Kepler Nvidia GeForce GTX 780

[3lfk1ng]

Home of the world’s fastest Graphics Card for 853 days and counting.

Anyone want to take a guess when this will be taken down? AMD might not let Nvidia take the title

[TRANCEFORMER]

New Pics:

Gigabyte

gigabyte_gtx680_1.jpeg
gigabyte_gtx680_7.jpeg
gigabyte_gtx680_8.jpeg
gigabyte_gtx680_4.jpeg

http://en.expreview.com/2012/03/21/g...ted/22042.html


MSI:

msi_gtx680_lpl04.jpg
msi_gtx680_lpl03.jpg
msi_gtx680_lpl01.jpg
msi_gtx680_lpl02.jpg

http://en.expreview.com/2012/03/21/m...red/22047.html

[GoldenTiger]

New Pics:
[/url]

Thanks but I've seen them all before in this thread.

[Motiv]

Home of the world’s fastest Graphics Card for 853 days and counting.

Anyone want to take a guess when this will be taken down? AMD might not let Nvidia take the title

well not until something beats the 6990.

Saying that, I'm quite looking forward to NVidias new card. If they have a mid-range card, that takes less fuel, is quieter, costs the same and is att the same performance as an AMD chip, I'll be upgrading.

only time will tell.

[Akrian]

Well it would've been the greatest fail ( such as Bulldozer) if 680 failed. I mean it's been almost 3 months since release of 7970, Nvidia was bound to make something faster to justify the time it took them to come up with the answer. At any rate, I'm more interested as to how good of an OC this card is, and what performance gain will the OC show. 7970 can go +300 mhz without too much of a hustle, giving a solid boost over the base 925mhz core clocks and a nice performance increase as well.

If Nvidia released the card 1.5 months earlier and had 3gb memory with it.....Witcher 2 pushed 1.5gb v-ram at uber settings at 1080p resolution ( one of the reasons I switched from my 580 3-way sli setup). I bet at 3-monitor setups with 5760 x 1200 res, those 2gb wont be enough either =\. On the other hand gratz greens on getting their new toy.

[Charles Wirth]

All THG links removed by request of THG, so please no more THG links.

[The Jesus]

Really now? No need to try to claim someone's lying/trying to intentionally mislead here.

Fine, pretend it's 24%... that's still a lot less than 40-50% over the 580. My point stands just fine without all the hyperbole. Why do you think the GTX 680 will suffer at 2560x1600 exactly? It has the same bandwidth as a GTX 580. Even at those #'s it's still 21% faster than a Radeon 7970 if the very-real-looking leaks pan out.

Umm, because the 680 has 2GB of VRAM opposed to the 7970's 3GB...

All THG links removed by request of THG, so please no more THG links.

Lol, not surprised.

[pr@$r1g]

LOL GTX 680, any how if its priced at 499 i will not buy, again after 6months there would be a big price drop killing even the resale value like it happened with GTX470.

Its better to go with 7870..... everything playable at expected performance.

[gumballguy]

Why has everyone forgotten that MSI has confirmed a 1500mhz 7970 card? That's a heck of a lot faster than stock.

Obviously, nvidia's overclockability needs to be considered. But I don't think anyone in their right mind expects a 50% overclock from nvidia (we shouldn't have expected it from AMD either).

End result: All these "amd has no answer" comments are rubbish.

Looking forward to the reviews and the response from amd. And it's about time for me to buy, so I'm quite prepared to buy nvidia if proves better than amd at 3*1920*1080.

[UrbanSmooth]

Why has everyone forgotten that MSI has confirmed a 1500mhz 7970 card? That's a heck of a lot faster than stock.

Obviously, nvidia's overclockability needs to be considered. But I don't think anyone in their right mind expects a 50% overclock from nvidia (we shouldn't have expected it from AMD either).

End result: All these "amd has no answer" comments are rubbish.

Looking forward to the reviews and the response from amd. And it's about time for me to buy, so I'm quite prepared to buy nvidia if proves better than amd at 3*1920*1080.

The 7990 (dual card) is coming in April.

As a triple surround gamer (possibly future 5x1 120/240hz) , I'm looking to the big guns, not the middle dog. So, for me, it's three GTX 780 (685s?) with 4GB or more, all water-cooled, of course.

Looks like this upgrade is coming late 2012, early 2013.

[tajoh111]

Links to such a card? On air.

http://www.hardocp.com/article/2012/..._card_review/3

They only reached 1190mhz with over volting and this is a MSI lightning series which has some of the best electronics components and cooling out there ?

1500 mhz if actually reached, how stable is it, how many hundreds of 7970 did they go through to find that cherry picked sample. Because from what I have seen, even on forums, 1500 on air is a number I haven't seen and definitely not stable.

