AMD committed to process technology schedule

During a meeting with analysts yesterday, AMD reiterated its plans to complete its transition to 65nm process technology and to introduce its first 45nm chips on time, as eWeek reports. AMD says it is “on schedule” to fully ramp 65nm production at both Fab 36 and its new Fab 38, the overhauled version of its 200mm Fab 30, later this year. eWeek quotes AMD Director of Manufacturing Technology Tom Sonderman as saying that AMD plans “full, 65-nanometer microprocessor manufacturing by the middle of this year”—presumably meaning that 90nm chips will stop being produced by then. As for 45nm production, Sonderman says AMD still plans to deliver 45nm processors by the middle of next year. Pilot lines for the AMD’s 45nm process technology are “already running in Dresden [the home of AMD’s fabs in Germany],” Sonderman adds.

Comments closed
    • blastdoor
    • 12 years ago

    AMD has had low clocked 65nm chips since the end of 2006. But if you go to newegg, you’ll see that of the 170+ different AMD chips they sell, a grand total of 6 are 65 nm chips, and none of those are among the fastest models.

    We also know that the die size of these 65nm AMD chips are not nearly as small as one would expect.

    So it would seem that AMD won’t be selling their “real” 65nm chips until the second half of this year.

    I wonder if we’ll see a similar pattern at 45 nm — introduce a small number of low clocked chips that somehow aren’t really quite as small as they ought to be in the middle of 2008, and then in 2009 bring out the “real” 45 nm chips.

    • green
    • 12 years ago

    [Edit] scratch that
    found the TR article on it
    was mid-2008

    • Deli
    • 12 years ago

    i haven’t seen amd so ‘gung-ho’ about their coming process tech.
    must be a good sign. and seeing how IBM is doing well with SOI with their Power6, then it MAY also look good for AMD’s process tech.

      • Flying Fox
      • 12 years ago

      They need all the good news they can get or generate themselves. They are not doing so hot in the stock market and it is affecting their ability to print more money to even keep the company running.

      • MadManOriginal
      • 12 years ago

      Either that or it’s a bad sign 🙁 ‘Don’t have the goods? Hype the future.’

      One of the most telling lines in the blurb to me was this:
      q[

        • just brew it!
        • 12 years ago

        AMD can’t afford to build an entire new fab every time a new process debuts. They’re a tiny fraction of Intel’s size… building a new fab every couple of years would’ve bankrupted them years ago.

          • MadManOriginal
          • 12 years ago

          Of course I understand that, just something that popped in to my head when I read the sentence about renaming a fab.

            • Flying Fox
            • 12 years ago

            It’s AMD’s convention of naming their fabs. Basically a new building or a building with new equipment/production line will be renamed.

      • SPOOFE
      • 12 years ago

      The most gung-ho people I’ve met were strung out ex-felon smack-freaks that could hustle twenty dollars and two cigarettes out of a vagrant to pay for their requisite two cans of Steel Reserve and still have enough left over to go in on a half-G. Sometimes desperation brings out a certain manic energy.

      Not to say AMD is in this state: The second-most gung-ho people I’ve met were very successful, reliable go-getters.

    • Peffse
    • 12 years ago

    Wait, they don’t have a 65nm processor out yet, but they truly believe they will be at 45nm by next year?

    This is gonna be the same case as the 80nm/65nm 2900XT.

    Why buy now if they are gonna make it obsolete so soon?

      • FunkeeC
      • 12 years ago

      Uhh they have plenty of 65nm cpus available…just not the next gen core…

      • flip-mode
      • 12 years ago

      LOL, I’m running a 65-nano AMD chip right this minute

      • Sargent Duck
      • 12 years ago

      They launched 65nm chips back in November I believe. You just haven’t heard anything about them becuase nobody really cares about the Athlon right now.

        • RadiationMan
        • 12 years ago

        They launched in December, but weren’t around till late January early Feburary.

