I was reading this letter on The Inquirer (before you jump for your guns - I'm going to slay their opinion shortly) about how the 4x4 is the PC equivalent of a 400 horse powered Merc and not a Toyota and I had a violent convulsion of disagreement resulting in this post:
http://www.theinquirer.net/default.aspx?article=36319
(Pls read the article before reading the rest of the post)
Unfortunately, the dynamics and business models of the car business and the microprocessor business are two different things...at least this point in time (I'll talk more about how they could be converging in a later post).
1. The car industry is hyper-segmented with multiple models serving niche markets. A couple of hundred thousand units make a product/brand a viable proposition due to how their design and manufacturing allows them to optimise across. Further more, cosumer decision making works across multiple axis from speed, mileage to all the bells & whistles inside the car which allow easy brand/product differentiation.
The microprocessor business however has two primary vectors of comparision. Performance and now power consumption as a recent phenomena. Hence, everyone is expecting more performance with all the benefits of reduced power consumption (lower electricity bills, less heat, smaller form factors, etc).
2. In 2005 the Mercedes car group sold ~1 million units. Of which their top of the line models contributed only 72k units. In the microprocessor business the same does not hold true. At least for now. If AMD were to operate at the same scale as Mercedes and sell only 72k of it's top of the line product then it would be meeting with it's debtors pretty soon.
Now, I know I'm pushing it a bit using these two as a comparision but it's to make a point. The business models and scale of both these industries is vastly different. Can you imagine someone walking into Hector's office and telling him they will sell only 72k units of their top of the line platform.
3. The Merc is what it is because it is not only powerful, but also a thing of beauty. Unfortunately, the 4x4 is a loud, ugly beast. Which is why I think it's more a Hummer if we have to draw a parallel. Allows you to drive everywhere - on and off road but is expensive, pig ugly and guzzles gas faster than an 8 year old can eat candy.
The only bright spot for AMD if they follow GM's model is at least the sales of Hummer's seem to be increasing YoY. They've sold 65k of them so far in 2006.
Subscribe to:
Post Comments (Atom)
16 comments:
Follow GM's model (didn't know GM had a model), you must be joking. GM is just a healthcare provider that happens to make cars. Nobody should be following their lead..er meander..er..ahh failure that's the one I was looking for.
Had to come back quickly to say that I hope your not actually comparing unit sales of $100,000 autos to top of the line cpu's that cost, at most, $1,200. You must be able to see the absurdity in doing something like that. That's why I trust you're just using it as an example.
Anonymous 1 - that's pretty funny dude. Wish I'd thought of it.
Anonymous 2 - yes, you're right. I'm using the analogy and not comparing actual units to units across. Basically, the consumer decision making process to buy a 1200$ CPU when the mainstream price is around 150$ is similar to the decision making for someone to buy a 100k Hummer when the mainstream price of a car is 20K USD. What this means is both the car & microprocessor markets are triangular shaped. With the lowest cost products forming the largest volume at the base and the most expnsive at the top forming the smallest volume.
Dude... at the risk of offending you, you sort of jumped the shark like Sharikou does your current post.
While I agreed with everything you said about the differences of the auto and microprocessor industries, I feel you missed the point of what the author was saying. (I can't believe I'm stooping to defend something written on The INQ.)
The point he was making wasn't about sales or advertising nor was it really about performance. If you re-read the article he alluded to how he felt about his own experience and his opinion. He acknowledged the downsides of the 4x4 and felt the extra cost from power consumption was worth the experience of pride in ownership... for him. I won't and I suspect a lot of others will not share his same experience, but I think you missed the mark you typically strive for with this latest post.
Sheepshagger - that's cool man. Honest feedback is appreciated...I don't feel offended by it. Perhaps I should have made the point I was trying to more forcefully.
I agree Fuad was talking about the experience and the pride of ownership. My point was I would have related to his comment if he were driving a Hummer. To equate 4x4 to the same experience you get from driving a Merc is not apt. If you think about my points 1 & 3, they are basically talking about purchase behaviour and as you say pride of ownership.
However, where The Inquirer missed the point is the PC is no longer just about speed. It's about speed and power consumption. And power consumption is not about the freakin' electricity bill alone. What that power consumption means is smaller, cooler and sexier PCs. Which the 4x4 is not right now. It's not a Merc...it's more like a Hummer. As a result, while the Inquirer may deride all the folks who are sniping at the 4x4, they missed the most fundamental point which is it's not enough to be faster any longer.
Am I clearer about where I'm going with this now?
thanks,
I think all of you missed the point.
