Gigabyte GTX680 retail pictures

[SolMiester]

Really, dude ? The 580 came 6 months after the 480. So using that as a baseline the 780 as GK110 could come in Sept. So my saying it could take till then or maybe 3 months later qualifies me as a shill and AMD PR ? I'm hoping for a hail mary and nv having made some sort of breakthrough and delivering a perf. increase consistent with their past deliveries with the 680, because I am dying for an upgrade in my machine. I must be a lousy shill then, or else I'd have my free 7970s from AMD and wouldn't care.

I'd report that crap, but people are getting banned left and right for that stuff now. Too many more and there will be no one left here to disucss anything with, so whatever.

But, really dude ? lol

Groove, no one said you were anything...I SAID that what a shill or PR guy might say...
2013 is not 3 months away either....so yeah...really dude!

 

[Imouto]

Man, if that OCN thread holds any truth this card is going to be such a letdown.

 

[Tempered81]

Man, if that OCN thread holds any truth this card is going to be such a letdown.

It's not a letdown. It's faster than the 7970, smaller and more efficient, ie a remarkable winner. Even clock-for-clock against the 7970, I'll wager it still has at least a 5% lead.

7970 OC at 1005/1500 (default CCC)

GTX680 stock at 1006/1500 with boost

7970 OC at 1005/1500 (with AMD optimized Tesselation selected in CCC)

http://www.xtremesystems.org/forums...orce-GTX-780&p=5071228&viewfull=1#post5071228

 

[BallaTheFeared]

It's just showing Nvidia is stronger with Tess, which isn't at all surprising given how badly GF100 dominated 5xxx and is even better than 6xxx at it.

I don't think we're going to see the 40% advantage here play out in many other titles.

I'd actually really like to see both run no tess at all so we can get an idea of which card is faster with textures since that's all that's left pretty much.

If Nvidia can beat AMD in the texture area, they'll one up AMD in many of the Gaming Evolved titles.

 

[n0x1ous]

It's just showing Nvidia is stronger with Tess, which isn't at all surprising given how badly GF100 dominated 5xxx and is even better than 6xxx at it.

I don't think we're going to see the 40% advantage here play out in many other titles.

I'd actually really like to see both run no tess at all so we can get an idea of which card is faster with textures since that's all that's left pretty much.

If Nvidia can beat AMD in the texture area, they'll one up AMD in many of the Gaming Evolved titles.

I could be wrong, but I thought the HD 7xxx Tess performance was massively improved to the point that it wasn't a clear win for NVidia on that anymore. Which if that were true, would make these results really impressive for the 680 if they are real

 

[Diceman2037]

??

z68 board... The Intel chip might be listed first. It does it with my 9800GT being listed in Heaven, and actually prevented me from running BM:AC in DX11 mode because it was the first listed in device manager.

Also:

Balla is correct, the IGP doesn't have to be in use for certain apps, Heaven x.o included, to display it as the primary device. 3dmark used ot have the same issue till they updated the detection driver to account for adapters with displays connected.

Infact, whether you have restarted after a driver update, particularly if its the first time the card has had drivers installed can bring this out in many applications.

 

[BallaTheFeared]

I could be wrong, but I thought the HD 7xxx Tess performance was massively improved to the point that it wasn't a clear win for NVidia on that anymore. Which if that were true, would make these results really impressive for the 680 if they are real

7970 is faster than the 580 at tess, but it's clear tess is still bottlenecking the 7970 in Heaven, simply lowering it raises fps... So the bottleneck is still the tess units.

Not that the 7970 is 5870 class in that area, the performance of tess on the 7970 is more than enough for current game titles. This however, is low res, extreme tess Heaven, even Crysis 2 tess pales in comparison.

 

[Diceman2037]

Isnt that guy using a Core I3 2100 to test?

no, the op is the one with the i3, lols is using a benching system.

 

[Diceman2037]

The guy is trolling people hard. Notice the long-time member who gave a decent and respectable post used a screenshot of his Heaven result.

The just-joined guy trying to debunk is using the HTML result file opened in his browser, which you can just open up and edit to show whatever you please. I don't get why no one in that thread does not just ask him for a screenshotted result from within the benchmark, or as he is avid to keep offering unrealistic results, a canned Crysis or 2033 bench run.

Notice for his B:AC results how he just gives you a text blurb and no SS. The guy is full of turd. I have an account over there, but I'm too lazy to recover the password and name as I don't remember it. Someone ask him to give something that is a little more concrete and see if he does or avoids it.

