HD6970 vs gtx580 benchmarks

N

nest2001

n00b
Joined
Oct 15, 2004
Messages
31
coolpc-gv-hd6970-hd6950-24.gif

coolpc-gv-hd6970-hd6950-25.gif

coolpc-gv-hd6970-hd6950-26.gif

coolpc-gv-hd6970-hd6950-27.gif

coolpc-gv-hd6970-hd6950-28.gif


Full review:
http://www.coolpc.com.tw/phpBB2/viewtopic.php?f=52&t=57400&st=0&sk=t&sd=a
 
How big a correlation is there between 3DMark performance and gaming performance? (In this case)
 
Why would you guys trust a supposed "review" (LOL) that uses Lost Planet 2 as one of its only benchmarks?
 
XacTactX said:
How big a correlation is there between 3DMark performance and gaming performance? (In this case)
Click to expand...


probably none, if any...not really any games out there using DX11 to its fullest.
 
Lost Planet 2 is an EXTREMELY Nvidia friendly benchmark. I believe 580 was beating 5870 by on the order of 50% at least in LP2. So, yeah...

And the other two benches are synthetics. And it already seemed known that 6970 may be a dog in 3Dmark while a winner in gaming benchmarks. So this review seems to be...fail.
 
martinje said:
I dont know about that fully, the 6870 beats the 470 at 2560 x 1600. In the other resolution it only loose by 4 fps or so..

http://www.tomshardware.com/reviews/radeon-hd-6870-radeon-hd-6850-barts,2776-11.html

I think Nvidia will probably come on top this series.
Click to expand...


I agree. But the gap seems to be closing with every new gen.

It just seems that this time, people expected AMD to blow Nvidia away, which wont happen. I guess because people see Cayman as all new while 580 is just Fermi refined, plus the Caymen delays, they expected Caymen to beat 580.

Nvidia will come out on top I think, but NOTHING like that LP2 bench makes it look...probably by 5-10% is my guess.

And then AMD will have Antilles/6990 in the 1st quarter, which I think will be the fastest single "card" over anything Nvidia offers, just like 5970 is now.
 
Or AMD/Ati could degrade their signal quality a bit to give their 6850/70 a !0% increase at default installation.


What I am talking about, they did do that. DoH
 
Terrible review. No 3dMark11 Extreme preset and using Lost Planet 2 as a basis for your performance? lol..
 
The big reason the 6970 benchmarks seem disappointing is that the 6870s trounced the 460s and even matched the 470s forcing Nvidia to pricecut both their 460/470 lines. I think a lot of people were hoping for the same kind of perfomance gains from the 6970 especially with the highly touted change in shader technology combined with the new architectural front end that had made the 6870 such a tasty value at the $240 pricepoint. In comparison the 6970 benches have not been promising to say the least.

Let's HOPE it really is because the 6970 was running at 190W mode on outdated drivers rather than that being its true performance. Because just matching a 570 isn't enough imo, the 6970 must at least trade blows with the 580 to be considered a success after being released later. Can't wait to see REAL 6970 reviews come tomorrow. And hopefully they address the different power settings/performance issue as well.
 
vjcsmoke said:
The big reason the 6970 benchmarks seem disappointing is that the 6870s trounced the 460s and even matched the 470s forcing Nvidia to pricecut both their 460/470 lines. I think a lot of people were hoping for the same kind of perfomance gains from the 6970 especially with the highly touted change in shader technology combined with the new architectural front end that had made the 6870 such a tasty value at the $240 pricepoint. In comparison the 6970 benches have not been promising to say the least.

Let's HOPE it really is because the 6970 was running at 190W mode on outdated drivers rather than that being its true performance. Because just matching a 570 isn't enough imo, the 6970 must at least trade blows with the 580 to be considered a success after being released later. Can't wait to see REAL 6970 reviews come tomorrow. And hopefully they address the different power settings/performance issue as well.
Click to expand...


And what if it's 10% slower than 580 at 399? I think most people would take the $100 savings.

