• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

The first "proper" Kepler news Fri 17th Feb?

If it was meant to be their midrange and they upgraded its rank, they could easily undercut the 7970 price and get hordes of people buying it instead.

But that won't happen.
 
If it was meant to be their midrange and they upgraded its rank, they could easily undercut the 7970 price and get hordes of people buying it instead.

But that won't happen.

Not if they feel at the price demand will outstrip production. They may well drop the price when they ramp up production and bring in their higher-end cards.
 
Stop and think how much smaller this is than the GTX 480/580 and this becomes the best NVIDIA launch. Now imagine this same chip scaled up to 580 sizes (i.e. more shaders).

That's what I'll wait for when 28nm production is in full swing and prices have come down.

I'm fully aware of all this and that's why it's so poor, full die size Kepler gk100 would have been awesome, instead we have the worst x80 part ever because it is in fact a x60, it makes fermi look good ffs!
 
+1.

We all know the 7970 clocks to at least 1125Mhz - if the 680 clocks high then we have a decent card en route...
Agreed. Very few (if any) recent NVidia cards have provided less than a 10% overclock. Even GTX480's provided a good 15% or so. Running too close to the red line tends to result in lots of RMA's so I expect there will be some headroom remaining. The fact that the starting point is 1006MHz, it has a smaller die, and the card is rumoured to run cooler than 7970's, indicates that 1100+ should be possible from GK104. It may overclock worse than Tahiti, or it may be better. Personally, I doubt there will be much in it.
 
Last edited:
Yeah, as much as I'm impressed that a mid range part is better than the top end AMD, I'm still thinking I want GK110. When's it slated for? August? THAT will be something that'll kick serious arse in BF3 on my 30" monitor... :D
 
Very few (if any) recent NVidia cards have provided less than a 10% overclock. Running too closed to the red line tends to result in lots of RMA's so I expect there will be some headroom remaining. The fact that thw starting point is 1006MHz, it has a smaller dies, and the card is rumoured to cooler than 7970's, indicate that 1100+ should be possible from GK104. It may overclock worse than Tahiti, or it may be better. Personally, I doubt there will be much in it.

With you. However the magical 1200Mhz clocks that people were getting out of the 7970's might rear their heads again - I noticed that MSI have just released their R7970 Lightning - it's got a default clock of 1070, but with those beefy PCB's and how they bin chips for those cards, 1200Mhz+ isn't entirely impossible...
 
I'm fully aware of all this and that's why it's so poor, full die size Kepler gk100 would have been awesome, instead we have the worst x80 part ever because it is in fact a x60, it makes fermi look good ffs!
I disagree strongly with NVidia's naming for this card, but there have been many worse NVidia cards (the 5800, 6800 and 7800 NVidaa cards spring to mind).

GK104 appears to be a GREAT GPU. However, because of it's potential greatness NVidia seem to be getting the name and price all wrong. They may well diminish their new "greatest ever" graphics card by playing these games. It could have been the best selling, highest performing (relative) mainstream GPU ever. Instead many people will avoid it because it is not a true top end graphics card, even if the preformance figures and crappy pricing indicate otherwise.

In summary, it looks to be a truly great and eveloutional GPU which should push NVidia well ahead of AMD both in architectural a featiure terms. However NVidia are probably going to shoot themselves in the foot by pricing it too high and naming it incorrectly (picking a fight with the wrong opposition). It is like putting the greatest ever Super Middleweight into the same ring as the Heavyweight champion. The 7970 may make it look average.
 
Last edited:
So bassicly from what I can make out, the shader cores can overclock themselves higher (I.E up to 1.4 Ghz) to help even out frame rate with this adaptive V-Sync. Interestinggggg.
 
Agreed. Very few (if any) recent NVidia cards have provided less than a 10% overclock. Even GTX480's provided a good 15% or so. Running too close to the red line tends to result in lots of RMA's so I expect there will be some headroom remaining. The fact that the starting point is 1006MHz, it has a smaller die, and the card is rumoured to run cooler than 7970's, indicates that 1100+ should be possible from GK104. It may overclock worse than Tahiti, or it may be better. Personally, I doubt there will be much in it.

It could be a lot more complex with turbo boost. We might see overclocked 680's with decent max FPS gains when manually overclocked but the min FPS gains to be low because turbo boost dynamic clocking has done that already.
 
So is the HD7970 going to be the VRAM MASTER?? :p

TBH,the VRAM quantities are getting silly now. They seem more like the amounts seen on Professional 3D cards. Now,if only we could hack them like in the past!!:D
 
I disagree strongly with NVidia's naming for this card, but there have been many worse NVidia cards (the 5800, 6800 and 7800 NVidaa cards spring to mind).

GK104 appears to be a GREAT GPU. However, because of it's potential greatness NVidia seem to be getting the name and price all wrong. They may well diminish their new "greatest ever" graphics card by playing these games. It could have been the best selling, highest performing (relative) mainstream GPU ever. Instead many people will avoid it because it is not a true top end graphics card, even if the preformance figures and crappy pricing indicate otherwise.

In summary, it looks to be a truly great and eveloutional GPU which should push NVidia well ahead of AMD both in architectural a featiure terms. However NVidia are probably going to shoot themselves in the foot by pricing it too high and naming it incorrectly (picking a fight with the wrong opposition). It is like putting the greatest ever Super Middleweight into the same ring as the Heavyweight champion. The 7970 may make it look average.

Have to agree to an extent from what I've seen so far it would appear to be what would be a £300 card with an overclock... unless its 25% faster than the 7970 atleast I can't agree with the potential £400+ price tag.
 
Sorry, I was just asking if it isn't inferior, would it still be dumb?

My point being, the benches I have seen (maybe untrue) put the 680 in the lead. This is why I asked you that.

No - if it is the same performance and same price = not dumb
- if it is better performance and same price = not dumb either


My point was that the chap said "same/better/worse for same price" he would get the Nvidia. I was wondering why if it turns out to be inferior he would still get it over the 7970 for example. As I don't buy into the 'ATi/AMD drivers are terribad' argument having had both ATi and Nvidia over the years. Just wondered if there was another reason...
 
GK104 appears to be a GREAT GPU. However, because of it's potential greatness NVidia seem to be getting the name and price all wrong. They may well diminish their new "greatest ever" graphics card by playing these games. It could have been the best selling, highest performing (relative) mainstream GPU ever. Instead many people will avoid it because it is not a true top end graphics card, even if the preformance figures and crappy pricing indicate otherwise.

I'm inclined to agree here...

This GPU at ~$300-$350 (i.e. around £250), named GTX660Ti and targeted as an mainstream GPU, would have sold by the bucketload and would have done wonders for Nvidia's reputation.

Clearly their sales analysis suggested this wasn't the way to go, for whatever reason. Limitied initial supply would be the most obvious factor (no point charging $300 for a product that will clear all your available stock at $550), though there are so many other variables to consider it's impossible to say with any certainty.
 
I wonder how long the GTX600 series will last though?? Will it be a quick stop gap until the GTX700 series are launched??

I suspect that GK104 (i.e. the current GTX680) will receive a re-branding to "GTX 760Ti" once GK110 is released.

So yes - I imagine that the 6-series will have a fairly short lifespan (~6 months or so).
 
Back
Top Bottom