Forgot your password?
typodupeerror

Intel Admits To Falling Behind AMD 244

Posted by Zonk
from the they've-still-got-a-good-run-going dept.
Vicegrip writes "CNN is carrying a Fortune story covering an analyst meeting held on Thursday. There, CEO Otellini admitted Intel has fallen behind AMD with lost market share, technological leadership, and recently profitability. Intel also announced cuts to 1 Billion in spending." From the article: "Intel's market share recently slipped below 80%, and Otellini strongly emphasized the need for market share gains in all his remarks. On the other hand, he also suggested that Intel's recent market share losses (to AMD, whose name was not mentioned) were in line with historical variations which tracked to Intel's product generations."
This discussion has been archived. No new comments can be posted.

Intel Admits To Falling Behind AMD

Comments Filter:
  • by foundme (897346) on Thursday April 27, 2006 @05:44PM (#15216222) Homepage
    It's not common for a CEO to make such admission, which can only mean one thing -- they already have plans to regain those market shares. Or is it the classic "Avis: We Try Harder"?

    It's like poorer people tend to avoid being seen as poor, while wealthy people almost always say they are poor.
    • I think you're right. With Merom and Conroe just a few months away (supposedly, but I don't think they're going to pull a Microsoft) I think Intel will regain a large part of the market share they've lost. I've heard lots of good things about Merom and Conroe and nothing about any AMD plans for future products. This just points to that fact even more.
      • by DrMrLordX (559371) on Thursday April 27, 2006 @06:01PM (#15216355)
        AMD's response will be a chip known currently as K8L. Whether it will launch on AM2 or AM3 is anybody's guess, but it will supposedly come out sometime in 2007. Intel will have at least 5-6 months with the performance crown since AMD can't beat Conroe with their current K8 processors on s939 or AM2.

        The question is: can Intel retain the performance crown once they gain it? The last time Intel was the top dog performance-wise was back when the Pentium IV 3.2C was their flagship desktop part. That lasted until K8 hit the streets.
        • by LarsWestergren (9033) on Friday April 28, 2006 @04:25AM (#15219291) Homepage Journal
          Intel will have at least 5-6 months with the performance crown since AMD can't beat Conroe with their current K8 processors on s939 or AM2.

          This statement presumes that
          a) The Conroe will launch when promised.
          b) That it is a real launch with the product in stores, not just a paper launch.
          c) That the performance will be as great as promised.

          And all that remains to be seen, right?
    • by fm6 (162816) on Thursday April 27, 2006 @08:19PM (#15217376) Homepage Journal
      Or is it the classic "Avis: We Try Harder"?
      The "we try harder" thing wasn't just a slogan — they actually ran the company that way. A long time ago I read a book called Up the Organization, by Robert Townsend, the guy who ran the company then. It's full of stuff that makes you say, "I wish my boss were that smart," but none of what Townsend did then would fly in today's business world. For example, he once refused to let the Directors give him a modest raise, because he felt the performance of the company didn't justify it. Imagine any current CEO doing anything like that!
  • Chipsets??? (Score:2, Interesting)

    by WebWeasel2006 (947837)
    From TFA Otellini did offer one excuse for its poor performance in market share in the past year - a shortage of "chipsets." Having only bought AMD for over 5 years now I have never even tried to buy an Intel based motherboard, is this staement true has anyone had purchasing problems with Intel based M/boards?
    • Re:Chipsets??? (Score:4, Informative)

      by Sebastopol (189276) on Thursday April 27, 2006 @05:59PM (#15216346) Homepage
      Intel generally sells a chipset with every CPU. A holding in chipset shipping halts the majority of CPU billing. If anyone has been following this market, Intel hit a shipment problem at the end of the fourth quarter of 2005 and was not able to meet it's CPU/chipset shipments. Nothing about the quality, just poor planning at the chipset fab.
      • Re:Chipsets??? (Score:3, Interesting)

        by tomstdenis (446163)
        Yes and no.

