Forgot your password?
typodupeerror
Intel Technology Hardware

Intel Designs Faster, 3D Transistor 141

Posted by samzenpus
from the nice-fins dept.
lee1 writes "Intel has found a way to keep on the Moore's Law track by making smaller, faster and lower-power computer chips by building 3D transistors. They are already manufacturing microprocessors using this new design, called a FINFET (for fin field-effect transistor), which incorporates a small pillar, or fin, of silicon that sticks up above the surface of the chip. Intel said that it expected to be able to make chips that run as much as 37 percent faster in low-voltage applications and use as much as 50 percent less power. Products based on the new technology may appear some time later this year."
This discussion has been archived. No new comments can be posted.

Intel Designs Faster, 3D Transistor

Comments Filter:
  • by tsa (15680)

    Is it possible to use that fin for cooling as well?

  • You'll still have to cough up an extra $200 for the privilege of using all the transistors in your Intel hardware though. Or maybe this will bring an end to them segregating things like HW virtualisation based on how deep the users' pockets are.

    • Oh, please. For consumer parts, the best price/performance by far is offered by Intel, specifically the i5-2500K. Intel doesn't price-gouge on consumer chips.
      • Re:37% faster! (Score:4, Insightful)

        by h4rr4r (612664) on Wednesday May 04, 2011 @04:24PM (#36028332)

        The i5-2500K is $230 for $15 less you can get a Phenom II X6 1100T Black Edition. Depending on what you are doing, say encoding video the Phenom is the way to go. Your best by far price/performance is just fanboy talk.

        • Re:37% faster! (Score:4, Insightful)

          by smelch (1988698) on Wednesday May 04, 2011 @04:28PM (#36028384)
          I bought one of those Phenom II X6 1100T Black Editions. Having six cores is pretty nice when you run a web server, file server and capture video and play starcraft II at the same time. I got blasted for not going intel, but I've had no problems at all with it and the chip was pretty cheap.
        • by alen (225700)

          no, if you use a quicksync enabled app to encode video then Intel is faster than everyone

          • by h4rr4r (612664)

            Which x264 does not support it so useless. For good reason as well, the encoder you are talking about is all in hardware and not very good.

            • Yeah, because the professionals use x264, right? The quick sync encoder is quite good, actually, but yeah you can choose between an imperceptible difference in quality or an additional 100X time to do the encoding.
              • by lymond01 (314120)

                I make sure to buy the Best Buy Monster cables to get the best video presentation possible.

                Kidding. Purchasing a 50 foot HDMI Monster cable at 5% over cost was something like $16. I believe they charge over $100 retail.

        • The i5-2500K is faster for the applications that most people are running which in many cases are single-threaded. Even when it comes to x264 encoding where the Phenom is faster, it's only by like a 10% margin at best. Not to mention that the Phenom is also more power hungry than the i5.

          • by h4rr4r (612664)

            Most people are not running only one application at a time.

            The Phenom is more power hungry but the GPU anyone would buy for a setup like this is going to dwarf the small difference in TDP of the CPUs.

            • Most people are not running only one application at a time.

              That rather depends on your definition of running.

              Yes they may have a number of apps loaded into memory but the chances are most of those apps are sitting idle waiting for user input events with only one or two doing actual computing work at any given time. Further in my experiance most desktop apps only have one thread doing significant work. The result is for general desktop use four fast sandy bridge cores are going to be a far better choice than six slower phenom II cores.

            • by jon3k (691256)
              You realize the 2500K is a quad core CPU right? And you understand the concept of preemptive multitasking right? So explain to me how you came to the conclusion that for most consumer use-cases that 6 slower cores are better than 4 faster cores?
        • by PitaBred (632671)

          At stock speeds, sure. But my i5-2500K will clock at 4.4GHz at stock voltages, which puts it's encoding pretty well above the 1100T, overclocked or not. And any single-threaded tasks (of which most of what anyone does is) will be much better served by the 2500K.

