Can't load Fudzilla

Teemax · Jan 13th, 10, 05:14 AM

Just some 2 months ago on Fudzilla:
Fudzilla - Nvidia Fermi launch slips into January 2010
(show case in Dec 2009, launch in Jan 2010)

Today, Fud's tone on Fermi is sharply different:
Fudzilla - Cypress yields not a problem
Masquerading inside a "Cypress yield" article is a wealth of new predictions for Fermi

Quote:

...Fermi chip is struggling with yield issues as our sources suggest that the actual yields are as low as 20% and are not improving quickly according to sources close to the fabs...
...the numbers suggest that each chip would cost an astounding estimated $200 per chip which pegs the card with a sticker price of about $600...
...only about 20% faster than Cypress, while Hemlock smokes it...
...Because of the situation a mole we know is suggesting that Nvidia may limit the sales of Fermi to consumers and instead use the chips for the Tesla and Quadro products where prices and margins are much better...

When did Fudzilla hired Charlie's cousin?

These claims, if even remotely true, spell a disaster for the upcoming Fermi.

Share|

psolord · Jan 13th, 10, 05:52 AM

Crap, this is the first time in my life (ok the second) that I actually want and need an Nvidia (two of them really) and now they are going to screw it?

What I want from GF100 is 20% faster than Cypress, be able to overclock just as much as Cypress "XT" (again that's 20%), keep resonable temperatures and even more reasonable fan noise.

The funniest thing is that I wished the "GTX 380" to be no more than 350 euros, since the 5870 already costs 300 euros and the price is going to drop! OMG!

iamverysmart · Jan 13th, 10, 08:03 AM

I just want a fast card, 5850's have come back in to stock here, but they are still more expensive then launch.

9VARZ · Jan 13th, 10, 10:00 AM

Seems to me that Nvidia has always made a lot more money in the "professional" segment rather than consumer cards anyway. Small setback for them (albeit made somewhat larger thanks to excessive media coverage again).

Megalord · Jan 13th, 10, 11:21 AM

I reckon its sharply different due to more information feedback. They didn't have the yield rate numbers back then. And anyway, yield rate decreases exponentially with larger chip size all else equal, so that number is hardly surprising, esp considering that Nv yield rate for GT200 has been amazingly low. Time will tell. I don't care about yield rate or such information as long as i can get my hands on the card without coughing out big $$$.

jason1980 · Jan 13th, 10, 11:44 AM

20% yield is damn jialat!

power666 · Jan 13th, 10, 01:05 PM

It has already been announced that the Telsa cards will only use 448 of the 512 shaders on the chip. With relatively poor yields and if there is no consumer card, it makes you wonder if there will be a variant that will have the full 512 shaders enabled. In fairness, the shader count on the Quadro parts haven't been revealed yet.

Teemax · Jan 13th, 10, 04:14 PM

Quote:

Originally Posted by 9VARZ

Seems to me that Nvidia has always made a lot more money in the "professional" segment rather than consumer cards anyway. Small setback for them (albeit made somewhat larger thanks to excessive media coverage again).

The professional market has much higher margin, that is very true.

However, it's typically the consumer market that bears the cost of development for NVIDIA. Relying solely on the professional market isn't sustainable; at least not yet.

FATRAT · Jan 14th, 10, 01:43 AM

Why I'm I not surprised? it sure looks like nvidia screwed it up this round, but lets not forget that its just rumors and you never know if rumors are true or not.

Brandon1987 · Jan 14th, 10, 06:09 PM

Hmm... Microsoft and IBM depended on professional market and the enterprise market, I don't see why it is not sustainable though.

The company behind the movie, Avatar, uses Nvidia graphics to do the CGI. Medical community uses Nvidia cards far more than ATI to do their protein folding etc. The Astrophysics community also uses Nvidia cards to do computing. In fact, the US military is also trying get consumer-based graphics card to do their research and development. One Nvidia graphics card cost less than a fraction of a custom built supercomputer that give the same amount of computing power.

At this rate of pushing, I think the professional market will be larger than that of the consumer. Besides, how many consumers are willing to spent that kind of money for the top end cards other than companies doing computation work, less those extreme gamers and people who care only about benchmarks.

So, ATI can have all the market for all I care....

Btw, the high cost of professional card, mainly goes back to the software support. Quadro drivers and tesla drivers are typically more stable, release far more slowly than consumer drivers...maybe all of us should buy quadro and tesla cards

Teemax · Jan 14th, 10, 06:53 PM

Quote:

Originally Posted by Brandon1987

Microsoft and IBM depended on professional market and the enterprise market, I don't see why it is not sustainable though.

The problem is Microsoft and IBM are software companies; while NVIDIA is still principally a hardware company. BTW, if you're citing "Microsoft" for the professional market, I have a strong suspicion you have no clue who the big players are.

Yes, some people are buying NVIDIA cards for professional use. But it doesn't mean that the current market is big enough to sustain NVIDIA. Consumer GPU market is still NVIDIA's bread and butter. But I guess losing it is not a big deal for you, eh?

Sexmaster2k · Jan 14th, 10, 07:10 PM

Quote:

Originally Posted by Teemax

The problem is Microsoft and IBM are software companies; while NVIDIA is still principally a hardware company. BTW, if you're citing "Microsoft" for the professional market, I have a strong suspicion you have no clue who the big players are.

Yes, some people are buying NVIDIA cards for professional use. But it doesn't mean that the current market is big enough to sustain NVIDIA. Consumer GPU market is still NVIDIA's bread and butter. But I guess losing it is not a big deal for you, eh?

And the big players are?

Brandon1987 · Jan 14th, 10, 09:08 PM

Quote:

Originally Posted by Teemax

The problem is Microsoft and IBM are software companies; while NVIDIA is still principally a hardware company. BTW, if you're citing "Microsoft" for the professional market, I have a strong suspicion you have no clue who the big players are.

Yes, some people are buying NVIDIA cards for professional use. But it doesn't mean that the current market is big enough to sustain NVIDIA. Consumer GPU market is still NVIDIA's bread and butter. But I guess losing it is not a big deal for you, eh?

Ah, well maybe you can educate me on who the big players are. Losing Nvidia for consumer market? Well, we got ATI there, there is nothing to be afraid. At least those ATI fan boys can have their fun.

jason1980 · Jan 14th, 10, 09:36 PM

Nvidia got bitten by ATI?

watzup_ken · Jan 14th, 10, 10:10 PM

If this is true, its a lose lose situation for us as the consumers. I am however, not surprise about the yields as well as cost of the chips simply because Nvidia has been very ambitious about making very complex chips. Surely it benefits us, but given the price, not many people will want or may not even be able to afford it. And given such a complex design, its difficult to see it trickle down to the lower end products. This is probably the case for GT200 as you may have noticed that Nvidia has all these while been using a very old chip to form the line up for the lower end cards. There is def nothing wrong in doing so since GT200 is just an evolution of the older chip, but I don't think this solution is not sustainable in the long run. It is quite apparent now if you look at the current GT240 against the new HD5670. Nvidia is aiming to be both a hardware and software company, but at their current state, they are nowhere here nor there. I do feel that they are losing their focus at this point in time.