Speed limit: 4GHz

Posted:
in General Discussion edited January 2014
Let me start by saying that i am a PC and Mac user. i know the superiority of mac, but i would like to learn all OSs. anyway, this may be of some extra significance since apple is moving to intel chips. (btw, this isnt a thread to discuss that change)



about a year ago i built my first PC. at the time the fastest processor (clock speed) was a P4 3.8GHz. Just about 4 years earlier my family purchased a rather speedy 733MHz. Speeds jumped leaps and bounds between then and a year ago, so why the sudden stall? The fastest P4 is still a 3.8. Sure they've tweeked it many ways-- hyper-threading, extreem edition, dual core, ect., but still no clock speed increase.



Heres one theory i've heard... a while back someone said there was a law saying no computer could be built at or above 4GHz. i imediately dismissed it as B.S. now i'm starting to wonder.



theory two... i've also heard that even at current speeds, the electrons have to be moving at speeds scientists once believed to be literally imposible (thats physics way over my head so dont ask for any more details)



Last point of interest... i have an inside source at IBM who told me something interesting. first i'm sure you've all heard that IBM sold there PC division to Lenovo. he said the reason is IBM is working on something new. A whole new way of computing, nothing like the PCs we know today. I'm sure you've also heard of IBMs new cell processors. He implied that its far more than that, but couldnt say more.



Could any of these things point to a limmit to PC speeds? i never thought it was possible but now it looks like it could be true. does anyone know anything more about any of this? is so, please share.

Comments

  • Reply 1 of 20
    powerdocpowerdoc Posts: 8,123member
    There is no physical laws preventing a chip to go further than 4 ghz.

    But engineers has found that increasing the clock speed was not the best way to have more powerful cpu.

    There is 2 issues with very high clock speed : heat and watt consumption.

    Intel gave up their 4 ghz P4 for this issue, and are going to give up the net burst technology (the one of the P4).

    The next géneration chips will be based upon the pentium M design, with a much better watt/power ratio.



    The future of chip technology is multicore, and Intel has scheduled to produce in the future a 32 core chip.



    I don't know what is IBM is planning to do, but you must be aware, that there is plenty of research to create non silicium chip. For example one of the field of research is nanotubes. If my memory is good, IBM is investing a lot of time here, and has been able to create a nanotube transistor.

    I bet that your friend was speaking of this. A nanotube chip, will have billions of transistors in a very limited space.



    Some others engineers are working on quantic computer, but we are even far away from seing one working.
  • Reply 2 of 20
    I am takin' a great risk by posting this information here, but the problem wi' fast processors is much mare sinister than ye's a' imagine. There is a guid reason for legally imposed limitations and it is tae my shame I encouraged Apple tae debut a YottaHertz chip in it's first Intel machine.





    Jist think o' a those superfast electrons and fundamental particles hurtling roond those circuits. Think RHIC. Think o' the inevitable collisions, releasing rogue quarks. Monopoles! Negative Strangelets!





    (I apologise for posting something similar tae this elsewhere, but it has here assumed a new relevance and urgency.)



    I hae said too much as it is. I'm awa tae hide for a while noo.
  • Reply 3 of 20
    ipodandimacipodandimac Posts: 3,273member
    Quote:

    Originally posted by Powerdoc

    There is no physical laws preventing a chip to go further than 4 ghz.

    But engineers has found that increasing the clock speed was not the best way to have more powerful cpu.

    There is 2 issues with very high clock speed : heat and watt consumption.

    Intel gave up their 4 ghz P4 for this issue, and are going to give up the net burst technology (the one of the P4).

    The next géneration chips will be based upon the pentium M design, with a much better watt/power ratio.



    The future of chip technology is multicore, and Intel has scheduled to produce in the future a 32 core chip.



    I don't know what is IBM is planning to do, but you must be aware, that there is plenty of research to create non silicium chip. For example one of the field of research is nanotubes. If my memory is good, IBM is investing a lot of time here, and has been able to create a nanotube transistor.

    I bet that your friend was speaking of this. A nanotube chip, will have billions of transistors in a very limited space.



    Some others engineers are working on quantic computer, but we are even far away from seing one working.




    "nanotube" sounds so old though...
  • Reply 4 of 20
    powerdocpowerdoc Posts: 8,123member
    Quote:

    Originally posted by ipodandimac

    "nanotube" sounds so old though...