        Every new chip has a chance of requiring a new chipset but usually the chipsets are backwards compatible amongst a line of processors. For instance, a 945 chipset will run a 775-pin Prescott originally destined for a 915 chipset. If you got a 945, 955 or 975 you can essentially run every 775-pin processor Intel makes. If you bought a 915 you're SOL. [e.g. myself]

        If they had a standard FSB (*cough* *cough* Hypertransport *cough*) they wouldn't have to tweak the damn thing with every new CPU.

        Not
  • by Jason1729 (561790) on Thursday April 27, 2006 @05:48PM (#15216249)
    They've been screwing over their customers for 15 years. With stuff like the spying serial number, tpa, etc, they've had an attitude of buy what we tell you or get lost. Not to mention price manipulation. They hold back each new iteration until prices slack off on the current product. AMD beat them to the 1 GHz punch because intel was holding back their own 1GHz chip to squeeze more profit. After AMD beat them, they released theirs 2 days later.

    Now that it's coming back to bite them on the ass, I think it's wonderful.
    • by Sebastopol (189276) on Thursday April 27, 2006 @05:51PM (#15216278) Homepage
      Pfff...

      With stuff like the spying serial number, tpa, etc,

      Which is why AMD implemented the exact same thing, right up to virtualization "secure" computing.

      intel was holding back their own 1GHz chip to squeeze more profit.

      Shame on a corporation for making a profit. AMD is so pure and virginal white, they'd never do something this dastardly.

      • by ScottLindner (954299) on Thursday April 27, 2006 @06:04PM (#15216381)
        Shame on a corporation for making a profit. AMD is so pure and virginal white, they'd never do something this dastardly.

        The OP didn't say there was anything wrong in making profit. The OP said it's wrong to use methods that are deemed illegal in our laws, and are typicaly not wise in a competitive market. Obviously Intel thought they still owned us all like M$ does... but AMD caught them with their pants down. That's the point of the OP. Intel had it coming to them for their arrogance. And they got exactly what they asked for.

        No where in those facts... is the OP indicating that making a profit is a bad thing.
        • The OP said it's wrong to use methods that are deemed illegal in our laws, and are typicaly not wise in a competitive market.

          The OP said no such thing. It was a rant about things Intel has done (considered evil) that AMD has mirrored (considered... I don't know. Acceptable?). Apparently you fall in the 99% of /.ers who see and hear exactly what they want, real world be damned.
          • I never called Intel evil or said what they did was evil. I guess you hear and see what you want, real world be damned.

            It is a very important distinction that Intel did it first even if AMD copied it. Every grocery store in my area has those loyalty discount cards, I'm pretty much forced to use them. I do use them without complaint, but I haven't set foot in the first store to introduce them in the 7 years or so since they did.
        • The OP implied intel deliberately held off the 1GHz to "squeeze" more profit.

          First, that is incorrect. Intel lost the race to 1GHz fair and square.

          Second, both AMD and Intel hold off releasing products and play all kinds of pricing games to squeeze their customers without killing them (because without a customer, you make zero dollars). Holding products in the wings to meet billing goals is not uncommon. To imply it is illegal, especially in this context, is wrong on both charges.

          However, given the inten
          • To imply it is illegal, especially in this context, is wrong on both charges.

            I never said it was illegal or wrong. However it is acting against the customer's interest, and that will make the customers less loyal. When that results in the customers jumping ship when a viable alternative shows up, all I can say is they're reaping what they sowed.
      • by Changa_MC (827317) on Thursday April 27, 2006 @06:10PM (#15216421) Homepage Journal

        Shame on a corporation for making a profit. AMD is so pure and virginal white, they'd never do something this dastardly.

        Any corporation causes damage to the market when they try to leverage their monopoly to maintain the monopoly. If AMD ever crushes all their competition (no that I'm worried) they will also slow innovation. Far better to have multiple competing companies, where you can buy whatever you want. Perhaps we could call it a "free market," since you can shop around.

