Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
×
IBM

New IBM Plant Will Mass Produce .1 Micron Chips 378

Posted by Hemos from the build-'em-small dept.
Ruger writes "AP News is carrying this story about IBM opening a new plant in upstate New York. What's most interesting about the story is that IBM will be producing .1 Micron Chips rather than the usual .25 or .18 produced by Intel and other chip makers, or .13 Micron chips they currently make for their PowerPC chips."
This discussion has been archived. No new comments can be posted.

New IBM Plant Will Mass Produce .1 Micron Chips More

New IBM Plant Will Mass Produce .1 Micron Chips

Comments Filter:
  • Waiting for G4 Newtons with the new process :)
    This should give the MHz deprived (But MIPS/FLOPS enriched) PowerPC line a boost in the PR speed department.

    • Re:Supertiny G4's (Score:4, Informative)

      by ivan256 ( 17499 ) on Thursday August 01, 2002 @10:09AM (#3992222)
      IBM does not make G4s. They don't have a license for Altivec. They already make quite speedy G3s, but you don't see them in consumer products that are marketed based on Mhz.
      • IBM's G3's are in Apple's iBooks.
    • IBM doesn't make G4's, for the Altivec licensing issue, which I believe they lost in a tiff with Motorola and Apple with the first gen G4's. IBM will like use this new process to make the next gen Power server chip and rule the RISC roost. I think HP's latest PA-Risc chip, kinda scares them. IBM also uses its fabs to etch chips for other companies, so a much faster Transmeta or AMD chip is possible, or maybe even a GPU.

      Whatever they do use this new fab for, Apple will probly be the last to benefit.

  • Why .1 micron? (Score:1, Interesting)

    by 3.5 stripes ( 578410 )
    And not .10?

    I know it's a stupid question, but I prefer a little consistency, .14 .13 .12 .11 .1 just doesn't seem right.
    • I thought the exact same thing when I read the article summary. More numbers after a decimal place inuitively make the number look smaller to most. I have the feeling that the marketing department will see that extra digit added on for consistency.

      Of course I never claimed to be a math major.
    • Maybe they can't really be sure it's .10 micron, and that it may be .14>x>.06 or something.

      (Of course, I'm only joking.)

      Scott

    • Once we get to .10 microns, we've reached another power of ten. So, 100 nanometers would be a better description, and we can ditch the decimal places. Next year we can talk about 99nm and 98nm parts.
      • Next year we can talk about 99nm and 98nm parts.

        Doh, I should have said "90 and 80nm parts". That would be slightly more interesting.

        Of course, this would also give the marketing droids a heck of alot of fun. They'd advertise that they have a 99nm process while their competion has a whopping 100nm process.
    • because "point one" looks smaller than "point ten" at first glance
    • It really shouldn't matter at all. If you can't tell the difference between .1 and .10, then the story is probably not going to interest you that much anyway. Why don't we write .130, or .1300000000... its because the 0's are useless information that just takes up space.. kinda like these posts =P
      • The number of digits indicated expresses the precision of the measurement. If I say that it's 80 degrees outside, I'm probably using my own human perception of the temperature, and if it's really 77.2 or 82.4, then I still gave a correct - if somewhat imprecise - reading. I only had one significant figure, and if you round these values to have just one significant digit, they come out to the same thing. If I declare the temperature to be 80.0 degrees, and you don't think I used a thermometer, you're rightly going to tell me that I'm a moron, because I don't have the ability to sense temperature with that sort of precision. If I did have an accurate thermometer which read 80.0, then I narrow the range of reasonable possible temperatures greatly.

        There are all sorts of rules that nobody learns anymore about how to propagate error by doing your math with significant figures - the result gives you an order-of-magnitude idea of how wrong your result might be. It's the old scientist's version of garbage-in, garbage-out. Likely, Intel's marketroids don't understand this distinction - the process is probably closer to .10 micron than .1; maybe even .100 is a better measurement (I know nothing about chip fabrication, of course.)