    X86 sound even older, though
  • Reply 5 of 20
    skatmanskatman Posts: 609member
    Quote:

    theory two... i've also heard that even at current speeds, the electrons have to be moving at speeds scientists once believed to be literally imposible (thats physics way over my head so dont ask for any more details)



    And I heard that if you flap your ears really fast, you can actually fly!
  • Reply 6 of 20
    dcqdcq Posts: 349member
    Quote:

    Originally posted by skatman

    And I heard that if you flap your ears really fast, you can actually fly!



    Just wait until the new optical CPUs come out. Then you'll see how fast they can make light go...
  • Reply 7 of 20
    Bah, Quantum Tunneling computers will bypass that pesky light barrier soon enough.
  • Reply 8 of 20
    wingnutwingnut Posts: 197member
    There is no "4.0ghz barrier." If you check the OCer's databases, many have gotten their P4s well past 4.0ghz. However, Intel ran into some serious thermal issues with netburst, so they aren't planning to release anything greater than 4.0ghz down the road. Actually, part of the P4 does exceed 4.0ghz already. The ALU runs at 2x the clock, so a 3.0ghz P4 actually has parts of it running at 6.0ghz. However, the overall logic runs at the rated speed.
  • Reply 9 of 20
    skatmanskatman Posts: 609member
    Only SSE2 units run at 2x the main clock.

    ALU runs at 1X the clock.
  • Reply 10 of 20
    hirohiro Posts: 2,663member
    Quote:

    Originally posted by Electric Monk

    Bah, Quantum Tunneling computers will bypass that pesky light barrier soon enough.



    Unfortunately, they will leak to fast to read. Sure, find the correct answer instantaneously, then loose it before you can coerce it out of the observation without changing it.



    But if you really want to be correct Quantum Tunneling is a reason for power dissipation in current CPUs, not responsible for future uber-fast computations.
  • Reply 11 of 20
    wingnutwingnut Posts: 197member
    skatman,

    Hmm, I have read otherwise in a few places. \
  • Reply 12 of 20
    Quote:

    Originally posted by Hiro

    Unfortunately, they will leak to fast to read. Sure, find the correct answer instantaneously, then loose it before you can coerce it out of the observation without changing it.



    But if you really want to be correct Quantum Tunneling is a reason for power dissipation in current CPUs, not responsible for future uber-fast computations.




    Jeez, try and make one ridiculous remark...



    In real life; the dramatic advantage of quantum computers is currently known to exist for only three problems: factoring, discrete log, and quantum physics simulations.
  • Reply 13 of 20
    hirohiro Posts: 2,663member
    Quote:

    Originally posted by Electric Monk

    Jeez, try and make one ridiculous remark...



    In real life; the dramatic advantage of quantum computers is currently known to exist for only three problems: factoring, discrete log, and quantum physics simulations.




    Yes, and unfortunately for quantum computers there is a problem. Just published last week, newscientist.com:



    Quote:

    Quantum computer springs a leak



    ATTEMPTS to build quantum computers could run up against a fundamental limit on how long useful information can persist inside them. Exceed the limit and information could just leak away, making computation impossible.



    A quantum computer manipulates stored information in quantum bits, or qubits. Because a qubit can be in twostates at the same time, and can be entangled with other qubits, a quantum computer can carry out multiple calculations simultaneously. But the entire system is delicate: during a computation the qubits have to be isolated from their environment, because any outside disturbance can cause "decoherence" and spoil the calculations.



    Coherence is harder to maintain in larger qubits containing more particles, because there is more potential for interaction with the surroundings. To try and limit this effect, researchers are pursuing ways of making microscopic qubits. These can be made using superconducting circuits on silicon chips or with quantum dots, which are essentially droplets of semiconducting materials that contain free electrons. In principle, qubits can be made out of individual electrons and photons.



    But physicists Jasper van Wezel, Jeroen van den Brink and Jan Zaanen of Leiden University in the Netherlands have shown that efforts to engineer quantum computers around ever-smaller qubits may face significant obstacles. "We have proven that there is a universal decoherence rate for qubits," says van den Brink. This means that quantum information will inevitably be lost after a certain time, even without any external disturbance. Rather than remaining in a superposition of two states, a qubit will spontaneously collapse into one state or another (Physical Review Letters, vol 94, p 230401). "When we discovered this we were stunned," says van den Brink.



