NVIDIA GeForce GTX 680 Kepler 2 GB

[alexsubri]

I want to look into this as well. It doesnt make sense that the 7870 would be faster

for instance

http://images.hardwarecanucks.com/image//skymtl/GPU/HD7870/HD7870-36.jpg

It's the drivers guys ,,,its a risk we all took when purchasing the 7xxx series

 

[Melvis]

I have done my own tests and contradict everything you say. It's pure myth. And W1zz's demostration of driver (non-)improvement is not good for you? haha, not when it doesn't suit you right?



I cannot say every other review. I cannot speak in absolute terms, that's not falsable, but yes, every review that I've seen other than TPU, have the HD7870 being a lot slower than HD7950.

Something tells me you haven't because if you did then you would of got the results like i did, i think i even wrote it all down yrs ago, if your lucky i might show you later. (if i can find them)

Regarding what Wizz demonstrated this was before those cards/drivers so i cant tell you, but i know just from my own experience from having these two cards < that almost every driver update that i have used (not all but most) i have seen an increase in performance in the games i play, who better to know then the person who owns the cards/setup right?

Well if thats the case then thats fair enough, its a glitch then, and im happy to go with that, i just stick with TPU reviews of GPU's as there the best out there and i go along with what Wizz says, he is the master after all

Once again only time will tell 6 or so months from now when the drivers have been all worked out for BOTH of these cards will tell us how much faster the GTX680 realy is compared to the 7970 at the moment i take it all with a grain of salt but either way the GTX680 is still a great step forward for Nvidia and TBH i was hopeing to go back to Nvidia soon and this is making me seriously considering it. ( does that make me a fanboy?)

 

[Benetanegia]

Well what can I do if you want to believe in Unicorns.

Sigh. Like I said, probably 1-2 games got an improvement and you attributed it to all of them, as if it was something that happened across the board. 2 months later, new drivers, new games, another increase in a couple games, across the board claim, rinse and repeat. Every single time happens the same.

And like I said I cannot convince you, but for me and I supose 95% of people out there a W1zzard's analysis like I posted has far more weight than myths and one single guy's beliefs of what's happened. Like I showed, when the HD4870 launched the HD3800 were still much slower than 8800. That is fact, so there were no significant improvements, at leasst over the improvements that the 8800 also achieved.

 

[Melvis]

Well what can I do if you want to believe in Unicorns.

Sigh. Like I said, probably 1-2 games got an improvement and you attributed it to all of them, as if it was something that happened across the board. 2 months later, new drivers, new games, another increase in a couple games, across the board claim, rinse and repeat. Every single time happens the same.

And like I said I cannot convince you, but for me and I supose 95% of people out there a W1zzard's analysis like I posted has far more weight than myths and one single guy's beliefs of what's happened. Like I showed, when the HD4870 launched the HD3800 were still much slower than 8800. That is fact, so there were no significant improvements, at leasst over the improvements that the 8800 also achieved.

Not as much as what i can do if you want to beleive in pigs that fly, o well. Grow up FFS.

And you just proved my point, across the games and in time there is imporvment you just said it ya self, so how can you not belevie me when i say this? you make no sense boy.

Well i guess your one of those ones that just hasnt tried for you self then i guess? not my problem if you dont beleive me or not, maybe before you start making claims based on someone elses opinon and not yours you should first try it for your self hmmm?

Fact is and is still my point from the start and you just agreed with me that newer drivers DO give improvement in performance regardless what anyone says, if you have ever owned a GPU and mainly a ATI one you would know this, but guess you havent? o well guess you wont know till ya buy one dude, have a nice day im done talking to kids

 

[KainXS]

Well what can I do if you want to believe in Unicorns.

Sigh. Like I said, probably 1-2 games got an improvement and you attributed it to all of them, as if it was something that happened across the board. 2 months later, new drivers, new games, another increase in a couple games, across the board claim, rinse and repeat. Every single time happens the same.

And like I said I cannot convince you, but for me and I supose 95% of people out there a W1zzard's analysis like I posted has far more weight than myths and one single guy's beliefs of what's happened. Like I showed, when the HD4870 launched the HD3800 were still much slower than 8800. That is fact, so there were no significant improvements, at leasst over the improvements that the 8800 also achieved.

Whoever has had these cards knows your right(I HAVE), why argue

just ignore it man, don't make a big deal of this.

 

[eidairaman1]

Again the GTX 680 is extremely well priced but didn't turn out to be the ultra powerful beast some were anticipating.
Ether way @ $499.99 the 680 is a better value than the 7970 at the moment.

Yes the 7970 is limited by drivers right now but I am sure Nvidia can make improvements. They are both good cards and the 680 should be faster considering it was released 4 months later.