          I love AMD. I have two of their GPUs in my desktop here, my last build was a Phenom II 955. But right now, AMD doesn't even come close in the raw price, price/performance or performance/watt categories. I wish they did, and I really hope Bulldozer ge

          • by h4rr4r (612664)

            Most folks don't overclock, I only do that to my phone. Sure most applications are single process/thread but you normally run many at once.

            I agree that at many price points Intel is untouchable right now, but at the i5-2500k that price is just barely in Intels favor.

            I am waiting to build a new machine right now, if bulldozer flops it will be another Intel for me. My last machine was a prebuilt Q8300 that I got a steal on.

            I wish I could use AMD GPUs but their linux driver sucks from what I have heard. Also f

            • by PitaBred (632671)

              If you have a 2500K, you're gonna overclock. Otherwise you should have saved $25 and gotten the i5-2500, which is identical except for the unlocked multiplier. Same with the 1100T. Should have gotten a 1090T if you're not going to overclock. I'm still wondering how you say it's just barely in Intel's favor though, because the ONLY benchmark that the 1100T comes close to it on are POV Ray and some video encoding benches where it wins by a very small margin, and that's very seldom the only or even primary thi

              • Otherwise you should have saved $25 and gotten the i5-2500, which is identical except for the unlocked multiplier.

                There are a number of differences between the K and non-K varients. The K variants have a better GPU which is why I bought one. I have no idea how much better because no-one seems to want to benchmark K vs non-K varients but as I want to drive high res monitors (not for games) I figured I may as well go for it.

        • by jon3k (691256)
          if you want to cherry pick really low utilization use-cases how about we compare AES performance? the reality is for 99% of consumers the intel chip will absolutely blow AMDs offerings out of the water. dollar for dollar, watt for watt or any other metric you'd like to use.
      • by pablomme (1270790)

        Are you aware of what is in a Core Solo processor [wikipedia.org]?

        • by Jackazz (572024)
          What self-respecting slashdotter is buying a Core Solo processor?
        • by gman003 (1693318)
          Are you aware that AMD is even bigger on the whole "this processor has a defect in one core so we'll disable it and sell it cheaper" bit? They have dual-core processors that are just quad-core chips with half the thing disabled. Usually, those parts are disabled for valid reasons, but sometimes, if demand for the cheaper ones is too high (or they've got less defects than expected) they have to disable working parts. I know of several people who bought X3s and re-enabled the last core, giving them a quad-cor
          • by aiht (1017790)

            Are you aware that AMD is even bigger on the whole "this processor has a defect in one core so we'll disable it and sell it cheaper" bit? They have dual-core processors that are just quad-core chips with half the thing disabled. Usually, those parts are disabled for valid reasons, but sometimes, if demand for the cheaper ones is too high (or they've got less defects than expected) they have to disable working parts. I know of several people who bought X3s and re-enabled the last core, giving them a quad-core processor at triple-core price. I also know that many of those people eventually reverted the chips back to triple-core, as the last core caused significant crashing.

            Then obviously those cores were disabled for a good reason. That hardly backs up your point about choosing to disable working parts.
            However, your point is still valid. I am currently using an x2 550. The third core works fine, the fourth core is bad. I have been happily running with three cores enabled for over a year now.

      • by cforciea (1926392)
        Sandy Bridge is the first time in a long time that Intel was remotely competitive with an AMD chip at the same price point. Many of us assume this bears some relation to the fact that early indicators put AMD's Bulldozer architecture on top of Intel's top end systems, so it is possible that we're going to see a role reversal because Intel needs to be more competitive in the mid-range market to keep in the game. This is, of course, all speculation, but it isn't hard to go back to pre-sandy bridge and see tha
      • Exactly. i5-2500k is a MONSTER of a chip for $200. The only reason you would go above this is for specialized media encoding. You dont NEED hyperthreading for anything other then specialized operations and time sensitive stuff. For that you should be expected to pay a premium. I would also like to add that the 1155 i7 is only $100 more.
  • lots of blather about the 'next dimmension' of computer chips... 3d Inside or some such. Still, nice to see that moore's law still holds... at least for now...
    • by SomePgmr (2021234)
      Still, nice to see that moore's law still holds... at least for now...

