Fermi needs A3 silicon? Launch in Q1 2010

[thilanliyan]

http://vr-zone.com/articles/nvidia/nvidia-fermi-in-q1-2010/8006.html?doc=8006

http://hardocp.com/news/2009/11/02/fermi_working_samples_for_ces63

So there won't be ANY availability until February/March? I thought I read somewhere that it was supposed to be a soft launch at the end of this year?

EDIT: They may need A3 silicon:
http://www.fudzilla.com/content/view/16342/1/

http://vr-zone.com/articles/-rumour...to-be-a3-revision-q2-2010-/8014.html?doc=8014

 

[zagood]

Soft launch will probably be enterprise/tesla.

 

[crazylegs]

I know you have to take anythign you read on teh internet with a pinch of salt, but the last two sentances cracked me up:

''Worth mentioning also NVIDIA has yet to speak one word to us about the gaming abilities of the Fermi GPU. It makes us think that NVIDIA does not even know yet.''

...

 

[thilanliyan]

Updated with Vr-zone link which states Jen Hsun Huang has said it will only be here in Q1 2010 and not late this year as some thought.
http://vr-zone.com/articles/nvidia/nvidia-fermi-in-q1-2010/8006.html?doc=8006

 

[BenSkywalker]

Updated with Vr-zone link which states Jen Hsun Huang has said it will only be here in Q1 2010 and not late this year as some thought.

Do you have the link to that quote? The one on the link you posted states-

At a conference with financial analysts, JHH said, "Next year it is going to be an interesting first quarter because, in fact, we will need more wafers than ever in Q1. The reason for that is because – and I mean more 40nm wafers than ever in Q1 –we are […] fully ramping Fermi for three different product lines: GeForce, Quadro and Tesla".

That doesn't mention anything about starting production. Not stating what you said was inaccurate, but it certainly isn't in that quote.

 

[thilanliyan]

He doesn't explicitly state it but doesn't that basically mean we can't buy it until Q1 next year? If it was going to start production this year wouldn't he have said that?

I suppose Tesla could launch this year but even that would be in short supply if what he says is true. Also compounding the situation is the continuing 40nm problems at TSMC.

 

[BenSkywalker]

He doesn't explicitly state it but doesn't that basically mean we can't buy it until Q1 next year? If it was going to start production this year wouldn't he have said that?

He's talking to investors, he legally isn't allowed to state they are going to start selling any products unless he's sure(only way he can get around that is if something catastrophic and unseeable happens, not something that covers a bad spin/fab issues since both are well known). If you read his quote he doesn't state when they are going to go on sale, just that they will fully ramp production in Q1(which likely means they see small scale production starting prior to that). I'm fairly certain if they are going to get small quantities of Tesla out they will also push out review boards for GeForce products also, not that we are likely to see availability, but with the reality of the current situation at TSMC it isn't like borderline vapourware is going to offend too many people atm

 

[BFG10K]

I’m actually looking forward to this card. I absolutely don’t care about the GPGPU side of things, but I’d love a solid upgrade over a GTX285 for gaming.

That and I want to see what the 5xxx series is truly capable of after a driver update, so I can retest my 5770.

 

[terentenet]

I too want to see what Fermi looks like in gaming. I need to upgrade the 285's and I want to see Nvidia's effort before I decide.
I hope, if the cards launch in Feb/March, we will see benchmarks this year, before Christmas.

 

[Idontcare]

He doesn't explicitly state it but doesn't that basically mean we can't buy it until Q1 next year? If it was going to start production this year wouldn't he have said that?

I suppose Tesla could launch this year but even that would be in short supply if what he says is true. Also compounding the situation is the continuing 40nm problems at TSMC.

He's talking to investors, he legally isn't allowed to state they are going to start selling any products unless he's sure(only way he can get around that is if something catastrophic and unseeable happens, not something that covers a bad spin/fab issues since both are well known). If you read his quote he doesn't state when they are going to go on sale, just that they will fully ramp production in Q1(which likely means they see small scale production starting prior to that). I'm fairly certain if they are going to get small quantities of Tesla out they will also push out review boards for GeForce products also, not that we are likely to see availability, but with the reality of the current situation at TSMC it isn't like borderline vapourware is going to offend too many people atm

Caution is definitely warranted in our layman's forum level interpretation of the meaning/implication/boundary-conditions of the phrase "fully ramping Fermi"...to me this says nothing regarding the timing of any stage that is less than "full" ramping...so if Q1 is "fully" ramping then presumably there is a phase in which they are less than fully ramping but nevertheless do have product coming out of the pipe and that product could be sold if they so desired.

At to that the legalities of Jensen's every word given the target audience and I say reading anything into that statement is activity fraught with likely misinterpretation because those statements are engineered to avoid communicating the facts on the ground.

I would not be surprised if Dirk were to hold a similar CC and tell his investor community that Cypress is still ramping and won't be "fully" ramping until next quarter either as presumably AMD hopes to have more 40nm cypress wspm next year than they currently have. (so by definition then AMD hasn't fully ramped Cypress either)

 

[GaiaHunter]

Do we know how many 40nm wafers do they need currently?

I bet it isn't too hard for any 2010 quarter to beat their current number of 40nm wafers.

Only thing I can read is that no cards available for the consumer this year.

 

[nemesismk2]

This is disappointing news because I was looking forward to seeing what kind of performance the Fermi can do. As a positive atleast there should be more games with DX11 Support when it is released.

 

[jvroig]

I’m actually looking forward to this card. I absolutely don’t care about the GPGPU side of things, but I’d love a solid upgrade over a GTX285 for gaming.

That and I want to see what the 5xxx series is truly capable of after a driver update, so I can retest my 5770.

