[Arun]

Hey everyone,

In my eternal lazyness, I'll just quote GPU:RW (
http://www.notforidiots.com/GPURW.php )

nVidia's NV40 has taped-out at an uncertain date. It is thus either very early August or mid/late July.

<continued with NV36/NV38 info - visit the URL for more>

Okay, so it seems nVidia is pretty much on schedule so far. Assuming they don't run in any major problems, the NV40 might even manage to be in our PC for Christmas.

Also, in a completely unrelated note, I've recently learnt about a new nVidia policy which, IMO, is just as good if not better than ATI's "We launch within 30 days of retail avaibility". And they are enforcing it. I cannot go into the details here though.

Back to the NV40 - my biggest question right now would be "How traditional is it?" - It cannot be a real ILDP yet, but who knows, it might be even more traditional than the NV30. I'm still betting on less traditional though, but... who knows?


Uttar

[Nebuchadnezzar]

Yepee.

[Heathen]

We launch within 30 days of retail avaibility

As opposed to the 30 weeks for the NV30

Well all we need is some concrete info on the nv40 and trustworthy drivers and I can start thinking about whether to buy Nvidia again.

However I will be waiting for solid info on the R420 before making any decisions.

[Blade]

If NV40 proves to be impressive enough, it might be my next card. I'm more interested in seeing ATi's NV40 competition first, though. ATi would be my blind pick for "better card".. seeing as how far ahead they've gotten. ArtX was a huge acquisition.. the 9700 is faster than the much newer GeForceFX 5800. If that trend continues.. nVidia stock = bye bye..

[Josiah]

awesome, I was starting to think that if NV40 doesn't tape-out soon there's no way it will be launched this year. now (knock on wood) they should have first silicon sometime in October, and hopefully some kind of availability for December

[RussSchultz]

Originally Posted by Josiah

awesome, I was starting to think that if NV40 doesn't tape-out soon there's no way it will be launched this year. now (knock on wood) they should have first silicon sometime in October, and hopefully some kind of availability for December

1st silicon could be back as soon as 6 weeks from tapeout.

[gkar1]

Just as NV30's could have been
What's that everyone says? "I'll believe it when I see it"

[Chalnoth]

Originally Posted by Blade

If that trend continues.. nVidia stock = bye bye..

nVidia has shown to be a very shrewd company. There's much, much more to the market than just the high-end. It would take a large number of hardware screw-ups for nVidia to really take a major financial dive.

[Lezmaka]

Where did it tape out? IBM or TSMC?

[SvP]

Jun 14, 2002; Chalnoth: Supposedly NV30 has already taped-out

Aug 14 2002: anandtech: NV30 taped-out just recently

Aug 8 2003: NV40 has taped-out

[Ailuros]

Aug 14 2002: anandtech: NV30 taped-out just recently

That wasn't wrong at all back then, trouble is that the samples that returned from the labs where far away from being alive.

But it's still not relevant to this generation, even if there should pop up problems I doubt that they'd be even close to the magnitude of what the NV30 faced.

It should be noted though that a successful tape out is the ideal case.

There's much, much more to the market than just the high-end.

Can't say I've been personally impressed with their mainstream offerings (NV31).

[Heathen]

But it's still not relevant to this generation

Maybe not, but how do we know that isn't going to happen again? Yes I realise that nvidia will have the Nv30 debacle firmly on their minds but I assume that the NV40 is a hideously complex design in terms of transistor count (and actuallayout if it has anythijng to dowith previous generations - likely considering their track record), but at the end of the day nothing has come back from the foundry and assuming Xmas may be over ambitious.

It would take a large number of hardware screw-ups for nVidia to really take a major financial dive.

Have you seen their recent financial returns?

[bdmosky]

Originally Posted by Heathen

Have you seen their recent financial returns?

Have you?

http://news.com.com/2100-1047_3-5061321.html?tag=fd_top

[Richthofen]

n1 profit and revenue gain.
I would't worry much about the margins. They can afford that

[Chalnoth]

Originally Posted by Ailuros

Can't say I've been personally impressed with their mainstream offerings (NV31).

The low-end (NV34) is more important. These products primarily sell to OEM's, where features are the most important thing. The NV34 certainly has a large number of advanced features.

[Richthofen]

Originally Posted by Ailuros

Can't say I've been personally impressed with their mainstream offerings (NV31).

well but the mainstream offering from the competition is not better and that is all what counts.
In the end ATI can't get a home run here.
And you have to keep in mind that they can't make enough 0.13 micron mainstream offerings due to Nvidia blocking a lot of TSMCs capacity.

In an environment where the market leader has similar or better products the smaller will not profit that much.
If AMD wants to gain market share they have to be a lot better than Intel.
If they have similar or worse products they will keep their position but nothing more.
So in that case Nvidia can live with its mainstream offering.
NV34 is where they have pretty massive gains and that will continue until ATI can delivier a competitive and massivly available DX9 class product in that price range!
RV280 is not competitive at all.