        *ponders the irony of using Farenheit degrees to explain scientific measurement*
    • I wasn't so much suprised by the missing 0 as much as the .10 target. Intel and AMD are aiming at .09 for their next generation so why is IBM not doing the same?

  • Returning to the fold. (Score:3, Informative)

    by GodInHell ( 258915 ) on Thursday August 01, 2002 @10:02AM (#3992175) Homepage
    'Bout time IBM got back into upstate NY.

    I remember when I was just leaving the area, the last of the local plants finally scaled back to just a matinance group, the whole area died. IBM was the heart and soul of quite a few towns in New York, and they didn't do very well when it left.

    -GiH
    • One company dominated towns are a horrible idea, always have been. IBM is a public company and if it benefits the stockholders to leave a town, they are legally obligated to do it.

      If you don't have a diverse economy that can take an IBM or a GM leaving, you have to fight like hell in the good times to grow one because if you don't the towns will shrivel up and die when they inevitably leave.

      • You could live in Rochester, NY and then you could be dominated by three failing companies - Bausch & Lomb, Xerox, and Kodak.
      • IBM is a public company and if it
        benefits the stockholders to leave a town, they are legally obligated to do it. [emphasis mine]

        I hear this argument a lot. That publicly held companies have a legal obligation to "benefit sotckholders" or "maximize profits." Is there really a legal basis for this? Are there civil statutes that say companies must do whatever it takes to make money for stockholders? Or is this legal obligation based in contract law where the stockholder will/can sue if the company makes decisions that appear to adversley affect them?

        I have a hard time believing that the DA, SEC, or FTC would go after a company that made unprofitable business decisions. Anybody know?

        • I think the stockholders are the law ... if they are not satisfied they'll vote agaist the current C*O's, and they loose their jobs ;)

        • That publicly held companies have a legal obligation to "benefit sotckholders" or "maximize profits." Is there really a legal basis for this? Are there civil statutes that say companies must do whatever it takes to make money for stockholders? Or is this legal obligation based in contract law where the stockholder will/can sue if the company makes decisions that appear to adversley affect them?

          IANAL, however my understanding is, yes, stock holders can sue the Board/CEO if they believe that that they are not working to maximise profits, and therefore stockholder value.

          Al.
        • It's not a law as such, but the stockholders can sue the board if they feel that all possible steps weren't taken in the best interests of shareprices / earnings.
        • The sharholders are the owners of the corpoiration. Accordingly, the law states that memebers of the Board and corporate officers have a fiduciary duty to act in the best interests of the shareholders. They may be sued by the shareholders for failure to do so, and moderately frequently are.

          IANAL
    • 'Bout time IBM got back into upstate NY.

      Actually, this isn't a new plant opening up. It's just a new assembly line in the existing East Fishkill plant. My mom works for the Microelectronics division there, and she points out that the line is referred to as the "300mm line", not the "0.1 micron line".
  • I wonder how easy it would be for lightning to fry these chips?

  • To the naysayers... (Score:2, Insightful)

    by colmore ( 56499 )
    Weren't we supposed to hit some sort of quantum limit before .1 Micron? What are the current guesses on how much smaller we can get?

    I wan't to be reading my email and playing nethack on a petaflop machine by the time this decade is out!
    • I wouldn't have done it if it weren't for your .sig, but come on man, wan't? What were you thinking?
      • What were you thinking?

        Let's see... I posted that before noon, so it was probably "Why the hell hasn't the caffeine kicked in yet."

        I'm perfectly aware that wan't isn't correct. It's a bizarre mistake that I catch myself making all the time, and I can't explain it. And otherwise, I'm pretty good with apostrophes; I don't put them on plural's like a lot of people do around here. But for some reason "wan't" is a really really stupid mistake that I've been making for a long long time.

    • Re:To the naysayers... (Score:3, Informative)

      by jmv ( 93421 )
      What are the current guesses on how much smaller we can get?

      Usually, the current guesses are about twice smaller than current technilogy :-)

      Seriously, there are two (in fact more) limits: there's the smallest transistor possible that works correctly and there's the smallest features size we can mass-produce with reasonnable (well, it's already unreasonnable...) cost.