Honestly this should bring pricing to a decent range.

 

[xenocide]

It's not drivers. W1z must have found a glitch with Crysis 2 that rendered it at a lower resolution or quality setting for the 78xx testing. Either that or the Drivers he was given had some odd tricks up their sleeves to make the cards appear to run better than they should.

Seriously, what does AMD have to gain out of gimping their flagship GPU?

 

[eidairaman1]

It's not drivers. W1z must have found a glitch with Crysis 2 that rendered it at a lower resolution or quality setting for the 78xx testing. Either that or the Drivers he was given had some odd tricks up their sleeves to make the cards appear to run better than they should.

Seriously, what does AMD have to gain out of gimping their flagship GPU?

IDK honestly and I dont care as card prices over 400 bux is ridiculous (400 bux got you a Radeon 9700 Pro AIW in 2002)

 

[xenocide]

IDK honestly and I dont care as card prices over 400 bux is ridiculous (400 bux got you a Radeon 9700 Pro AIW in 2002)

http://146.142.4.24/cgi-bin/cpicalc.pl?cost1=400&year1=2002&year2=2012

2002: $400
=
2012: $506.20

Actually the GTX680 is the same value (when factoring inflation).

 

[eidairaman1]

ya but 400 bux is still 400 and 500 is still 500

 

[entropy13]

 

[sergionography]

Remember the brilliant RV770? This is history repeating with Nvidia as the one with the clever architecture.

GTX680 = Tahiti. So either card is worth buying in terms of performance. The reason to buy Kepler would be the better price, lower power consumption (i.e. GPU produce less heat energy), CUDA, fancy AA. As for features that were unique to ATI is also now available in NV, eyefinity and the lean/low power consuming architecture. I hope NV will never make another barbarian/joule guzzler GPU again.

This is really the perfect achievement from Nvidia! This is definitely a worthy buy, lets see if this will ignite the price war. I'll definitely get one when the price is a bit better than they are now.

What I seriously love about this card is that it is able to keep up despite being lean and consumes less power. Moreover, for once, this chip does not have a heat spreader like previous NV cards. This makes the job of adding water block a lot easier. When I mod my 560TI to water cooling, it was quite annoying that the spreader was in the way. I end up drilling holes and a whole lot of custom mod to get it to work.

The dynamic clocking system is a good feature. I actually tried my own custom dynamic clocking on my 4890. It worked but everytime the clock changes, there is a noticeable halt. This is similar to the problem faced by Kepler now. See the last paragraph:
http://techreport.com/articles.x/22653/9
Hopefully future drivers will fix this up because I hate those inconsistent framerate.

well its now kinda the other way around, kepler is kickass in gaming, in compute not so much, or atleast according the to benchmarks ive seen on anandtech, were yet to see more benchmarks about compute
in other words cuda isnt why you would buy nvidia this time.
as far as gtx680 is concerned i think its most appealing for those who just wanna put a card in and game, for overclocking i think the hd7970 still holds its ground, and with amd drivers it has room to improve
in other words this will translate into some interesting price wars! bring it on

Something tells me you haven't because if you did then you would of got the results like i did, i think i even wrote it all down yrs ago, if your lucky i might show you later. (if i can find them)

Regarding what Wizz demonstrated this was before those cards/drivers so i cant tell you, but i know just from my own experience from having these two cards < that almost every driver update that i have used (not all but most) i have seen an increase in performance in the games i play, who better to know then the person who owns the cards/setup right?

Well if thats the case then thats fair enough, its a glitch then, and im happy to go with that, i just stick with TPU reviews of GPU's as there the best out there and i go along with what Wizz says, he is the master after all

Once again only time will tell 6 or so months from now when the drivers have been all worked out for BOTH of these cards will tell us how much faster the GTX680 realy is compared to the 7970 at the moment i take it all with a grain of salt but either way the GTX680 is still a great step forward for Nvidia and TBH i was hopeing to go back to Nvidia soon and this is making me seriously considering it. ( does that make me a fanboy?)

you talk as if ur the only one who owned the card

and again as far as drivers are concerned its almost always AMD who has potential increase from driver updates mostly because everytime a new driver comes out it either is half baked or ruins something else while its at it, the only question is will we see these "perfect" drivers? and if so, when?
nvidia on the other hand has it down right usualy from the start, with minor tweaks here and there.

 

[burebista]

everything can be hacked. just a question of how difficult and who will spend time on it

My friend Ramiro tried something.
To sum his thoughts: the card is perfect for gamers out-of-the-box but crappy for benchers (crappy VRM and crappy power limit).