    • by XanC (644172) on Thursday April 27, 2006 @05:59PM (#15216342)
      AMD's GHz chips were available. Intel "released" them two days later, but good luck actually finding one!

      AMD beat Intel in that race any way you look at it.

    • by Salandarin (895819) <salandarin.gmail@com> on Thursday April 27, 2006 @06:18PM (#15216472) Journal
      Oh PLEASE, give me a break!

      I hate reading this kind of generic accusation on Slashdot. For whatever reason, most /.'ers seem to be under the impression that somehow, Intel is inexorably related to an evil scheme by Microsoft/SomeCorporation, Inc. to consistently screw over the consumer. Honestly, I've never experienced any direct "screwing over" from Intel. If the chip performs up to the specification that it says it will, then there's nothing wrong here. Just because you don't like their market tactics doesn't justify this kind of arrogance. Considering they still have 80% of the market, it seems that whatever they're doing is working, contrary to your opinions.

      While I like AMD's desktop chips more, one must be fully aware of the fact that AMD is no less of a corporation than Intel is, and want your money just as badly as everyone else. As others have mentioned and will mention, AMD's hardware tactics are no less questionable than Intel's, and thus, it becomes a moot point.

      • For whatever reason, most /.'ers seem to be under the impression that somehow, Intel is inexorably related to an evil scheme by Microsoft/SomeCorporation, Inc. to consistently screw over the consumer.

        And artifically doubling the prices of their CPUs isn't?
      • Honestly, I've never experienced any direct "screwing over" from Intel. If the chip performs up to the specification that it says it will, then there's nothing wrong here.

        Remember the pentium bug? Intel was not going to replace the defective chips, they claimed incorrect division results a tiny fraction of a percent off would not affect most users so there's no need to replace them.

        They only replaced the chips after a huge public outcry and threats of class action lawsuits.

        Those chips were not up t
    • That's some creative history. There was a race to 1 GHz, the Intel release was a paper one. They then rushed out the 1.13 GHz, which they retired after Tom's hardware and some other sources claimed it to be unstable. It didn't return until the Tualatin die shrink. In short, Intel was hard pressed regarding the performance crown even back then. They had been all since the original Athlon at something like 600 MHz was released. They countered with the original Coppermine (the first PIII that made sense, L2 at
    • AMD beat them to the 1 GHz punch because intel was holding back their own 1GHz chip to squeeze more profit.

      Yeah, I remember those oh-so-stable 1Ghz Intel chips.

  • by Sebastopol (189276) on Thursday April 27, 2006 @05:48PM (#15216253) Homepage
    My god, the spin is breathtaking:

    80% market share != Falling Behind

    50% market share == Falling Behind
    • Agreed. Intel's GROWTH is falling behind AMD's growth. Intel still has a firm grip on the market. Intel still sold more chips then AMD. Intel is only falling behind their previous performance. Which as mentioned in TFA, is still within their historical variance.

      -Rick
    • by fm6 (162816) on Thursday April 27, 2006 @06:16PM (#15216462) Homepage Journal
      On the contrary, in today's crazy financial markets, 80% market share is falling behind, when your market share used to be 95%. Of course, the real reason they're talking about market share is they don't want to talk about profits. Which were $12 billion last year, and are $9 billion this year. That's still a lot of money, but not acceptable to Wall Street, which expects certain kinds of investments to grow consistently, no exceptions.

      That's why they broke up Knight-Ridder, an extremely successful newspaper chain with 20% annual profits. Huge profits or not, other newspaper chains were doing even better. Sorry folks we need to see 30% profits or you're not doing your jobs.