      Right now, the most limiting factor is the second. The visible light is already much too big (wavelength) for lithography so they're using (AFAIK) ultra-violet, but one of the problems is that the smaller the wavelength, the harder it is to find a transparent material at that wavelength (glass doesn't work past a certain wavelength).
    • I thought the limit was that to reach .1 microns you had to use UV lithography instead of more traditional means.
    • There is a reasonably well researched report published each year called "International Technology Roadmap for Semiconductors", homepage is here [itrs.net].

      The executive summary of the 2001 edition predicts that in 2016 the drawn gate length for microprocessors will be 13 nanometers (0.013 microns).

      Now that we're on the verge of 0.1 micron transistors it is time to dump the microns unit and start using nanometers. The tables in the "International Technology Roadmap for Semiconductors" all use nanometers.

  • impressive (Score:3, Troll)

    by tps12 ( 105590 ) on Thursday August 01, 2002 @10:05AM (#3992192) Homepage Journal
    If this is true, it looks like Moore's Law could have a few years left in it, after all. In a few years, we may end up living in the future!

    Imagine a computer small enough to fit in your pocket. Imagine a computer in your car. Imagine a computer in your glasses! It sounds like science fiction, but it looks like IBM is actually seizing the bull by the horns and making it a reality.

    It's also interesting that they are doing this in New York. I thought all chip manufacturing was done overseas, where labor is cheaper. Perhaps IBM is getting some sort of government subsidy for creating American jobs. Or maybe New York has a good supply of chipmakers already, so they can find more skilled workers.

    Whatever the reason, it's good to see innovation marching along. This is the kind of activity that will get us out of the current recession. Good luck, IBM!
    • Imagine a computer small enough to fit in your pocket. Imagine a computer in your car. Imagine a computer in your glasses! It sounds like science fiction, but it looks like IBM is actually seizing the bull by the horns and making it a reality.

      Um.. To summarize,
      Computer in your pocket: PocketPCs/Palms
      Computer in the car: The volvo story that was run a week back
      Computer in glasses: oh so many stories posted about the wearable computers at MIT.

    • That's funny, I could have sworn I've seen people going to work everyday at the intel fab down the street from my appartment (Hudson, MA). In fact, rumor has it that this fab was the worlds first to churn out working parts on the .13 micron process.

      There are still plenty of fabs in the US. It's probably because the people who can make these tiny technologies actually work aren't cheap anywhere.
    • Imagine a computer small enough to fit in your pocket. Imagine a computer in your car. Imagine a computer in your glasses! It sounds like science fiction...

      I no longer have the exact quote, but it goes something like this...

      While computers today have over 18,000 vaccum tubes and weigh 1 ton, in the future computers may have as few as 1,000 vaccum tubes and weigh only 1/2 ton. --Popular Mechanics 1949
    • I appreciate your sentiment, but really -- I have a computer in my pocket! It's called a PDA. Most modern cars have computers in them, too! Maybe not the kind you're thinking of, though.

      I guess what I want to know is what's impressive about this advance? What will it actually mean as far as technology advances go? Faster chips? "Better?" Just smaller? What? How will this improve our lives? It's not immediately obvious to me, and I'd like to know!

