Sign in with
Sign up | Sign in

Opinion: Did We Expect Too Much of AMD Bulldozer?

By - Source: Tom's Hardware US | B 205 comments

After reading Chris Angelini's thorough review of AMD's first Bulldozer CPU, the FX-8150, I have to admit that I was somewhat confused. Didn't AMD claim a few weeks ago that the FX would be the world's fastest CPU?

Not that I was first in line to buy this processor; and not that I cried my eyes out over the test results. Chris' conclusion was that, for the price, he would rather choose Intel's i5-2500K than the FX-8150. I expected the FX to do better and it appears that most other potential buyers of a high-performance chip did as well. The most positive review conclusion I read yesterday was that FX-8150 is a "mixed bag." In PR and marketing, that is not what you are aiming. AMD, how did that happen? Why did we expect so much and why did we get - my apologies - so little?

I am not going to chime in on the hardware architecture. It is what it is and if you want to debate whether this is a true 8-core chip or not, and why it occasionally outperforms Intel's chip and loses in other benchmarks, Chris has all the information you need. What I don't quite understand is why AMD laid a foundation that would go on to lift the expectation layer and why the company actively participated in a scenario that moved performance expectations into fantasy territory, instead of correcting the public expectation layer.

The problem AMD is facing now is the fact that the FX-8150 was the launch product for Bulldozer, which was originally announced more than four years ago. By underwhelming reviewers and, ultimately, its loyal customer base, the entire Bulldozer image has been damaged and it will take a lot of effort to recover. Imagine this had been a quarterly earnings report and how the financial community would chastise AMD for yesterday's results.

I have no idea who made the decisions to position the PR and marketing for the FX series, but it is marketing that has to accept the blame. On September 13, AMD circulated images of the highest-clocked CPU and its Guinness World Record achievement with a 8.429 GHz CPU. You could argue that this resulted since the AMD CPU does not have a cold bug and allowed a ridiculously low temperature via liquid helium , but it was the marketing trend that built up the expectations for an insanely fast CPU.

The icing on the cake was a press release issued the very same day. It was entitled "AMD Showcases World's Fastest CPU". Who would have thought, based on that statement, that the FX-8150 would end up just on the upper end of Intel's mainstream product line? If you ignored all reporting and simply went by what AMD said, I would have expected a CPU that blew Intel's higher end CPUs such as the Core i7-980X out of the water.

I was told that there are different ways how to interpret "world's fastest CPU" and that expression may have only referred to the clock speed. I don't buy that, by the way, because it was AMD that explained to me in 2000 that clock speed and performance cannot be directly compared. They even put me in a Chevy Corvette Z06 (which represented the Pentium 4 at the time) and a Dodge Viper (which represented the Athlon) on a racetrack to provide an explanation for their reasoning. So, why would we suddenly describe performance with clock speed again?

Of course, with the FX-8150, there was also the case of a press preview event prior to launch where hardware reviewers had the first opportunity to see the finalized CPU in action. However, at that particular event, only multi-threaded applications were allowed and reviewers did not have an opportunity to look at the single-threaded performance. In such a case, you already know that something is up, but you scratch your head over the communication strategy, because it was obvious that there may be problem with single-threaded performance. Reviewers would come back and bite AMD with their reviews on launch day.

Why would you hold back that information and risk that the expectations in the product will not be met? The entire launch of the FX-8150 is close to a trainwreck and caused some discussions here what the reasons may have been. Did we expect too much? And if we did, why? Did the community build up unreasonable expectations or did AMD? I feel, and it seems that there is a certain consensus, that AMD at least participated in this trend and pumped expectations with the ultra-long development time, hand-picked benchmarks and especially its Guinness World Record one month ago. At the very least, AMD is guilty of not correcting expectations that had spun out of control. It could have been much clearer about the positioning of the CPU.

The launch somewhat reminded me of the introduction of Intel's first Core 2 (65 nm Conroe) processor back in 2006, the processor that brought Intel's desktop CPUs back from the dead and went on to nearly kill AMD. In our initial briefings, I remember Intel as being extremely conservative with its performance estimates. In the end, Conroe arrived with a performance boost that was 80 percent higher than the company initially had promised. It was Intel's conservative communication that caught AMD completely on the wrong foot. Just three months before the launch of Core 2, AMD told me in interviews just how confident they were that the Athlon X2 would keep or immediately regain the performance crown from Intel, should Intel in fact capture it. I was not sure if AMD already knew how strong Conroe was and their answers were marketing bubbles, or if they actually fell into Intel's trap.

