Discussion in 'Article Discussion' started by arcticstoat, 28 Mar 2011.
Can open, worms everywhere.
People are missing the point here. Of course it matter's who has the fastest graphics card, thats what its all about. You don't see Ferrari trying to build the fastest production car then being happy "its as good but we are not sure" as a Maserati.
These two companies pride themselves on providing you, the customer with the best your money can buy and if both sides cliam to produce the most powerful card then of course it muct be tested and proved, not just to show off but if anything so the losing side can go back and bring something even better out.
Plus all the inovations they make at this end of the scale are what trickle down to the cheaper cards making them even better value.
innovations made for fastest GPU gets trickled down to cheaper cards.
innovation made for dual-GPU card are simply how to fit them onto a board and how to cool them. zero actual GPU innovation that is of any use to cheaper cards.
this multi-GPU on business is getting ridiculous by the day. this is why i read bit-tech and CPC, for their ever lasting single GPU support, instead of mixing messy multi-card graph getting mixed in the review.
Fear is the mind killer. And heat is the board killer. The one thing ive learned over the long years is that x2 cards get too hot and aren't fun when it comes to overclocking your rig. Plus stay away from nvidia's current flagship single chip cards as they tend to be too hot. Plus dual card setups have a tendancy to fry what ever card is in the secondary slot(but this could happen as much as 2-3 years after the dual install) Ah im glad i got that off my chest
Hang on hang on... Let me get your argument. The only way that the GTX 590 (which is two throttled back 580's) will beat the 6990 (which is two throttled back 6970s) is by unlocking the clocks on the 590 to make it two full fat 580's? And only then will it make the 6990 "crumble".
Urrmmm... As Ron Burgundy said "you know that doesn't make sense"
G o to both of reviews, 6990 and 590. Ofcourse AMD wouldn't be happy with these reviews. Battlefield: Bad Company 2 1,680 x 1,050, nvidia 590 review came after 6990 review, and amd is clearly ahead, but is not listed in the 590 review. and then there two games that bit-tech did not compare with the 2 cards. Arma 2 and Just Cause 2. So they are just comparing 3 games only.
In Black Ops AMD Radeon HD 6990 4GB is worse than 5850. That's interesting too. + the fps for AMD's card in 2 reviews is totally different.
And I don't have the slightest idea why they are testing cards on Dirt 2. Because it's just not for radeon. It's like comparing which will be the best nvidia card, there is no competition with amd on that game.
-What are the other components of the PC? They may have an impact.
-If you're using synthetic benchmarks, which ones? Different benchmarks may stress different features.
-If you're using games, which ones? Different games have been shown to heavily favor one side or the other. Just look at Dirt 2 and Nvidia, or STALKER and AMD.
-What resolution are you running at and with what settings? Different cards can have different strengths and weaknesses which may become apparent with different settings.
-Are you looking at average or minimum frames per second.
-What about anti-aliasing quality? Using a prettier method may give a better experiece even if it doesn't give as many FPS.
And countless other factors which I can't be bothered to think about.
I'm gonna go ahead and cosign this.
I love how AMD just happens to conveniently exclude real-world gaming benchmarks, especially since they've been known to have driver optimizations for those benchmarks.
It's really poor of them to do this. They realize they're outgunned, and their first reaction is to complain.
590 = Yes it'll play Crysis, Crysis 2, Metro 2033, Arma II and even Minecraft
6990 = Yes it'll play Crysis, Crysis 2, Metro 2033, Arma II and even Minecraft.
590 = Yes it'll bloody expensive.
6990 = Yes it'll bloody expensive.
590 = So very dependent on having drivers tweaked to the game / benchmark
6990 = So very dependent on having drivers tweaked to the game / benchmark
590 = Beats 6990 in some tests
6990 = Beats 590 in some tests
590 = For willy waving by those that own one
6990 = For willy waving by those than own one
590 = For willy waving by those that make them
6990 = For willy waving by those that make them
They seem pretty similar to me...
And I do love AMD's disclaimer...
"Dave Erskine is the Senior Public Relations Manager for Graphics Desktop at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions."
