programmable memory that requires an electric current to set it moves them apart. If they are a computational device this is also becomes true.

 

A capacitor requires a programming current as well, why is this the second time now that you've brought up that same faulty logic? A capacitor does not have an intrinsic value, in theory it's two plates will never actually be the same so it has no intrinsic state, this does not discount the parasitica in a real world capacitor, however the physical makeup of a memristors is multiple orders of magnitude lower. A capacitor stores it's date in an electric field, a memristor is the material finite measurable resistance.


Capacitors are also used as computational devices in as much they are the set points (the gates) of CMOS logic and it's kin.

You've shown no logical seperation between the two devices mode of operation.
This type of passive was theorized back in the 70s but production methods hadn't come up with any way to realize a function device.


Again simply by stating that memristors can be used as storage elements in programmable arrays proves nothing about their active/passive state, as capacitors are used in the EXACT same way current with flash memory.

If a memristor is an active device then so is a capacitor.

If you can by pure logic refute that I would LOVE to see it.

 

The value of the cap does not change, a 0.1µF stays at that value. What is programmable about that? A resistor is a fixed value, again, not programmable, ditto on inductor. A memristor changes and holds last value, which is set by an electrical current. How is this analogous to a capacitor, coil, or resistor? I don't think my logic is flawed, but I do think you like to argue.

I notice we've both been spelling it wrong, but I added it to my spellchecker. All fixed (just like my cat).

 

http://www.memristor.org/emerging/3...discovery-of-memristor-video-lecture-invented

Please note time 2:45

Please also note 3:40

4:30 is another good quote.

 

Pertinent content starts at 5:00 feel free to sit for the rest of the 47 minutes of it's length. Good stuff at 10:00 as well.

 

Pertinent content starts at 5:00 feel free to sit for the rest of the 47 minutes of it's length. Good stuff at 10:00 as well.

Totally cool. This makes the next 10 years look pretty interesting in the semi-fab world if no major road-blocks are found.

http://www.google.com/url?sa=t&sour...uOPOGMTOXj30R1tnw&sig2=kRltBEBAjCXSLJMHzeHvuw

 

I"m sure there will be roadblocks, I imagine they've hit at least a half dozen, and either worked around them or are figuring out the practicalities, that's what research is for. The future of electronic systems is more than a little bright that's for sure. This is only one thing being worked on too, the electronics market is being massively developed.

I would like to say that even if the memristor is considered a passive component the devices that they're going to be developed for are active devices, so the whole active vs passive thing is you say tomatoe I say tomatoe.. (that analgoy really doesn't work online)

Regardless of any of that it's amazing technology being developed now.