Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Fastest computer

  1. Jun 30, 2003 #1
    there must be a theoretical limit of how fast a computer can be, since the maximum speed is the speed of light. so what would happen after that? Everyone will just buy one computer with the best stuff, i.e. graphics card (1*10^999999999999 MB) and ram, and pass it on to the next generations forever?
     
  2. jcsd
  3. Jun 30, 2003 #2
    One thing you could do if you got a fastest processor is just start putting them in parallel to make yet faster clusters. In fact they could just keep adding this parallel capability within a single unit, so even if you got the fastest fabrication procedure you might still be able to parallize it indefinently.
     
  4. Jun 30, 2003 #3

    paralell processors is your answer.

    If a processor is processing at the speed of light then just use paralell ones forever until each one is computing the least possible smallest amount of information........


    And bTW - no computer will ever calculate anything near the speed of light. It would cause the mass to be so much it would fall through the earth.
     
  5. Jun 30, 2003 #4

    megashawn

    User Avatar
    Science Advisor

    I thought a quantum computer would calculate over the speed of light, simply because of the nature of quantum uhm, stuff.

    I think there is a limit to hardrive space. But even then, just buy another harddrive.
     
  6. Jun 30, 2003 #5
    I don't see what you all mean by calculate at speed of light. What does this mean?
     
  7. Jul 1, 2003 #6
    how can something process faster than light? (is it even possible?)
     
  8. Jul 1, 2003 #7
    If we can slow down time in the future we could contain the light computer inside a time field after giving it something to do and it would calculate it faster relative to our time. taadaa.
     
  9. Jul 1, 2003 #8
    Another way to speed the computing process is to shorten the time it takes to input data. Right now computers are very fast, but it still can take hours to input data for a spread sheet or write a letter, or draw a building using CAD.

    I think that the future of computing is making the interface invisible to the user, by just talking, hand drawing, or thinking the commands. We have voice recognition now, but it still needs refinement; drawing pads are okay, but the computer has trouble deciding what it is seeing; and cache has the ability built in to anticipate which pieces of program to hold that will be used again soon, but this is still limited.

    The human to machine interface is where real leaps in speed still need to occur. Do we need a lightspeed computer when the average user can only type at twenty or thirty words a minute?
     
  10. Jul 1, 2003 #9

    megashawn

    User Avatar
    Science Advisor

    I agree with ya, but I also think that the interface will come along with the technology.

    I do know that one of the major reasons why computers have problems interpreting analog information (drawing, face recognition, etc) is because computers process digital information. Life is analog, so unless we can figure out a way to build analog computers (us) that process on the same speed as current pcs, I'd say we will reach a limit in a short time as to just what a computer can do.

    And that, is why more power is needed. You know, voice technologies have increased like crazy in the past 5 years. Why? I'd say due to the massive jump in cpu speed, increase in fsb, better ram, etc.

    And I'm pretty sure Microsoft is working on technology to give total control to your computer through a mind link type device. Maybe it was on new scientist or something of the sort.
     
  11. Jul 1, 2003 #10

    FZ+

    User Avatar

    I think the limit to be reached is not what the hardware is capable of - superconductors, qubits etc allow computers that are staggeringly powerful. The limit would be from us - what we want, what we can take, what information we can feed it.

    The fastest computer in the universe is just a heap of junk if you have no programs to run on it. And the output from computers would very likely exceed what the human body can experience. An excess of detail, perhaps?
     
  12. Jul 2, 2003 #11
    Yes. I agree.
     
  13. Jul 3, 2003 #12
    I've read that scientist are working on making a Quantum Computer that can perforn a task in seconds that takes a supercomputer years to do. I don't know if that's anywhere near the speed of light, but it is pretty fast.
     
  14. Jul 3, 2003 #13

    FZ+

    User Avatar

    Actually, I'd like a reply to grady's question too... how do you relate processing speeds (which is usually given in terms of processor clock cycles per second, Hz) to the speed of light? If you are talking about the carrier of information within the chips, then you would notice that the "speed" of computers has never changed - the distance travelled to go through each switch has. In the same way, c wouldn't be a practical barrier to computing, as it can be circumvented with advanced switchs and miniaturisation. Quantum computers don't get their speed from going over c. They do it because they don't have the practical weaknesses of silicon computing, and that qubits allow multiple data storage, and perhaps even fuzzy logic computing, allowing switches that are much more efficient.
     
  15. Jul 3, 2003 #14
    it defintly woulden go that high not becuase of physics but becuase we are too slow to notice like i really cant tell the diffrenace betwwen a 1.8 ghz and a 2.4 ghz
     
  16. Dec 30, 2003 #15
    Bus bottleneck

    i think the main limiting factor for traditional computer model will be the bus paradigm. no matter how fast the cpu is, no matter many cpus are connected in parallel, it will always bog down to how the data that is to be processed transported to the processor and the processed result returned for display/stored/etc.

    is there a way to make N-N connections amongs computer components? there are numerous tha have been studied (mostly in the family of hypercubes). however i imagine that someday (or someone has already implemented it?) the wired connections will be replaced by lasers (and further on by protein-based/biological computer?). the connections topology will be in 3-D where the restriction of overlapping electrical circuits will non-issue. switching from one connection to another will be performed with mems or other to-be invented optical switching methods.

    granted these computers will be much more complex than today's ones. but so is einstein's relativity (and later quantum mechanics) compared to newtonian mechanics.
     
  17. Jan 2, 2004 #16
    I would like to stress that there is no limit to the calculation speed of a computer, as the speed of processing is not related to the speed of light.
     
  18. Jan 3, 2004 #17

    russ_watters

    User Avatar

    Staff: Mentor

    There may be no limit because of the speed of light, but there are limits for other reasons - like the size of the atom. There are only a certain number of transistors you can pack into a silicon chip. And then there's heat dissipation...

    5 years is what I give it before we see a major slowdown in Moore's Law (which predicts doubling of the number of transistors and by implication computing power every 18 months).
     
  19. Jan 3, 2004 #18
    Someone said you can't calculate at the speed of light. This is incorrect. There is absolutely no reason why you can't have a chip passing signals of light rather than current. The computer would NOT become too massive and fall through the Earth. The computer itself does not move at C, only the light does, and since light moves that fast anyway, it won't make any difference.

    And as stated, not only can we have it using light, but we can do it all in parallel. Parallel processing is why the human brain is still way faster than any computer on Earth.
     
  20. Jan 3, 2004 #19
    Sorry, Let me rephrase that, as the computing speed of the computers are not dependent on c, therefore, the calculation speed of a computer will not be limited by c. But as russ_watters said, there are other elements that will limit the calculation speed of a computer.

    By the way, I would like to ask adam, why would any one want to build a computer that sends light signals?
     
  21. Jan 4, 2004 #20

    russ_watters

    User Avatar

    Staff: Mentor

    Since high end storage and transmission media are light based and light travels faster than electonic signals, its a desirable direction to go.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Fastest computer
  1. Fastest sort method? (Replies: 8)

Loading...