    Worryingly, the time limit for decoherence seems to grow shorter as systems get smaller. Zaanen says that for some of the most promising qubit technologies the limit would be about 1 second. It's not a problem at the moment, he says, because researchers are fighting to get coherence times up to around a microsecond. "But this fundamental limit is getting within reach."



    "This is very interesting," says physicist Peter Zoller of the University of Innsbruck in Austria. But the real implications for quantum computing will only become clear with further work, especially experiments that would measure these effects, he says.



  • Reply 14 of 20
    skatmanskatman Posts: 609member
    Quote:

    skatman,

    Hmm, I have read otherwise in a few places.



    Technically, you're partly right.



    ALU on the netburst architecture has 2 simple instruction units capable of the doing 2 instructions per clock.

    However, load/ store and more complex instructions are handled by units working at the main clock speed.

    If you look at the overall average efficiency of the ALU (if you could draw a black box around the entire ALU and measure it's effective clock speed) it is almost 1/ main clock.



    The reason simple ALUs are working at 2X clock speed is to compensate for bubbles created in the long pipeline due to branch mispredictions, data dependence (non optimal scheduling) and such.
  • Reply 15 of 20
    wingnutwingnut Posts: 197member
    I'll take your word for it!
  • Reply 16 of 20
    skatmanskatman Posts: 609member
    Quote:

    Originally posted by Wingnut

    I'll take your word for it!



    I don't recommend that. 8)
  • Reply 17 of 20
    groveratgroverat Posts: 10,872member
    Mac O:



    A chilling portrait of the future, indeed! \
  • Reply 18 of 20
    junkyard dawgjunkyard dawg Posts: 2,801member
    Quote:

    Originally posted by Mac o' the Isles

    I am takin' a great risk by posting this information here, but the problem wi' fast processors is much mare sinister than ye's a' imagine. There is a guid reason for legally imposed limitations and it is tae my shame I encouraged Apple tae debut a YottaHertz chip in it's first Intel machine.





    Jist think o' a those superfast electrons and fundamental particles hurtling roond those circuits. Think RHIC. Think o' the inevitable collisions, releasing rogue quarks. Monopoles! Negative Strangelets!





    (I apologise for posting something similar tae this elsewhere, but it has here assumed a new relevance and urgency.)



    I hae said too much as it is. I'm awa tae hide for a while noo.




    LMAO!



    Yeah, I got some "for your eyes only" shit here, too. You see, the Dubya is 'fraid that if computers get too fast, then the terrarists will win. He says you don't need computers, just GOD and he will tell you what to do and how to rid the neighborhood of evil. Then the terrarists lose. So the CPUs must not be too fast or Joe Bin Laden could design an evil nukular bomb and we'd only find out in the form of a giant 'shroom over a major US city. Just imagine doing 'shrooms at over 5 GHz d00d!
  • Reply 19 of 20
    Groverat. How are you?



    Mr Dawg! my auld freend.



    Guid points if I may say so, but consider this too...



    Firstly this 32 bit, multi Gigahertz thing is moonshine. All domestic computers are simply auld 8-bit 6800s disguised wi' fancy silicone implants and falsified benchmarks. A' that 3D stuff and number crunching is little mare than 8-bit legerdemain.



    Secondly - domestic machines operate on the antiquated binary system. What THEY dinnae want ye tae know is that THEY have denary logic units and greater, where each bit can carry 10 (or maybe now 1000)states.



    Once again I hae said too much. I hope I havnae endangered the safety o' everybody on this entire board wi' these disclosures. My sincere apologies if THEY come knocking on your doors.
  • Reply 20 of 20
    powerdocpowerdoc Posts: 8,123member
    Quote:

    Originally posted by Mac o' the Isles

    Groverat. How are you?



    Mr Dawg! my auld freend.



    Guid points if I may say so, but consider this too...



    Firstly this 32 bit, multi Gigahertz thing is moonshine. All domestic computers are simply auld 8-bit 6800s disguised wi' fancy silicone implants and falsified benchmarks. A' that 3D stuff and number crunching is little mare than 8-bit legerdemain.



    Secondly - domestic machines operate on the antiquated binary system. What THEY dinnae want ye tae know is that THEY have denary logic units and greater, where each bit can carry 10 (or maybe now 1000)states.



    Once again I hae said too much. I hope I havnae endangered the safety o' everybody on this entire board wi' these disclosures. My sincere apologies if THEY come knocking on your doors.




    I fear this is the end of this community















































Sign In or Register to comment.