AMD: R9xx Speculation

Thsese are AMD marketing slides , they are chosen specifically to represent best outcome possible , I suspect they used 2500x1600 with 8XAA just to run the competition into FB limitation (pretty obvious in Crysis and Metro).

Using 8X AA is probably a good thing to do though , I think AMD just improved 8XAA performance again (enhanced ROPs) .

Judging from the benches of the guys posting on Hardwareluxx.de, HD6970 is better than a GTX580 even on Crysis at 1920x1200 AA4x. ;)
To me it's not bad at all... It seems like by cherrypicking benches an HD6970 could challenge a GTX580, so it is probably a bit slower across the line (5-10% below), but closer to the top NVidia single gpu card than HD5870 and HD4870 were.
It could also be that the HD6950 is a worse shot for the money than HD5850 and HD4850 were (too close to an HD6870 to be 100€/$ more expensive).
 
against the 5870 it's benchmarked at 25x16, against the 480 it's benchmarked at 19x12.
driver rev for the 480 is 260.99, all cayman and cypress benchmarks are done with 10.11.

If these slides are true, Neliz, how do they justify your claim that "GTX 580 can trade some blows with HD6950?".

Im still finding all this veeeeery fishy.
 
I really want to believe that all the slides are fake and that Cayman is much faster then a GTX580, but I don't think that's gonna happen.

Fake slides don't indicate anything about performance one way or another. I can believe the slides are fake and still believe 6970 is slower than the 580.
 
Ill call it fake because:

1 - Would a marketing slide say "on average"?
2 - On HD6950 heading is too close to graphic top.

They dont look very professional to me.
Well, these are outdated. The new ones are more professional (starting at 90%) ;)
 
against the 5870 it's benchmarked at 25x16, against the 480 it's benchmarked at 19x12.
driver rev for the 480 is 260.99, all cayman and cypress benchmarks are done with 10.11.

In AMD slides? :oops:
This means that there could be room for improvement even on what shown on marketing slides? :?:
That seems pretty pointless from AMD if it is so... I mean, shouldn't marketing material show best case scenarios?

Well, these are outdated. The new ones are more professional (starting at 90%) ;)

Hmm, so are you saying that these slides are real but older, when 10.12 performance was still not known?
But what's the point of showing results on official slides with drivers which are still not fully finalized?
 
Last edited by a moderator:
So how do you explain it showing 1120 SP 6850s for example?
It can read SIMD configurations in order to differentiate between a HD 5870 and 5850 or GTX 460 and "SE". But if not backed up by it's database, it cannot know how many ALUs are present within each SIMD. At least as far as I know.
 
Maybe these leaked ones are just the first batch of slides from AMD.

They then decided to improve them (e.g. by comparing 59xx to nv's 5xx gen) hence the launch delay.



Sorry, I'm possessed by the silly season :D
 
Yes, certainly possible. Right now, only thing I have is my most recent GTX 580 result and the alleged score from the guy who already could buy a HD 6970. I am switching back and forth between three cards this weekend and my trusty HD 5870 is certainly getting it's stab at Computemark as soon as I am done with the rest.

HD5870 on the leaked Catalyst 8.79.6.2 RC2, ComputeMark 2.0, 1920x1200, Extreme: 233

Driver bug, i presume.
 
Back
Top