[RussSchultz]

Originally Posted by gkar1

Just as NV30's could have been
What's that everyone says? "I'll believe it when I see it"

Are you suggesting that first silicon did not come back ~6 weeks after the tapeout of NV30? Or are you rolling your eyes at something completely different?

[PaulS]

He's probably talking about fully functionning silicon, rather than just first silicon.

[WaltC]

I may have missed something here, but how does quoting a rumor from a rumors web page change the character of the "information" from "rumor" to "objective fact"...? The problem with rumors is that most are proven false with time and if not wholly false then at best materially incomplete.

Perhaps instead of "rumors" the page might better be described as "wishful thinking"...?

[Ailuros]

Originally Posted by Chalnoth

The low-end (NV34) is more important. These products primarily sell to OEM's, where features are the most important thing. The NV34 certainly has a large number of advanced features.

OEMs yes and checkboard features yes. Granted it's the only low end offering that has MS/AF, the rest of the featureset seems to be mostly decorational, albeit a driving force for OEM sales. It's a far better offering in that sense than the NV17.

From a marketing standpoint NV detected the weak spot of the Radeon-line and targeted there. I'd think that ATI will be concentrating even more on the budget line in the next round and ironically if the high end part should be the real winner in the NV line , then the marketing yadda yadda will switch the importance immediately to the high end.

I do listen to recorded press conferences too.

[Ailuros]

NV34 is where they have pretty massive gains and that will continue until ATI can delivier a competitive and massivly available DX9 class product in that price range!

I always stood for that and won't be changing it now. I replied to a remark stating that high end is not all there is. Low end gains the biggest number in sales and probably revenue, but it's not all there is either.

ATI has won some serious impressions with the R3xx line and the market leader to have to pounder on it's low end offering as the spot where they really are able to make a significant difference sounds rather pathetic to me.

[Richthofen]

Well but the problem is they didn't gain much out of their good R3xx.

More than 9 months have passed now and Nvidia still mainstains its position. Of course with lower margins due to lower ASPs but they can afford that with their revenue and 1 Billion in cash.

The point is if ATI looses some ground on the next generation and even if it is only a littlebit, than they will loose a lot of their desktop business.

The market forgets failures of market leaders pretty fast and it needs a hell of time until you see significant changes in the overall market.
On the other side that does not happen if the smaller competitor fails.
In that case you would see siginificant market changes in about 3 to 4 months.
That is ATIs problem. They gained not enough to back up if something goes wrong.
If their next product line is slower than Nvidias and even if it is only by a small percentage, then they will run into problems.
In that case people would tend to buy market leader products.

[Heathen]

Have you?

Yes thanks,

Slightly old example but it does show what they were raking in: http://sanjose.bizjournals.com/sanjo...5/daily60.html

No with ATi revenues up and both Intel and ATi gaining market share, nvidia's returns don't look so good to me. Granted they are still 'on top' when it comes to market share (although revenues are relatively static) & plenty of cash (an easily lost commodity when developing GPUs) so that should stand them ion good stead.

The market forgets failures of market leaders pretty fast and it needs a hell of time until you see significant changes in the overall market.
On the other side that does not happen if the smaller competitor fails.
In that case you would see siginificant market changes in about 3 to 4 months.
That is ATIs problem. They gained not enough to back up if something goes wrong.
If their next product line is slower than Nvidias and even if it is only by a small percentage, then they will run into problems.
In that case people would tend to buy market leader products.

I take it you glass is half empty then.

[Josiah]

Originally Posted by WaltC

I may have missed something here, but how does quoting a rumor from a rumors web page change the character of the "information" from "rumor" to "objective fact"...? The problem with rumors is that most are proven false with time and if not wholly false then at best materially incomplete.

Perhaps instead of "rumors" the page might better be described as "wishful thinking"...?

well, Uttar is usually right, and he rates the credibility of this information as "very high." it does fall in line with Nvidias roadmap and their supposed Comdex launch. granted, only time will tell!

[duncan36]

Originally Posted by Richthofen

The market forgets failures of market leaders pretty fast and it needs a hell of time until you see significant changes in the overall market.
On the other side that does not happen if the smaller competitor fails.
In that case you would see siginificant market changes in about 3 to 4 months.
That is ATIs problem. They gained not enough to back up if something goes wrong.
If their next product line is slower than Nvidias and even if it is only by a small percentage, then they will run into problems.
In that case people would tend to buy market leader products.

Well one company has a history of being able to produce good solid chips with market leading thermal characteristics and one company isnt as of late.
The fact is that right now ATi is the market leader and Nvidia is playing catchup.
This type of shift takes time, but it definitely started with the R300, now ATi has better margins, better product, better yields, better everything.
It was only a matter of time until ATi passed Nvidia, theres nothing magical about Nvidia just like there was nothing magical about 3dfx when Nvidia passed them.

The window of opportunity is there for Nvidia to retake the throne, however as we've seen with 3dfx its tempting when you're in a bind to come up with outrageous technological solutions that dont work well instead of simple steady process improvements.