          • Deli
          • 12 years ago

          pretty sure they will continue making the majority if their chips with 65nm SOI even by end of 2008. Also pretty sure that the first 45nm chips will be for servers (Shanghai) and then move down to desktop. So 65nm won’t exactly be ‘wasted’ – 2 years old.

    • Sargent Duck
    • 12 years ago

    Just words. Valve said they were commited to launching episodic content every 6 months. The company producing Duke Nukem forever probably at one time said they were commited to releasing that game. Intel probably at one time as well said they were commited to seeing the P4 scale to 10ghz. And yet, here we are. Yes, I’m being overly pesstimistic, but you never know if AMD will run into Fab troubles, or process troubles until they actually reach that point.

    However, here’s hoping everything goes smoothly and trouble free for AMD.

      • alex666
      • 12 years ago

      I totally concur. As consumers, we all need AMD in the game (and anybody else for that matter), not just Intel. I love my 6600 C2D, the first Intel chip I’ve purchased since the 486 days. The C2Ds are killer chips, but like others have said, Intel probably would have never designed them if AMD hadn’t kicked their butt for a couple of years. Without competiton, we all suffer, even the most diehard Intel and AMD fanboys. So here’s hoping that AMD gets back on track and makes some killer chips in a timely fashion.

        • SPOOFE
        • 12 years ago

        I don’t buy into the “We need competition” myth as it pertains to AMD. I can just as easily claim that if AMD had folded years ago, some other entity would have filled the power vacuum which would have spurned chip development even further. Oh, how glorious of a world we’d be living in if it weren’t for AMD! And isn’t it mighty convenient that I naturally have no way of proving this!

          • BobbinThreadbare
          • 12 years ago

          There is huge barrier to enter in the CPU market, and a long list of failed companies and designs. I think it’s far more likely that AMD has helped consumers than hurt us.

            • stdPikachu
            • 12 years ago

            Quite. When you see how many billions are spent on R&D in this business you’ll see how difficult it is. AMD alone must have spent millions of man hours into researching mundane things like cache, let alone all the million other components that form a modern x86 CPU.

            If AMD went under, you wouldn’t see a viable x86 competitor for half a decade, at least. Heck, just the fact that a company as large as AMD went under the wheels of Intel’s juggernaut would be enough to put anyone else off for life.

          • green
          • 12 years ago

          you could also claim that if amd had folded a decade ago that now all we’d have is intel churning out a p6 architecture with only speed improvements, a larger cache and an extra core slapped on…… oh….. wait a minute……

          i’m more of the opinion that if amd had died off we’d see more architecture battles rather than clock / core battles (ala x86 vs power) as there’s always ibm to compete against. they’re still in the game as well considering that’s where amd’s getting bit and pieces of their process tech

          as a side note i realised why i thought amd said they’d have 45nm on shelves at the start of 2008:

          y[http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2986&p=8<]§ so exactly what those "45nm's" coming out of fab36 are is a bit of a mystery if it'd still be at most 6 months before it was on the market

            • Whybee
            • 12 years ago

            It is possible that we would have a mature Itanium in our desktops and an entirely new desktop operating system rather than Vista.

            You could argue that AMD has been holding us back all those years. So much for competition.

            • Buub
            • 12 years ago

            Itanium is not a good general-purpose processor. It is a good niche processor that does a few things really well and lots of other stuff less well. I’m glad we don’t have it in our desktops.

            And how does the choice of processor have anything to do with Vista?

            I find it amusing all the people here arguing how limiting choices could be good for the consumer. Looking at broader market trends, when have reduced choices benefited the consumer? And how often have they hurt the consumer? Seriously, I’m expected to believe we’d be better off if Intel had no credible competition?

            • Whybee
            • 12 years ago

            OK, static scheduling proved to be not terribly effective in increasing ILP, but you should also keep in mind that Itanium did not get nearly as much resources as the x86 line. If Intel could push Itanium into the mainstream market it would lead to much better compilers and more software, and more R&D money for the hardware development.