So under those terms, dear old Dave Erskine can say whatever he likes, pick and choose his "facts" as it's not AMD saying this, it's just one guy.
this kind of disclaimer is under all blog posts so i'm not sure what you're getting at.
Here's hoping for a new graphics card manufacturer in the near future, because AMD and nVidia are starting to get long in the tooth - not mention taking the p1ss. Who cares which is faster? Well, obviously they do, but how many sales will a public hissy-fit gain for them? 0.001% increase or something?
We need a modern day equivalent of the times when we had S3, Matrox, 3dfx -- ie more competition!!
Run tests at 2560x1600 (minimum) on a wide variety of popular current games (from the last 3 years) and average out who scores better overall.
Although I prefer nVidia, I have a feeling at those resolutions the 6990 would come out on top thanks to the extra 512MB VRAM on each GPU..
I would really love to see a detailed analysis on which is actually the OVERALL fastest card. As they are quite close
not if your motherboad has 3 slot spacing for SLI/xfire - aka virtually all reasonable SandyBridge Mobos, and some higher end 1366's (some even have the two main slots 4 spaces apart due to quad capabilities)
You know exactly what he's getting at.
The fact is AMD are not saying this, it's just their PR guy having a hissy fit.
Very nicely put
Consider this 2 6950s unlocked ( aka 2 6970s) beat both for outright performance cost bearly £500 and dont sound like a vacuum cleaner in take Off mode
Anyone buying a £5-600 card is going to have an sli or crossfire board I don't really see the point in duel gpu cards
Go read amd forums about crysis 2 support as the game is crippled and bearly playable ac brotherhood is the same.
As for biggest games is that by sales?
World of Warcraft does not support duel gpu cards or sli or crossfire nor will It ever
Cod mw2 min fps above 100 on any card built in last years
Cod black ops same as above
Sims3 it runs on integrated graphics
Football manager same as above
That's by sales of pc games in last 5-6 years
Metro 2033 did not sell well
Crysis did not sell well
AVP was a truly awful game and did not sell well
They are the 3 most demanding games graphically released in the last 5 years .
Hawx the only real testilation game has nvidia ahead so far it's not even a close match 570 onwards beats any amd card
Bit tech review what people actually play. Not games nobody plays or else you could add those 4 games into results would not change much
Err so what version of Crysis 2 have I been playing then? I have 2 x 6870's and it's running extreme settings on a 1920 x 1200 res...
So far I have not noticed any "crippling" game play.
The usual TWIMTBP games?
Rollo the logic of you arguement would have Bit-tech simply saying in their gpu reviews
"You're all playing these games...
World of Warcraft
Cod black ops
....this HDxxxx or this xxxgtx gives you 60fps no probs. Enjoy." End of review!
Where as Metro, Crysis and, as you say AvP do test a gpu but aren't used in the reviews... from Bit-tech 20.10.10. review of the 470 AMP! ed. ...
"Ah, yes, but can it play Crysis? There, we beat you to it...
If thereâ€™s one game to prove whether a graphics card has enough performance to cope with any game you might throw at it, itâ€™s Crysis. Even though the game released in November 2007, it still remains one of the most visually stunning games around, with volumetric fog, crisp textures and more eye-candy than an opticianâ€™s sweet shop."
That was the last gpu review to use Crysis. Isn't the point of a test review to actually test the cards in a controlled and FAIR manner? However Dirt 2 remains as a bench mark to this day. Which is strange given...
"However, Dirt 2 is a game in which Nvidia cards perform very well and this means that the HD 6870 1GB is fighting off competition from the GeForce GTX 460 1GB rather than its price point competitor" ...Bit-tech's 22.10.10 review of the HD6870.
"In Dirt 2, a game in which weâ€™ve found Nvidia cards hold a significant advantage, the 6990..." ...from Bit-tech's 6990 review.
Which of Arma ll, Bad Company 2 or CoD:Black Ops is known to 'favour' AMD gpus?
A game that tests all cards is dropped, a game known to favour one gpu supplier is kept. Go figure!
Separate names with a comma.