    • The loss in market share shows that their technology has fallen behind, i.e. their price/performance are not as good as AMD. (Either that or Intel's marketing has fallen behind that of AMD, which I don't believe for a moment).
  • What? (Score:4, Interesting)

    by nnnneedles (216864) on Thursday April 27, 2006 @05:50PM (#15216268)
    "Intel's market share recently slipped below 80%"

    Around eighty % is still incredible, not least when you have a competitor like AMD. But I guess companies like Intel do what they can to instill fear in their employees to get them to work harder.
    • Around eighty % is still incredible, not least when you have a competitor like AMD.

      It's not the number that's most significant, it's the trend. Intel once had nearly 100% of this market, but now they are (or seem to be) on a slide down without a strategy to stop it.

      Does anyone have a chart/graph/pretty picture showing the marketshare decreasing for Intel and increasing for AMD?
      • Isn't marketshare one of those semi-subjective terms? It requires someone to 1) clearly define what the specific market is (server chips? desktop? mobile? is there a bright red line you can draw between them?), 2) to accurately estimate the total size of the market, and finally, 3) get the company to accurately report their sales figures in that market.
        • 1) yes
          2) yes
          3) yes

          But let's not let facts interfere with the debate!

          • Boy, you sure did win that debate.

            Since I was looking for knowledge and not a debate, does anyone have any substantive information? Are marketshare figures often something of an inconsistent guess that wouldn't be very useful if graphed (particularly because they tend to be skewed by self-interested parties)? Or are they consistent enough (like the market price of a barrel of oil) that it would be informative if graphed?

        • Sorry, I don't have your hard numbers, but subjectively, I can say that I was one of those "genuine intel" people who would have been afraid to buy non-Intel chips back in the 386/486/pentium days, and now nearly all my friends run AMD athlon or AMD64 systems.

          If you go to a large computer parts mega-store like Fry's, you'll find almost equal motherboard / cpu / system floorspace for AMD and Intel systems.

          Think about that... if display space in Fry's is a measure of interest and systems moved, then Intel's c
      • It's not only the decrease, it's where some of it is happening. AMD is the new golden child of the server market. Not only are they a more popular choice for most applications, their multi-core processors caused an explosion in the virtualization field by driving down the cost of building the massive servers to make it worthwhile for mainstream production use.

        Before the Opteron Intel was the only processor supplier for x86 server applications in most companies eyes. The market belonged exclusively to
  • With intel increasingly becoming dominant in mobile markets, particularly capturing the next-gen Apple market share, I think that cost-cutting is eminent. AMD holds a firm grasp on the fastest chips, and some of their 64-bit chips are available in notebooks geared toward power-users wanting desktop replacements. There is a big difference nowadays between the size, heat output, power consumption, and power, so chip makers need to emphasize on certain markets. Intel couldn't hope to maintain dominance fore
  • In other news... (Score:3, Interesting)

    by Shadow Wrought (586631) * <shadow.wrought@g ... m minus language> on Thursday April 27, 2006 @05:51PM (#15216279) Homepage Journal
    Intel admits that water is wet and ponies are cute.

    Intel's marketshare is deceiving because it is propped up on a number of "exclusive" contracts. Once those go away, and they will as AMD pulls away technologically and pricewise, Intel is going to see the market flipped in a very short amount of time.

    Intel Outside, not just a good idea anymore.

    • Re:In other news... (Score:4, Informative)

      by Sebastopol (189276) on Thursday April 27, 2006 @05:53PM (#15216291) Homepage
      and they will as AMD pulls away technologically and pricewise

      Not likely. AMD slipped ONE YEAR on AM2, and has nothing on the horizion for 2 more years.

      Intel has three fabs ramping to 65 then 45 nm, and two years worth of products that handily defeat anything from AMD.

      • Re:In other news... (Score:5, Interesting)

        by RzUpAnmsCwrds (262647) on Thursday April 27, 2006 @10:55PM (#15218218)
        Intel has three fabs ramping to 65 then 45 nm

        Intel has always had a process technology advantage over AMD. That never stopped AMD from shipping competitive products. Also, note that AMD's fab situation has gotten a lot better in the last year - with Fab 36 (and soon Chartered), AMD has the capacity to take on Intel in the market - something that they just couldn't do in the early Athlon days.

