NVIDIA Rushing in Stopgap HD 7970 Competitor This February?

[btarunr]

AMD's Radeon HD 7970 seems to have ruffled a few feathers at NVIDIA and it looks like the green team doesn't want too much market exposure for it. A fairly-reliable source at ChipHell learned that NVIDIA's GeForce "GTX 680" part could be launched some time in February. The source says that this part could be competitive with the HD 7970, though not exactly NVIDIA's fastest next-generation GPU in the works. So it has to be something other than the GeForce Kepler 110, that's reportedly slated for March-April. At least the tiny pieces of specifications trickling out seem to reinforce this theory. Graphics cards based on this part apparently have 2 GB of memory, and its core clock speed is reported to be 780 MHz.



View at TechPowerUp Main Site

 

[blibba]

Gtx560x2? Gtx585?

 

[Lionheart]

I might just hold off with the HD7970 for now ^_^

 

[Aceman.au]

Hold off until the 680 comes out. Wait for price drop!

Might have to go ATI again this series although I wanted to try Nvidia... Nvidia will come out overpriced and ATI will once again drop prices...

EDIT: The retarded battle continues

 

[xenocide]

I might just hold off with the HD7970 for now ^_^

I don't see why anyone wouldn't wait. If the HD7970 is the only thing on the market, the price is inflated since it faces no competition. Once Nvidia releases some options, and AMD gets the rest of their line out, the prices will stabalize and you will have time to find the best card in terms of Price\Performance in this round. Jumping on the HD7970 right away risks overpaying for a potentially underperforming product. Plus, GPU's get better with Driver releases, so buying a card right at launch tends to be problematic...

Looks like Nvidia are having issues... (just an observation)

By what reasoning? They are waiting until they are adequately stocked and able to replace their current offerings, I see no "issues" with this. Nvidia has nothing to lose by waiting. AMD rushed their cards out just to get some form of revenue going into the new year.

 

[arnoo1]

Also nvidia doesn't want a papet launch and that a good thing!
I can wait

 

[Red_Machine]

TBH, I'd rather nVIDIA waited until they were ready before releasing. Don't want another Fermi mk1 fiasco...

 

[JrRacinFan]

The retarded battle continues

Let them battle. Its good for the consumer/end user.

 

[Aceman.au]

By what reasoning? They are waiting until they are adequately stocked and able to replace their current offerings, I see no "issues" with this. Nvidia has nothing to lose by waiting. AMD rushed their cards out just to get some form of revenue going into the new year.

I just heard somewhere Nvidia was having trouble with power consumption/heat on this series that's all.

Probably should've quoted it... But I can't find it.

 

[Lionheart]

I don't see why anyone wouldn't wait. If the HD7970 is the only thing on the market, the price is inflated since it faces no competition. Once Nvidia releases some options, and AMD gets the rest of their line out, the prices will stabalize and you will have time to find the best card in terms of Price\Performance in this round. Jumping on the HD7970 right away risks overpaying for a potentially underperforming product. Plus, GPU's get better with Driver releases, so buying a card right at launch tends to be problematic...

I'm really tight with money right now bro, I want the card bad but gotta make do without ^_^

 

[20mmrain]

The way this sounds is that we have another GTX 480 on the way. Not the monster GTX 780 we have all heard about. Everything is showing that this will not be the big "Worlds fastest GPU we have heard about"
If you ask me if this is true the HD 7970 will still have a leg up. The odds are (like the GTX 480) this GPU will be a rushed version that runs way too hot and does not have the whole chip unlocked. If it is able to beat the HD7970 it will be barely and we will have another HD5870 vs GTX 480 competition on our hands.
I also think if this move turns out to be true that Nvidia's GTX 780 part is still a lot further off than April and closer to fall 2012 like we originally heard about.
This is all just speculation..... but also looking at history I think I might be on to something hear.

 

[_JP_]

Rushing is rarely a good idea. Especially in this market. Let's see what comes out of it.

 

[Red_Machine]

Not the monster GTX 780 we have all heard about.

Why does everyone seem to think it's going to be the 780? I've heard nothing so far that would make me draw that conclusion.

 

[Benetanegia]

This is probably GK104. I think all the confusion comes from the fact that some media talks about Kepler as in GK110, and only GK110, just like they said Fermi to refer to GF100, almost constantly forgetting that GF104/106/etc. were Fermi cards too.

