When will desktop speed reach tera-hertz(or whatever after GHz)

Discussion in 'Apple, Inc and Tech Industry' started by Mr. MacBook, May 1, 2007.

  1. Mr. MacBook macrumors 6502

    Joined:
    Feb 28, 2007
    #1
    I'm just wondering as i opened up a magazine of Laptop.

    A decade ago, 15MHz was blazing fast, complete with a couple mb/kbs of ram.

    Now we're 3.7GHz(overclocked Core2 Quad Extreme).

    Maybe in another decade, we'll reach 1000 Gigahertz and think of today as "useless computers?"

    Idk, im just wondering, cause i think it would be AWESOME. Can you imagine just having to shave off "a few" gigahertz per processor?
     
  2. FF_productions macrumors 68030

    FF_productions

    Joined:
    Apr 16, 2005
    Location:
    Mt. Prospect, Illinois
    #2
    I can only see the amount of cores really getting pushed. We'll see 8-cores and laugh at them 10 years from now.

    Clock speed isn't a big deal as it used to be.
     
  3. ironic23 macrumors 6502

    Joined:
    Feb 8, 2006
    #3
    I don't think it's because clock speed isn't as big a deal as it used to be. I think that Intel and AMD haven't found efficient enough solutions to bump the clockspeeds higher than 3.0++GHz without increasing the power input and thermal dissipations to unacceptable levels. Thus, they decided that instead of increasing the clock speed, they'll have multi-cores to make up for it.

    As far as cores go, Intel recently displayed their 80-core processor at some trade show (forgot when/where). If i recall correctly, an article cited Intel saying that having a processor with more than 8 cores becomes inefficient due to scheduling and caching. (correct me if i'm wrong)

    Hopefully we'll see higher clock speeds in the time to come, but IMHO, the problem lies in the software, not the hardware. more efficient ways of coding need to be implemented, but thats my own 2cents worth.
     
  4. SkyBell macrumors 604

    SkyBell

    Joined:
    Sep 7, 2006
    Location:
    Texas, unfortunately.
    #4
    I doubt if it'll ever get that far.

    And BTW, 15 MHz was fast in the late eighties.:p
     
  5. janey macrumors 603

    janey

    Joined:
    Dec 20, 2002
    Location:
    sunny los angeles
    #5
    Actually no. Someone's OC'd a P4 to 8179.89mhz. http://www.fx57.net/?p=540

    mmmm ln2 ftw :D :D

    BTW, slow processors are not useless. They do not use up lots of power nor do they produce lots of heat, making them ideal for certain uses (embedded hardware, stuff to use in calculators and watches and routers..name anything).
     
  6. RedDragon870503 macrumors 6502

    RedDragon870503

    Joined:
    Sep 6, 2006
    #6
    I am not a chip designer but I have a feeling clock speed will level off at a certain point and efficiency per clock cycle will reign supreme.
     
  7. Apple Architect macrumors regular

    Joined:
    Apr 4, 2007
    #7
    Clock speed

    Clock speed is a massive deal on a single threaded application stream. That said, what the processor can do in the single clock cycle is pretty damn important. Richer processing chips may solve some of this, but some basic maths problems can only be solved in a linear fashion.
     
  8. KD7IWP macrumors 6502a

    KD7IWP

    Joined:
    Mar 8, 2004
    Location:
    American living in Canada
    #8
    I don't believe we'll ever get past 4Ghz. The way we have been increasing speed is adjusting voltages and shrinking the die's that we use for the CPUs. We can't shrink the die anymore because we're at such a small level that quantum physics are showing us that electrons will jump paths.
     
  9. Jasonbot macrumors 68020

    Jasonbot

    Joined:
    Aug 15, 2006
    Location:
    The Rainbow Nation RSA
    #9
    If Moore's Law about amount of transistors per die or whatever doubling every so many months holds true we will keep seeing smaller more efficent processorsdoing the same job as their larger counterparts. Although personally I think that cores and clock speed will level off soon and processors designers will need to find better ways to increase the power of processors, maybe a complete overhaul of computing is required?
     