      Well I can't see into the future, but I have to say... it's encouraging that we've been saying that for a long, long time. :)
      • by cartman (18204)

        The NYTimes article on this topic had a quotation from an intel researcher, who claimed they're almost certain they can reach 10nm, probably by 2015.

        It appears that Moore's law has at least a few years left to it.

        • by Thud457 (234763)
          The generalized form of Moore's law describes society, not so much physics.
        • by tsotha (720379)
          Assuming progress is incremental There are a lot of things in the pipeline that might mean orders of magnitude smaller design rules. One of them will work out eventually.
        • Afaict the lattice constant of most semiconductors is about half a nanometer, other semiconductors are similar. That is 40 times smaller than 20 nm.

          In other words if 20nm stuff comes out in 2012 and features sizes halve every 3-4 years (equivalent to doubling transistor counts every 1.5-3 years assuming all else remains equal) so then arround 2028-2033 we will be down to feature sizes that are comparable with the size of the repeating pattern of that crystal structure. My understanding is that as we approac

  • its 2D (Score:4, Informative)

    by aepurniet (995777) on Wednesday May 04, 2011 @04:20PM (#36028296)
    they are not 3D, they are just thinner and deeper than the standard, we still dont see transistors on top of each other. the latice is still pretty much 2D. i ussually dont complain too much, but slash dot summaries are batting way below the mendoza line.
    • by bryan1945 (301828)

      Even the actual article wasn't very good- "streetscapes" and "phone poles"? Yeesh. I find myself coming here less often in the past year.

      • by Coren22 (1625475)

        No one reads Slashdot for the article itself, people come to Slashdot for the conversation. If you are trying to say the article is bad, you are doing it wrong.

    • by Anonymous Coward

      You are right that 3D is a bullshit term here; but this time slashdot is not to blame. Almost all the IT press is describing it in those terms. This is probably due to this nice headline from Intel's press release: "Intel Reinvents Transistors Using New 3-D Structure". Allow me to quote from ChipsZilla's website:
      "The three-dimensional Tri-Gate transistors represent a fundamental departure from the two-dimensional planar transistor structure..."
      It is just like Nostradamus predicted.

    • Re:its 2D (Score:5, Informative)

      by denobug (753200) on Wednesday May 04, 2011 @04:53PM (#36028616)

      they are not 3D, they are just thinner and deeper than the standard, we still dont see transistors on top of each other. the latice is still pretty much 2D. i ussually dont complain too much, but slash dot summaries are batting way below the mendoza line.

      No the structure is totally different. Look at how the source, drain, and the gate are arranged. Different geometry here.

      The article from AnanTech has the best explanation of the technique I have come across so far today:

      http://www.anandtech.com/show/4313/intel-announces-first-22nm-3d-trigate-transistors-shipping-in-2h-2011 [anandtech.com]

      • by emarsman (2110594)
        Haven't transistors always been 3D? You may draw them 2D in layout, but it's still a gate sitting on top of a channel with stacks of metal... it has L x W x H. I think that's called 3D. Silly Intel marketing.
        • by slew (2918) on Wednesday May 04, 2011 @05:34PM (#36029118)

          Haven't transistors always been 3D? You may draw them 2D in layout, but it's still a gate sitting on top of a channel with stacks of metal... it has L x W x H. I think that's called 3D. Silly Intel marketing.

          Actually integrated transistors have been "planar" for the most part (although there have been "vfets" and other types of 3d channels in the past)...

          In planar transistors, the field that chokes off the source/drain path has been mediated by a gate which is just on top of the channel on one plane. Imagine an iron on of a ironing board heating it up the board when you turn it on. Although the ironing board and the iron are both 3-dimensional objects, the interface in a "plane" and the heat diffuses across this plane. In this analogy with a planar transistor, the channel is dug into in the ironing board and the iron is the gate.