It may also be essentially true this time around. GK110 may actually be the only 100% Kepler part, while the smaller chips are Fermi-Kepler hybrids. Or maybe simply Fermi based cards, with different amount of shaders/rops/tmu.

For example due to the description by chiphell, this could easily be a "GF104 based" (SM with 48 SP) card with 1 extra GPC. Final specs:

576 SP
96 TMU
32 ROP
256 bit - coupled with 5000+ Mhz GDDR5, it could offer enough bandwidth without becoming a massive bottleneck.

I just heard somewhere Nvidia was having trouble with power consumption/heat on this series that's all.

Probably should've quoted it... But I can't find it.

A 5 euro bill here says you heard it at Semiaccurate, or from somewhere quoting it.

The way this sounds is that we have another GTX 480 on the way. Not the monster GTX 780 we have all heard about. Everything is showing that this will not be the big "Worlds fastest GPU we have heard about"
If you ask me if this is true the HD 7970 will still have a leg up. The odds are (like the GTX 480) this GPU will be a rushed version that runs way too hot and does not have the whole chip unlocked. If it is able to beat the HD7970 it will be barely and we will have another HD5870 vs GTX 480 competition on our hands.
I also think if this move turns out to be true that Nvidia's GTX 780 part is still a lot further off than April and closer to fall 2012 like we originally heard about.
This is all just speculation..... but also looking at history I think I might be on to something hear.

Makes no sense at all. Honestly, do people really think that Nvidia can't release a card 20% faster than previous generation, especially considering how troublesome was the previous generation?

Like I'm saying in any related post in the hopes of people gaining some common sense. GF110 is a 3 billion transistor chip and it's close to 4.3 billion transistor Tahiti performance wise (in fact it's just as fast clock for clock). Nvidia does not need to add GPGPU stuff that costs many transistors, unlike AMD that added 1:2 SP : DP ratio, 1D shaders, memory management stuff. Fermi already had all of this, Kepler is mostly evolutionary rather than revolutionary, so do people really expect Nvidia to release a 4-5 billion transistor chip that can hardly compete with their own previous generation? Really?

Also GTX780 name makes little sense. Granted GTX680 doesn't make a lot of sense for the upper mid-range/performance chip, but AMD made that move first with HD68xx and it's following suit with HD7xxx, so maybe Nvidia is just adjusting. Marketing guys are like that. GK110 will probably be named GTX690.

 

[Hokum]

TBH its going to have to either be a new cooler 580 which clocks higher with less power else everyone will just laugh at it, even IF it is faster...

The 7970 has combined performance and power saving like never before. Nvidia now have to compete on two fronts not just performance.

 

[entropy13]

I just heard somewhere Nvidia was having trouble with power consumption/heat on this series that's all.

Probably should've quoted it... But I can't find it.

Technically, you're correct that Nvidia is "having trouble"....IF AND ONLY IF you compare them to AMD's offering(s). But there's actually some improvement in that aspect for Kepler compared to Fermi.

 

[20mmrain]

This is probably GK104. I think all the confusion comes from the fact that some media talks about Kepler as in GK110, and only GK110, just like they said Fermi to refer to GF100, almost constantly forgetting that GF104/106/etc. were Fermi cards too.

It may also be essentially true this time around. GK110 may actually be the only 100% Kepler part, while the smaller chips are Fermi-Kepler hybrids. Or maybe simply Fermi based cards, with different amount of shaders/rops/tmu.

For example due to the description by chiphell, this could easily be a "GF104 based" (SM with 48 SP) card with 1 extra GPC. Final specs:

576 SP
96 TMU
32 ROP
256 bit - coupled with 5000+ Mhz GDDR5, it could offer enough bandwidth without becoming a massive bottleneck.



A 5 euro bill here says you heard it at Semiaccurate, or from somewhere quoting it.



Makes no sense at all. Honestly, do people really think that Nvidia can't release a card 20% faster than previous generation, especially considering how troublesome was the previous generation?

Like I'm saying in any related post in the hopes of people gaining some common sense. GF110 is a 3 billion transistor chip and it's close to 4.3 billion transistor Tahiti performance wise (in fact it's just as fast clock for clock). Nvidia does not need to add GPGPU stuff that costs many transistors, unlike AMD that added 1:2 SP : DP ratio, 1D shaders, memory management stuff. Fermi already had all of this, Kepler is mostly evolutionary rather than revolutionary, so do people really expect Nvidia to release a 4-5 billion transistor chip that can hardly compete with their own previous generation? Really?

