Slashdot Mirror

← Back to Stories (view on slashdot.org)

IBM Demonstrates High-k/Metal Gate Chips

Posted by kdawson on from the faster-and-thriftier dept.

Last summer we discussed twin announcements from Intel and IBM/AMD about a new chip manufacturing technology dubbed high-k/metal gate. Intel is using the tech to improve speed and power consumption in its 45-nm chips. IBM, along with its manufacturing partners, just demonstrated chips it says show that high-k/metal gate technology at 32 nm can result in performance gains up to 30% and power savings up to 50%, compared to 45-nm process. IBM plans to be manufacturing 32 nm parts by the end of 2009. (AMD is not using high-k/metal gate yet, but it has access to the technology by virtue of its agreements with IBM.)

20 of 72 comments (clear)

  1. pretty cool by voislav98 · · Score: 2, Interesting

    and it just goes to show that silicon is not dead yet

  2. M is for MOSFET by IorDMUX · · Score: 4, Informative

    I know that I've posted about this before, but...
    Huzzah! For the first time in 25 years, the name MOSFET ( Metal -Oxide-Semiconductor Field Effect Transistor) will correctly describe the device that goes by that name!

    (For those confused as to my jubilation, highly doped polysilicon replaced metal gates over 25 years ago. As a result most MOSFETS haven't actually had metal in them since.)

    --
    >> Standing on head makes smile of frown, but rest of face also upside down.
    1. Re:M is for MOSFET by Cyclon · · Score: 5, Informative

      Huzzah! For the first time in 25 years, the name MOSFET ( Metal -Oxide-Semiconductor Field Effect Transistor) will correctly describe the device that goes by that name!

      Sort of. The gate is still mostly poly, with a relatively thin metal layer below it. Also, the devices use a high-k material like HfO2 for the dielectric, with a thin silicon oxynitride mobility enhancement layer. There's a decent overview at Semiconductor International

  3. I hope AMD uses this technology by moderatorrater · · Score: 2, Interesting

    AMD earned my loyalty many times over the years, and now that it's fallen from the top of the price/performance heap, I feel bad buying another chip. This is the company who made the chip for my first computer, that made 64 bit mainstream, and made intel actually improve their products. They've done so much for the industry, it'd be a shame for them to continue taking a pounding like they have.

    Also, I own some of their stock. Go team!

    1. Re:I hope AMD uses this technology by OrangeTide · · Score: 3, Insightful

      Why would a large international company need your loyalty? You should buy whatever has a good price and good quality. Although you can argue that once AMD is no longer a mainstream processor vendor that Intel will raise their prices.

      --
      “Common sense is not so common.” — Voltaire
    2. Re:I hope AMD uses this technology by Saffaya · · Score: 4, Interesting

      You should buy whatever has a good price and good quality. I disagree in the sense that your comment is incomplete.

      The moral behaviour of the company making the product is to be taken into account, at least it is in my case.

      If company A has tried to screw me over several times (defective products), kept lying about it, and engages in generally anti-competitive behaviour, then I will buy products from company B instead.
      Even if B's product are less competitive by a certain margin left to my appreciation.

      Some will say that I am acting against how the market is supposed to work, that is not true.
      Business honesty and customer consideration IS valuable to my eyes, and I factor it in the final price of the product.

      Hence, I will pay more to buy B's product.
    3. Re:I hope AMD uses this technology by Chris+Burke · · Score: 5, Insightful

      The moral behaviour of the company making the product is to be taken into account, at least it is in my case.

      Some will say that I am acting against how the market is supposed to work, that is not true.

      Well if you listen to the die-hard capitalists (in particular the Randian strain of Libertarian), then basing purchasing decisions on the moral behavior of the company is your only valid way of preventing them from screwing you seven ways till sunday. Because any actual law that prohibited such immoral behavior would be at least as immoral as the behavior itself.

      And not so die-hard capitalists will also agree that not buying a company's products because of their behavior is a valid way to punish them, even if there are laws that also prohibit such behavior.

      Pretty much the only people who will say you are acting against how the market is supposed to work are die-hard sociopaths who don't care that some behavior is "immoral", and want you to keep lining their pockets regardless of what evil things they do in the name of making a buck, and the argument is solely a way of tricking you into ignoring your own moral outrage.