On some occasions in the past, AMD had a tendency to overstate processor performances and the outcome has never been favorable for the company. It just doesn't look good if you promote an unrealistic expectation, as the reviews will come in at some point a reveal the actual picture. The unfortunate outcome with the FX-8150 is that, for some users, this may be the best processor for the money they can buy. However, those users may now be in doubt and rather consider a competing Intel processor as the better and safe buy. A much more modest marketing approach with a clear product positioning would have helped AMD to launch Bulldozer in a more positive light.

Granted, Intel and AMD are not making their money with high-end processors that cost $500 or more. The money is made with volume processors. At $245, the FX-8150 is already an expensive volume processor, but it is well within reach for those who are considering to spend at least somewhere between $800 and $1000 on a complete system. However, to reach those users, AMD's marketing will have to rely on its Guinness World Record overclocking event and its ability to convey that eight cores are better than four. I am not sure if that will give AMD enough firepower against Intel.

So, what do you think? Did I expect too much from AMD? Or did AMD marketing go too far?

Display 205 Comments.
This thread is closed for comments
Top Comments
  • 41 Hide
    Lyden , October 13, 2011 8:14 PM
    No. AMD damaged themselves. Nuff said.
  • 28 Hide
    Inferno1217 , October 13, 2011 8:13 PM
    No we didn't expect too much from AMD. This might explain the change in CEO's as they knew Bulldozer was not going to be competitive. In some cases the Phenom II series out performs Bulldozer. I hope the rebound with 2nd gen of this cpu. Looking at benchmarks I also wonder if the 990FX chipset is utilizing Bulldozer's full potential although I may be wrong.
Other Comments
  • 28 Hide
    Inferno1217 , October 13, 2011 8:13 PM
    No we didn't expect too much from AMD. This might explain the change in CEO's as they knew Bulldozer was not going to be competitive. In some cases the Phenom II series out performs Bulldozer. I hope the rebound with 2nd gen of this cpu. Looking at benchmarks I also wonder if the 990FX chipset is utilizing Bulldozer's full potential although I may be wrong.
  • 41 Hide
    Lyden , October 13, 2011 8:14 PM
    No. AMD damaged themselves. Nuff said.
  • -3 Hide
    billybobser , October 13, 2011 8:15 PM
    the flagship is horrendous at the moment (it may be more applicable on multithreaded real world scenario's in the future.)

    But windows 7 cannot even address the cores properly, it's like reverse hyper threading.

    Being crap and future proof in electronics don't mix.
  • 20 Hide
    batman4u , October 13, 2011 8:15 PM
    i can explain you why AMD looks like another company from which we fell in love, ever since the stupid Hector ruiz crusade became and did all that stupid things like buying ATI and firing everyone within and focusing on the Fusion technology and not bothering to keep ahead on the CPU market , like they say in my language... the bigger the grasp the weaker you hold (spanish)

    AMD was not up for a ATI buyout and they bought ATI cause Nvidia broke their relationship of joining
  • 18 Hide
    beenthere , October 13, 2011 8:16 PM
    As people are starting to understand today Zambezi is a perfectly fine CPU for most people and most applications. It's current optimization is for heavy work loads where it does well. It's weal point is single threaded worlk loads. It's fine in games and most other apps. It is not the fastest CPU for those apps but it works just fine for most everything and is a building block for the next piledriver cores.

    Did some folks expect too much? Absolutely. Does Zamabezi OC better than Deneb or Thurban? Yes it most definitely does. Does anyone think that the FX-8150 is the fastest Zambezi that will be released? Not likely. Is Zambezi a decent performing/value CPU for the majority of consumers? Most definitely.

    While we would all have liked more from Zambezi, what AMD delivered is still a significant step forward on many fronts - especially for heavy processing loads, i.e server use - where Opteron 6200/4200's are selling VERY well. Let's not forget that AMD also has great demand for Llano APUs and trinity is not too far off, so consumers most definitely have more CPU/APU choices now than ever.
  • 25 Hide
    cryptz , October 13, 2011 8:16 PM
    i dont think you expected to much. the previous generation of amd cpu's beat the bulldozer in many of the benchmarks. that is illogical and unacceptable. i wouldnt fault marketing though, no one markets their chips with the slogan: "this is going to suck" im sure any attempt marketing made to be realistic was overturned by management.
  • 18 Hide
    wintermint , October 13, 2011 8:19 PM
    To be honest, I had high hopes for Bulldozer. However, I know that AMD was and still far behind Intel. Because of that, I was hoping for Bulldozer challenge Intel Nehalem but at a competitive price point. It's kind of near that..
  • 1 Hide
    Mastervivi10 , October 13, 2011 8:19 PM
    AMD marketing go too far, Simple and Efficient.
  • 16 Hide
    torque79 , October 13, 2011 8:20 PM
    The internet is loaded with conversations right now on if there's some bug or a bad combination of particular motherboards/video cards, but more and more of that seems to be getting eliminated, showing consistent disappointment for the AMD product. I have been a proud AMD fanboy for a long time, but after waiting for so long for a (delayed) disappointing product, I've already jumped ship and ordered my i5-2500k build. It's a sad day when AMD can't even provide what has always been their strong point, high performance/$.