    • Or it could be that we already have plenty of talented employees here at the east fishkill plant who know 200mm fine so converting to 300 won't be that bad.
      Disclaimer: I work at the east fishkill plant as a programmer for the 300mm testing systems.
    • It's also interesting that they are doing this in New York. I thought all chip manufacturing was done overseas, where labor is cheaper. Perhaps IBM is getting some sort of government subsidy for creating American jobs. Or maybe New York has a good supply of chipmakers already, so they can find more skilled workers.
      Not all of it is done overseas. Check out Fairchild Semiconductor's Manufacturing [fairchildsemi.com] or National Semiconductor [national.com] or Motorola [motorola.com] to name just a couple off the top of my head. I know some of them at least do the manufacturing over here, but assemble them over seas. I know both National and Fairchild only becuase I use to support the Suns and VAXen for production in the fabs.
  • Midly offtopic/funny.... EAST FISHKILL, NY?
    Anyway.
    First we got all this buzz about IBM buying PWC. Now, this about .1 micron chips. Upstate NY isn't exactly the center of the electronic world, but the goal was to do something _new_ & slightly daring. (and hey, they are selling/sold the plant in which IBM was born) They're taking a chance with the automated wafer production. (IANA Quality Control Specialist) -- But wonder about the types of problems they may run into. If they pull this off, they could make a nice comeback. Only question is...how strong is the demand for these chips *right now*, instead of "when the economy recovers" ?
    • It's an anglicization of the original dutch which has nothing to do with killing. There are a bunch of places called kill in metro NY and unless you know something of the history of the area you'll jump to the wrong conclusion.
  • What will overclocking chips made with this new wire size do to their heat output? No matter, it still can't outperform my current Athlon/space-heater.

  • Intel at .13? (Score:4, Informative)

    by timwhit ( 572715 ) on Thursday August 01, 2002 @10:06AM (#3992201)
    Isn't the northwood P4 produced at .13 micron? And the AMD Throughbred is also at .13? The header says that other chip manufacturers produce chips at .25 or .18 when this simply isn't true.
    • Right, but it depends on what type of chip IBM is going to produce. I expect many other chips are still being produced at .25 microns. So while the article was at best worded poorly, it may be an indication of the market IBM is planning on competing in.
    • Hmm, it seems not too many people read Digitimes. TSMC and UMC have both been at .13 for a long time and I was under the impression they were doing .90nm already. That's what all the talk about SOC bottlenecks on EETimes is about. Check out the TSMC home page for more on 90nm SOC plans.
      For more fun, according to Digitimes, TSMC and a German company have already begun work on a .60nm foundry in Singapore that is supposed to be in operation by 2Q 2003.
      To answer an earlier question about ultimate CMOS limits, IBM says 45nm is the limit for gates at 1V and below that you lose speed in data processing applications. Of course GaAs and InP might extend that limit, but then you introduce process issues. The lithography challenges while expensive are not the primary technologcical issues. This latter point is intriguing because it leaves the door open for applications where minute dimensions are more important than processing speed such as biochips. Check out the Univeristy of Michigan for some awesome preliminary work on production level implanted bionetworks.
    • I may be feeding a Troll, but here goes...

      It is true, some manufacture does currently occur at .25 micron. Obviously there is advantages to .13 micron such as, less heat production, smaller die size etc. and this is what the latest chip fabrication plants use.

      The older fabs still use much larger processes, e.g. .18 and .25 micron. I know that the P3's still being produced (pre Tualatin 733, 800 Mhz etc.) are on .18 micron processes, and I would guess that the K6-2's that AMD still produces are at least a .25 micron process, maybe larger.

  • Upstate, eh? (Score:1, Offtopic)

    by Glove d'OJ ( 227281 )
    East Fishkill is *so* no Upstate NY. Albany? That's upstate. Syracuse? Definitely upstate? Fishkill? No freaking way.

    East Fishkill is 1/2 way between NYC and the extension of the horizontal line that divides most of NY and PA.

    Dude, get your geography straight, or at least *look* at a map.

    (sheesh)

    WWJD? JWRTFM!
    • For somebody from Long Island, everything north of the Bronx is upstate.
    • For most people, especially those who don't live in NY, anything not in NYC is "upstate NY."

  • Embedded (Score:3, Interesting)

    by levik ( 52444 ) on Thursday August 01, 2002 @10:07AM (#3992209) Homepage
    So these new chips will probably be targeted mostly at embedded/handheld devices, where running at lower temperatures is more important than raw power, since you can't have a loud fan in a set-top box drowning out your home theater system, or a big heat-sink in a palm-top (for obvious reasons)...

    I kinda wish IBM whould come out with a x86 compatible chip to introduce some competition to the field. Their big name should give them enough of a leverage to allow them to enter the market, and I truly believe them capable of delivering a high-performance, low power product.

