Speed ceiling for cpu

Is there a theoretical limit to CPU speed? Will there be a point in time when Intel and AMD can no longer produce faster CPUs?
14 answers Last reply
More about speed ceiling
  1. i dont think theres a limit for the moment. not in theory anyways. they will continue to be more power efficient and more powerful but its possible they might op for more cores than higher clockspeed for performance in the future.

    who knows.
  2. There is a definite limit,
    as tech gets smaller it becomes faster with less voltage required and less heat output,
    the limit is likely in the single nanometer range but can we massproduce at that small a scale? not at the moment, we're still at the teens stage, but with a solid eye on going as far as we can,
    http://news.softpedia.com/news/Intel-s-16nm-Became-14nm-Advanced-10nm-Coming-in-2015-291969.shtml
    but having said that theres a lot of work going on with DNA Storage, Graphene and Crystal processing as well, so what we have now as top of the line will one day (scarily soon I think) be ancient tech of a bygone era,
    but as for saying the thirty first of february 2019 is the day we reach the point of no more, or the eighteenth of august 2014 is a math and development question,
    I can't calculate how fast tech behind closed doors is progressing or the timescales Intel and Amd have given themselves, but its happening and fast, I think we could easily see the limits of clockspeed on silicon within twenty years.
    **Edit to add link,
    http://www.tomshardware.co.uk/forum/304084-10-what
    Another Toms thread with some handy thoughts, I'd forgot Graphene in my list :)
    and a Crystal link,
    http://www.foxnews.com/tech/2012/04/26/tiny-crystal-may-hold-key-to-future-computers/
    Moto
  3. Hi :)

    Interesting question....

    I have two answers....

    1, the speed of thought is the limit as even if it was faster , you wouldn't be able to think so...

    2, 42

    All the best Brett :)
  4. I'm assuming you mean speed as in frequency, if you mean CPU performance, I suspect that will continue improving as long as there are people making CPU's.

    There is a point where silicone based CPU's quite simply cant go faster.
    Both AMD and Intel discovered that the hard way back in the single-core days. Both at the time were trying to push the fastest possible clock speeds from their chips and quite simply they couldn't do it. Eventually they went with the solution of having more cores rather than a faster single core.
    That might change in the future when some other material (graphite is likely) becomes dominant or new technology (remember an article a while ago that theorized that light could be used instead of electricity to move data in a CPU) comes into the field.
  5. Brett928S2 said:
    1, the speed of thought is the limit as even if it was faster , you wouldn't be able to think so...


    Imagine that, where the bottleneck to a computers speed would be latency in our perception of it. Light has to travel say 50cm from the screen to our eyes, eyes pick up the light and take some amount of time to convert to a signal, optical nerve firing its data the brain, then how long it takes to process the data, then a bit longer for the conscious mind to understand what it see's.
    If human perception was the bottleneck to how fast a computer can effectively work, that would be the point where we don't need faster computers.
    That's not even accounting for the time it would take to interact with the computer, moving (or "thinking" :o ) the mouse and clicking.

    Though chances are the machines will have rebelled by then, so its kind of a hypothetical situation :p.
  6. manofchalk said:


    Though chances are the machines will have rebelled by then, so its kind of a hypothetical situation :p.



    Hi :)

    As long as we don't let them have guns we are ok...lol

    All the best Brett :)
  7. The max atm on an Ivy Bridge CPU is 7.2Ghz. World record, I believe.
  8. I'm not exactly an expert on that subject but there is no way to tell if there's a limit. If you see that question on it's literal meaning, probably there isn't a limit to CPU speed when it cames to performance meanings, like manofchalk said.

    There's a very good and conductive material called graphene, which consists solely in carbon (its a carbon isomer). Its a very fragile material, so the scientists haven't find a way to use it to create CPU's yet. But that's just a matter of time. It is estimated that a CPU build with graphene would be five hundred times faster than any known processor. Still, it is most likely to be used as warfare technology and researches before becoming "domestic", and THAT could take years.

    Also, not all the CPU's are electrical. For example, our actual processor's are electrical since ever. Pentium I, II, III, IV, Dual Cores, Core i3 to i7. They all work on electricity. That not necessarily prevents them to be faster, but since they work on a binary logic (0 and 1 as with or without eletric current) they MIGHT have limited development over time. And that's one of the reasons why, all over the world, scientists are already creating atomic computers. These processors no longer work on a binary logic, since atoms might have infinite positions over a very small amount of time, they are using it's "spin" to create CPU technologies that are hundreds of thousands faster than graphene.
    These computers already exists, but they are on the size of an entire room. Also we don't know the atoms so well, and it is mostly instable.

    But, since the first computers in human history used to occupy entire buildings, i believe that there will be a day when a atomic processor fits a machine of the size of a smartphone, maybe even smaller. I'm just not sure if we will be alive until than. But these processors, just as the graphene technology, shows promise.

    Some speculates that it would be just as fast as a human brain, or even more. That's is hard, but possible. Which creates a dilemma of how dangerous could this be all over the scientific community. Just as manofchalk said, there are chances of machine rebellion. This can be seen as a joke, but also, it might not. I don't really remember which, but there's a University (maybe Stanford? Google It and you'll see) that created a research department to prevent and discuss machines over-intelligence that could put our existence to risk. That's crazy, we all know.

    I hope that this give you some clarification.
  9. The world record on CPU clock speeds is 8.429Ghz on an FX-8150. Needed a ridiculous amount of cooling, liquid helium to get to that speed. Probably wasn't even stable, all they did was a CPU-Z validation, no benchmarks.
    http://techcrunch.com/2011/09/13/amds-new-fx-processor-reaches-world-record-clock-speed/
    So its possible that CPU frequencies can be pushed a lot further than they are currently, but would require an ungodly amount of cooling to make it possible. The guys who did the overclock roughly guessed the CPU would have only been 10c above absolute zero. So chances are AMD/Intel stopped since it wasn't possible to make a marketable cooling solution for a chip running that fast.

    Heard that Quantum computing is starting to become possible now, if only in a lab environment.
  10. manofchalk said:
    The world record on CPU clock speeds is 8.429Ghz on an FX-8150. Needed a ridiculous amount of cooling, liquid helium to get to that speed. Probably wasn't even stable, all they did was a CPU-Z validation, no benchmarks.
    http://techcrunch.com/2011/09/13/amds-new-fx-processor-reaches-world-record-clock-speed/
    So its possible that CPU frequencies can be pushed a lot further than they are currently, but would require an ungodly amount of cooling to make it possible. The guys who did the overclock roughly guessed the CPU would have only been 10c above absolute zero. So chances are AMD/Intel stopped since it wasn't possible to make a marketable cooling solution for a chip running that fast.

    Heard that Quantum computing is starting to become possible now, if only in a lab environment.

    If it's not stable and benchmarked, it doesn't count. That's just cheating! :P
  11. Surprised they weren't running a lightweight version of Linux off like a 20 SSD RAID0 array actually. Less strain the processor is under and the less time it takes to get that validation, higher clocks you could probably get too.
  12. who knows, like manofchalk said, if they change the material the transitors and eletronics of the cpu chip are made of maybe, but it be smarter to optimize performance and have lower clockspeed that performs just as fast as a higherclocked cpu, for heat and energy reasons.
  13. manofchalk said:
    Surprised they weren't running a lightweight version of Linux off like a 20 SSD RAID0 array actually. Less strain the processor is under and the less time it takes to get that validation, higher clocks you could probably get too.

    That's like those people that use laser sites for target shooting lol.
  14. lol.
Ask a new question

Read More

CPUs AMD Intel