Bulldozer delayed til Q3?

[pxc]

Yes, I did. But he's a server guy, so I pushed it into his direction. He suggested that Zambezi is delayed not because of problems, but to make room for Llano.

I give him credit where due. He's a master chain yanker.

 

[Michaelius]

You're assuming that AMD has customers lined up for those "100" BDs. You're also not applying certain properties of microeconomics that are relevant to this type of supply decision. I find that a reading on the the Guns versus butter model - Wikipedia, the free encyclopedia would be apropos here (even though it's technically applies to macroeconomics.)

Let's use your numbers for a simple example. Say AMD can only crank out 1000 32nm chips. If there is demand for Llano that is >= to 1000 then they can sell everyone of those Llanos. But maybe demand for BD chips isn't so high. Maybe it's only 50 BD chips. So, if they made 100 BD chips they would not only not sell 50 BDs but there is also an opportunity cost in the form of Llano chips that could have been made (and also were not sold).

So if it takes, I dunno, $30 to make the chip and AMD sells Llano chips at $100 a piece then the total cost to AMD is $130 for every BD chip not sold (since we're assuming they can sell every Llano they can conceivably make). Technically the opportunity cost still applies to all those chips AMD made BDs instead of Llanos, but I thought is was easier to comprehend this way.

I dunno what AMD is up to, but this would be a scenario where not releasing BD in conjunction with Llano would make sense.

And what makes you thinking there wouldn't be enough demand for bulldozer if it supposedly kills i7 2600 in multi threaded performance?

 

[FreiDOg]

I give him credit where due. He's a master chain yanker.

Which seems to be a more useful skill than on-time product manager or reality based outlook provider at AMD these days.

 

[InternationalHat]

And what makes you thinking there wouldn't be enough demand for bulldozer if it supposedly kills i7 2600 in multi threaded performance?

I think he's simplifying a bit.

Supply is not just BD chips, it's yields per bin per wafer. Demand is not just BD chips either, it's demand for a specific model. OEMs will likely demand x number of each model where supply may not make that entirely economical for the first few months.

If Llano's yields, margin, and binning are more profitable than what they expect out of preliminary bulldozer silicon, they're going to refine the process until it makes sense.

And even THAT is a simplification. It's not cut and dry really. No one gives a shit about Bulldozer on the desktop aside from a very small segment of the market. The server chips and llano are probably an order of magnitude more important. Its performance is almost irrelevant when compared to other economic factors.

 

[Ryun]

I think he's simplifying a bit.

Pretty much.

The poster I originally quoted seemed to be under the impression that it didn't make any sense to not release BD with Llano when there are a number of scenarios where it would make fiscal sense. I don't work for AMD, and I'm certainly no economist so really any example I give is going to be simplistic. It's still plausible though.

Anyway, I'm sure we'll know more once it releases.

 

[nonameo]

Might I suggest that, even if bulldozer gets more $ per die for AMD, it might have poorer yields, so in the end $ per wafer might be more for llano.

Llano being easier to produce also makes sense because both the GPU and CPU architectures it is made from have been around for a while.

 

[flexcore]

I'm not sure what kind of arrangement AMD has for Llano, but AMD and GF have a new agreement so AMD pays only for good dies not per wafer starts. I don;t know what they pay but I take it this is due to poor yields at GF.

 

[nonameo]

I'm not sure what kind of arrangement AMD has for Llano, but AMD and GF have a new agreement so AMD pays only for good dies not per wafer starts. I don;t know what they pay but I take it this is due to poor yields at GF.

And I would imagine that the price is at least somewhat based on how well the wafers yield.

 

[Michaelius]

I'm not sure what kind of arrangement AMD has for Llano, but AMD and GF have a new agreement so AMD pays only for good dies not per wafer starts. I don;t know what they pay but I take it this is due to poor yields at GF.

Which isn't exactly perfect solution for long term production even if temporarily AMD throws the risk onto GF.

 

[JF-AMD]

I'm not sure what kind of arrangement AMD has for Llano, but AMD and GF have a new agreement so AMD pays only for good dies not per wafer starts. I don;t know what they pay but I take it this is due to poor yields at GF.

Not necessarily. All of the discussion of wafer yields is not reflecting reality. That is all I will say.

 

[Nightbird]

Not necessarily. All of the discussion of wafer yields is not reflecting reality. That is all I will say.

Delays do suggest it's more likely bad than good. It could be similar to the endless problems Nvidia had with Fermi, new architecture on new process is a pretty risky move. Even Intel with their superior fabbing avoids it, but I understand AMD's need to catch up quickly.

 

[flexcore]

I only know what I read, but it clearly changed the way AMD pays for the dies from GF. From per wafer to per usable die. The only reason I can see this being done is GF is having poor yields, to the point that it isn't viable for AMD. I'll try to find the release.

Edit: Here is an article on it. It's the INQ, but I have read it elsewhere also. $ per 45nm wafers and $ per xxx amount of good 32nm dies.

 

[JF-AMD]

Delays do suggest it's more likely bad than good.

Sandy Bridge E was supposed to be Q3
Sandy Bridge EP was supposed to be late Q3
Ivy bridge was supposed to be Q4

It should be interesting to see where these all shake out.

I only know what I read, but it clearly changed the way AMD pays for the dies from GF. From per wafer to per usable die. The only reason I can see this being done is GF is having poor yields, to the point that it isn't viable for AMD. I'll try to find the release.

Edit: Here is an article on it. It's the INQ, but I have read it elsewhere also. $ per 45nm wafers and $ per xxx amount of good 32nm dies.

Please don't take this the wrong way, but are you in the foundry or semiconductor business? If not, then consider that there could be a lot of other factors at play that you are not aware of.

 

[Deleted whining member 223597]

The thing is, Sandy Bridge E doesn't HAVE to come out on time. Sandy bridge is already taking down current AMD CPUs for the same or less(pricewise, some may cost more if you want lets say quad SLI/Xfire). And leaks don't seem to be showing Bulldozer beating sandy bridge. Bulldozer needs to be released this summer in my opinion.

 

[Nightbird]

Sandy Bridge E was supposed to be Q3
Sandy Bridge EP was supposed to be late Q3
Ivy bridge was supposed to be Q4

It should be interesting to see where these all shake out.

By no means did I suggest AMD is the only one suffering from delays. I hope whatever issues the chips are having, be it is design or yield, are resolved quickly for all sides. Because the companies don't tend to release delay reasons in any detail, it is understandable that people interested will speculate, usually using unreliable information sources. Unless AMD wants to break that trend, and there is no reason for it to, we'll continue to do so.