Also GTX780 name makes little sense. Granted GTX680 doesn't make a lot of sense for the upper mid-range/performance chip, but AMD made that move first with HD68xx and it's following suit with HD7xxx, so maybe Nvidia is just adjusting. Marketing guys are like that. GK110 will probably be named GTX690.

I never said they couldn't.... (and the HD7970 is not always just 20% faster I own two trust me. Sometimes much more.) what I am saying is that they are struggling. They shouldn't had any problem with releasing the GTX 580 correct the first time either but they did. Instead we got the GTX 480 which wasn't that much faster and ran hot as hell. Again we are in the same situation..... a new architecture Nvidia is releasing, AMD again released first, So Nvidia is rushing. Hey I love Nvidia.... it just seems to me this is what is going on.
As a matter of fact I remember having a similar argument with you last time. Right before the GTX 480 released. Look what happened.
Look I hope your right because that means I will have two more video cards to play with very soon. But it just seems like Deja Vu

 

[Hokum]

IMHO its starting to look a lot like intel vs amd but the other way around with AMD the top dog.

I'm an ATI man but I have nothing against Nvidia. I just ATi as i'm familiar with them.

 

[entropy13]

IMHO its starting to look a lot like intel vs amd but the other way around with AMD the top dog.

For the AMD v. Nvidia battle to look like Intel v. AMD, and with Nvidia "losing", Nvidia first have to release high-end cards that only competes with AMD's mid-range cards. Then those high-end cards must be priced roughly the same as AMD's high-end cards as well. Nvidia also have to start making their drivers worse than they are right now.

 

[Hayder_Master]

I think GTX680 will be another fail like GTX480 and GTX590.

 

[Benetanegia]

I never said they couldn't.... (and the HD7970 is not always just 20% faster I own two trust me. Sometimes much more.) what I am saying is that they are struggling. They shouldn't had any problem with releasing the GTX 580 correct the first time either but they did. Instead we got the GTX 480 which wasn't that much faster and ran hot as hell. Again we are in the same situation..... a new architecture Nvidia is releasing, AMD again released first, So Nvidia is rushing. Hey I love Nvidia.... it just seems to me this is what is going on.
As a matter of fact I remember having a similar argument with you last time. Right before the GTX 480 released. Look what happened.
Look I hope your right because that means I will have two more video cards to play with very soon. But it just seems like Deja Vu

Fermi and Kepler is like comparing apples to oranges* and who exactly says that Nvidia is having troubles or that they are late? No one really knows crap about this generation from Nvidia. The only bad news are originating at Semiaccurate and related sites, as always have (with different outcomes), and according to them, the chips that are going to be released in February-April timeframe have not even been taped out yet, which means that they will not release it until 2013, ha... ha... ha... Really don't assume anything you've read, not even good things and remember the days prior to 8800 too for your "Deja Vu" argument. All I'm saying is that there's as much of an evidence for Kepler being very good as there is for Kepler being a bad chip. And even the "bad" GF100 was faster than GTX295.

Everyone seems to prefer the doom and gloom version, so I'm just offering the other side of the coin, for balance and some common sense.
So following that logic of seing things from a wider perspective, they are NOT rushing anything from what I can tell, not with this chip anyway. The plan, at least according to rumors, has always been to release the performance part first and that's what they are doing. They want to release when there's enough stock to warrant a hard launch and that's what they are doing. Where's HD7970 really? Sure it's been officially released, but can you easily find one in stores? Nope. Where's HD7950? To me it looks like the one rushing was AMD, while Nvidia might be just stockpiling. Even with 6 months lead with HD5870, AMD didn't sold much more 40 nm cards than Nvidia by the time that generation ended. IN the worst case scenario, releasing second, but with a faster product ended up paying up (also take Quadro and Tesla into account). Why should they rush and try to release sooner when they don't need to?

*Fermi (mostly GF100) looked bad when compared to AMD offerings at the time, but it was a chip that added a lot of GPGPU stuff, and a lot of tesselation power. Sorry to bring the same argument again, but please look at Tahiti for a second. 4.3 billion transistor and only 30% faster than Cayman, 2.7 billion. Also only 50% faster than Cypress with more than double the transistors. GPGPU is goddamn expensive and Tahiti is as much of a proof as Fermi was back then. But think about that again. GTX580 is what Nvidia tried 2 years ago, on a new and problematic 40nm node. They tried a chip that is within 20% performance of Tahiti, 2 years ago, in a process that allows half the ammount of transistors as 28nm. That is why GF100 failed, but it did pay off in the end.

