Why does page 3 spend the first 3 paragraphs repeating exactly what was just written on pages 1 & 2? Sloppy writing to say the least - we do have memories!
Card looks totally insane though, the power requirements are ridiculous to say the least. I'll stick with single GPUs I think...
Originally Posted by r3loaded Performance is great, but it doesn't really put enough ground between itself and the GTX 580. The 590 will probably crush it when it's released.
Eh? In BC2 as an example it is nearly twice as fast at resolutions this card was designed for (2560x1600). How is that not putting enough ground between itself and the GTX580?
Please use games that cripple graphics cards such as Metro 2033. Running games like you have done shows nothing and makes the card look faster than it is hence the fact it was only 10% faster than the 580 except for a single game at a resolution most do not have.
Thats like getting a pair of 480s or 6950s/6970s and running at 1920x1200... very few do that as it is known to pretty much be pointless.
Power usage and noise was the big issue with the Nvidia 4XX series in their first generation but people still picked them up.
I'm still a little sceptical about multi cards on one card but that's mainly due to people who had them moaning. I've never owned one so I don't know for sure.
I'm sure AMD will sell a few of these until the data is in on the 590. If the 590 can compete on performance then it will sell more even if it was more expensive. If you are paying £540 or £600, that wont matter to that kind of buyer.
Its good to see AMD put a performance in though, even if it's tainted by the curse of the GTX 4XX series. It's a pity the rest of the line up didn't live up to the hype.
Originally Posted by andrew8200m Please use games that cripple graphics cards such as Metro 2033.
Sorry Andrew, but we don't test with Metro 2033 because it's a dull game that very few people play so it has very little revelance to the gaming market.
Originally Posted by Snips Power usage and noise was the big issue with the Nvidia 4XX series in their first generation but people still picked them up.
True, but not to the extent of blowing past PCIe power standards and basically requiring two separate rails. I think noise however is a given when you buy a high end card though.
Originally Posted by Lizard Sorry Andrew, but we don't test with Metro 2033 because it's a dull game that very few people play so it has very little revelance to the gaming market.
I agree here but this isn't a review for games. Its a review for a graphics card so showing its performance in the most demanding of games regardless of whether or not people like to play it should be a priority?
If there was a craze to play Quake the or the original dawn of war would you review cards based on those game just because people may like playing them? If you did it would probably be pretty pointless with every card newer than say an 8800gt maxing frames to a point that it would be indistinguishable even by using fraps.
Not overly bothered with this release - it is fairly obvious it was going to smash all benchmarks simply because of what it has under the bonnet. Comes at too much of a price on balance, power, cost and driver issues - not for me, although the figures are impressive.
Originally Posted by Snips I'm still a little sceptical about multi cards on one card but that's mainly due to people who had them moaning. I've never owned one so I don't know for sure.
I've never had any problems with my 4870x2. The results from various sites have these new card between 50% (but mainly 100%+) and 300% faster than it. At its release the 4870x2 was an amazing card and great value for its performance, so you can imagine how happy I am. Still, as I've said previous, we've still got Nvidia to present their equivalent.
Just far far far too expensive to be practical at the moment and i dont think it can justify its price....its almost the half the build cost of a decent gaming pc on its own..
That is defintely a beast of a card, although I still think 2x 6950's re flashed to 6970's is the way to go. Just shy over £400 too.
Quote:
Originally Posted by Lizard Sorry Andrew, but we don't test with Metro 2033 because it's a dull game that very few people play so it has very little revelance to the gaming market.
Whether i agree with Andrew or not, i can't help but feel the same way with the your current games list.
(honestly BT need to increase the number of gaming benchmarks, none of those games really showed much for the card except when the 2 cores bulldozed through 2560x1600 res).
Originally Posted by Lizard Sorry Andrew, but we don't test with Metro 2033 because it's a dull game that very few people play so it has very little revelance to the gaming market.
As the others say it's testing performance not game quality.
The other thing this would test much better is microstutter - you aren't going to see it if the min frame rate > 60fps, if you max out metro and bring the card to it's knees then you have a much better chance of finding out if it stutters.
Comments 1 to 25 of 107
ReplyI'd rather save some money and buy a 580.
Still reading though, this card is a powerhouse..
Card looks totally insane though, the power requirements are ridiculous to say the least. I'll stick with single GPUs I think...
Just waiting on the 590 tests now and then I can buy a card ...
Eh? In BC2 as an example it is nearly twice as fast at resolutions this card was designed for (2560x1600). How is that not putting enough ground between itself and the GTX580?
Please use games that cripple graphics cards such as Metro 2033. Running games like you have done shows nothing and makes the card look faster than it is hence the fact it was only 10% faster than the 580 except for a single game at a resolution most do not have.
Thats like getting a pair of 480s or 6950s/6970s and running at 1920x1200... very few do that as it is known to pretty much be pointless.
My entire rig probably consumes about the same amount of power in total.
Crazy fast though!
I'm still a little sceptical about multi cards on one card but that's mainly due to people who had them moaning. I've never owned one so I don't know for sure.
I'm sure AMD will sell a few of these until the data is in on the 590. If the 590 can compete on performance then it will sell more even if it was more expensive. If you are paying £540 or £600, that wont matter to that kind of buyer.
Its good to see AMD put a performance in though, even if it's tainted by the curse of the GTX 4XX series. It's a pity the rest of the line up didn't live up to the hype.
Sorry Andrew, but we don't test with Metro 2033 because it's a dull game that very few people play so it has very little revelance to the gaming market.
True, but not to the extent of blowing past PCIe power standards and basically requiring two separate rails. I think noise however is a given when you buy a high end card though.
I agree here but this isn't a review for games. Its a review for a graphics card so showing its performance in the most demanding of games regardless of whether or not people like to play it should be a priority?
If there was a craze to play Quake the or the original dawn of war would you review cards based on those game just because people may like playing them? If you did it would probably be pretty pointless with every card newer than say an 8800gt maxing frames to a point that it would be indistinguishable even by using fraps.
Food for thought.
I've never had any problems with my 4870x2. The results from various sites have these new card between 50% (but mainly 100%+) and 300% faster than it. At its release the 4870x2 was an amazing card and great value for its performance, so you can imagine how happy I am. Still, as I've said previous, we've still got Nvidia to present their equivalent.
That is defintely a beast of a card, although I still think 2x 6950's re flashed to 6970's is the way to go. Just shy over £400 too.
Whether i agree with Andrew or not, i can't help but feel the same way with the your current games list.
(honestly BT need to increase the number of gaming benchmarks, none of those games really showed much for the card except when the 2 cores bulldozed through 2560x1600 res).
As the others say it's testing performance not game quality.
The other thing this would test much better is microstutter - you aren't going to see it if the min frame rate > 60fps, if you max out metro and bring the card to it's knees then you have a much better chance of finding out if it stutters.
stupid stupid
-
« Previous
-
1
-
2
-
3
-
4
-
5
-
Next »
Discuss in the forums