% | $
Quotes you view appear here for quick access.

Advanced Micro Devices, Inc. Message Board

  • longnetflix longnetflix Dec 4, 2007 2:55 PM Flag

    BREAKING: K10 has a power/heat problem!!

    Here's the scoop. AMD has been using Fake-o-watts called "ACP" to quote power recently, comparing it to Intel's maximum power ratings (TDP). Why? Because AMD is underhanded, of course.

    But within their own power documents, two versions of the same document reveal the K10 power usage is out of control.

    TDPs are soaring, and that's why AMD doesn't want to talk about them. TDPs are what matter when designing thermal solutions for systems, of course.

    In the OLD document, version A, AMD has 3 TDP families for the K10:
    68W, 95W and 120W. These are what folks have expected.

    In the NEW document, version B, AMD has had to increase these drastically:
    79W, 115W (!!) and 137W(!!)

    They quote the same invented Fake-o-watts ACP values of:
    55W, 75W, and 105W for both.

    Amusingly, the "tech doc guy" made an error in "B" and forgot to update the "max power" comments to reflect the newly increased TDP values! Guess he really needed that vacation.

    This explains why AMD is going to require a horrible 140W TDP for the 2.6GHz Phenoms, and suggests continuing severe 65nm process problems for AMD. Their parts are literally burning up as they approach 2.6GHz.

    They're done for.

    SortNewest  |  Oldest  |  Most Replied Expand all replies
    • Yeah, that's pretty much close to the truth, I guess... Just by reading some of my favourite websites lately (check this out: or significant "lack of reviews" of some others ( had absolute zero Phenom reviews and is very reluctant to give any detailed comment on the HD**** cards), I'd say that people have been increasingly frustrated with AMD. Customers as well...

    • AMD problem is much more serious.
      use your mind, what is INTEL's largest fab doing right now.
      it stops making products...... why? it is converting into 45nm. INTEL's 45nm product are only coming out of a small testing fab until early next year. the big fab is converting and making no procucts right now. After the conversion is done. 45nm products will be able to sold at all price points. for most low end product, AMD is selling, INTEL's 45nm core 2 duo celeron will perform twice as good at the same price.....all those product will be extra products, because they are procing nothing right now. this means INTEL can really play with the little mouse any way he likes
      run now when you still can.

    • The above post is why I feel comfortable about my long position in AMD. Oh, I hate the short term movement as much as the next guy. But longs don't sell out on an unexpected dip. Nothing has changed long term, which is why I continue to expect $30-$40 share prices eventually.

      But back to my point. The guy that wrote the post I'm referring to most likely knows there is no issues with AMD's TDP, or ACP if you would prefer. Despite this knowledge, he makes a dramatic effort to "find out some bad stuff" on AMD, and even gets as far as having his post duplicated by theinquirer.

      There is clearly a concerted effort to bash AMD into the ground. Hector Ruiz, in his defense continues to focus on the business, rather than on the bashing, or the shareprice for that matter. I expect to see a strong Q4 report, with continued progress on the earnings picture. AMD is continuing to ship it's volume parts, on a strong 65nm SOI process, satisfying it's OEM customers. AMD likely gained share once again, because there aren't enough gamers in the world to buy high priced C2D's to make a difference.


      • 3 Replies to p_l_u_s_c_a_r_d
      • "There is clearly a concerted effort to bash AMD into the ground."

        Yes Plus, we know, it's all a grand old conspiracy against AMD. It runs so deep that it includes AMD's own design, process development, and marketing folks all the way up to Dr. Ruiz. They are all bashing AMD into the ground.

        "But longs don't sell out on an unexpected dip."

        If you would read this board with more comprehension and less hate for anything intel you would see that this dip was not unexpected. We have been predicting it for months.

        "Hector Ruiz, in his defense continues to focus on the business, rather than on the bashing, or the shareprice for that matter."

        Hector spends most of his time publicly bashing intel rather than leading his company to produce better products. What AMD needs is more leadership and less focus on external factors out of their control.

        "I expect to see a strong Q4 report, with continued progress on the earnings picture. AMD is continuing to ship it's volume parts, on a strong 65nm SOI process, satisfying it's OEM customers. AMD likely gained share once again, because there aren't enough gamers in the world to buy high priced C2D's to make a difference."

        You have such a knack for stating the opposite of reality that it is almost uncanny. If this is true then explain the drop in stock price... Oh, that's right... We get back to your conspiracy theory.

        It is truly unbelievable that you believe a word that you post.

      • By screwing up on the Barcelona and then Phenom launches, there is absolutely no way AMD will have a strong Q4 earning.

        Why do you think they got the cash infusion recently? AMD clearly knew they will bleed capital and need more to keep the business running.

        Now if AMD can finally deliver the "good" Barcelona processors in Q1/08 and in high volume, you can bet their Q2/08 earning will be strong.

      • universe?

        AMD's K10 has turned out to be a disaster, and their stock price, and dismal future prospects reflect that.

        They promised an Intel-killer, and have instead delivered a delayed, overheated, under-performing, poor-yielding, bug-ridden white elephant.