There's more than one way to be a good card.

But I def agree that if it only matches 570 it will be a failure. Especially since 570 basically=480, which would mean AMD only matched Nvidia's last high end. Needs to slot between the 570/580 at least, which it almost certainly will anyway.
 
If the 6970 is 10% slower than the 580, then it is only 25% faster than the 6870.

It would also mean the 6950 is only 10% faster than the 6870.

Consume 50% more power, and gain only 25% performance.... we'll see in 1 day if that theory pans out.
 
The real winner in my book will be the one that can have the highest 24/7 stable gaming clocks for price and performance. I really hope they OC like no tomorrow without the need for waterblocks or anything cause if I get one on zero day instinct will say to push it till it bleeds... and then some more. :)
 
heflys20 said:
Anyone notice that those cards have the same green stickers that were put on the engineering sample cards? Even more interesting is that the sticker is blurred out in several pics.

Look familiar?:

http://img6.imageshack.us/img6/3537/lowspeed6970.jpg

Just something to contemplate.
Click to expand...

Well, the stickers are both green, but the ones in the review definitely has something else written on them. Some kind of serial number, looks like. So I doubt that's the case.
 
Raap said:
Well, the stickers are both green, but the ones in the review definitely has something else written on them. Some kind of serial number, looks like. So I doubt that's the case.
Click to expand...

The fact that they blurred it out makes me suspicious.
 
It looks like Dan (who I bet by now has tested the Radeon cards) has just bought two GTX 580!! I guess the staff here voted with their money for the best card, and to me that looks like the best indication of performance. see his post here
 
Let me repost for those who are too lazy to check the link:

________________________________________
Dan_D said:
He's got some things which drive the price of his system up compared to yours. Dell 3007WFP-HC monitors. Those are $1,200 or more each. Three GeForce GTX 480's in 3-Way SLI. They aren't as expensive or as fast as GeForce GTX 580's now, but you've got two compared to his three. He also most likely bought them a little while ago when they were pricier. Again when new this all drove the price up. He's also watercooling his GPUs which adds quite a bit of cost to the equation. I'm not sure if your doing the same.

I've got a similar setup to yours myself. Water cooled 4.2GHz Core i7, 12GB Corsair DDR3 2000MHz modules, dual GeForce GTX 580's (well those are actually in transit, should be here tomorrow), 3 Dell 3007WFP-HC's, Obsidian case, Intel SSD. So mine should be faster than his too and just about as fast as yours. However, I'd wager that my machine was more expensive than yours. Cost and performance aren't always so closely related. My cost to build this system using the original price of all parts puts it up there quite a bit. Here is an example of my system's cost using the prices I paid for most things or their new prices at the time I purchased the items. I got deals on all the monitors, etc. so the actual out of pocket cost was less than this.

Intel Core i7 920 =$199.99
Koolance CPU-350AT =$79.99
Koolance Exos 2.5 =$374.99
Koolance LIQ-702 Fluid $14.99
Fittings + Quick Disconnects =$75.94
EVGA X58 3X SLI Classified =$429.99
2x Corsair Dominator GT 2000MHz DDR3 (2, 3x2GB Kit, 12GB total) =$659.98
Intel X25-M 80GB Gen 2 =$264.99
2x Western Digital Caviar Black 1TB =$199.98
Sasmsung DVD-R/RW Drive (SATA) =$34.99
Corsair Obsidian 800D =$299.99
Thermaltake ToughPower 1200 watt =$349.99
2x Galaxy GeForce GTX 580 SLI =$1033.60
120mm Yate Loon case fans (x3) =$11.97
Miscellaneous Cable Extensions (ATX/EPS12v and 8-pin extensions) =$50

----------------------------------------------------------------------------------------------------------
3x Dell 3007WFP-HC 30" LCD monitors =$3,899.97
Das Keyboard Professional =$139.99
Logitech G9x Laser Mouse =$69.99
Belkin Nostromo N52 Tournament Edition $49.99
RatpadZ XT =$FREE
15ft. DVI and USB cables (For the displays) =33.68

