NVIDIA GF100 & Friends speculation

[Bouncing Zabaglione Bros.]

I think it's hilarious that Nvidia PR is using Twitter as a medium. I'm not sure which aspect of Charlie's article they're mocking though since he said a lot of stuff in there.

Nvidia won't engage with Charlie or what he writes on a company level, yet they have their prominent employees publicly making statements on it.

It's a way for Nvidia to unofficially say stuff they want to, without being held to it. Plausible deniability, like all those "unofficially leaked" slides where they can say what they want, and then claim "no comment", "unofficial source", "personal view of a single employee not endorsed by the company", etc.

 

[Rebel44]

Oh he did? Got a link? I want to see if he was as smug then as he is being now.

He was source for these articles (in czech)
http://pctuning.tyden.cz/component/content/article/1-aktualni-zpravy/15278-stripky-o-fermi-aka-gf100
http://pctuning.tyden.cz/component/content/article/1-aktualni-zpravy/15545-geforce-380-uz-brzy
http://pctuning.tyden.cz/hardware/g...vidia-fermi-analyza-nove-generace-gpu?start=6

I wasnt able to find article where they promised launch in November (search on that site isnt very usefull).

 

[Squilliam]

Did not the Nvidia CPU also call for 'Zero defect vias?'. If Dave alludes, Anand reveals and the Nvidia CPU mentions, then what Charlie says on the matter does fall into a zone of a lot of credibility.

I don't wish to speculate on the availability of the GPU come next month or whenever the thing actually ships or its performance. However I do wonder when Nvidia will come into a competitive position?

If Cypress wasn't all that it could have been, then Northern Islands may very well be more of a jump than it ought to have been. I don't see 2010 being at all kind for Nvidia, but even then will 2011 be any kinder once Fusion etc are released?

 

[trinibwoy]

What do you mean by that? The functionality/new features + the vastly better DP are more important than ~20-30% performance.

Perhaps, but any such changes have much more far-reaching effects over at Nvidia than they do at AMD given the narrower focus at the latter.

Depends what GF104 is. It should be TSMC's next process, i.e. 28nm. It should have been ready for production at the end of this year and TSMC is saying that it will be.

Im expecting half-GF100 on 40nm.

I wasnt able to find article where they promised launch in November (search on that site isnt very usefull).

Ok, I just find it very silly of them to be strutting around on twitter if they are just bullshitting.

I don't wish to speculate on the availability of the GPU come next month or whenever the thing actually ships or its performance. However I do wonder when Nvidia will come into a competitive position?

According to JHH on today's conf call 40nm supply is heavily constrained. I don't know if that means constrained for Nvidia or constrained industry wide. If it's the latter I wonder how it's affecting AMD.

I don't see 2010 being at all kind for Nvidia, but even then will 2011 be any kinder once Fusion etc are released?

In the enthusiast community maybe but they sound bullish overall across all their businesses. If things go as planned Tegra will more than replace any losses in the IGP/value segment. Remember Fusion is limited to AMD products so it doesn't have much bearing on Nvidia unless AMD captures more of the PC market.

 

[MfA]

Fusion is not limited to the PC market ... betting on x86 being unable to migrate into a space has never been a winning bet before, I don't think this time will be an exception.

 

[trinibwoy]

Not sure what you mean. Where is Fusion headed and in what form? Also, where has x86 ever broken into besides servers?

 

[Florin]

Not sure what you mean. Where is Fusion headed and in what form? Also, where has x86 ever broken into besides servers?

Well, supercomputers and ultraportables I guess..

 

[Jawed]

Perhaps, but any such changes have much more far-reaching effects over at Nvidia than they do at AMD given the narrower focus at the latter.

What effects? What is it about a slightly slower than originally planned card that has the far-reaching effects you keep alluding to but not enumerating?

Im expecting half-GF100 on 40nm.

I was thinking of it as the replacement. Whoops, that should be GF102...

Jawed

 

[trinibwoy]

Well, supercomputers and ultraportables I guess..

I was referring to supercomputers when I said servers and that was more a result of low manufacturing cost compared to other solutions, not due to the superiority of x86 itself. Ultraportable PC's you mean? It's the same platform in a different form factor. If x86 makes it into smartphones or something that would count as breaking into a new market.

What effects? What is it about a slightly slower than originally planned card that has the far-reaching effects you keep alluding to but not enumerating?

Seems pretty cut and dry to me. Nvidia has three important businesses that would be affected - Tesla, Geforce, Quadro. AMD only has one. Is that distinction not a big deal in your eyes? Granted JHH is telling everyone that it's all about the software when it comes to Tesla and Quadro.

 

[Jawed]

My Psion 3c has an x86 CPU.

http://en.wikipedia.org/wiki/Psion_Series_3

Jawed

 

[Jawed]

Seems pretty cut and dry to me.

So, what are the effects?



Jawed

 

[Grall]

I'm wondering, where would it leave Nvidia from a high-end perspective for pretty much all of 2010 if Charlie's doomsday article turns out to be correct (at least in the sense that Fermi's so hard to produce it essentially won't make an impression in the market).

I mean, Anand claimed months ago already that GT200b products have been EOL'd, so any cards still being sold would have been produced with what chips are still available in the channel, but that supply will be dwindling - perhaps gone already. And the year has just started.

So will Nvidia have to resume GT200b production to have at least something to meet Cypress with

 

[trinibwoy]

My Psion 3c has an x86 CPU.

http://en.wikipedia.org/wiki/Psion_Series_3

Jawed

8Mhz...nice!

So, what are the effects?



Jawed

I'm not sure if you're just being a pest or haven't actually worked for a large corporation but the more businesses a decision touches, the more complicated it becomes. I'm sure that's within your ability to comprehend.