    • Re:Embedded (Score:2, Interesting)

      by handorf ( 29768 )
      or a big heat-sink in a palm-top (for obvious reasons)...

      I never understood this attitude. Just make it so that the processor speed scales with cooling ability. Most of the time a handheld PC will be attached to a 40-60kg heat sink with fairly good conduction properties (water cooling!). Why not take advantage of the situation?!?
    • I kinda wish IBM whould come out with a x86 compatible chip to introduce some competition to the field.
      right... to compete with who? chipzilla? They did that several years ago and found it wasn't profitable... very wisely got the heck out of that dead end.
      Their big name should give them enough of a leverage to allow them to enter the market, and I truly believe them capable of delivering a high-performance, low power product.
      They already do, it's called the PowerPC and is quite popular in the embeded market, it also rocks on the desktop, as any Mac user will attest. You just have to get your mind out of the wintel monopoly and several new avenues open up. (and before any one complains that you can't buy ppc mobos like you can x86, that's not true anymore, they are starting to show up again.)
  • Since 0.09 micron starts is not so catch they'll probably call the next reduction in feature size 90 nanometers.

    Ouch.

  • Current 0.13/300mm technology went into pilot production more than a year ago, and only Intel (or maybe AMD) seems to be able to get reasonably good yield on it now... Some even called 0.13/300mm the biggest vendor conspiracy ever. Well, good luck to IBM and may the step from 0.13 to 0.10 be less painful than 0.18/15 to 0.13 mircon.
  • Here's a story right from an upstaty NY newspaper:
    Press and Sun Bulletin [pressconnects.com]

    IBM used to be a big thing where I come from. It started in my hometown of Endicott NY. Over time it's slowly moved away. But I'm glad to see they still stay in upstate NY somewhat.
  • It's not the size that matters baby, it's how you use it!

    ----------

  • I'll be damned... (Score:2, Informative)

    by powerlinekid ( 442532 )
    "Look ma, I'm on slashdot"... well not exactly but I actually work there. I program the testing systems so that the engineers can run test on the wafers. The ribbon cutting was pretty cool, CEO Sam was here and so was George Pataki. Nothing like sitting in the conourse for lunch and seeing a massive black helicopter fly overhead. Got a free hat out of it... to be entirely honest this is a big deal but business here really isn't going to change. We've been porting our testing system from the old design to the 300mm for awhile now and theres been alot of restructuring of the departments such as moving people to the new 300mm ones etc.

  • What will they make? PowerPC? (Score:3, Insightful)

    by crow ( 16139 ) on Thursday August 01, 2002 @10:24AM (#3992372) Homepage Journal
    The real demand for using the smaller feature size is in two areas--low power and high performance. In the low-power market, you have all sorts of consumer electronics like cell phones. In the high performance, you're talking CPUs. Personally, I would love to see them build PowerPC chips.

    From the article, it sounds like they'll be operating the plan under contract from other companies, so it will most likely be making chipsets for pagers and cell phones.

    Of course, the market can be expected to change significantly between now and when the plant is actually ready to build chips.
    • It's a fab plant, it can make anything they are sent the masks for. This week it maybe the new Power 4 chips for high end servers, next week it might be small embeded chips for a cell phone. maybe in between they'll run an engineering test batch of some cool new, as yet unnamed, next generation chip for the chip developers to trial.

  • IBM vaporchip (Score:3, Funny)

    by mapmaker ( 140036 ) on Thursday August 01, 2002 @10:28AM (#3992405)
    AP Reporter: Wow! 0.1 microns! How small is that?

    IBM marketroid: That's almost as small as some gas molecules. In fact, you could say these new chips are just VAPOR.

  • Additional Coverage (Score:3, Informative)

    by leibnizme ( 264472 ) on Thursday August 01, 2002 @10:40AM (#3992496)
    If you want to have more details about this fab, check out:

    IBM's news [ibm.com]

    Yahoo Story [yahoo.com]

    NY Times (free reg, blah) [nytimes.com]
  • From the last line in the article
    Worldwide sales of all chips are expected to total $143 billion in 2002, $177 billion in 2003 and $213 billion - a 20.9 percent increase - in 2004. Another slowdown is expected by 2005.