Total before tax and shipping costs =$8,275.01

So despite the fact that you've got a higher clocked system than I do, mine was WAY more expensive. The OP's system is as well. Now that doesn't mean that we spent foolishly or you bought smarter. That's a different argument. You aren't running 3x30" displays like we are. We can argue that we get a better gaming experience but the cost is much, much higher. Add to that the fact that cases, drive configuration, etc. drive the price up and only make a difference when it comes to how you use the system as well as your personal preferences. The picture thus gets skewed even more. My point is that your system was more economical, but not necessarily better in all areas. Once you hit certain price points for certain things you end up spending a lot more for modest gains. Whether or not that's worth it is up to the individual buyer.
Click to expand...
 
Last edited:
bezant said:
It looks like Dan (who I bet by now has tested the Radeon cards) has just bought two GTX 580!! I guess the staff here voted with their money for the best card, and to me that looks like the best indication of performance. see his post here
Click to expand...

Or maybe he just wanted the GTX 580'? It also sounds like he bought the cards retail, since the 6970 isn't out yet.
 
Kyle and Dan_D are both using 580 SLI (Galaxy)...are they both using the same system for testing purposes or do they both just happen to like the Galaxy 580 cards?

Kyle was a big proponent of Eyefinity early on so the fact that he switched over to Nvidia says a lot
 
bezant said:
No, he usually goes for the best!
Click to expand...

Yes... and the 580 is the best right now, because the 6970 isn't out yet. He's the motherboard editor anyway, not the video card editor.

If the 6970 ends up faster than the 580 he'll just sell the 580's.
 
bezant said:
No, he usually goes for the best!
Click to expand...

Like I also said, it sounds like he purchased retail cards, since he includes the price. How would he able to purchase any 69xx products when they have yet to be released?
 
AceCR42 said:
Yes... and the 580 is the best right now, because the 6970 isn't out yet. He's the motherboard editor anyway, not the video card editor.

If the 6970 ends up faster than the 580 he'll just sell the 580's.
Click to expand...

Yeah, I knew some people cannot pay attention to details...

He just ordered them and they are in transit.

Cards are out and reviewers have signed NDA's....
 
heflys20 said:
Like I also said, it sounds like he purchased retail cards, since he includes the price. How would he able to purchase any 69xx products when they have yet to be released?
Click to expand...

If the Radeons were better he would have waited 1 day... Duh!
 
He could have built that system a month ago when the 580's launched. In fact, he could have had that entire system for 480 SLi, and then upgraded to 580 SLi at launch.
Unless at some point in that thread he gave a build date...

edit;
Oh wait, ignore me.
dual GeForce GTX 580's (well those are actually in transit, should be here tomorrow)
Click to expand...
Bad news for the 6970 no matter how you slice it. Nice detective work. :)
 
Well, since he seems to be the type who cares more about performance than price, all that really tells is that the 6970 won't be faster, at least not generally. Or, possibly, they trade blows but he prefers SLI over CFX.
 
UtopiA said:
He could have built that system a month ago when the 580's launched. In fact, he could have had that entire system for 480 SLi, and then upgraded to 580 SLi at launch.
Unless at some point in that thread he gave a build date...
Click to expand...

dual GeForce GTX 580's (well those are actually in transit, should be here tomorrow),
Click to expand...

Meaning he ordered them recently, likely after they received the 6970's for testing.
 
bezant said:
It looks like Dan (who I bet by now has tested the Radeon cards) has just bought two GTX 580!! I guess the staff here voted with their money for the best card, and to me that looks like the best indication of performance. see his post here
Click to expand...

We have a winner.

Nice work on this. I'd be really surprised if you aren't 100% correct.
 
Dan doesn't test video cards, he tests Motherboards, and he has almost ALWAYS had a 2/3x sli setup.

Brent does the video card testing.
 
You must log in or register to reply here.
Back
Top