  10. stoid macrumors 601

    stoid

    Joined:
    Feb 17, 2002
    Location:
    So long, and thanks for all the fish!
    #10
    The problem with higher clock speeds is that at some point, the electrons won't be able to travel (even if moving at the speed of light) across the chip within one clock cycle, and then you wind up with a menagerie of interference problems when electrons from one clock cycle are moving through the chip at the same time as from the previous cycle.
     
  11. ChrisA macrumors G4

    Joined:
    Jan 5, 2006
    Location:
    Redondo Beach, California
    #11
    I don't think we we ever get there. Technology does not work that way. What happens is when a new industry starts up at first there is a rapid advance and then after 50 or so years things settle down. It's not that advances can't continue but that after a while there is not need and no economic insentive to do so. We are sill in the first 50 years of small computers. It started in the late 70's.

    Look at the example of passenger aircraft. Lots of advances from the early 1900's and then once we got large jets like the Boeing 707 only incremental changes. It turns out the sub-sonic jet is about what works. You can go faster but the cost of doing so does not justify the saving of a couple hours in flight time.

    We are beginning to see this in computers too. We can make them much faster but why? If the machine runs 80% idle whay make it go even faster and run at 90% idle? Again it's cost. I think multi-core moderate speed CPUs are what will turn out to be the most economical to build. In a mature technology (like airplanes, cars, refrigerators, building materials and so on. What determines what's sold is more due to economics than technology. Small computer technology is not yet fully mature but is becoming so.

    Moore himself said that "Moore's law" has driven by economics, not technology. The number of transistors double in 18 months because that is the level of R&D spending and capital investment that is "about right" and sustainable. But that R&D and capital spending depends on an expanding market. Exponential growth can not continue forever. Intels spending on R&D and new fabs can't exceed the total World GWP But if current trends continue it will which tells us current trends can't contiinue, hence no teraherz computers

    There is also physics. Making the clock rate faster means making the CPU smaller. Making it 300 times smaller means you are building with smaller parts but attoms can't be shrunk. You can't make parts half the saze of an atom so at some point after you can't make the parts smaller you have to use few parts.
     
  12. Mr. MacBook thread starter macrumors 6502

    Joined:
    Feb 28, 2007
    #12
    i bet when we were at 300MHz for a couple thousand, none of us thought we would get past 1-2GHz, and we didnt know what dual-core was.

    And who knows, years ago, 300MHz processors probably takes up as much power as your Mac Mini.
     
  13. CanadaRAM macrumors G5

    CanadaRAM

    Joined:
    Oct 11, 2004
    Location:
    On the Left Coast - Victoria BC Canada
    #13
    But it won't (I know you agree)

    In order to reduce the process size, you have to manipulate smaller and smaller elements. For example, photolithography has been a primary way to produce circuitry on chips. But it is already to to point where the wavelength of regular light is longer than the design details of the circuits, and the process has to be reengineered with shorter and shorter wavelengths of radiation. The current state of the art is 29 nm in the lab and later this year 45 nm in production. However the cost of production rises dramatically as you have to use more exotic processes.

    Other problems with continued shrinking of the process are
    leakage, a host of resistance and capacitance issues, heat dissipation, and greater problems in purity in the materials. Once you get really small, 'normal' macro-world physics (which by and large relies on the average properties of large collections of atoms) starts to get distorted by atomic-level forces, and variations among individual atoms.

    http://en.wikipedia.org/wiki/Moore's_law#Future_trends

    In Gordon Moore's own words:
    "In terms of size [of transistor] you can see that we're approaching the size of atoms which is a fundamental barrier, but it'll be two or three generations before we get that far—but that's as far out as we've ever been able to see. We have another 10 to 20 years before we reach a fundamental limit. By then they'll be able to make bigger chips and have transistor budgets in the billions."
     
  14. Jasonbot macrumors 68020

    Jasonbot

    Joined:
    Aug 15, 2006
    Location:
    The Rainbow Nation RSA
    #14
    If what you say is true, and it probably is, where by we see transistors smaller in size than the particles that travel through them why not have physically larger processors? Or does that make heat dissipation and processing less possible.

    When nano technology comes around (if it ever does) I can see atom sized particles being used in computing. That would surely put anything to shame:)
     
  15. janey macrumors 603

    janey

    Joined:
    Dec 20, 2002
    Location:
    sunny los angeles
    #15
    .
     
