A serious question

On/off is far easier to work with.

>serious question
no way

>>>/sqt/

>>57461768
Transistors work on on and off states.
Electric current either passed through them or not.
Hence all the basic operation are based on "on or off" = "1 or 0".

>>57461780
I know but what will happen if we use this on todays processors and quantumm PCs?

>>57461806
Quantum processing doesn't operate the same way.

>>57461768
There was development on analog CPUs a long time ago. It was stopped because fixing such things was a titanic feat.

>>57461797
They have to be a different explanation. My teacher asked I didn't find anything in university sources

>>57461797
I don't know what a transistor is the post.

We were trying analog computers too before we figured out digital was the way to go. It's a practical matter. We got op amps for our trouble so it wasn't all bad.

It's just easier to add another bit to data representation than it is to design a circuit than can distinguish and properly process twice as many voltage levels.

>>57461829
Read up on analog computers. Seriously op amps were a big deal even before they got put on a chip and sold for dirt cheap.

>>57461768
Because it's easier to distinguish between an on/off signal than between various different voltages, specially when you want to keep them low.

>>57462095
This. Analog computing requires much more robust manufacturing methods, higher voltages, and would require a great deal of compiler and pipeline redesign. And the benefits aren't all that clear.

When quantum computing takes off, we will probably see a great deal more interest in analog-style logic.

>>57461806
>>57461824
Quantum computing operates a similar way, instead of "on/off" you can "observe" different quantum states of electrons and determine your "data" that way. So essentially instead of "on/off" you get "on/off/left/right" and whatever else you can determine from an electrons position, so instead of 1 and 0 you get 1,0,2,3
That's why quantum computing is so hot, it'll multiply all processing with almost no extra effort. We just have to get quantum electron determination down to a room-temperature, long-lasting consumer level, and it's expensive to read electrons for data

>>57461768
why not -1 0 1

>>57463716
There's no such thing as negative.

>>57463748
There's no such thing as memes

>>57463716
01110111 01101000 01111001 00100000 01101110 01101111 01110100 00100000 00101101 00110001 00100000 00110000 00100000 00110001

>>57463716
why not -2 -1 0 1 2?

>>57463859
makes sense compared to 0 1 2 3 4 5

>>57461768

because transistors. you have to understand that CPUs really only work because transistors can be made so small.

But I think binary is a natural choice for a number system anyway because 1 is a physical constant (kind of). Also I couldn't imagine what we would do if we didn't have the bitwise operations

>>57462171
multiple voltage levels is not the same thing as analog.

analog would be continuous voltage as opposed to discrete levels

here's something with more than 2 voltage levels, which is not analog
https://en.wikipedia.org/wiki/Ternary_computer

File: osamafinger.gif (83KB, 304x400px) Image search: [Google]

83KB, 304x400px

because of cheap jews

trinary is objectively superior

>>57461768
Why todays sedans uses 4 wheels instead of 5, 6, 7 etc... ? I know it's possible but why not?

>>57461982
>We got op amps
what?

>>57465756
Witches post amongst us!

>>57467267
Operational amplifiers

>>57461829
>university sources

wtf you're on a university?
Where are you from?

>>57463576
no, quantum computing works completely differently from traditional computing, what you described is just a computer that does base 4 instead of base. it's not just a different base.

File: sbahj-are-you-serious.gif (282KB, 768x273px) Image search: [Google]

282KB, 768x273px

>>57461768

File: setun.jpg (196KB, 992x625px) Image search: [Google]

196KB, 992x625px

>>57463716
>>57463748
https://en.wikipedia.org/wiki/Ternary_computer

Triple state could probably be done, or even x number of multi state.

Traditionally we use 0 and 1 for on, off. However I think with today's advances, we can create 0, 0.5, 1, etc given that we probably have tech to control precise number of electrons. Or maybe I'm overthinking, these types of electron regulators maybe either too expensive or doesn't scale well at nm scaling.

they already do, mathematics is just a lie so you can change the base of your number system

>>57462171
>And the benefits aren't all that clear.
Analog machines are faster, but they can be constructed only for one purpose, and digital machines are a lot cheaper now.
I never heard about programmable analog machines.

>>57469631
Anything with a third state needs some degree of precision, which is a nuisance if you're trying to put a billion of them on a chip that costs a few bucks.

With two states, you drive the output either as high as possible or as low as possible. Then the input only needs to distinguish between "low" and "high"; the threshold level could be off by a volt and it wouldn't matter.

>>57469742
Analogue computers were "programmed" with patch cords (i.e. physically changing the wiring) and turning knobs (changing capacitor/resistor values).

They don't "execute" programs in the way that digital computers do. They're basically configurable electronic circuits, used for solving systems of differential equations. You create a circuit whose behaviour is described by the equations you want to solve, then the circuit's outputs are the solution.

>>57461768
Computerphile gives a simple explanation.

https://www.youtube.com/watch?v=thrx3SBEpL8

>>57469631
what you want is a -1 state, so you can represent negative values and magnitudes naturally, the current IEEE signing shit is a god damn fucking nightmare that complicates execution units and programs considerably

that's the benefit

>>57470986
>the current IEEE signing shit is a god damn fucking nightmare that complicates execution units and programs considerably
>implying you work with bits today

Anyway, it's too late. It's like Chinese stuck with hanju, we're stuck with binary. We have enourmous pile of computers to support, we barely can handle such maneuvre.

>>57463576
That's not at all how it works you dumbass, and those aren't the problems at all you fucking ignoramus

Why do you comment on something you have no understanding of whatsoever?

The concept of quantum superposition does not allow for observations of non-discrete values (you write 1,0, 2, 3 which is technically discrete but belies the fact that you don't understand at all what the fuck 0,1 signify), that's the whole point of the wave function collapse hypothesis.

>>57461768
Why is AMD still using pins?

The new stuff they're teasing is still using pins.

>>57471489
is there a problem with pins? FX has flat bottom on the chip and the pins in the socket if that's what you mean