I've heard (well, read) you say this before, so you really believe that the 5xxx series are currently intentionally crippled until they see Fermi's ability? Didn't an AMD/ATI engineer already respond with something like "given that a video card is less than 30% of the factors in game performance, it's already amazing that we managed to achieve even just a 33% increase", which kinda closes that hope?

Don't get me wrong, I'm on your side, I'm hoping the same as you, I'm just wondering if that's actually possible (given the response given by the AMD guy) and if it's actually a viable tactic by AMD.

As for Fermi's Q1 launch... well, what's new? Ever since they "faked" that board and Charlie hyped it up, I resigned myself to expect Fermi by next year. I'm holding off on a 5xxx purchase so I can see how Fermi stacks up first. (Hahahah right. Actually, no one's holding off on anything - I just can't get a card, and now it's official that prices are up and stocks are down :biggrin: )

 

[shangshang]

with all this drama, hype, and anticipation of Fermi... if NV doesn't outright bust down the doors of the AMD 5870 series... I mean if it only bests the 5870 by 1.1x-1.2x times... then it's semi-failure. Gotta beat the 5870 by at least 1.5x, but preferably by 1.8x-2.0x, in order to get some respect here.

 

[Markfw]

I gues I must be out of it. What is tesla ?

 

[theAnimal]

I gues I must be out of it. What is tesla ?

http://www.nvidia.com/object/personal_supercomputing.html

 

[SlowSpyder]

with all this drama, hype, and anticipation of Fermi... if NV doesn't outright bust down the doors of the AMD 5870 series... I mean if it only bests the 5870 by 1.1x-1.2x times... then it's semi-failure. Gotta beat the 5870 by at least 1.5x, but preferably by 1.8x-2.0x, in order to get some respect here.

Yea, if Fermi comes out a full 6 months after the 58x0 launch and is 10-15% faster on average, I'm not sure that you could really call it a 'win' for Nvidia. I mean, at that point they'd have the quicker part, but hardly faster to the point that I see many people dropping their 58x0 part for Fermi. I wonder how long it'll be until we see a 58x0 refresh, not to mention the 59xx part that should be out well before Fermi.

 

[scooterlibby]

I wonder how long it'll be until we see a 58x0 refresh, not to mention the 59xx part that should be out well before Fermi.

I definitely wonder about that too, but am not familiar enough with the process to speculate as to whether they might have it by Q12010. It would be nice, though, to see a refresh by then.

 

[BenSkywalker]

I definitely wonder about that too, but am not familiar enough with the process to speculate as to whether they might have it by Q12010. It would be nice, though, to see a refresh by then.

That would be horrific from a business standpoint. The biggest cost associated with video cards isn't the die size, it isn't the RAM or the PCB, it is R&D by a long shot. If you push out a refresh before you have come remotely close to recouping the R&D costs on the first part you are in essence throwing away millions of dollars. Given the current issues at TSMC it doesn't make any sense to get a refresh out prior to several months after TSMC has fixed their yield issues at the earliest(barring enormous competitive pressure that puts them in a position to lose money per unit sold which I can't fathom within the next couple of quarters).

 

[Lonyo]

Yea, if Fermi comes out a full 6 months after the 58x0 launch and is 10-15% faster on average, I'm not sure that you could really call it a 'win' for Nvidia. I mean, at that point they'd have the quicker part, but hardly faster to the point that I see many people dropping their 58x0 part for Fermi. I wonder how long it'll be until we see a 58x0 refresh, not to mention the 59xx part that should be out well before Fermi.

As long as the Fermi launch isn't as terrible as the HD5xxx launch it's a win for NV
All these delays aren't hurting them as much as they could, given the TSMC problems.
If yield issues had been solved and AMD could flood the market, NV would be in a bit more trouble, but as it is, it looks like NV might have a chance of ramping up when TSMC claim they will have sorted their problems by.

Even if NV wanted to, I doubt they could ramp up before Q1.

 

[Genx87]

with all this drama, hype, and anticipation of Fermi... if NV doesn't outright bust down the doors of the AMD 5870 series... I mean if it only bests the 5870 by 1.1x-1.2x times... then it's semi-failure. Gotta beat the 5870 by at least 1.5x, but preferably by 1.8x-2.0x, in order to get some respect here.

Why? Twice the speed of the competition?? When was the last time that ever happened?

 

[shangshang]

Why? Twice the speed of the competition?? When was the last time that ever happened?

8800gtx

 

[Genx87]

8800gtx

Was twice as fast as ATIs counterpart?

 

[GaiaHunter]

That would be horrific from a business standpoint. The biggest cost associated with video cards isn't the die size, it isn't the RAM or the PCB, it is R&D by a long shot. If you push out a refresh before you have come remotely close to recouping the R&D costs on the first part you are in essence throwing away millions of dollars. Given the current issues at TSMC it doesn't make any sense to get a refresh out prior to several months after TSMC has fixed their yield issues at the earliest(barring enormous competitive pressure that puts them in a position to lose money per unit sold which I can't fathom within the next couple of quarters).

[speculative mode] Well since according to some "sources" 58xx is only 2 RV770 with DX11, maybe there isn't much R&D to recoup, considering that 6xxx is rumored to appear by the end of 2010... [/speculative]

 

[Obsoleet]

If fermi is indeed a large, complex pwnage chip that NV says it is, then this is going to be very bad. They're competing for space at TSMC, who is also having trouble with 40nm, and if Fermi is complex as I stated we're looking at low yields on top of it all.

The best prediction is to count Nvidia out until 2011, maybe longer because just because Fermi is looking like a misstep, doesn't mean NV will have great luck with it's successor either.