Wolfire Games Forums

Assaultman67 wrote:The fact that it can solve the segments instantaneously means it can solve the exact solution for the problem instantaneously.

A problem. Infinitely long. Solved instantly.

Any problem you can think of, no matter how long, can be solved instantly as long as the answer converges somewhere.

Unless I am misunderstanding something here, how can a problem infinitely long converge somewhere?

it just does ... it converges to a value.

such as if you were to summarize

1/2^n as n goes to infinity.

1/2+1/4+1/8+1/16+1/32+1/64+1/128+1/256+1/512+1/1024+1/2048 ... yadda yadda yadda

Converges to 1, not infinity.

The joke about the mathematician never reaching the door is completely based on this. The mathmatician is too busy going half the distance to the door and adding it to what he has already done. However, the distance from him to the door is a constant and not infinite.

If that fucking blows your mind, look up gabriel's horn. The formula for the surface area diverges to infinity, but the volume that the "horn" can hold is a constant value. However, the horn is also infinitely long.

So ... I can fill the "horn" to a point where it gets full. But I cannot paint the "horn" because it has infinite surface area.

If i were to use my theoretical quantum computer, I could calculate the exact volume instantly, but get a buffer overflow for calculating surface area.

But this happens all the fucking time in computing. A problem is too complex or infinitely long even though it converges to a certain value when expressed discretely so an approximation is made.

Shit, forms of euler approximation appears all the time in video game code due to restrictions in computational power.

But a quantum computer ... wouldn't ever need to approximate.

We're talking game physics 10,000x more accurate with AI as smart as people and a quantum computer could run it as fast as your monitor could refresh. It would still puke if you divided by 0 though

(shit, actually im not so sure anymore if this is possible. simply because im not sure how you would code "loop infinity" ... ah. you could just make a case where the previous iteration matches the current one. Technically not "infinity" due to rounding error, but close enough. Actually, that wouldn't work because it could round off into a loop :/)

Wait wait wait, regardless of how quickly indivitual expressions can be caculated if a problem is infinate you never stop, thus with no end you never get an answer.

Ah, ok. The article on the horn helped.

So, how does one go about this quantum computing?

Zelron wrote:Wait wait wait, regardless of how quickly indivitual expressions can be caculated if a problem is infinate you never stop, thus with no end you never get an answer.

Yes and no.

the problem is infinitely long, but the components of the problem take no time to calculate so ...

0 seconds per component * infinite components = 0 seconds.

Here's the real problem.

The processor could hypothetically process an infinitely long problem, but we don't have a way to store an infinitely long variable to be processed.

This causes a convergent problem appear to be divergent in some cases with limited memory.

Go figure.

So quantum computers would not be rated based on speed, but rather resolution of their results.

Now i need to figure out quantum memory.

Jacktheawesome wrote:Ah, ok. The article on the horn helped.

So, how does one go about this quantum computing?

Simple, process data, send data back in time, process again, send back in time until results have been achieved.

Of course all we will see is outputted results based off of our input and we will not see this whole "time loop".

Now, provided your theortical quantum computer ran C++, what would be on the screen if you gave it this program?

SteelRaven7 wrote:Well, the fact that the series converges to f(x) as is essential, since it's an approximation that with infinite terms, IE infinite computer power is equal to the sought value is essential in the described problem.

Yeah, I'm just namedropping things for funsies!

EDIT: Well this was flawed, I'll fix it in the morning...

I said it converges, instead of going to infinity, which, like you said, is what matters. Also, does this mean a computer could react to our mouse movements before we move the mouse?

Speaking of too much money

Assaultman67 wrote:

Jacktheawesome wrote:Ah, ok. The article on the horn helped.

So, how does one go about this quantum computing?

Simple, process data, send data back in time, process again, send back in time until results have been achieved.

Of course all we will see is outputted results based off of our input and we will not see this whole "time loop".

Simple 0.o

Well honestly, isn't the definition of quantum computers the fact that it uses quantized states of something to account for a number of bytes, or rather a number? And the fact that this might allow logic gates with much higher clock frequencies, maybe even none at all compared to what's possible today? If so, the non-causality described in that article a few pages back isn't vital to the construction of quantum computers, but hey, it'd be an asset .

Basicly, messing with time isn't required for quantum computers.

Here I made it myself

Freshbite wrote:http://www.humblebundle.com/ GO GO GO !!

I watched the video.
Oh god the puns!