In terms of specific effects within Nvidia, how would I know?

 

[Grall]

My Psion 3c has an x86 CPU.

The Atari Portfolio, as featured in Terminator 2, had an x86 CPU also, and was released in '89...

Still, I don't think either of these really count. Do you?

 

[Squilliam]

According to JHH on today's conf call 40nm supply is heavily constrained. I don't know if that means constrained for Nvidia or constrained industry wide. If it's the latter I wonder how it's affecting AMD.

Probably TSMC overall. AMD aren't exactly pushing out 40nm products at a record rate and most of their high volume chips are pretty small. Its probably effecting AMD in terms of being able to price their boards lower. If they priced Cypress any lower they would probably not be able to keep up with demand.

Its interesting that the relative price between the HD 5850 and 5870 has narrowed slightly since launch. $395 - $439 for 5870 and $289 - $339 for 5850 when the launch prices were $259 and $359 respectively considering the range has increased less proportionally than the overall price which has gone up by $40 for both SKUs. It implies that perhaps the yields on the top parts have improved which increases the supply of the top SKU and perhaps reduces the supply of the bottom SKU for the Cypress chip.

In the enthusiast community maybe but they sound bullish overall across all their businesses. If things go as planned Tegra will more than replace any losses in the IGP/value segment. Remember Fusion is limited to AMD products so it doesn't have much bearing on Nvidia unless AMD captures more of the PC market.

It does have a bearing on their core business which is still Tegra + Geforce. If they cannot sell as many Geforce chips then Tegra cannot be as profitable. They have practically abandoned most of their Intel/AMD chipset business and they don't yet have any firm foothold in the mobile space.

 

[Jawed]

8Mhz...nice!

More importantly: instant-on, batteries run for months in stand-by and the UI/software is slick. It's normal not to re-boot it for years

I'm not sure if you're just being a pest or haven't actually worked for a large corporation but the more businesses a decision touches, the more complicated it becomes. I'm sure that's within your ability to comprehend.

I'd say the ramifications of a chip that's 6 months late are far more significant on these businesses than the imaginary repurcussions of a chip that's "on-time" but 70% of internally-only expected performance. Particularly when these businesses you're alluding to are businesses where "AMD doesn't compete" and where "performance doesn't really matter, it's the software stupid".

Jawed

 

[trinibwoy]

It does have a bearing on their core business which is still Tegra + Geforce. If they cannot sell as many Geforce chips then Tegra cannot be as profitable. They have practically abandoned most of their Intel/AMD chipset business and they don't yet have any firm foothold in the mobile space.

Do you mean Tesla+Geforce? I'm not sure why you think they will be materially impacted by Fusion. Let's assume that Nvidia can't get any cards into an AMD Fusion based PC. How much are they actually losing given AMD's share of the market? Note thatAMD will also lose a lot of discrete sales due to Arrandale and Sandy Bridge etc.

Tegra is targeting a far larger market than AMD based PC's and has much higher margins than the low-end discrete segment. If it's successful I don't see why it can't completely replace that revenue and then some.

I'd say the ramifications of a chip that's 6 months late are far more significant on these businesses than the imaginary repurcussions of a chip that's "on-time" but 70% of internally-only expected performance. Particularly when these businesses you're alluding to are businesses where "AMD doesn't compete" and where "performance doesn't really matter, it's the software stupid".

You're starting off on the assumption that they could get out a smaller chip in a materially shorter timeline. As I mentioned earlier, it's true that Quadro and Tesla are all about software but that's not to say hardware changes can be made on a whim with those lines given a secondary thought in the process. If anything, the drama surrounding Cypress should give you a new perspective on how involved these decisions can be and in that scenario you didn't have multiple competing product lines to worry about (in terms of priorities).

By the way, it could very well be the case that "only 70% of internally-only expected performance" would have been a laughably conservative goal.

 

[Squilliam]

Do you mean Tesla+Geforce? I'm not sure why you think they will be materially impacted by Fusion. Let's assume that Nvidia can't get any cards into an AMD Fusion based PC. How much are they actually losing given AMD's share of the market? Note thatAMD will also lose a lot of discrete sales due to Arrandale and Sandy Bridge etc.

Tegra is targeting a far larger market than AMD based PC's and has much higher margins than the low-end discrete segment. If it's successful I don't see why it can't completely replace that revenue and then some.

Its also the lower end on the Intel side and the IGPs as well. So they lose in mobile discrete (fusion is an enthusiast level product there) on the AMD side as well as the lower level IGP and low level discrete cards from Intel.

I hope that Tegra performs well for them but at this point I would not bank on its success. Theres a lot of competition in that segment.

 

[trinibwoy]

Its also the lower end on the Intel side and the IGPs as well. So they lose in mobile discrete (fusion is an enthusiast level product there) on the AMD side as well as the lower level IGP and low level discrete cards from Intel.

Yep, that whole segment is essentially busted in a year or two. I don't think Fusion is gonna be anything near an enthusiast mobile part though. What's the current rumour - Redwood class? You have to compare it to the discrete mobile parts that will be out at that time.

I hope that Tegra performs well for them but at this point I would not bank on its success. Theres a lot of competition in that segment.

Yeah, time will tell.

 

[aaronspink]

P
In the enthusiast community maybe but they sound bullish overall across all their businesses. If things go as planned Tegra will more than replace any losses in the IGP/value segment. Remember Fusion is limited to AMD products so it doesn't have much bearing on Nvidia unless AMD captures more of the PC market.

How well have things been going as planned at nvidia? People are pinning way to much hope on tegra, in a market that has historically killed non-standard architectures, and pushed down to bargain basement pricing.