            And it has a lot to do with Vista because we could have a much better desktop operating system if Microsoft didnt have to think about all those compatibility issues. Starting from scratch now and then would be a very good idea.

            Another thing to keep in mind is that in a sense Intel always has competition – it has to compete against the installed base of Intel’s CPUs.

            So my point is that 1) things AMD does maybe good for consumers in the short term but bad in the long term. 2) intel would still have to offer better CPUs even if it did not have to compete against AMD. It is a complex picture and I do not think AMD’s bancruptcy will necessarily be a tragedy.

            • flip-mode
            • 12 years ago

            B.S. Whybee, and I’m surprised to hear it from you. A lack of competition wouldn’t likely mean a halt to progress but it would likely mean slower progress and it would certainly mean less price competition. So that’s bollocks #1.

            Bollocks #2: It could easily be argued that Itanium was Intel’s (flawed) response to competition and that if there were no competition then there would never have been any Itanium. Except at the time Itanium was conceived (early 90s) the competition wasn’t AMD but “big tin” processor makers. AMD was barely a pimple on Intel’s ass when the Itanium was conceived.

            Bollocks #3: Microsoft’s OS’s aren’t being held back by compatibility issues with various CPU architectures – I haven’t seen one single report or blog or analysis or whatever suggesting that is the case (though I haven’t searched for one so….) I’ve heard a million other reasons MSOSs are slow to evolve / innovate, but that’s never been one of them. And BTW, did you notice that Microsoft has no meaningful competition, and often when MS does “innovate” that those innovations are often inspired by the nearest thing it has to competition – Apple?

            Held back by competition my shoe. Sheesh…

            • Whybee
            • 12 years ago

            Yeah, dont take it too seriously. Sometimes, I like to take positions that are hard to defend and have fun arguing my point ;-). Anyway:

            1) In an industry with huge economies of scale (like semiconductors or OS development) competition CAN be bad. In economics it is called natural monopoly. So it is not B.S, it is a point you can discuss – people write PhDs about it.

            2) Intel used to be paranoid, looking for new opportunities – so I would not say Itanium was a flawed response to competition. Static scheduling combined with a very wide superscalar architecture was a reasonably good architectural idea that did not work out because of a number of factors. One of those factors was AMD64. Itanium got caught in a vicious circle: limited software support – limited market demand – limited R&D funds and so on. If it werent for AMD Itanium could be very succeful. For desktop aplpications higher ILP could actually be more important than for servers because you do not have so many threads.

            3) Microsoft is held back by the backward-compatibility requirements for software (I am not talking about CPU architectures). It is not really so difficult to write a good new OS, given MS resources. A completely new instruction set, if it became a de-facto standard, would force MS to write a new OS from scratch – it would have to innovate and it would not be limited by all those constraints. So we all would benefit.

            • flip-mode
            • 12 years ago

            If just for the sake of discussion then that’s cool. People may write papers on this stuff but it has to be taken on a case by case basis.

            As far as MS writing an OS from scratch – I wonder how much money I should wager that they’d try to port a bunch of code and still end up futzing up an opportunity for perfection anyway. I’ve seen some examples of Windows 3.1 icons that still linger in Windows Vista: bwah-ha-ha-ha-ha!!!

            • Buub
            • 12 years ago

            As flip-mode alluded to, Vista is not tied to the x86 architecture. That’s all MS is releasing the desktop for, but it’s built on the same foundations as Windows Server, which also runs on Itanium. There goes your no Vista argument.

            Windows Server (and in fact Workstation) have run on other architectures in the past, too, including PowerPC, Alpha, and MIPS. They have also flirted with running it on Sparc. The Windows “NT” server base is very well architected for platform-neutral development.

            • Whybee
            • 12 years ago

            y[

            • Buub
            • 12 years ago

            You’re right — it’s primarily a financial business decision. Itanium hasn’t earned its keep. Microsoft actually had a stronger portfolio for the Itanium early in its life, but Itanium just didn’t live up to its hype, so MS cut back on its offerings.