        Many of us saw this coming over a year ago. The fact that you didn't, and rode your stock down to $9 does not mean that the stock is going back to $30-$40 just because you want it to.

    • First off, Intels TDP is not the max power consumption. Think of it more as a design number, kinda like a design temp for a locality when figuring for AC.

      AMD used to use their max draw, but they decided to go with the flow and change the numbers to match Intel's standard.

      No news here, move on...


    • INTEL new 52 week high, AMD 4 year low.

      This company is done, BK shortly.

      • 2 Replies to ostk.ptrnr
      • Purely for Amusement.
        The NY times article this morning refers to Barcelona as "Energy-Efficient". LOL.

      • yourbestfriend,

        nope if ATI continues to falter I believe Microsoft would ask Nvidia to produce a graphics solution for them. If I where Microsoft I would be doing that ASAP and not wait until they are shivering to think that Sony may scrap the entire PS3 using the unsupportable Cell and last generation Nvidia Graphics sans Unified shader and go with a more traditional CPU and a graphics processor from the G80 era instead of the crap they purchased from Nvidia two years ago that no really Graphics intensive games can be released on.

        Any time I read of a game that is multi ported and is released upon all consoles and PC at the same time I realize even though the pictures in the magazine's may make the game look great if it is released on the PS3 as well as the Xbox 360 I know the game is either a graphical lemon or the game graphics have been turned down for the handicapped PS3.

        If you believe AMD is having problems with its flops, the PS3 makes anything that AMD has done wrong seem like AMD did it right. It cost the media division chief and the CEO their jobs at Sony.

    • queer_lies_from_the_atyguy queer_lies_from_the_atyguy Dec 6, 2007 5:07 PM Flag

      Sound more and more like some of those ATI xbox360* quasi-engineers were involved here. BTW.....

      *Who will MSFT be going after for that 1 BILLION $$$ set aside to fix all of those overheating ATI graphics chip issues?

      AMD is going NOWHERE with ATI baby!!!! STRONG DUMP!

    • well, if you need AMD's documents to tell you that their 65nm processors have higher TDP than Intel's 45nm processors, you have lost the game already.

      TDP is always a function of process maturity. It is scaled linearly up to a point, then scales in a parabolic curve due to current leakage. That's why you see Intel release a bunch of 45nm Penryns in 80W up to 3Ghz, and 3.2Ghz goes all the way to 120W.

      In fact, when AMD prepared the Phenom line of processors, they were going to release the 9500, 9600 on the 95W TDP, and 9700 on 130W TDP. That should be the clue that 2.6Ghz 9900 would approach 150W TDP or more.

      Let's assume that AMD uses the same process on their X2s and barcelonas. That means you can estimate how much TDP barcelonas will take buy doubling the TDP of their dual core parts. Their dual core EE line has a TDP of 45W for 2.1Ghz X2, standard line has TPD of 65W. Doubling that gives you the current picture of between 90W-130W. It is probably more for barcelona due to the added transistors for the 2MB of L3 cache and added TLB logic and virtualization features.

      So yes, they are going to suffer until they can produce the barcelonas on 45nm SOI, which is at least 9-12 months from now.

      Something funny about the ATYT acquisition: Hector Ruiz thought he could produce the ATI GPUs on his spare 65nm SOI capacity. The ATI HD3800 series are produced by TSMC for AMD. That was another clue that their 65nm process sucks.
      Imagine when fusion comes :) How can you fuse something you cannot partly produce in the first place? LOL

    • Its a good thing AMD took that watt meter down from times square, it would have been running backwards for over a year.

      (If you dont know what I am talking about, AMD put a large sign in Times Square that had ever-increasing numbers claiming it was the extra money spent on electricity by using Intel processors instead of AMD. AMD took it down before the Core 2 launch...LOL!)

      • 1 Reply to V1ad_The_Impaler
      • Yea its old new for anyone that has been reading my post here for a while that AMD had nothing above 2.3 HZ that they could put together to form a four core chip. I nailed It right on the nose if this article is true 2.6 GHz equals two pairs of cores each using 55 Watts which equals 110 watts out the gate. What AMD did not tell everyone and is the reason why they can not get 4 high clocked cores on one chip is that there is a 2o% to 30% multicore penalty to be paid when combining cores. Do you think Intel could not do the same as AMD put four cores on one chip if they wanted to? Intel realized with the multicore penalty their products also would be off the charts if frequency was a concern for the consuming public.

        So going back to our out of the gate 110 watts add 20% to that equal 133 watts minimum anything higher gives you that 140 watts that this rumor is suggesting.

        Anyway I would put our collection of bashers here on this board against any other site on the web for ridding the air of all the smoke AMD tries to blow up every ones asses

    • yourbestfriendintheworld yourbestfriendintheworld Dec 6, 2007 4:04 PM Flag

      This thread just made the Inquirer:

      You're famous, LNF!

    • AMD has long since lost the ability to be honest with the public on anything. Add this one to the list.

7.50-0.10(-1.33%)Oct 25 4:00 PMEDT