        AMD has always been conservative in launching new processes, and it has benifited them in the past. Intel's 90nm process turned out to be the nail in the Prescott coffin, but AMD's 90nm launch resulted in CPUs that clocked much higher, used less power, and cost less money.

        nothing on the horizion for 2 more years
        K8L, for one. Dual-core Turions. 65nm in 1Q 2007. Quad-core in 2007.

        two years worth of products that handily defeat anything from AMD

        Ah, another Intel Conroe fanboy. While I'd agree that Conroe is looking quite good, note that Athlon 64 is not sitting still. Even a simple die shrink may allow AMD to put out 3.4-3.6GHz parts, which would be quite competitive with what we're currently seeing from Conroe.

        I would certainly hope that Conroe has a performance advantage over AMD64, though. No desktop or server part that Intel has put out in the last two years has been competitive from a performance standpoint with Athlon 64. The dual-core Xeon parts are a joke (and everyone in the industry knows it), the Pentium D gobbles down power and can't match Athlon 64's performance at half the wattage, and even Intel's low-end Celeron D is killed by the cheaper Sempron.

        It's only rarely about performance anymore. Most PCs sold do absolutely nothing 95% of the time. It's about power usage, availability, the strength of the chipsets and the price of the chipset and CPU.

        AMD chipsets are cheaper than Intel chipsets. Semprons are cheaper than Celeron Ds. Unless that changes, AMD is going to continue to destroy Intel's marketshare in the low-end and mid-range PC business. Only Dell is keeping Intel alive in the low-end market now.

        Take a look in any retail store. You see more AMD than Intel. That has never been the case before - AMD has never had this kind of shelf space. They've never had this much fab capacity. They've never had this much acceptance in the corporate world.

        That alone should have Intel very, very worried.
  • by moore.dustin (942289) on Thursday April 27, 2006 @05:53PM (#15216296) Homepage
    Intel may be losing market share right now, but as the article mentions, Intel plans on changing that and I think they have the power to do so. Intel still has the best name recognition on top of a steady advertising campaign over the last several years. They certainly have the power to take back what was lost and continue advancing in other markets.

    AMD on the other hand is showing steady and strong growth. It has solidified its place in the market and has shown that it not only can compete, but will continue to innovate and compete with Intel.

    In the end, something like this is only good for the consumers. Intel admitting losses to AMD will raise the bar of both companies so they can continue to compete, bring us better technology, faster.

  • by staryc (852301) <melissavoegeliMN ... m minus math_god> on Thursday April 27, 2006 @05:56PM (#15216321)
    Step one: Admitting your problem.
  • by XanC (644172) on Thursday April 27, 2006 @05:57PM (#15216328)
    Thanks to Intel, Bumfuckgrad has expandability built in!
  • I think Intel may not have much to worry about, with the recent headway that their Israeli Labs have been making with the development of the Pentium M, Centrino, and now the Core Duo/Solo chips.

    It's no wonder why there was some market loss when all the could produce out of the US Labs were overclocked P3/P4 chips that produced too much heat to be useful in mobile applications.
    • Hey, Banias/Dothan ARE overclocked P6 chips. The changes have been gradual, but I think there is no doubt about that. They've kept the number of execution units and many other characteristics. The cache and branch predictor have indeed changed, but it did so many times before in P6 (even within the same moniker [Pentium !!!], like Katmai to Coppermine).
  • Intel Internal Memo (Score:5, Informative)

    by Anonymous Coward on Thursday April 27, 2006 @05:58PM (#15216334)
    Here is an Intel internal memo from the CIO's office. Intel is going to be firing (aka "redeploying") nearly 1200 people.