[The Jesus]

Links to such a card? On air.

http://www.hardocp.com/article/2012/..._card_review/3

They only reached 1190mhz with over volting and this is a MSI lightning series which has some of the best electronics components and cooling out there ?

1500 mhz if actually reached, how stable is it, how many hundreds of 7970 did they go through to find that cherry picked sample. Because from what I have seen, even on forums, 1500 on air is a number I haven't seen and definitely not stable.

Reviewers never push the cards as far as we do. They could only get 1425 on the memory? That's complete BS. People hit 1800+ (given that is the error correction zone where its negating the benefits) all the time on air.

[turtle]

Maybe Nvidia will wait until 8000 series release then the gk110 will come out as gtx780 and gk104 will be rebadged to gtx760 ...

I think most of us agree that will be the end-game...subject to release order.

It could be something like:

gk104/106/107 are released. (Q1/2)

AMD releases Trinity smack dab between gpu generations. (Q2/Q3)

AMD replaces Tahiti with gk104-like chip. (Q3/4)

nvidia releases gk110, rebrands everything...maybe respins and/or faster/lower power gddr5 for skus. (Q3/4)

AMD releases high-end chip and/or dual-chip solutions to compete with gk110 and low-end chip similar to gk106. (Q4-Q1)

AMD releases Kaveri...which IGP 'oddly' matches a rebrand of 7750. (Q2-Q3)

Rinse. Wash. Repeat.

I totally understand that TSMC's 28nm is a uniquely hot commodity that needs time to ramp/grow in production resources, and it is probably the first in a new reality of how products will come to market on a new process. I also fully respect and appreciate what both companies are doing for efficiency and yields (perhaps because of that). That said, it kind of makes one yearn for yesteryear when big chips caused a quicker turn-over of greater amounts of logic lower down the food-chain faster. Where-as before we may have had a 7900ish/gk104 battle at the mid-range in the coming months, we are now forced to wait another whole product cycle for the exact same performance(~/watt), perhaps in a slightly different configuration, at a more palatable price tag. It seems clear nVIDIA was ready to have that fight now and AMD, whom typically thrives on that concept, took the former nvidia route to maximize profit...nvidia ain't gonna argue and will revert back to taking the inflated income. It makes me ponder how much of the decision-making is yield-oriented, and how much is marketing/capitalization knowing they have to use the same process for two cycles. Each product supports and refutes the opposite's theory not only of each other, but of their company's own past.

The 7900 series is (imo) directly influenced by both, but which is more influential is the question. Sure, you have a bigger chip (with tahiti) with more redundancy for higher yields, faster time to market, clockspeeds that beat the old competition and have some higher-end features (for high-rez, gpgpu) the midrange won't, but on a pure gaming logic level you also have the clockspeed and real tdp (188w at stock, ~225w overclocked under avg load) that make it clear they will sell a similar product sporting similar power consumption and less pci-e connectors down the line as an 'improvement in efficiency'...or pretty much exactly what gk104 is.

In the end, I blame Gordon Moore and nVIDIA's past over-zealous engineering for my jaded outlook.

[SKYMTL]

Reviewers never push the cards as far as we do. They could only get 1425 on the memory? That's complete BS. People hit 1800+ (given that is the error correction zone where its negating the benefits) all the time on air.

No. The error correction zone is fluid from one ASIC to the next and is based upon the overhead differences between one set of ICs and another. Even if the memory starts throwing errors at a 4MHz overclock, it will keep error correction in place throughout every subsequent clock speed increase.

[tajoh111]

Reviewers never push the cards as far as we do. They could only get 1425 on the memory? That's complete BS. People hit 1800+ (given that is the error correction zone where its negating the benefits) all the time on air.

They probably got a bad sample or average sample. They got 1300mhz with the gigabyte, so it obvious they know how to overclock as much as the average consumer. MSI I doubt wants to push this as hard, as most consumers who overclock do. That just means a future RMA request is far more likely.

Is this where the 1500mhz rumor came from?

http://www.xtremesystems.org/forums/...t=#post5070673

I posted an update about the 7970 Lightning: the card will be 3GB memory with 1500 MHz rated chips, the cooler will be new generation Twin Frozr 4, it will be the second-fastest clocked custom 7970 (so maybe Asus or Gigabyte will have a faster 7970) and, the best of all, the hole in the backplate has a cool purpose: there you mount something called "GPU Core reactor", esentially a small PCB with tantalum caps that should provide adittional filtering for the GPU core voltage; if you want to use multiple cards though you need to remove it, as it takes some space.