      --

      The enemies of Democracy are
    4. Re:I hope AMD uses this technology by Anonymous Coward · · Score: 3, Insightful

      Well...if you just need a good reason to continue supporting AMD, here's one:

      The fastest FSB on any Intel laptop chip today is 800MHz. The Slowest (and ONLY) FSB on any AMD chip today is 1600MHz, or 1.6GHz. This means that even if an intel chip can process data 60% faster (and currently even their best chip is only 40% faster than AMD's worst chip and even with a biased, paid for test. A real test of a comperable AMD lends about a 5% greater speed with a 2% margin of error.) the Intel system can still only get data in and out of the processor at 50% of the speed of it's AMD counterpart. What does this mean?

      It means if you're trying to do a little processing to a lot of data (i.e. watch a movie, manage a database, etc) the AMD will vastly outperform the Intel. If you're trying to do something which has to do a LOT of processing to just a LITTLE data (video editing if you have no good GPU, or playing a game with amazing graphics and a small field of view so that it doesn't process much at once) then the Intel will beat the AMD. Strangely enough, this typically means that Intels run games like Crysis better than an AMD, and yet if you use bittorrent and you have at least an 8MBit connection, not using an AMD means you had better dedicate the system to nothing but bittorrent - trying to multitask when you're trying to write half a meg a second on an Intel will make you want to blow your brains all over your keyboard.

      The truth is both have advantages and disadvantages and at the end of the day, neither does everything perfectly. I know many people who still keep a desktop around just as a gaming rig because the heat problems (and price point) for a gaming laptop is too high. Personally, I have an HP dv9420us and an ASUS EEE. I would never play anything beyond MAYBE Red Alert 2 on the EEE, and if I turn the settings on CNC3 above "Very Low" it will grind to a halt (well...3 FPS is more or less a halt.) But that's not the AMD in it - it's the damn GeForce 6150 Go. If it had an 8 series, only then would it be the AMD. Right now, I have Microtorrent, Pidgin, Firefox, WMP (still haven't gotten around to moving to Winamp), LimeWire Pro, and 78 other processes running. My hard drive write is now averaging over 1.2 Megabytes a sec. Any noticable slowdown from just Firefox? Nope. On an Intel, I couldn't say that, but I could probably run CNC3 at "Medium" and still maintain 30FPS, which though low, would suit me just fine. My point is that computers are still totally purpose-built. There is no all-in-wonder. Until a computer can flip a switch and go from a 2.8GHz quad core with 8GB of memory and 1 hour of battery life to a 1GHz low voltage CPU with 1GB of memory and 7 hours of battery life then people will always have to just buy two systems.

      So I suppose my point is just that you have good reasons to continue to buy AMD. AMD's processors are slightly behind the curve and they recently laid off 10% of their workforce but people don't quit buying ford trucks just because they close down one plant and add GPS capability a year after everyone else. If a ford truck still outpulls a chevy then people keep buying fords regardless, and if an AMD ships with a faster FSB then most often an AMD will outpace an intel in everyday use several times over. If you really must have your GPS or play Crysis then go with a chevy or an intel.

      All that said, with AMD's aquisition of ATI a while back I expect them to announce CPU's that can do many of the functions of their GPU's pretty soon. I just bought an HTC x7501 that has an ATI graphics coprocessor. And it's little more than a cell phone. The sole thing that sold me on my old Nintendo Gamecube was the ATI sticker on the box. ATI is to graphics what Intel is to processors - they both specialize on desktops but dabble in all manner of other arenas. I have seldom seen any game on an orginal XBOX look half as nice as any game on a Gamecube. Hell, even Pikmin had better looking effects than Ghost Recon, and I love Ghost Recon and hated Pi

    5. Re:I hope AMD uses this technology by Trogre · · Score: 2, Insightful

      I would like to go one step further and claim that in this corrupt age, the only way to vote with any effect on your rulers whatsoever is not with a ballot box, but with your wallet.

      IMHO, that's the secret, chaps. Power's all in your wallet. Use it wisely.

      --
      "Nine times out of ten, starting a fire is not the best way to solve the problem." - my wife
    6. Re:I hope AMD uses this technology by TheRecklessWanderer · · Score: 2, Interesting
      Good for you!