  16. Mpulsive81 macrumors 6502

    Joined:
    Jun 7, 2006
    Location:
    McKinney, TX
    #16
    I had a Commodore 486 that ran at 22MHz, so yeah I remember those days :)

    And yes I'm sure there are some old folks who can 1-up me there, but that was my first computer so :p
     
  17. trule macrumors 6502

    Joined:
    Mar 16, 2007
    #17
    The POWER PC based CELL processor has 9 cores, one normal POWER PC unit and 8 specalised computing units, these 8 units have a direct connection to main memory (no cache sharing with the PPC core) and a high bandwidth bus. It requires purpose built software design however the aproach of having specalised cores operating under loose control of a normal core is very very effective - it addresses the issues that effect Intel style multi-core designs but it is specalised and only useful in CPU intensive tasks...like video encoding/decoding.

    The POWER6 core is running at speeds above 5GHz...
     
  18. The General macrumors 601

    Joined:
    Jul 7, 2006
    #18
    Processor power is measured it flops, not Hz.

    The Ghz war is over.
     
  19. QCassidy352 macrumors G3

    QCassidy352

    Joined:
    Mar 20, 2003
    Location:
    Bay Area
    #19
    my first computer was my family's LC III, 25 mhz. When my dad bought it his coworkers told him it was overkill to have something that powerful for home use. :D
     
  20. BlakTornado Guest

    BlakTornado

    Joined:
    Apr 24, 2007
    Location:
    Washington, OH
    #20
    To answer the question "When will desktop speed reach tera-hertz(or whatever after GHz)"

    the answer is, in macs, several months after PCs :p

    Noone can deny that some PCs have better processors than the current iMacs :p

    I hate PCs but noone can deny that PCs are further in the hardware aspect, despite Apple's outstanding OS.
     
  21. CanadaRAM macrumors G5

    CanadaRAM

    Joined:
    Oct 11, 2004
    Location:
    On the Left Coast - Victoria BC Canada
    #21
    What are you on about, man? How many PCs are shipping with 8 cores of 3 GHz Xeon?
     
  22. mkrishnan Moderator emeritus

    mkrishnan

    Joined:
    Jan 9, 2004
    Location:
    Grand Rapids, MI, USA
    #22
    Well, fast electronics don't really go much faster than nanosecond class pulses, as far as I know. At 2.0GHz, electron currents in the processor are on the order of 1 / 2x10^9 = 5x10^-8 or 50 nanoseconds. I don't think solid state electronics can support a pulse length much more than a factor of 10 shorter than that... meaning that 10s of GHz are possible, and low 100s of GHz, but not much more than that. I don't think you can physically get 1 THz using solid state electronics? Optics, yes.
     
  23. CANEHDN macrumors 6502a

    CANEHDN

    Joined:
    Dec 12, 2005
    Location:
    Eagle Mountain, UT
    #23
    I don't think we'll see much higher clock speeds either. Before we know it, light will be the method of transfer and processing and clock speed will be irrelevant.
     
  24. BlakTornado Guest

    BlakTornado

    Joined:
    Apr 24, 2007
    Location:
    Washington, OH
    #24
    In general, you can buy cheaper PCs with same or better specs for less than an equivalent mac.

    Sure, there aren't (m)any 8 core 3GHz PCs but not many people can afford 8 core 3GHz Mac Pros.

    And, in general, the hardware for Mac's are a bit behind PCs. (spec-wise. Not design-wise)
     
  25. pseudobrit macrumors 68040

    pseudobrit

    Joined:
    Jul 23, 2002
    Location:
    Jobs' Spare Liver Jar
    #25
    How so? Because you can't spec an iMac with Conroe? Who cares? Look at what Dell offers for $5200: An overclocked Core 2 Extreme with 4GB of RAM maximum. Oh, the video card is better than the Mac Pro's.

    Step up to the servers and it doesn't get any better either. Same with the laptops. I was browsing through Dell's selections yesterday and was astonished with how uninspiring the hardware choices and prices were across their lineup compared to the current Macs.
     

Share This Page