    April 26, 2006

    IT's first steps in 'relentless pursuit of efficiency'
    Understand IT's response to Otellini direction

    As Paul Otellini outlined in his April 25 Exec Connect Webcast, Intel faces significant challenges in Q2 and beyond as we adjust to meet our new business climate. With 2006 revenue now projected to be approximately 3 percent below 2005 and spending reduced by approximately $1 billion, Paul has given clear direction that Intel must "re-structure, re-size, and re-purpose" in a "relentless pursuit of efficiency." As a result, all organizations, including Intel Information Technology (IT), have been given new spending and headcount targets and a corporate effort has been launched to determine how to streamline Intel to make our company more agile and more competitive long term.

    IT's response

    IT's response is clear. We must rapidly adjust our programs and headcount to meet our new Plan of Record (POR) spending and headcount targets. And we must streamline our organization to maximize the services and programs we can deliver to the company while executing on our commitments to keep Intel running and to deliver on new capabilities.

      Our immediate first steps are to stop selected programs completely, scale back other programs, and remove unnecessary overhead across the organization. The identification of these programs will be completed in roughly 30 days.

    To create a more efficient IT, we must develop a more compact, less hierarchical organizational structure, reduce process and operational overhead, and selectively reduce jobs in some skill areas. Identifying specific actions in this area will take 60 to 90 days to complete.

    Redeployment will be necessary

    Taking these steps will require the redeployment of some IT personnel. Redeployment, or the effort to move employees to areas of greater return when there is a change in business conditions, is a standard practice that allows Intel to remain competitive and increases opportunities for employees. We anticipate some redeployment will begin before the end of this quarter and all redeployment will be started by the end of Q3.

    I know this is a difficult time and redeployments are not always easy. And I know these actions are absolutely necessary to improve our competitiveness and effectiveness as an organization. You will hear more at the coming business update meetings (BUMs). I urge everyone to attend.

    If you were unable to watch Paul's live Webcast, replays are available at the Webcast Exec Connect Events Site.

    Thank you,

    XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
    • All those kisses at the bottom - that's a nice touch. Seems like quite a close and intimate company to me - not like the evil mega-corp people portray it to be...
    • I mean what the hell is this?

      "We anticipate some redeployment will begin before the end of this quarter and all redeployment will be started by the end of Q3."

      Why not just say it like a man?

      "We will be redeploying many of you from your Aeron to your SOFA just in time for you to file for unemployment insurance before taking back the Christmas presents you were hoping to buy for your parents, so that your car is not repo'd as you struggle like mad to pay your rent or mortgage in the wonderful Bay Area."

    • Paul has given clear direction that Intel must "re-structure, re-size, and re-purpose" in a "relentless pursuit of efficiency." As a result, all organizations, including Intel Information Technology (IT), have been given new spending and headcount targets and a corporate effort has been launched to determine how to streamline Intel to make our company more agile and more competitive long term.

      Translation: Paul has no idea how to fix the problem but he needs to make the board think that he's actually doin
  • Okay.. (Score:3, Interesting)

    by Anonymous Coward on Thursday April 27, 2006 @05:58PM (#15216338)

    Your lips keep moving, but I can't hear what you say. [anandtech.com]

    Intel has such a long way to go at the highest end. They need to move away from their silly, old fashioned CPU <-> Northbridge <-> RAM architecture. I think it's telling that Otellini blamed "chipset" shortages for some of their market share loss, whatever the hell that really means. Intel is going to eventually have to sacrifice its chipset business to stay competitive. Nothing will change that. The memory controller has to be moved on-die. HyperTransport is here to stay and it will wipe the floor at the high end.

    It's not just getting rid of NetBurst-- high IPC is great --but the more you have cores and sockets contending for memory access, the worse it will get for a shared FSB. Get your head out your butt Intel and fix the design.
  • Do you see what a little competition can do for a market? Everybody expected AMD to die years ago until they got to 1Ghz before Intel and now they are around the 20% mark for marketshare. Intel has responded by bringing out a load of new technologies and have IMO been dropping prices on products far more regularly.
  • by Salandarin (895819)
    Is it possible this is a PR stunt to aid Intel's image in light of the antitrust trial? If Intel admits that it has lost significant market share, it makes AMD look that much less helpless.
  • CEO Otellini admitted Intel has fallen behind AMD with lost market share, technological leadership, and recently profitability.

