Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[mastermenthe]

At last, we have computers out that can run Dwarf Fortress in it's full glory, without any FPS drop!
http://www.dwavesys.com/en/dw_homepage.html

Anyone want to try their hand at translating Dwarf Fortress from binary code into qubit code?
http://www.dwavesys.com/en/dev-tutorial-intro.html

[Gigaz]

At last, we have computers out that can run Dwarf Fortress in it's full glory, without any FPS drop!
http://www.dwavesys.com/en/dw_homepage.html

Anyone want to try their hand at translating Dwarf Fortress from binary code into qubit code?
http://www.dwavesys.com/en/dev-tutorial-intro.html

As far as I know, no quantum algorith for path-finding is known which is more efficient than the classical ones. But you could easily add quantum dwarfs to your fortress which enter a room through two doors simultaneously and create funny interference patterns.

[No1]

Well the problem is that any path finder needs to check all nodes at worst. The only thing that can be made better is heuristics that reduce the check on unnecessary nodes, that is it guesses what nodes might not be relevant. I do not know if Toady uses any heuristics but I would imagine that there are at least some implementation of A* as it is a efficient and simple path finder.

I've been on a lecture about quantum computers and as far as I know they would exceed in breaking encryption and other number heavy calculations but not much else. Of course there are calculations in DF that might be done faster but path finding is as far as I know the biggest user of CPU time.

[zombie urist]

I read the URL as dwarves sys and immediately thought that they specifically designed a computer to play DF.

[Tharwen]

Well the problem is that any path finder needs to check all nodes at worst. The only thing that can be made better is heuristics that reduce the check on unnecessary nodes, that is it guesses what nodes might not be relevant. I do not know if Toady uses any heuristics but I would imagine that there are at least some implementation of A* as it is a efficient and simple path finder.

I've been on a lecture about quantum computers and as far as I know they would exceed in breaking encryption and other number heavy calculations but not much else. Of course there are calculations in DF that might be done faster but path finding is as far as I know the biggest user of CPU time.

I think he's said before that he uses A*. Don't take that as fact though...

[mastermenthe]

I read the URL as dwarves sys and immediately thought that they specifically designed a computer to play DF.

Considering what these things are capable of, they might as well have.
Who knows, maybe a quantum GPU might even be able to handle Dwarf Fortress 3D!

[veok]

I read the URL as dwarves sys and immediately thought that they specifically designed a computer to play DF.

Considering what these things are capable of, they might as well have.
Who knows, maybe a quantum GPU might even be able to handle Dwarf Fortress 3D!

DF ain't usin' most of its processing cycles on graphics.

I'd wager that the graphics system is one of the *least* processor intensive systems (esp. since spritesheets), and don't expect that would change assuming reasonably simple 3d graphics.

[Stargrasper]

The big boon of quantum computers is the quantum bits can have more than two states, that is, we move out of binary bits.  That allows us to do some really amazing things.  I'm not sure how much this would help DF, but porting it to this kind of architecture would take decades.

[MagmaMcFry]

The big boon of quantum computers is the quantum bits can have more than two states, that is, we move out of binary bits.  That allows us to do some really amazing things.  I'm not sure how much this would help DF, but porting it to this kind of architecture would take decades.

Nope, we're still in binary. And even if we moved out of binary, it would not be much better, because human logic is mostly binary. If you had a computer that ran in ternary or even decimal, it could add larger numbers slightly faster, nothing more.


I've looked at the programming tutorials of the D-Wave 1, and wow, that computer seems pretty useless. Programs only have a certain probability to output the correct solution, and 128 qubits are much too few, given the architecture. But not even all of these qubits can be used; about 20% of these will be faulty and can't be used. Then some qubits have to be linked to each other, giving you about 13 usable qubits. Finally, you can't run algorithmic programs using these qubits, so all a program can actually do is locally minimize a quadratic equation in 13 variables, getting the wrong solution half of the time. And all you need to do that is a 10-foot cube with insane energy consumption. That's even worse than the first non-quantum computers. But the worst part: It doesn't even use quantum superposition, the stuff that's so great about quantum computer theory. In fact, the whole fucking system could probably be redesigned with simple memristors instead of super-cooled niobium strips and still fit onto a microchip.

Hooray?


I'll rebuild one of those quantum computers in DF. Shouldn't be too hard. Really.

[Stargrasper]

The big boon of quantum computers is the quantum bits can have more than two states, that is, we move out of binary bits.  That allows us to do some really amazing things.  I'm not sure how much this would help DF, but porting it to this kind of architecture would take decades.

Nope, we're still in binary. And even if we moved out of binary, it would not be much better, because human logic is mostly binary. If you had a computer that ran in ternary or even decimal, it could add larger numbers slightly faster, nothing more.


I've looked at the programming tutorials of the D-Wave 1, and wow, that computer seems pretty useless. Programs only have a certain probability to output the correct solution, and 128 qubits are much too few, given the architecture. But not even all of these qubits can be used; about 20% of these will be faulty and can't be used. Then some qubits have to be linked to each other, giving you about 13 usable qubits. Finally, you can't run algorithmic programs using these qubits, so all a program can actually do is locally minimize a quadratic equation in 13 variables, getting the wrong solution half of the time. And all you need to do that is a 10-foot cube with insane energy consumption. That's even worse than the first non-quantum computers. But the worst part: It doesn't even use quantum superposition, the stuff that's so great about quantum computer theory. In fact, the whole fucking system could probably be redesigned with simple memristors instead of super-cooled niobium strips and still fit onto a microchip.

Hooray?


I'll rebuild one of those quantum computers in DF. Shouldn't be too hard. Really.

I haven't looked into the cited machines.  Are they really quantum computers?  I'll look into them later.  My comments and knowledge come from my academic knowledge of computer science and physics, my fields of study.  I supposedly know a few things about quantum physics and quantum computing.

[YetAnotherStupidDorf]

I haven't looked into the cited machines.  Are they really quantum computers?  I'll look into them later.  My comments and knowledge come from my academic knowledge of computer science and physics, my fields of study.  I supposedly know a few things about quantum physics and quantum computing.

Just wait hundred or two years and you will have decent quantym computer that actually computes. Considering that quantum computers can have gains only in few specific cases (cryptography), I do not think they will be ever as popular as classic binary computers.
So, if there is no way to gain from other cases, I predict there will be quantum chip on motherboard or future equivalent used for quantum cryptography etc.

[MagmaMcFry]

Actually, most useful problems that can be translated to a problem the quantum computer can solve can also be solved with a normal computer in about 0.1 seconds with a simple brute force algorithm.

[No1]

Actually, most useful problems that can be translated to a problem the quantum computer can solve can also be solved with a normal computer in about 0.1 seconds with a simple brute force algorithm.

What do you mean? Do you mean that the problems that can be done on quantum computers today can be done by brute force on a normal computer? The only thing I know they have done on a quantum computer is to factor a number, I believe it was 15 so yes that could have been done really fast on a normal computer.

[MagmaMcFry]

No, the factorization of 15 was done on a real 4-qubit quantum computer, with quantum superposition, entanglement and such.

[YetAnotherStupidDorf]

Actually, most useful problems that can be translated to a problem the quantum computer can solve can also be solved with a normal computer

Yes, quantum computer cannot solve problems that normal computer cannot solve.

in about 0.1 seconds with a simple brute force algorithm.

Nope. Entire point of quantum computers is that they can solve certain class of problems many orders of magnitude faster than equivalent classical computer.

You know, first cars could be outrun by walking human, too. Surely useless invention, right?