 

[qubit]

This does indeed sound like the GK104 that's been mooted recently. Should be interesting.

 

[treboRR]

Fermi and Kepler is like comparing apples to oranges* and who exactly says that Nvidia is having troubles or that they are late? No one really knows crap about this generation from Nvidia. The only bad news are originating at Semiaccurate and related sites, as always have (with different outcomes), and according to them, the chips that are going to be released in February-April timeframe have not even been taped out yet, which means that they will not release it until 2013, ha... ha... ha... Really don't assume anything you've read, not even good things and remember the days prior to 8800 too for your "Deja Vu" argument. All I'm saying is that there's as much of an evidence for Kepler being very good as there is for Kepler being a bad chip. And even the "bad" GF100 was faster than GTX295.

Everyone seems to prefer the doom and gloom version, so I'm just offering the other side of the coin, for balance and some common sense.
So following that logic of seing things from a wider perspective, they are NOT rushing anything from what I can tell, not with this chip anyway. The plan, at least according to rumors, has always been to release the performance part first and that's what they are doing. They want to release when there's enough stock to warrant a hard launch and that's what they are doing. Where's HD7970 really? Sure it's been officially released, but can you easily find one in stores? Nope. Where's HD7950? To me it looks like the one rushing was AMD, while Nvidia might be just stockpiling. Even with 6 months lead with HD5870, AMD didn't sold much more 40 nm cards than Nvidia by the time that generation ended. IN the worst case scenario, releasing second, but with a faster product ended up paying up (also take Quadro and Tesla into account). Why should they rush and try to release sooner when they don't need to?

*Fermi (mostly GF100) looked bad when compared to AMD offerings at the time, but it was a chip that added a lot of GPGPU stuff, and a lot of tesselation power. Sorry to bring the same argument again, but please look at Tahiti for a second. 4.3 billion transistor and only 30% faster than Cayman, 2.7 billion. Also only 50% faster than Cypress with more than double the transistors. GPGPU is goddamn expensive and Tahiti is as much of a proof as Fermi was back then. But think about that again. GTX580 is what Nvidia tried 2 years ago, on a new and problematic 40nm node. They tried a chip that is within 20% performance of Tahiti, 2 years ago, in a process that allows half the ammount of transistors as 28nm. That is why GF100 failed, but it did pay off in the end.

WOW u really hit the nail! Really good point of view! I like!

 

[ensabrenoir]

The stop gap gotta have an incredible low price otherwise ...whats the point. Everyone will just wait. Better idea is to release super overclocked current gen cards at deep dis,ounts. Make amds new seem close to thier current gen. Down side is it gives off a were not prepared vibe

 

[the54thvoid]

Why do so many folk have to put on either red or green socks to discuss this? Even someone saying, "I'm just pointing out facts" is tending to lean to one side.

When NV paper launched the 480 they were ridiculed by many (as it was already running late). AMD have near enough paper launched the 7970 as availability is low for many brands. Also, the GTX 580 as it is is what, 12-13 months old?
People talk as if Nvidia is down and that couldn't be farther from the truth. They still have a card that performs within -5 to 30% (review & game dependant) of a card that is a year later and on a far smaller production node.
Think about it for a second. The GTX 580 was launched in retail December 2010 on a 40nm node. It runs fairly cool for it's performance and without too much noise. Now we are a year later, with a production size that is 30% smaller than 40nm (for the maths bastards out there 40-28=12 and 12/40 =0.3) and it performs about 0-30% better, again situation dependant.

Do people really have such red coloured vision that they believe NV can't release a 28nm card that will equal the 7970 in performance and not run hot?
GF100 was resurrected by GF110. Some compute functionailty (what Ben's talking about) was dropped and it resulted in a manageable 512 core card.

To counter Ben though the original rumours suggested a bottom up launch for NV to get the process right. The last time I checked, the Kepler behemoth gpu isn't scheduled to arrive until 2013. GK 104 is the card for now (a high end but not top end single card).

But folk are also then forgetting AMD do plan a dual Tahiti card in Q1 2012 which will be enormously powerful..and expensive.

So, to summarise, just lets wait and see what happens. We have two tech companies still vying to be the best, so let them fight it out and we'll all benefit.

And yes, my 7970 purchase plans are now on hold (was waiting for water blocks anyway!).