          In finFET, the gate surrounds the channel on 3 side. Imagine now a tube on an ironing board and the iron has a notch cut in it so the iron surrounds the tube on three side. When you turn on the iron, heat diffuses across all around the notch instead of interfacing on 1 plane. This is "3D" or finFET instead of planar. In this analogy with a finFET transistor, the channel is the tube on the ironing board and the iron with a notch is the gate.

          As you might imagine, the finFET architecture should have a better capability to turn on and off the channel since there is field is wrapping around the channel instead of just being applied to one side (okay that's simplification, but you get the idea).

          Intel took, this finFET idea and added another twist with a "3", called tri-gate (or tri-channel depending on your point of view). This congolomeration of two independent ideas that both revolve around the number 3 is the kind of thing that drives marketing people to be silly ;^)

          And for those that don't understand w/o a car analogy, imagine the difference in traction you get with bald tires on ice (planar where tire is the gate and ice is the channel), vs snow tires on dirt (finFET where the tire is the gate and dirt is the channel)...

          • by Anonymous Coward

            Perhaps a better car analogy would be a planar transistor is a two-cylinder single-port injector engine and the intel 3d-tri-gate is a six-cylinder multiport injector engine...

          • Car analogy FTW! :D

        • by Belial6 (794905)
          No kidding, if they could make them in only 2 dimensions, now THAT would be impressive!
      • they are not 3D, they are just thinner and deeper than the standard, we still dont see transistors on top of each other. the latice is still pretty much 2D. i ussually dont complain too much, but slash dot summaries are batting way below the mendoza line.

        No the structure is totally different. Look at how the source, drain, and the gate are arranged. Different geometry here.

        The article from AnanTech has the best explanation of the technique I have come across so far today:
        http://www.anandtech.com/show/4313/intel-announces-first-22nm-3d-trigate-transistors-shipping-in-2h-2011 [anandtech.com]

        Well... the geometry is still pretty similar. What they've done is raised the source/drain up and wrapped the gate around it, which means they have increased gate area, and can pretty much further increase it as much as they want, without changing the surface area required by the transistor.

        However, it seems to me that this is going to increase the gate oxide leakage, which is a significant part of the power burned non-usefully by FETs, and also increase the gate capacitance, which means the gate drive cu

    • by lee1 (219161)
      Allow me to rise in defense of my summary. The press article that I summarized described the transistors as "3D", as did the people from Intel that they used as sources. Your quarrel, which is valid, is with them, not with the summary. Also, Intel's use of the term "3D" can be justified; even though the array of transistors might be planar, each transistor employs a 3D geometry in a nontrivial way. And, who or what is mendoza?
  • Millions of free accelerometers! (For high-G's anyway.)
  • D'oh - they should have thought of that much earlier! Get perpendicular - soon also available for SSDs. http://www.youtube.com/watch?v=-xPvD0Z9kz8 [youtube.com]
  • What is indeed news is that intel is fielding them first.

    Well, while it is nice a slashdot article has finally been written about FinFET's - there may already have been one, I just can't remember - these devices have been widely guessed to be a part of the 22 nm technology node for quite some time. (see: http://www.itrs.net/ [itrs.net] and http://en.wikipedia.org/wiki/22_nanometer [wikipedia.org] ).

    They offer more effectivity for your gates as the field is not coming from one, but from 3 sides to the channel. That means a bit more s

    • These aren't (common) FinFETs. They're trigate, and a little different than conventionally defined FinFET. See here [intel.com].
      • Well, the original FinFETs were twingate (as were some similar transistors before the term FinFET was coined for them), intel invented trigate. In the end, this still is the first kind of FinFET coming to market.

        I just think inventing them in the first place is more inventive then refining them.

        My point of them having been quite a long time in the making still stands. Just note the age of that press release (2002).

        Oh dear. They called them 3D-Transistors even back then.

        • I agree with your tautology that inventing them is more inventive ;)

          But what's actually meaningful is deriving benefits from something. You can go off and invent all kinds of crazy shit in a lab, but if nobody ever does it it's meaningless.

          • Yes. You may note, that was my original point. Intel fielding them before AMD and the rest does is what the real news is about here.