    Can they *really* predict this stuff so far ahead??

    • Can they *really* predict this stuff so far ahead??

      No, and if they could they would work on wallstreet, not for IBM or any other non financials company =)
    • They can't really predict so far ahead, but the cycle has been so reliable that it is sort of a self fulfilling prophecy.

  • One of the Laws related to Moore's asserts that the cost of a state-of-the-art fab line doubles every three years. The article says that this line is costing $2.5B. IIRC, new lines in 1999 cost about $800M, so this would appear to be pretty close to the prediction. The potentially bad news in this is that by 2011, a new fab line will cost $20B, which is probably more than anyone except large governments (or Microsoft) can afford. By simple calculation, 100M working devices produced over the lifetime of a $20B fab line must cost $200 each just to cover the initial cost of the fab.

    I don't really care about Moore's Law itself (transister count doubles every 18 months), but do care about the corollary that says instructions per second per dollar doubles every 18 months. Can we keep that corollary going without Moore's Law itself (and the attendant economic fab limitations)? Asynchronous circuit designs? Parallel processing? Alternate cheaper fabrication for a 1B op/second processor?

  • ARGH! (Score:2)

    by shepd ( 155729 )
    "While companies like Intel Corp. and Samsung Electronic Co. already manufacture 12-inch wafers..."

    "...will be the first IBM chips to be made on 300mm wafers of silicon"

    Don't mix metric and imperial measuring systems.

    Doing this is like SHOUTING. Well, maybe not. Its more like a cellphone ringing in a theater.
  • AMD and Intel both have .9 in their near future plans. That production should be occuring within a year or so, I think. I don't really see what the point of this story is, companies build plants all the time.
  • So this technology allows speed ups of about 70%, all other factors being equal. Will bring chips to around 3.5 GHz or so.
  • AMD and Intel are right there. Consider ...

    http://www.amd.com/us-en/Processors/ProductInforma tion/0,,30_118_608,00.html

    and

    http://www.intel.com/ebusiness/products/roadmap.ht m

    Both of these show .09 in the next year. I'm all for giving a company its due, but lets not leave the other players out. Maybe even, *gasp* go for a complete story.

  • GamePro Magazine has an article in it's latest issue speculating that these .1 micron chips IBM is producing will be cell based and will form the basis of the PlayStation 3.

  • "New" IBM Plant (Score:5, Informative)

    by wilsonjd ( 597750 ) on Thursday August 01, 2002 @12:07PM (#3993179)

    Just for the record, I work at IBM in East Fishkill, NY (but not in Chip manufacturing.) I do NOT speak for IBM.

    IBM has been in East Fishkill since the early 60s, manufacturing chips and packaging (MCMs, etc.) for mainframes. IBM employs over 10,000 people in Poughkeepsie (10 miles away) and East Fishkill. The "New" plant is a new chip fabrication line in an old building (building 323) that used to be used to make bipolar chips for mainframes. They have been working on this new plant for over two years, and it is already producing sample chips. Normal production is scheduled to begin in the fourth quarter (btw, the current production from this plant is 0.13 micron, but in the future, it will move to sub 0.1 micron processes.)

    IBM is using this plant as a high-end foundry. In other words, customers will design high performance chips that will be manufactured here. They are already working with some high-volume customers (Nintendo, Sony, etc.) Customers will also include IBM chip designers (mostly IBM servers.)

    Oh, and on the whole upstate, downstate issue: People who live in upstate New York consider us downstate. People who live in downstate New York consider us upstate.

    And, as Gov. Pataki said yesterday, the Hudson Valley is much nicer than Silicon Valley. We have trees.

  • The article says "thinner than 0.1 micron".
    The industry was working for a while on
    90 nm (0.09 micron) tech so I guess this is
    what they have there.

Slashdot Top Deals

Any sufficiently advanced technology is indistinguishable from magic. -- Arthur C. Clarke

Close