I want a real upgrade, but wouldn't let an obvious troll hold my hand, lead me down the garden path and slip it to me in the bushes. :sneaky:

:hmm: except an ex admin from there is confirming the higher score as valid.

 

[Imouto]

It's not a letdown. It's faster than the 7970, smaller and more efficient, ie a remarkable winner. Even clock-for-clock against the 7970, I'll wager it still has at least a 5% lead.

As I said I'm sticking to that thread info. As for me it's the only close to legit sample out there.

 

[Tempered81]

As I said I'm sticking to that thread info. As for me it's the only close to legit sample out there.

Yes me too. This bench comes from that thread

 

[blackened23]

Balla is correct, the IGP doesn't have to be in use for certain apps, Heaven x.o included, to display it as the primary device. 3dmark used ot have the same issue till they updated the detection driver to account for adapters with displays connected.

Infact, whether you have restarted after a driver update, particularly if its the first time the card has had drivers installed can bring this out in many applications.

Intel HD3000 will not appear in your device manager unless are you using lucid virtnu software on a z68 board. It is not in my DM - p8z68 board. Secondly, HD3000 will never show regardless if you run Heaven DX11 in fullscreen (at least it didn't with the older version) I seriously doubt many people use it because its a buggy piece of crap that lowers your discrete GPU performance. (benchmarks on AT or TH prove this, which is why most don't use it)

I dunno, it just seems sketchy. We'll find out soon enough though, if its true it looks like new toys this week for me. Then I will bug balla to figure out the ways of water cooling.

 

[chimaxi83]

Looks like the OP of the thread over at OCN is under a Gigabyte NDA now

 

[blastingcap]

That’s not necessary true. It could be clever engineering and/or manufacturing like Ivy Bridge. Ivy Bridge is faster than Sandy Bridge while using less power.

Is Ivy Bridge “not powerful enough to need more power”? Nope.

I personally like the low TDP. I think the thermal envelope has been pushed too hard in GPU space, and they need to start focusing more on performance/watt like Intel does. Intel’s products became amazing once they adopted this philosophy.

+1 I completely agree.

 

[Crap Daddy]

... What. I don't even know where to begin, so I'll just leave it at that.

You can't take a joke? This thread, as all regarding Kepler are too heated up.

 

[Aeiou]

I hope it really does end up with 15-20% on the 7970.

not that i care which is better, i want PRICES to go down, or at least the gouging to stop.

 

[Arzachel]

You can't take a joke? This thread, as all regarding Kepler are too heated up.

Sorry, the internet has trained me to always expect the worst

 

[AtenRa]

http://forum.oktabit.gr/content/φήμ...τες-γραφικών-της-nvidia?page=28#comment-91321

 

[Quantos]

The length difference between 2447 and 2595 looks a bit much. We don't know the settings used of course, but do you guys know how that compares to 7970 / 580?

Also, what's wireless OC? :hmm:

 

[Imouto]

Yes me too. This bench comes from that thread

I might add "posts coming from the OP in that thread".

 

[MrK6]

The length difference between 2447 and 2595 looks a bit much. We don't know the settings used of course, but do you guys know how that compares to 7970 / 580?

Also, what's wireless OC? :hmm:

Interesting results. Quick googling pulls up these Unigine 2.5 scores with GTX 580 and 7970 comparisons:
http://www.pcper.com/reviews/Graphi...cs-Card-Review-Tahiti-28nm/Unigine-Heaven-v25
http://www.vortez.net/articles_pages/gtx580_vs_hd7970_the_ultimate_face_off,16.html

I don't think there were any performance changes between the benchmark revisions, so hopefully the results are comparable. Anyway, from the links it seems that the difference between the GTX 580 and 7970 goes anywhere from ~23% all the way to 40%, so it's tough to say. Judging by the high numbers the MSI handout gives, I think they're running at a very low resolution/tessellation setting.

The score boost seems to be ~5%, which would be the same as the clock boost to ~1050MHz we've seen before, so good scaling there.

 

[BFG10K]

“All solid capacitor – no explosion”.

That’s a good feature to have, LOL.

 

[Arzachel]

“All solid capacitor – no explosion”.

That’s a good feature to have, LOL.

Yeah, had to smile at that "feature" The lifetime bar for office usage is also silly. Who would buy a GTX680 to browse web?

 

[Tempered81]

Gonna run my GTX680 in 2D for 57 years, see if it dies on me.

 

[MrK6]

Yeah, had to smile at that "feature" The lifetime bar for office usage is also silly. Who would buy a GTX680 to browse web?

Because I want to future-proof my next build, so I need something that will run Adobe Flash 43 in 2050 :awe:.

This is looking good though, hopefully reviews come up soon.