If people are referring to this, this is obviously the memory and not the chips themselves. As googling 1500mhz on air yields nothing. Even on water, I haven't really heard 1500mhz really being thrown around. This doesn't change that 1500mhz clocked 7970's are not an option as a sell-able product which is what I am primarily talking about. It would take 1 in thousands binning which is not an option.

[To(V)bo Co(V)bo]

I think we should all remember that its been known for a while nvidia wasnt planning to release their high-end till september (end of year) time frame. IMHO I think gk110 is as late, as fermi was late to market. Im sure Gk110 was suppose to be released now or even earlier on the roadmap. Nvidia isnt holding back a gk110, it just plain isnt ready to go yet. Nvidia just got EXTREMELY lucky with the performance of gk104, and the underwhelming power of ati's 7 series . Its simply amazing what they were able to pull off with just a performance based design. Im just as sure though that ati has another design ready go when needed. The 7 series was intended for 32nm not 28nm, ati didnt want to scrap a whole generation of products that have been worked and developed for months if not years and lose revenue. If anything the 7 series is a bonus generation to ati. It buys them alot of time to rework the intended 28nm designs.

[paulbagz]

Can't wait to bury this thread lol.

-PB

[TurboDiv]

Release already. Im tired of my 580. I want to bully some 7970 owners and troll the hellz out of them with this new weapon !

[DilTech]

Why has everyone forgotten that MSI has confirmed a 1500mhz 7970 card? That's a heck of a lot faster than stock.

Obviously, nvidia's overclockability needs to be considered. But I don't think anyone in their right mind expects a 50% overclock from nvidia (we shouldn't have expected it from AMD either).

End result: All these "amd has no answer" comments are rubbish.

Looking forward to the reviews and the response from amd. And it's about time for me to buy, so I'm quite prepared to buy nvidia if proves better than amd at 3*1920*1080.

... MSI confirmed a card with 1500mhz ram, NOT 1500mhz core. If ANY partner pulled off a 1500mhz CORE 7970 as an actual card for sale on newegg, I will personally bake them cookies and WALK them to their place of business. From there, I will go get them a glass of milk. Why would I do all of this? Because clearly they are santa claus himself.

[gumballguy]

Links to such a card? On air.

Is this where the 1500mhz rumor came from?
http://www.xtremesystems.org/forums/...t=#post5070673

The information didn't describe whether it was on air or not IIRC.

I dont think it was from that thread. I've been hunting for it. It was a photograph of a printout (company letterhead kinda paper, it looked reasonably legit to me) with a list of cards they would be releasing. It detailed clocks and video output ports for each card.

It was a thread on xs, because I don't read other sites. But I just can't find the bloody thing now.. wondering if it could have been removed.

(Very easy to remember this particular pic, and no I dont believe I'm mixed up with the mem clock. I thought it was an absurd value and was stoked to see it).


Edit: still trying to find it but running out of lunch break. So I give up. Just assume I'm wrong.

[Solus Corvus]

Just assume I'm wrong.

Ok.

[tajoh111]

I think 7xxx series will buy them some time but they are still stuck with GCN. I have a feeling that AMD will needs to exceed the 400nm2 and jump into or near the 500nm range to compete with gk110, which might be against their design philosophies. If Gk104 performance indicates anything, we might get a gk110, that is 40-60% faster than a 7970; that is if gk104 is 10% faster in real world testing than the 7970 and the die size of gk110 is 60-70% bigger. GCN is AMD's new architecture and it is here to stay at least one more generation. It is called GraphicsCoreNext for a reason.

Increasing the frequency will help of course, but 1305mhz 7970 consume in excess of 300 watts and 62 more watts than a power hungry gtx 580. The biggest deterrent for high clocks is chip reliability over time and binning of course. Also there have been rumors of Keplar clocking quite, similar to what 7xxx has been doing so this won't help that much either. Considering the high clocks at stock voltage, i think this rumor is actually likely.

http://www.hardocp.com/article/2012/..._card_review/9

A 360mm2 high clocked a card would probably still lose to gk110. Losing the performance per watt/die size at the high end has alot of long term consequences. Unlike the lower end, they can't rip out unnecessary components for GPU compute because they still want to tackle this market. A big performance based pure gaming part(uncompromised gaming performance with no GPU compute features) doesn't make sense for any company. When your competitor has the more efficient architecture and is willing to make chips much bigger than you, it will be crazy hard to catch up. AMD has to either have another architecture altogether or be willing to make big ass chips like Nvidia. I would happy with scenario 2 but I don't think AMD will do this.

Instead AMD may be more content with second place again and I think is a more likely scenario, try to compete for the bang for your buck area and avoid the monolith. AMD's market share is much smaller than Nvidia's in the professional market and without the assurance of revenue coming from the professional market(as Nvidia is the industry standard here) there is much less incentive to make the monolith monster dies which funds these projects and cards.