    I DO think software is holding it back, that there's not enough multithreading in use. This will probably change in the next couple of years, but I want to play games out there right now unfortunately.
  • 13 Hide
    krinkles619 , October 13, 2011 8:23 PM
    a bit of both, really. i think amd marketing really pushed the idea that bulldozer was going to dominate sandybridge in anything that required good virtual processing (i.e. graphics). and i think that, while it may have been partly amd's fault, we the userbase thought that this totally new processor and processor design would be able to at the very least compete with sandybridge, if not beat it in gaming at the least.
  • 5 Hide
    Anonymous , October 13, 2011 8:23 PM
    I think the marketing did give us high expectation, however I feel that especially if you check out the reviews if you compare things to the previous generation of the Phenom II a lot of us expected things to be somewhat better considering where the X6 and X4 could reach. If in some of the test it hadnt appeared that the previous architecture could nearly keep you there we may not have been so upset. This is where I feel a lot of expectation came from. Who knows, maybe the next generation will be better and give us the performance per dollar and better like the Phenom II chips did. I'm still a fan of AMD but getting an Intel does seem to look like a better purchase for a high performance cpu.
  • -3 Hide
    Anonymous , October 13, 2011 8:25 PM

    this tells a different picture.. and i would agree, why not pair an amd proccesor with an amd top of the line gfx card...
  • 10 Hide
    batman4u , October 13, 2011 8:26 PM
    i can go along with this architecture as you most say that its "enough" but Enough doesnt go for 250 bucks, come on!! 130 maybe
  • 5 Hide
    Anonymous , October 13, 2011 8:28 PM
    I have been waiting eagerly and been disappointed. AMD should have done something to temper the claims circulating the web. I would not have minded if expectations had been set at the start, but after so much build up I doubt there are many fans happy with the outcome (other than Intel ones).
  • 4 Hide
    Anonymous , October 13, 2011 8:31 PM
    I think AMD and the community share the blame equally. AMD's engineers had released enough details in architectural talks to hint that the per-thread performance would not be much better than Phenom II. These details were subtle, but they were there. AMD's marketing and the community's desire for an Intel-killer got the best of us all. AMD should have been stressing that this is a server processor which may not improve desktop performance very much. With everyone hoping/expecting it to compete with the i7 2600K, there's no doubt why all the reviews are so unfavorable. When looking at the reviews, the FX-8150 isn't great, but it is decent. It should probably be a little cheaper than the i5 2500K, not more expensive.
  • 9 Hide
    stm1185 , October 13, 2011 8:32 PM
    Is Zambezi a decent performing/value CPU for the majority of consumers? Most definitely

    When you look at the performance in applications the majority of consumers would use its lagging quite a bit behind the Core i5 and i7 chips, even last generation. iTunes, Excel, Photoshop all vastly favor Intel architecture. The most popular PC games like WoW or Star Craft 2 also perform far greater with the Intel architecture.

    Its only in a few situations where the FX chips pulls ahead, and none of them are what a mainstream user would likely be doing.
  • 22 Hide
    geekapproved , October 13, 2011 8:35 PM
    The problem is it's not really a 8core and should never have been marketed as one. Really it's more like a 4 core with hyperthrading. Windows and other apps don't recognize it's cores and jumble the threads out of order which is why it performs slow in bencharks that rely on "in order" threading. It seems more of a software issue than a cpu issue. Even Microsoft said Windows said it's not optimized for those kind of cores.
    I really don't get what AMD was thinking. I would rather they had just cancelled BD than release a dissapointing product.
  • 2 Hide
    spentshells , October 13, 2011 8:37 PM
    It was exactly what I said it would be, just below the first gen I7's
    Some other reviews showed it being a lot better than I thought,
    I have to see a review using all amd CPU and GPU's Im sure it will fair better as it did in reviews that did not use the release motherboard.
  • -7 Hide
    captaincharisma , October 13, 2011 8:37 PM
    LOL the fanbois expected too much from bulldozer. they were hyping it up so much it was like there own 2nd coming of Christ
Display more comments