      I used to buy a lot of maxtor drives. Probably 2 or 300 a month. Then I tried to return one, and they gave me a hard time.

      Now I buy a lot of Seagate and WD Drives. No Maxto Drives.

      Since Seagate and Maxtor have jumped in bed together, I just buy WD Drives.

      --
      Mean what you say...say what you mean.
    7. Re:I hope AMD uses this technology by CajunArson · · Score: 2, Interesting

      This shit gets modded insightful? I had long suspected that to get an insightful mod on this website all you had to do was follow the formula of: 1. bash a company that has earned Slashdot's 2 minutes of hate award (Intel here); 2. blather on for more than 4 words to give the illusion of thought; 3. Throw in some non sequitur numbers to look like you know math.

      I don't have time to take apart every number in this stupid troll but, first of all: AMD does not use FSB's at all, and you have no idea what the "1.6Ghz" number you are quoting actually means... you are (ironically) trying to throw around frequency numbers to make the bigger number look better for AMD (wow sounds a lot like that "megahertz myth" everyone on Slashdot adores so much). Take a look at how badly the Phenom cache architecture works sometime, you'll see that core to core communications on those chips are only about 2% faster than when Intel chips have to use the FSB, and when Intel chips in dual core are using shared cache they communicate roughly 3 times faster than the best-case for the AMD chips. Hypertransport IS useful... but only for 4+ socket server systems where there are real advantages. On desktop machines & laptops the differences are negligible, and there can even be disadvantages when talking to a graphics card on a desktop machine (but I don't think the AMD powerpoints you were working off of mentioned that did it?)

        Second of all, you claim that if an Intel machine has to move a whopping 1 whole MEGAbyte of data a second around it somehow gets irreparably crippled to the point of not even being able to use a mouse or something. Lets actually run some numbers, even using your artificially low laptop ratings for the Intel FSB (desktops are much faster, and you are also failing to count the faster FSB used on Penryn laptops that are already on the market). Let's see here: 800 Mhz FSB * 8 bytes per clock (64 bit data bus) comes out to 6.4 Billion bytes per second. Your "massive" 1.2 Mbyte/sec transfer is therefore completely choking off... wait for it... about 0.02% of the bandwidth. OH NO MY COMPUTER IS GOING TO FREEZE IF ONLY 99.98% OF THE BANDWIDTH IS AVAILABLE!!

          Oh wow, you apparently can count processes and found out that your computer runs more than one program at a time? Guess what, I ran firefox (version 2 mind you not the nice & faster version 3) on a 2.4Ghz P4 just fine, and I could even run other programs at the same time! Now, I was using Linux of course, but you are either 1. lying, or 2. completel y incompetent when you say that somehow it is impossible to run more than one program "smoothly" on anything but an AMD system.

          Third (and this takes the fucking cake): You compare the intenionally low-power EEE PC from Asus that INTENTIONALLY uses very low power, very slow chips from Intel because it is targeting the ultra-mobile market, and then because this laptop, which is INTENTIONALLY designed to be slower & low power does not perform as well as a laptop that costs 3 times as much and comes with 4 times the RAM and has a massively more expensive and power hungry CPU and graphics processor in it. Yes you stupid fuck, I actually looked up the model number you posted. I'm sure you thought that you could put it right next to a mention of the EEE PC and think you could fool everyone into thinking the only difference was an Intel vs. an AMD chip. The only thing you've managed to prove is that there are AMD fanboys brainwashed enough to believe that because the EEE PC exists, it is physically impossible for anyone to use Intel chips in something that might run faster.

          That's all I have time for. You are either truly brainwashed, or are smart enough to know the Slashdot system to getting "insightful" mods when you are literally posting objectively proveable lies.

      --
      AntiFA: An abbreviation for Anti First Amendment.
  4. Yeah... by jd · · Score: 2, Insightful

    But Germanium-Arsonide is a much-neglected technology that could do with more investment, as it should do much better than silicon. Graphene is another technology that risks being ignored for as long as silicon is a viable option. I'd far prefer chip companies to be pushing the boundaries with materials that should offer far more extreme performance. Nonetheless, any progress is good progress.