            Nevermind the tautology. I'm tired and drunk. Sorry. ;-)

            But a link to the inventors would have been ... well ... informative.

            And the kind of FinFET technology used is more a matter of what is more effective in your semiconductor process. Whether it's twingsate, trigate, quadgate, whatever. There's a good chance what they published years ago is not all that similar to what

  • What other trending marketing buzzwords can we apply to unrelated technology to get attention? Social Transistors? Transistors 2.0? Cloud Transistors? Tablet Transistors? iTransistor. What would truly be remarkable would be to develop a 2-dimensional transistor. Infinite layers of them with no height added to the die - and it'd still be slim enough to fit in a tablet!
  • What's next? A Twisted Transistor?
  • by slapout (93640)

    3D? Aren't all transistors 3d? Along with every other physical item?

    • by mark-t (151149)

      Transistors are only 3d in the same sense that a drawing on a piece of paper could be called 3D owing to the thickness of the paper it is drawn upon.

      The silicon upon which the transistor exists is, of course, 3d... but the transistor itself is simply etched into its surface and treated with chemicals to produce the desired effect.

      • by Darinbob (1142669)

        But multiple layers requires 3D. It's more than just a drawing because some bits are on top of other bits.

        • by mark-t (151149)
          In which case you are talking about something with more than one transistor.
          • by Darinbob (1142669)

            No, you can't have a wire that has width cross another wire that has width and still be 2D.

    • by tyrione (134248)

      3D? Aren't all transistors 3d? Along with every other physical item?

      The point of 3D is to have 3 distinct dimension and in this case, 3 distinct dimensions where electrical current is controlled.

      • by strack (1051390)
        wow dude. you managed to use the strict definition of something to point out a trivial fallacy in the article, then use that to make a patronising and trivial correction to make yourself seem smart. ooo. you used the word 'distinct' twice. in the same sentence. it must be the only 'fancy' word you know. i got a idea, unless you have something real to add to the conversation, shut the fuck up.
    • by TopSpin (753)

      It's 3D in the sense that it has more depth than the previous version. The original Castle Wolfenstien [wikipedia.org] (1981) had very little depth relative to later incarnations, such as Wolfenstein 3D [wikipedia.org]. The 3D aspect of the new Intel transistors improve power consumption by reducing current leakage. The new transistor may therefore have much smaller feature size. Likewise, Wolfenstien 3D improves game play by reducing the players perspective of the map and increasing anticipation.

      You will notice that both the origin

  • Although this provides a superior transistor, there is a drawback from the IC designer's standpoint, if the illustrations are accurate. The designer of an IC with planar FETs can control both the width and the length of the active region. With these FinFETs, only the length is under the designer's control. Width is fixed (it's the vertical dimension in the illustrations). The only way to get stronger drive is to parallel transistors, whereas in planar design the transistor is simply made wider.
    • This is a real disadvantage from an IC designer standpoint, I would expect that this is not as much of restriction as they are already contained by in the extreme limits imposed on the design by double (possibly triple in this case) patterning to get down to 22nm.

      I did see a TSMC presentation earlier this spring at SPIE showing scaling down to the 1xnm node where they actually had little "pipes" of channel where the gate wrapped all the way around (Think of the gate line in a FINFET going through the "fins"

  • It now makes much more sense that Apple is having Intel manufacture the next generation IOS chips [slashdot.org]. Based on information from Ars [arstechnica.com], the finFET shines at lower voltages. So Apple is going to be using Intel not just because they can manufacture smaller than everyone else, but also because the resulting chips will be faster or take less power than anyone else.
  • Faster *and* lower-powered, from *Intel* ?

    Truly, the end is nigh.

  • isn't this just a mulitgate finFet. That structure has been around for a long time.

  • Are there articles from nine, four, three, two, and one years ago predicting this breakthrough? To someone not in the business, this seems to have come out of the blue--already in production by the time the news breaks. How long have they been working on this?

There are worse things in life than death. Have you ever spent an evening with an insurance salesman? -- Woody Allen

Working...