You cannot associate the QuadFX to anything other than a over spec'd family van. The mere presence of the Intel quadcore being capabable of doing more at half the power draw makes QFX completely useless and embarrasing to own. If anything, Kentsfield is the Hummer and I don't mind being seen having one. Owning QFX only shows your extent of fanboyism even in the face of unanimous benchmark beating. I certainly wouldn't want to be seen owning one. I would pity anyone owning one either.
Frankly, I think Inquirer is funded by AMD. They make it very obvious by keeping bad news abt Intel on their front page while the good news vanishes quickly. The opposite is true for AMD.
In other news seems like the 65nm shrink has enabled the new AMD's to be overclocked by a whole 200mhz more on stock air from 2.8GHz to 2.94GHZ. Well atleast it is consuming less power than the 90nm chips. Another great release in a long list of great releases for the consumer this year from amd.
I find it hilarious to suggest that the INQ is biased in AMD's favor. I am constantly correcting things that the INQ says about AMD which are typically incorrect against rather than for AMD. I've corrected articles written by Fuad.
To suggest that 4x4 is a big mistake because of power draw should make a useful prediction. Presumably, the people speaking against it must feel that 4x4 will fade away when AMD releases its own quad core for the desktop in Q3 07.
However, I think that Intel will have to follow suit and have its own dual socket quad core. And, an Intel dual socket quad core would be just as bad on power draw. I'm wondering if Intel does release a dual socket quad core desktop system if all these negative comments will switch to rationalizations about why Intel's power draw doesn't matter.
BTW, the common element in all of these derisive comments about 4x4 is the logical fallacy of hasty generalization. This is where you suggest that something in a specific case is invalid because it can't be applied to a general case. Simply put, 4x4 is not intended to be a common system but then FX, EE, and X never have been.
I was curious about whether anyone was having trouble logging in here. However, it seems like if someone was that they could post a comment about this anonymously. I was unable to login here and on sharikou's blog until I upgrade my blog to beta.
Yeah,
This freaking switch to google caused me so much pain!
Scientia - I think the consideration for AMD will be whether they should continue to support 4x4 as a mainstream desktop solution. Undoubtedly they could retain 4x4 and position it into server or workstations. At least, that's what I'm principally thinking about when I wrote this post. However, I beg to differ that AMD's marketing engine is not positioning this as a response to Intel being first to mainstreamish desktop quad core. They are definitely using this as a marketing tactic.
BTW - Intel did follow AMD's lead at and IDF announced a similar effort but targetted at servers. It was during Gelsinger's speech if I re-call correctly.
Well, the problem with Intel following 4x4 is board size. 4X4 is tight and just barely fits. A similar Intel board would not be able to fit because of the size of the Northbridge chip. This means that most of the ATX cases would be out and Intel would have to use an extended ATX case. I assume that there must be something to this or else AMD wouldn't have tried so hard to make this board so small.
Servers wouldn't be an issue. Dual socket K8, Prescott Xeon, and Woodcrest systems are available today. There doesn't seem to be any reason for any new plan. Intel's current weakness is in 4-way but they won't have a C2D 4-way chipset until late 2007.
I would say that board size may be the only thing preventing an Intel equivalent to 4X4. However, it still seems that Intel would need something equivalent by the time K8L FX is released to avoid losing the top spot.
However, it still seems that Intel would need something equivalent by the time K8L FX is released to avoid losing the top spot.
Intel doesn't need to.
By the time desktop K8L is out, the 4x4 platform will be dead. I can't think of anyone adopting this platform without being ridiculed.
Even if one 4x4 motherboard does survive extinction, it would probably perform slightly better that an intel Quad core only on some higly threaded benchmark while showing no gain at all to most. I think we've already seen how software isn't providing the scaling that theoretically should be there.
The twice the power draw would still be a problem in this ever energy concious market.
At some point once they know how well Barcelona performs against Penryn, AMD will have to make the decision whether to invest substantially more behind 4x4 or focus their resources behind Barcelona.
As I keep saying, they have to manage their cash carefully. In case you didn't see this, they will incur a 350 million $ transition cost for the ATI acquisition which Bob Rivet said would impact their 1H07 results. I will re-visit my forecast that AMD will swing to a loss in Q107 at some point because now that is looking more and more likely to me.
Roborat, while for you 4x4 makes no sense, even at the quad core level, I've explained before that I could make it worth it with the work I do. Many enthusiasts who are probably beyond your level could probably do the same thing non-professionally.
You also pretty much ignore scientia and sharikou180's comments on why they may or may not keep 4x4, so you should probably apologize to them. AMD dropping 4x4 at this moment would be short sighted, though them releasing it, and then drawing away attention from it for a while would not be, as they could give it a bit of a rebirth with 65nm chips and or barcelona. Thus, you'd have to deal with my first argument.
Post a Comment