    And how does this admission make Apple feel? They've hung their star on Intel being the best out there. Just a wild guess, but Steve Jobs is probably not commenting on this revelation right now.

    • by nonlnear (893672) on Thursday April 27, 2006 @06:44PM (#15216649)
      But what's the problem? Now that apple has transitioned to x86, they could introduce an AMD lineup any time they want to without any technical problems at all. Apple might lose a discounted rate on Intel chips*, but their customers historically are willing to pay a premium for their product anyways.

      * I have no idea what their pricing agreement with Intel is, so this may or may not be the case.

    • I actually think they've hunge their star on intel having the best mobile procesors, and the best mobile prosessor roadmap. AMD is creaming Intel (relatively speaking) in the server markets, not on laptops (or small form factor shiny apples).
  • There is a good possibility that some other player may come out of nowhere and pass up both Intel and AMD. The reason is that current processors are based on a computing model that has not changed drastically since the days of Charles Babbage and Lady Lovelace. They are all optiminsed for the algorithm. Problem is that basing software on the alogorithm is the primary reason for the current software reliability crisis. Switch to a signal-based, synchronous software model and the problem will disappear. Will
    • That's a damned good point.

      However, there are several key obstacles to making this happen.

      a) fab capacity and validation infrastructure
      b) component suppliers
      c) chipset support
      d) software support

      If you make it x86 compatible and it can slap into an AMD or Intel socket, you can drop c and d. If you don't do anything to extreme with die size or requirements, you can drop b.

      The hard part is the billions of dollars in fab capacity required to be profitable. ASPs are very thin, and cost to build a fab is very bi
    • Before you mod this guy up, read his contribution history and the links in his sig and under his name. This guy (Louis Savain) is a crank. As well as being a physics crank who seems to think that Stephen Hawking, Richard Feynman, Kurt Gödel, and Albert Einstein are "crackpots" [rebelscience.org], he's apparently also a computer science crank.
    • Switch to a signal-based, synchronous software model and the problem will disappear.

      Yes, that's likely to happen. Two words: Chicken, egg.
    • Dude, do you make _any_ posts not about your new software model? I just took a look at your posting history, and it all seems to be about that.

      If you keep this up, you'll just get marked as a troll, and it'll be harder for you to get your message out.


      How about writing a demo app instead? ;)

  • This is no surprise. Intel dedicated its' best people to the IA64 (Itamium) Project, and so lost about 6 years of x86 CPU microarchitecture development. Had that effort succeeded, Intel would be well in the lead. But neither the complier nor the clock made it.

    So it didn't, and this was actually somewhat predictable. Intel has tried these flyers about every 10 years: first, IAPX432. next i860/960. Now IA64. This was the most costly by far. I'm a little surprised they admit it, but possibly that's b

  • by this great guy (922511) on Thursday April 27, 2006 @07:07PM (#15216813)

    Wow, just wow ! Did you guys see that in the article:

    <<
    According to some industry experts, Google is now assembling so many of its own servers that it may be the third or fourth-largest server maker in the world.
    >>

    I think that a lot of companies could reduce their expenses by doing the same thing than Google: instead of buying expensive hardware, warranties and support from IBM/HP/Dell/Sun, they could hire people to design, build and maintain their own IT infrastructure. I think it makes sense for any shop with 1000+ machines. Think about it again:

    • A 24x7 support contract from BigITCompany is good, but a team of your own technicians already working for you on-site is even better.
    • BigITCompany doesn't sell AMD server ? Your own team of technicians can build any server customized to your own needs and won't try to sell you unnecessary parts/services, since it is in their interest to save you money (since they work for the same company than you :P).
    • Tired of waiting for hours on the phone with a BigITCompany support guy to replace a stupid broken fan on a desktop machine ? Your own team of support technicians will never make you wait.
    • ...

