Hornet
Supreme [H]ardness
- Joined
- Oct 4, 2005
- Messages
- 6,624
So, the only interesting thing will be the 9600 which is a completely new architecture.
WTF
WTF
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
GeForce 9800 GTX in March or Feb eh ?
I wonder how many SP it will carry? Some on B3D think 192 SP @ 2Ghz.
NV will always spend vast amounts on R&D, whether they release the fruits of that labour onto the market is another question.Nvidia will not spend money on R&D unless ATI trumps it. It was made plain and simple in 07 and continues in 08. If ATI went under right now they would probably not release anything in 08. I wonder what their engineering department is doing or has been doing the last year.
a 9800 GTX will simply be a G92 SKU with higher clocks.
there is easily headroom to increase the core to 720 and the memory to 2400.
I would not even dignify this as a 'refresh' of G92.As for Larabee,I think the pounding that will come will be inevitable.But also years away.
What's really sad about this is that AMD might actually have the faster card in March at this rate. We know Crossfire generally scales much better than SLI (up to 100% performance increase where as SLI never comes close to that) so a 3870 X2 might actually be as fast or faster than a 9800GX2. Honestly, a 30% increase after a year-and-a-half or almost two years? Come the fuck on...
http://enthusiast.hardocp.com/article.html?art=MTQzOSwxLCxoZW50aHVzaWFzdA==You're assuming that the 9800 GX2 is not as fast or faster than two 8800 Ultras, which we don't know for sure yet.
The 9800 GX2 is said to be at least 30% faster than a 8800 Ultra.
Having a hard time reading ? It says "at least 30%". That's worst case scenario. Real World performance is what will settle it and we don't know it yet.
So the worst case would be 30% faster than a single 8800Ultra but the best case, it can be faster than two Ultras? But looking at the SLi scaling, matching the two 8800Ultra in SLi is not too hard actually
a 9800 GTX will simply be a G92 SKU with higher clocks.
there is easily headroom to increase the core to 720 and the memory to 2400.
If you mean mid-low range, I agree that a third player will make a difference. But if you mean high-end, you need to think again. Intel is NOT going to defy NVIDIA in the high-end GPU market for a while. You can quote me on that.
So what it comes down to is, they are shuffling around their product line a bit to better compete with ATI's midrange parts, and holding back everything else due to a lack of competition.
NV, you did so well with the 8800GT, then you pooched availability. Now that the 8800GT is getting more common, we're all waiting for the next-generation parts. I don't see them. I purchased my 8600GT (for $120) almost 9 months ago, and I'm looking for something new to replace it with. I don't want a 120W card (330W PSU, MicroATX case), I don't want to pay $260, I don't have a 30" monitor, and I don't play Crysis. I want a nice, sensible, $150, 75W card with the kind of performance you delivered on the old 8800GTS.
ATI's Radeon HD3850 is looking pretty darn attractive at $170 now, considering that you can't even buy a GeForce 8800-series card (256M or 512M) for under $250 on the Egg right now.
Nvidia will not spend money on R&D unless ATI trumps it. It was made plain and simple in 07 and continues in 08. If ATI went under right now they would probably not release anything in 08. I wonder what their engineering department is doing or has been doing the last year.
GeForce 9800 GTX in March or Feb eh ?
I wonder how many SP it will carry? Some on B3D think 192 SP @ 2Ghz.
For those waiting for a new architecture, why would do think a new would come ?
Seems that you are either new to this or never really bothered to read the evaluations of new cards. New architectures appear, when:
1) The previous one was a failure and thus new cards need to based on something better. This means shorter time frame between new architectures.
2) The process goes as usual, where the current architecture is successful and can be used to base, at most, the next couple generations, while R&D efforts are being done in the background, for something new, that should debut in 2-3 years.
So just over a year as passed, since NVIDIA revolutionazed the graphics industry with G80 and it was and still is, quite successful. I'm sure NVIDIA is already focusing resources on a new architecture, but the next couple of generations, including the 9 series, will still be based on G80 and its derivatives i.e. G92.
Anyway, on to the article, this is a roadmap that doesn't surprise me. The lack of competition from AMD, doesn't really forces NVIDIA to do anything. Obviously, we, as consumers, don't like it. I sure am one. But from a money making company perspective, this is the perfect time to make huge profits, while not investing a whole lot of money. I also believe that, if NVIDIA had the chance to crush AMD/ATI, they would, but why would they, if this exact situation, let's them keep making profits, without much costs ?
As Kyle suggested, and if R700 is indeed scheduled for early this year, NVIDIA is surely holding back some cards. Given their advantage in the GPU market, they can afford to play the waiting game and surprise the competition with something much better.