 

[Rahmat Sofyan]

I wonder, on this review, FXAA is on or off?
and how about TXAA?
whether TXAA automatically activated, and therefore the GTX 680 so fast?
or indeed GTX 680 really fast, because we already knew nVidia bring new AA on kepler, but on any review I can't find anything how to use TXAA or something else about it.

any advice?

 

[Melvis]

you talk as if ur the only one who owned the card

and again as far as drivers are concerned its almost always AMD who has potential increase from driver updates mostly because everytime a new driver comes out it either is half baked or ruins something else while its at it, the only question is will we see these "perfect" drivers? and if so, when?
nvidia on the other hand has it down right usualy from the start, with minor tweaks here and there.

lol naa and TBH both the cards wasnt even mine, i just build them and test them, so i did and thats the results i got. the 3850 matched the 8800GT in most of the games i benched using the Newest drivers from ATi at the time and leaving 6-8months old drivers with the 8800GT. BUT once i updated the drivers for the 8800GT yea it was game over the 8800GT walked all over the 3850. (windows XP was used)

Thats exactly right, and as far as perfect drivers go, that will never happen. Only improvments till us the user is happy with it with what games we play.

 

[W1zzard]

W1z must have found a glitch with Crysis 2 that rendered it at a lower resolution or quality setting for the 78xx testing

i think that's what was happening for hd 78xx. i restored the original disk image and crysis 2 gains went away. will look more into this, but second gtx 680 is arriving today, so sli review has priority

I wonder, on this review, FXAA is on or off?
and how about TXAA?
whether TXAA automatically activated, and therefore the GTX 680 so fast?

txaa can not be activated through the driver, it requires game developers to actively integrate it in their program code.

fxaa was off, i never touch any settings in the driver for my reviews, which is what most people do

 

[Rahmat Sofyan]

i think that's what was happening for hd 78xx. i restored the original disk image and crysis 2 gains went away. will look more into this, but second gtx 680 is arriving today, so sli review has priority



txaa can not be activated through the driver, it requires game developers to actively integrate it in their program code.

fxaa was off, i never touch any settings in the driver for my reviews, which is what most people do

now it's clear.GTX 680 is the fastest one!!!

thanx W1zz

 

[W1zzard]

Just got a package from ASUS:

Spoiler

 

[m1dg3t]

Can i haz when done pl0x?

 

[entropy13]

That empty box looks quite heavy.

 

[MxPhenom 216]

so much for GTX680 failing at overclocking

http://www.overclock.net/t/1232947/hwbot-gtx-680-takes-record-1900-mhz#post_16783937

oh jesus Zotac is releasing a 2GHZ GTX680 varient....I doubt this will happen.

http://translate.google.com/transla...http://digi.tech.qq.com/a/20120322/002197.htm

 

[m1dg3t]

so much for GTX680 failing at overclocking

http://www.overclock.net/t/1232947/hwbot-gtx-680-takes-record-1900-mhz#post_16783937

Def nice clock's! bta already posted a link to that yesterday IIRC. Also you wont see any end user's getting that high, best we/they can hope for is what ever the DOC paramater's allow

Plus that was done with the add on VRM pwr board

 

[jamsbong]

well its now kinda the other way around, kepler is kickass in gaming, in compute not so much, or atleast according the to benchmarks ive seen on anandtech, were yet to see more benchmarks about compute
in other words cuda isnt why you would buy nvidia this time.

You're right, I just went through Anandtech's review. There was a long detailed explanation comparing Fermi with Kepler architecture. In summary, Kepler remove many of the non-gaming features to make it go fast and efficient. Surprisingly, ATI has gone in the opposite direction with GCN being more complex to adapt itself for general computing capability. Note that Kepler's double precision FLOP is lower than Fermi!

I guess this also means that Tahiti will never be as efficient in gaming as Kepler. Oh well...

 

[Recus]

Here is new tech demo preview:

Bokeh depth of field
APEX clothing
Tessellation wrinkles?

http://product.pcpop.com/000337190/Picture/005200798.html#005200828

 

[Crap Daddy]

The only thing left for the 7970 is the ability to overclock in a more flexible way and in theory reach higher speeds than the GTX680. But in a more practical way this is a bit hard to reach.

While overclocking on the GTX680 is limited by certain factors, 1250-1300MHz are attainable under the restrictions imposed by NV although we must take into consideration that these clocks are not fixed, they change all the time under load depending on certain conditions. So a comparison clock/clock with Radeon is almost impossible. On the other hand, 1200 MHz are pretty hard to reach with a 7970 and the cost in thermals and power consumption are huge.

Take a look at this graph. GTX680 is clocked at 1156 base, 1209 boost, highest read in Afterburner 1275. The 7970 was clocked at 1180 with 1.25v