            The point is that Windows is quite capable of running on other architectures, but as you pointed out, MS isn’t going to do it just for the helluvit, because it is expensive.

            In fact, if I’m not mistaken, MS expects the host of that architecture to shoulder most of the cost of maintaining Windows for that platform. My take on that was formed when MS stopped Windows development for Alpha. I read between the lines that DEC no longer considered on-going maintenance of the Windows port to be cost-effective, so the arrangement was discontinued.

            • Whybee
            • 12 years ago

            Returning to the subject of application compatibility. Dont you think this is now a major problem for desktop OS development?

            EDIT: I also wonder what happens to buffer overflows and other similar nasty bugs when MS ports software to IA-64. After all Itanium has some form of non-execute protection and a lot of other security mechanisms. You have any idea?

            • green
            • 12 years ago

            argh i knew i shouldn’t have continued the ‘what if’ situation
            my point was that it’s silly to say we ‘need’ company X for competition is silly
            especially when it translates to ‘their product might be bad but i’m buying for the sake of competition’
            at that point it’s just ridiculous as it’s rewarding a company for bringing out bad products

            but can does someone have any info on what the 45nm coming out of fab36 will be around the start of 2008?

            • Whybee
            • 12 years ago

            You made a good point about P6, though. If you expected just a LOL in response, well…

            If you ask me, I have no idea about AMDs’ 45nm plans. It is anyway pointless to talk about this right now because nobody (including AMD) knows how may respins it will take to get the final silicon.

            • Whybee
            • 12 years ago

            y[< People may write papers on this stuff but it has to be taken on a case by case basis<]y So in this particular case why do you think competition is good - or is it just an item of faith? Without AMD, Intel would be able to satisfy all the demand at little extra cost, and maybe even at a lower cost since it will be able to cut marketing expenses. This may well translate into lower prices for everyone. In terms of prices, Intel still sets the level of prices and AMD has to follow. In the Prescott era, AMD was quite happy to keep Athlon prices high, and didnot put too much pricing pressure on Intel. In terms of innovation, AMD is mostly conservative and does not generat nearly as much innovative research as intel. So why do you think AMD should stay alive? It spends about $5B annualy to do things that Intel would be able to do essentailly for free, it has limited influence on pricing, and it is more conservative than innovative. P.S. y[< If just for the sake of discussion then that's cool.<]y Have you ever heard about a one-handed economist?

      • willyolio
      • 12 years ago

      the big difference, however, is that AMD has to commit to this schedule or they’re dead. if they keep underperforming against intel, they’ll lose all the marketshare they’ve gained in the past few years, and eventually be too far behind to be competitive, if Intel keeps the pressure up.

      Valve can survive if they scrapped all the HL2 episodes altogether and go back to releasing a big game every once in a while… it would piss some people off, but they’ll live, no problem. but look what happened to intel when the P4 didn’t scale up as expected. at the very least, they had size, momentum and a strong marketing department to keep things going.

      i can expect AMD to have far higher level of commitment to 45nm than, say, Valve with HL2 episodes.

      • alex666
      • 12 years ago

      SPOOFE, please note that I qualified my statement about AMD with the phrase “and anybody else for that matter”. But the fact remains, AMD was starting to kick Intel’s butt with some great processors. They were the only other player and they were taking it to Intel. Personally, I’d love to see 3 or 4 chipmakers out there, much as we have with memory, mobos, and hdds (though, unfortunately, no longer so with GPU makers). But the real point I was trying to make was that competition benefits us all.

    • king_kilr
    • 12 years ago

    The building I work in has a sign that says welcome AMD in the lobby, I am tempted to go and look for them and ask when Agena/Barcelona/K10/Phenom/Whatever the hell it is called will be relased.

    • Lord.Blue
    • 12 years ago

    So much for Intel’s “AMD is 18 to 24 months behind us.” quote.

      • eitje
      • 12 years ago

      hah – you’re tanking one company’s quote about a competitor based on that competitor’s opinion of themselves. well done!

Pin It on Pinterest

Share This