Is a Power State Other Than 0 and 1 Possible?

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

NanoStuff

Banned
Mar 23, 2006
2,981
1
0
You're asking about quantum computers :) You have 0 or 1 or both at the same time. Hard to compare to digital computers, but you can generalize that as three states.
 

CTho9305

Elite Member
Jul 26, 2000
9,214
1
81
Originally posted by: grant2
Originally posted by: CTho9305
You can significantly decrease the voltage if you're willing to run slower.
That's what I said isn't it? (your quote cut off most of my post).

You said
Modern ICs operate at the minimun voltage necessary for stable operation (1.5v or whatever).
Then you mentioned that a trinary computer would need to run at a higher voltage and slower, so I didn't interpret your first sentence as, "minimum voltage to run stable at the target speed".
 

xtknight

Elite Member
Oct 15, 2004
12,974
0
71
Originally posted by: NanoStuff
You're asking about quantum computers :) You have 0 or 1 or both at the same time. Hard to compare to digital computers, but you can generalize that as three states.

Interesting. I have no idea what quantum computers even are, but if you could have both, couldn't you also have one or the other, yielding four states?

Both off: 0
Left one on: 1
Right one on: 2
Both on: 3

Or am I thinking too digitally?
 

f95toli

Golden Member
Nov 21, 2002
1,547
0
0
Quantum computeras are not digital, but NanoStuff's explanation wasn't entirely correct either.

A QC is essentiall a computer built using quantum bits known as qubits.
A qubit is 2-state system where we can denote the states |0> and |1>, i.e. it is still essentially a binary computer.

The difference between a qubit and an ordinary bit is that whereas a bit can be in either
state 0 or 1 a qubit can also be in a SUPERPOSITION of those states, e.g. a state

a|0>+b|1> where a,b are are numbers.

Depending on how you interpret this you can say that a qubit is in "both states at once" but there are a number of problems with that interpretation, it is usually better to think of it as a "hybrid" between the two states.

There are a numer of algorithms that use this to great effect, Shor's algorithm for factorizing number into primes is the most famous but there are others.

There are various ways to implement qubits: The "prototype" is a en electron in a magnetic field where spin up/down are the states; NMR experiments use various molecules. There are also a number of solid state implementations; mainly superconducing electronics (charge- and flux- qubits) but also some semiconducting versions using e.g. quantum dots.


 

NanoStuff

Banned
Mar 23, 2006
2,981
1
0
Originally posted by: xtknight
Originally posted by: NanoStuff
You're asking about quantum computers :) You have 0 or 1 or both at the same time. Hard to compare to digital computers, but you can generalize that as three states.

Interesting. I have no idea what quantum computers even are, but if you could have both, couldn't you also have one or the other, yielding four states?

Both off: 0
Left one on: 1
Right one on: 2
Both on: 3

Or am I thinking too digitally?

You're definitely thinking too digitally :), I'm referring to a single 'transistor', not a pair. I mean quite literally a single switch being on and off at the same time. It's not something that can really be imagined in newtonian physics. The best analogy would be to compare it to a regular digital transistor that happens to have a value of 0 and 1 at the same time, not one or the other, or better yet, the same transistor being in two places at once, each instance having a different value, which as you might have guessed is quite impossible, but nevertheless, that's quantum physics for you. It's called superposition if you're up for a read. Having three states is probably not the most accurate description, the fact is it's two states at once, which state gets interpreted being uncertain but having a higher probability than the other.
 

shortylickens

No Lifer
Jul 15, 2003
80,287
17,081
136
I dont normally come in here but in case anyone is curious I work at a semiconductor plant in Oregon.
I was also a Navy Electronics Technician for 9 years and before that I was in my high-schools electronics club. Actually, I was the president all 3 years.

And before anybody even takes a guess, YES, I played Dungeons and Dragons.

Getting back to the point at hand, with what we currently know about transistors we can only have two states for a bit of data. On and off. Or high and low. Commonly refered to as a 1 and 0 for data purposes.
You cant hold a bit of data with only one transistor. You need two of them, at least. The voltages we have from TTL and CMOS are not written in stone. It used to be between 0 and 5 volts and depending on which type you were using, a given voltage might get you a 1, a 0, or an undefined, which was no good and usually resulted in errors. But you can not get three genuinely seperate states in data. It all has to be ones or zeros because of how the transistors are tied together and hold voltage, which in turn translates to data later on.
In advanced devices like CPU's, they run on much less than a volt. But you still only get 2 usable states.

Also, as an aside, changing from a 1 to a 0 and 0 to 1 causes the transistors to use power, which is why the more times per second you change those states, the more heat you build up.
I.E. the more powerful your device is (CPU, RAM, whatever) the hotter it gets. Thats why we try to use lower voltages and less current in each new generation, to keep the heat reasonable so the devices can be cheaper and used easier.
Also less power draw for portable devices.

I have very little idea on how quantum computers work, but after listening to most of the discussion here on anandtech I can honestly say we shouldnt be too excited just yet. That type of computer theory needs a lot more practice before its useful to the industry.

EDIT: Sorry.
I was thinking DC versus digital in my previous comment. You could get any number of distiguished voltages if you had enough transistors, but (as was already pointed out) all the extra transistors/gates needed to handle that more complicated bit would slow down everything else. You would lose all that performance. Modern processors have billions of transistors already. We have hit many limits as is because we cant make them smaller and fit more of them on a die. At my fab we are using UV lasers and its still not enough to just keep getting faster and faster.
Also as was pointed out, the major change to the math involved in changing 0,1 to 0,1,2,3 would offset the supposed performance increase.
Its kind of pointless. Better to keep it as ones and zeros and just figure ways to make the whole process more efficient.
 

Snooper

Senior member
Oct 10, 1999
465
1
76
Intel has been doing this for years. Look up info on there "strataflash" flash memory. Basically, they are storing one of four distinct voltages (states) in the gate, which effectively means each flash cell is storing two bits of data instead of a single bit. It requires a fairly complex system to analyze the cell and determine what the actual programmed state is from the stored charge (voltage) on the gate.