    To any non-believer: Google does exactly this, and it works very well for them. So why not starting to do it at your company ?

    • Building servers would require enough people to build said servers.

      Where I work, they barely have one pc technician to 500 people. And we support the entire infrastructure; not just the PC sitting on the desk.

      Plus, corporate America likes to be able to place blame. As long as it's not within the company. Aw hell, who am I kidding, most of the fingerpointing that goes on happens behind closed doors....

      I need to start my own company. So I can fingerpoint!!!
  • by Sebastopol (189276) on Thursday April 27, 2006 @08:00PM (#15217220) Homepage
    Please, no more myths about Intel power hungry designs. Core Duo fixes this for good:

    AMD Athlon64
    Freq: 2.0 GHz
    Tcase: 70 degC
    TDP: 89 W

    http://www.amd.com/us-en/assets/content_type/white _papers_and_tech_docs/30430.pdf [amd.com]

    Intel Core Duo Destkop (Yonah)
    Freq: 2.16 GHz
    Tcase: 100 degC
    TDP: 31 W (scale to 70deC like AMD puts you at about 20W).

    http://download.intel.com/design/mobile/datashts/3 0922102.pdf [intel.com]

    • I agree, but to be fair:
      1. Yonah is a 32-bit only chip. Driving more wires/pins/etc in 64-bit mode internally and externally burns more power. I doubt that the K8 core shuts off the upper 32 bits of various logic/flip flop/RAM/CAM structures while in 32-bit mode...if anyone has real information on this, that would be appreciated, but that's probably not public knowledge.
      2. Yonah is built on a 65nm process. Transistor sizing from 90-->65nm gains you significant power reductions and performance increases.
    • And.. AMD Turion on same PDF has Thermal Design Power of 24W @1800MHz... why are oyu comparing desktop CPUs with mobile ones?
  • Does anybody else think perhaps a motive behind Intel's admission to this is to give themselves a more credible look in the antitrust allegations that are against them?
  • he also suggested that Intel's recent market share losses... were in line with historical variations which tracked to Intel's product generations."

    corporate-speak translation- "Yeah we always lose money when we make shitty product and bad decisions, this ain't new."
  • by mcrbids (148650) on Friday April 28, 2006 @05:40AM (#15219462) Journal
    I am a system administrator, and CTO.

    I manage servers, with tens of thousands of users, all over California.

    ALL of the servers I've deployed in the last 2 years have been AMD, with a heavy bias towards Opterons. For me, one of the key advantages AMD has over Intel is the "last resort" advantage.

    If I deploy Xeon servers, and something goes terribly wrong, I can't go to a local retail outlet and buy any hardware that would work - Xeons are not binary compatible with X86. Local tech shops here in my hometown (Chico, CA) don't have Xeon anything. But they DO have Athlon/64s in droves!

    So, if I deploy an Opteron server, I *know* that I can get an Athlon/64 that's binary compatible with my system images from the local l337 Gam3rz computer store with aliens and funky lights, but that's binary compatible with my rackmounted servers. No matter what, I have something I can count on in less than a single working day. I've had to fall back to this in the past, so I'm ready to in the future.

    This gives me a worst-case recovery time of about 4 hours during business hours. (the only ones that really, really count)

    Xeon is compatible with... Xeon. At best, in a worst-case scenario, I wait 48 hours to get some kind of support in small town, CA. Ouch!!!! No way this is acceptable.

    4 hours vs 48 hours. Not a hard decision... So Opteron/Athlon/64 it is, then...

    And I don't mind that it's both faster and thousands $$$$ cheaper!

The universe seems neither benign nor hostile, merely indifferent. -- Sagan

Working...