Nvidia has serious yield problems with the GT200.

[cegras]

Had the 8800 for only 6 month, sold it for $525 and put the 6800 GS I had laying around. It's not like I sold my 8800 for $100 just to trade up. I'm glad I waited this long and not rush into getting some HD3 series crappola.

Calculate with something to back it up? Calculate what and what's there to back it up with? Other than someone "having fun".

Everyone? Is that why so many don't think this is "credible" news?

calculation and predictions IS NOT exactly knowing. Again, where is such thing as nVidia is having issues "actually"? Care to post a link or "everyone" should know because "someone" is saying so OR........

We should know because we know how to "calculate" and "Predit"?

Which one will it be?

Well, someone did do a calculation earlier the board. I assume he's well versed in that theory, and although it is probably a 'rough' calculation (he admitted he didn't know how good TSMC's process was), he based his numbers on 'best/worst' case scenarios from AMD/intel, which are probably better at this than TSMC. And he still didn't get very encouraging numbers.

We don't know, but we have a pretty good guess, and our common sense.

K404: Thanks for agreeing with me : P

[Aberration]

What we also dont know is the contract negotiated between TSMC and nVidia. The yeild could mean nothing to nVidia, and everything to TSMC.

Besides, ATi/AMD uses TSMC also. So I would expect the defect density to be similar.

[_Lone_Wolf_]

We also don't know how redundant GT200 is

Wouldn't this be the key component in any yeild calculation?

The architects would surely know the defect density for their target process node and design to compensate for it, 65 nm is a mature node? I've posed a few yield related question on Aces and RWT and several of the responses included the point that defect density is currently a secondary factor.

[Morais]

lol...gotta love people puching each other here.I smell a burried thead

[Warboy]

lol...gotta love people puching each other here.I smell a burried thead

Good, I request a Close!

[spursindonesia]

I vote to keep this thread OPEN, this is a good discussion fanboism aside. nVidia can only hope on their G92b to save their bottomline for the next two quarters, because the GTX series willl be their achilles heel before it goes into smaller process node.

[safan80]

Good, I request a Close!

I 2nd that request. Yield problems don't mean jack because there's plenty of stock availble for anyone that wants one.

[savantu]

Wouldn't this be the key component in any yeild calculation?

The architects would surely know the defect density for their target process node and design to compensate for it, 65 nm is a mature node? I've posed a few yield related question on Aces and RWT and several of the responses included the point that defect density is currently a secondary factor.

I think you missed the discussions when AMD 0.5dd/cm^2 was put forward.Defect density makes the difference between world class processes and "good enough" ones.

I think I've stated pretty clearly in my original post :

....Does that tell us how man working chips NVIDA salvages ? No.The chip has a lot of redundancy since it is build from hundreds of parallel very simple cores.

What it tell us , is that NVIDIA gets 10-14 GTX280 ( or what's it's called ) premium chips per wafer.The rest are lesser models with fewer shader/vertex/etc units.

Out of those 10-14GTX280 , some might fail running at the required frequency or in the envisioned TDP.So I'd venture to say that they get less than 10 full fledged chips per wafer.

I think this is as clear as clear can be.

[savantu]

What we also dont know is the contract negotiated between TSMC and nVidia. The yeild could mean nothing to nVidia, and everything to TSMC.

Besides, ATi/AMD uses TSMC also. So I would expect the defect density to be similar.

No sane foundry guarantees a certain yield IMO.They tell process performance , time to market and capacity available.Based on that , and probably a few trial runs , the customer decides what to do.

Besides , it's not like Nvidia or ATI have a choice.TSMC is the largest foundry in the world and has the best process tech after the top guys.The choice is rather limited.

[Sr7]

I think you missed the discussions when AMD 0.5dd/cm^2 was put forward.Defect density makes the difference between world class processes and "good enough" ones.

I think I've stated pretty clearly in my original post :



I think this is as clear as clear can be.

The beautiful thing about not owning your own fabs is that it's likely NVIDIA doesn't have to pay for the bad chips. They just don't get as many chips to sell as quickly, which is the only reason they'd care about yields. The rest is on TSMC

[DilTech]

This thread has come to an end... I was going to let it go, but it's literally become a battle of "what if's" and he-said she-said.