    --
    It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
    1. Re:Yeah... by andreyvul · · Score: 5, Informative

      It's not germanium arsenide, it's Gallium (III) Arsenide: http://en.wikipedia.org/wiki/GaAs
      On the plus side, this means that solar-powered chips (i.e. transistor and photovoltaic cell on same die) will eventually exist.

      --
      proud caffeine whore
    2. Re:Yeah... by Anonymous Coward · · Score: 2, Informative

      Absolute bollocks. Si-Ge is used all over the place. Germanium was used for the first diodes and transistors. It is an abundant element.
      http://en.wikipedia.org/wiki/Germanium
      Learn something before spouting nonsense out of your dicksucker.

  5. High K is nothing new by LM741N · · Score: 4, Funny

    I've been getting my seeds from the Netherlands for some time now.

  6. Barbie sez, math iz hard! by pla · · Score: 5, Informative

    high-k/metal gate technology at 32 nm can result in performance gains up to 30% and power savings up to 50%, compared to 45-nm process.

    Really revolutionary announcement there...

    Power consumption scales with the square of gate size. (32*32)/(45*45) = 0.51, or 50%.

    Clock speed scales linearly with gate size. 32/45 = 0.71, or 29%.


    Not to minimize the fact that these gates reduce leakage enough to actually get those gains, but the drop in gate size alone (all other factors equal) would give the same numbers.

    1. Re:Barbie sez, math iz hard! by Chris+Burke · · Score: 4, Informative

      Not to minimize the fact that these gates reduce leakage enough to actually get those gains, but the drop in gate size alone (all other factors equal) would give the same numbers.

      Yeah, but that's a pretty big freaking deal. Leakage current has come to dominate (or at least become a factor as significant as switching current), and is actually increasing as the technology shrinks. Going to 32nm, without paying a penalty in increased leakage, is quite an accomplishment.

      Or put another way, all other factors being equal (including the dielectric), a drop in gate size alone would not result in the same numbers.

      I know you said you aren't trying to minimize their accomplishment, but by saying it simply follows a basic equation when that equation has been failing for the past few generations (because it's not correct at these scales), you're doing just that.

      --

      The enemies of Democracy are
  7. Take a look at the literature by JSchoeck · · Score: 3, Interesting

    Graphene is not at all nearly ready to even build reliable, well-performing transistors with it. I'm in a research group that is trying to implant a gate electrode into Silicon-carbide with a Graphene layer ontop, but that's still basic research. If it should really work with good yield and that also in an industrial process, then we can talk about Graphene-based CPUs.

    And by the way: it's spelled "Arsenide"

    1. Re:Take a look at the literature by Kjella · · Score: 2, Insightful

      True, it's not ready now, but research and development budgets are finite and therefore the more that is spent on silicon, the less you can spend on graphene and the longer it will take for graphene VLSI to be a practical day-to-day thing. My big concern is that, as is the case with nuclear fusion, the amount spent will be too small in comparison to the amount required to produce useful (in the marketplace) results. That assumes graphene will overtake silicon, as if that's a certainty. Silicon has scaled extremely well now for 30 years and all the people thinking "this has got to stop soon" have been proven wrong time and time again. If the silicon improvements bottom out, one also has to ask how much more is there to gain? Are the the same fundamental limits going to hit graphene? You got to ask when they're showing 300-atom (32nm) thick layers now with plans for going near 100 atoms. Can it really be packed that much tighter? Can it really operate at a higher frequency without using more power as power/watt is very important? All I'm saying is that it's no point making massive investments into something that may turn out worse or even or just slightly better than silicon.

      Actually I think your fusion power is a great example, it's not for lack of research as there's been many years to let it mature, there's been oil crisises that have really boosted interest in alternative power and if you've paid any attention to the oil prices and the possibility of "peak oil" which has delivered funding. Fusion is a great idea that just isn't partical, it's like five year-olds with tons of energy but it's impossible to make them do any real work. Theoretically better is great, in theory. But while I don't want companies to try to turn the clock back, being so far ahead of the technology that it's not actually usable isn't a good thing either.
      --
      Live today, because you never know what tomorrow brings
  8. Comment removed by account_deleted · · Score: 2, Funny

    Comment removed based on user account deletion