Unless they pull a shark out of their ass or GK110 simply is a terrible products which loses everything that makes gk104 run well, 8970 could be very much slower than the GK110. If Fermi wasn't rushed and gtx 580 was the chip released initially, the 6970 would obviously still be slower even though it was second generation. From what it sounds like, with a product to sell, Nvidia is in no rush like Fermi to get gk110 on the market. Partners have something to sell in the form of gk104, gk107 and I am not sure about gk106. I can imagine Nv are likely trying to get it right the first time as their competitors don't even have their professional market products out and they would rather avoid another gtx 480 debacle(henced why gk100 was scrapped). I don't think anyone is sure what AMD will do at this point, but there is no simple solution and I have a feeling there is going to be a lot of compromises made by the company and hopefully price drops this generation.

All this speculation is based on a lot of rumors but I am also following the history of what both companies have done in the past and of course chip planning is done years in advance.

[TurboDiv]

Look at its eyes for 5 seconds....



Do you feel the power ?

March 22...

[DStealth]

Nice scores.

Wake me up when 2560x1600 OC and max OC results vs 7970 OC start appearing, then we will see the real champion

+1
And min FPS, not only Average ...

@TurboDiv : LOL ... you filthy canadian

[zerazax]

I think 7xxx series will buy them some time but they are still stuck with GCN. I have a feeling that AMD will needs to exceed the 400nm2 and jump into or near the 500nm range to compete with gk110, which might be against their design philosophies. If Gk104 performance indicates anything, we might get a gk110, that is 40-60% faster than a 7970; that is if gk104 is 10% faster in real world testing than the 7970 and the die size of gk110 is 60-70% bigger. GCN is AMD's new architecture and it is here to stay at least one more generation. It is called GraphicsCoreNext for a reason.

Increasing the frequency will help of course, but 1305mhz 7970 consume in excess of 300 watts and 62 more watts than a power hungry gtx 580. The biggest deterrent for high clocks is chip reliability over time and binning of course. Also there have been rumors of Keplar clocking quite, similar to what 7xxx has been doing so this won't help that much either. Considering the high clocks at stock voltage, i think this rumor is actually likely.

http://www.hardocp.com/article/2012/..._card_review/9

A 360mm2 high clocked a card would probably still lose to gk110. Losing the performance per watt/die size at the high end has alot of long term consequences. Unlike the lower end, they can't rip out unnecessary components for GPU compute because they still want to tackle this market. A big performance based pure gaming part(uncompromised gaming performance with no GPU compute features) doesn't make sense for any company. When your competitor has the more efficient architecture and is willing to make chips much bigger than you, it will be crazy hard to catch up. AMD has to either have another architecture altogether or be willing to make big ass chips like Nvidia. I would happy with scenario 2 but I don't think AMD will do this.

Instead AMD may be more content with second place again and I think is a more likely scenario, try to compete for the bang for your buck area and avoid the monolith. AMD's market share is much smaller than Nvidia's in the professional market and without the assurance of revenue coming from the professional market(as Nvidia is the industry standard here) there is much less incentive to make the monolith monster dies which funds these projects and cards.

Unless they pull a shark out of their ass or GK110 simply is a terrible products which loses everything that makes gk104 run well, 8970 could be very much slower than the GK110. If Fermi wasn't rushed and gtx 580 was the chip released initially, the 6970 would obviously still be slower even though it was second generation. From what it sounds like, with a product to sell, Nvidia is in no rush like Fermi to get gk110 on the market. Partners have something to sell in the form of gk104, gk107 and I am not sure about gk106. I can imagine Nv are likely trying to get it right the first time as their competitors don't even have their professional market products out and they would rather avoid another gtx 480 debacle(henced why gk100 was scrapped). I don't think anyone is sure what AMD will do at this point, but there is no simple solution and I have a feeling there is going to be a lot of compromises made by the company and hopefully price drops this generation.

All this speculation is based on a lot of rumors but I am also following the history of what both companies have done in the past and of course chip planning is done years in advance.

That's a ton of speculation on GK110 with little known

Given the efficiency in gaming of the 7870's, and how much smaller it is than the 7970's, its very possible that GK104 to GK110 is a similar situation. GK104 is small and efficient at gaming, GK110 is big but much more compute friendly.

Yes, AMD might have to go into the > 400 mm^2 range, but at the end, they're playing the same ballgame

GCN is new, but that doesn't mean it's inefficient. The 78xx's show that they're quite good, and can be quite efficient power and die size wise if you strip out a lot of the excessive compute stuff. Would not be surprised to see GK104 reviews echoing the same thing